AAPM REPORT NO.

71

A PRIMER FOR RADIOIMMUNOTHERAPY AND

RADIONUCLIDE THERAPY

Report of Task Group #7

AAPM Nuclear Medicine Committee

Daniel J. Macey
Lawrence E. Williams (Chairman)
Hazel B. Breitz
An Liu
Timothy K. Johnson
Pat B. Zanzonico

April 2001

Published for the

American Association of Physicists in Medicine
by Medical Physics Publishing
DISCLAIMER: This publication is based on sources and information believed to
be reliable, but the AAPM and the editors disclaim any warranty or liability based
on or relating to the contents of this publication.

The AAPM does not endorse any products, manufacturers, or suppliers. Nothing
in this publication should be interpreted as implying such endorsement.

Further copies of this report ($10 prepaid) may be obtained from:

Medical Physics Publishing

4513 Vernon Blvd.
Madison, WI 53705-4964
Telephone: 1-800-442-5778 or
608-262-4021
Fax: 608-265-2121
Email: [email protected]
Web site: www.medicalphysics.org

International Standard Book Number: 1-888340-29-0

International Standard Serial Number: 0271-7344

 2001 by the American Association of Physicists in Medicine

One Physics Ellipse
College Park, MD 20740-3843

All rights reserved. No part of this publication may be reproduced, stored in a

retrieval system, or transmitted in any form or by any means (electronic, mechan-
ical, photocopying, recording, or otherwise) without the prior written permission
of the publisher.

Published by Medical Physics Publishing

4513 Vernon Blvd., Madison, WI 53705-4964

Printed in the United States of America

 TABLE OF CONTENTS
Terms for Internal Dosimetry ............................................................................vii
I. RADIONUCLIDE THERAPY PRINCIPLES...........................................1
A. Objectives and Scope of the Primer ....................................................1
B. Dose Estimation Relationships............................................................1
C. Properties of the S Matrix....................................................................3
1. Non-Penetrating (Charged Particle) Radiation ...............................3
2. Penetrating (Photon) Radiation.......................................................3
D. Historical Background.........................................................................4
E. The Tracer Principle ............................................................................4
F. Treatment Planning for Radioimunotherapy .......................................5
G. Toxicity and Efficacy...........................................................................5
II. PROTOCOL DESIGN FOR RADIONUCLIDE THERAPY....................6
A. What Biological Data are Required ....................................................6
B. Design of Dose Estimation Protocol
for Radioimmunotherapy.....................................................................6
C. Dose Estimation in Humans from Murine
Biodistribution Data ............................................................................7
III. CLINICAL ASPECTS OF RADIOIMMUNOTHERAPY TRIALS .........9
A. Patient Eligibility.................................................................................9
B. Administration of Radiolabeled Antibody.........................................10
1. Intravenous Injections ...................................................................11
2. Intraperitoneal Injections ..............................................................11
3. Intra-Arterial Injections ................................................................13
4. Intralesional Injections..................................................................13
C. Patient Monitoring After Radiolabeled Antibody Administration.....14
D. Repeat Injections of the Same Antibody ...........................................15
E. Protein Preloading .............................................................................15
IV. ACQUISITION OF PHARMACOKINETIC PATIENT DATA..............16
A. Biological Data of Interest.................................................................16
1. Blood Activity...............................................................................16
2. Whole Body Retention..................................................................16
3. Other Normal Organ and Tumor Activity .....................................17
4. Minimal Levels of Uptake for Visualization.................................18
B. Uptake Measurements in Vivo ...........................................................18
1. Attenuation Corrections ................................................................18
2. Scatter Correction Techniques ......................................................21
3. Use of Geometric Data from Anatomic Imaging..........................22
4. Quantitative SPECT Imaging .......................................................24

iii
 C. Organ and Tumor Volumes................................................................25
1. Palpation .......................................................................................25
2. Imaging Methods to Determine Organ Mass................................25
V. DATA ANALYSIS METHODS...............................................................26
A. Activity and Percent Injected Activity (PIA) ....................................26
B. Justifications for Data Modeling .......................................................26
C. Types of Models ................................................................................27
1. Individual Organ Curve Fitting via
Multi-Exponential Functions ........................................................27
2. Compartmental Modeling of the Entire Physiological System.....28
D. Which Data to Analyze......................................................................28
E. Physical Decay as a Clearance Mechanism in the Model .................30
F. Integration of Area Under the Curve (AUC) .....................................31
1. Mathematical Form.......................................................................31
2. Multiple Exponential Functions....................................................31
3. Compartmental Model ..................................................................31
4. Variation of the Radionuclide .......................................................33
VI. DOSE CALCULATION METHODS AND PROGRAMS .....................33
A. Normal Organ Dose Estimates in Standard Man...............................33
1. MIRD 11 Pamphlet S Values ........................................................33
2. MIRDOSE2 and MIRDOSE3.......................................................34
3. Bone Marrow Dose Estimation.....................................................35
B. Patient-Specific Normal Organ Doses...............................................37
1. Target Organ Mass Correction ......................................................37
2. Convolution Dose Estimates.........................................................38
3. Voxel Source Kernel Method........................................................39
C. Tumor Dose Estimations in Radioimmunotherapy
and Radionuclide Therapy.................................................................39
1. Alpha and Auger Emissions..........................................................40
2. Beta Emissions..............................................................................40
a. Uniform Uptake.......................................................................40
b. Non-Uniform Uptake...............................................................40
3. The Tumor à Algorithm ...............................................................41
4. Tumor Mass Determination ..........................................................41
5. Computational Algorithms............................................................42
6. Photon Contributions to Tumor Dose Estimates...........................43
7. Clinical Computations (Beta and Photon
Contributions Combined)..............................................................44
8. Dose to Tumors That Are Not Imaged ..........................................44
9. Treatment Strategy and Results ....................................................45

iv
VII. THYROID CANCER THERAPY - THYROID
THERAPY WITH 131I..............................................................................46
VIII. BONE PAIN PALLIATION WITH RADIOPHARMACEUTICALS.....48
A. Clinical Situation ...............................................................................48
B. Radionuclides ....................................................................................48
C. Dose Estimation.................................................................................49
IX. RADIATION PROTECTION CONSIDERATIONS ..............................50
A. General Radiation Safety...................................................................50
B. Discharge of Patients with Radioactivity from Hospital ...................51
X. SUMMARY OF RADIOIMMUNOTHERAPY AND RADIONUCLIDE
THERAPY IN CLINICAL PRACTICE ..................................................53
A. Treatment Planning............................................................................53
B. Clinical RIT Results ..........................................................................53
C. Limitations of the RIT Method..........................................................54
D. Other Internal Emitter Therapies.......................................................54
REFERENCES ..................................................................................................55
APPENDIX A. Human Residence Time Estimates Made
With Murine Data ....................................................................................64
APPENDIX B. Tumor Dose Estimation...........................................................65
APPENDIX C. Organ Data Acquired and Dose Estimates ..............................66
APPENDIX D. Red Marrow Data and Dose Estimation..................................68

v
 LIST OF TABLES
Table 1. A Partial List of Radionuclides Used in Internal Emitter Therapy ....2
Table 2. Typical Data Acquisition Protocol (CC49 Antibody) ........................7
Table 3. Pre-Study Patient Evaluation for Radioimmunotherapy..................11
Table 4. Non-SPECT Methods to Determine Organ Uptake in Nuclear
Medicine ..........................................................................................23
Table 5. Radiation Safety Guidelines for Patients Receiving Radionuclide
Therapy ............................................................................................51
Table A1. Sample Human Residence Time Estimates Made
with Murine Data for111In-mT84.66.................................................64
Table C1. Partial List of Anterior and Posterior Counts Obtained
with a Whole Body Gamma Camera ...............................................66
Table C2. Fractional Injected Activities (FIAs) for 111In-cT84.66 ...................67
Table C3. Residence Times for 111In-cT84.66 and 90Y-cT84.66.......................67
Table C4. Estimated Absorbed Doses for cT84.66...........................................67

LIST OF FIGURES
Figure 1. Routes for administration of radiopharmaceuticals to a therapy
patient .............................................................................................12
Figure 2. Measurement of whole body retention of activity with probe
detector ...........................................................................................17
Figure 3. Acquisition of transmission and emission images. ACF is the
attenuation correction factor ...........................................................20
Figure 4. Photon spectrum from a gamma camera for 111In...........................22
Figure 5. Mammillary and catenary model subsystems.................................29
Figure 6. Five-compartment model for the intact antibody cT84.66 .............32
Figure 7. Outline of the general dose estimation process ..............................35
Figure 8. Schematic of red marrow dose estimation......................................37
Figure 9. Steps in tumor dose estimation. ACF refers to attenuation
correction factor .............................................................................45
Figure 10. Typical platelet data from four representative 131I-CC49 RIT
patients ................................................................................................46
Figure 11. Treatment strategy for RIT based on limited dose (200 rad)
to marrow ..............................................................................................47
Figure 12. Whole body and organ retention data for a 131I-LYM-1 patient .....52

vi
 Terms for Internal Dosimetry

Term Symbol Traditional Units SI Units

Administered radioactivity A0 µCi, mCi MBq, GBq

Anterior counts Nap number number

Attenuation correction factor ACF

Biological half time T1/2b h h

Body thickness T cm m

Cumulated activity à µCi-h number

Effective half time Teff h h

Geometric mean count GM number number

Organ or tumor thickness  cm m

Percent Injected Activity PIA ( not decay-corrected) number

Percent Injected Dose PID (decay-corrected) number

Physical half time T1/2p h h

Posterior counts Npa number number

Remainder body RB

Residence time τ h h

S values (target t, source s) St←s rad/µCi-h cGy/MBq-h

Sensitivity of Anger camera ε counts/µCi-s counts/MBq-s

Total (Whole) Body WB

vii
I. RADIONUCLIDE THERAPY PRINCIPLES

A. Objectives and Scope of the Primer

Recent trends in radionuclide therapy with radiolabeled antibodies and bone-
seeking radiopharmaceuticals have promoted renewed interest in providing
more accurate radiation absorbed dose estimates. A wide range of radionuclides
(Table 1), attached to a variety of targeting agents, have been used in these trials.
For many of the protocols, radiation absorbed dose estimates are relied on to
decide whether a patient should proceed to therapy and for prescribing the
amount of activity that should be administered. The motivation for this primer
is the provision of absorbed dose estimates for individual patients within the lim-
itations of the associated methodology. Although the MIRD Pamphlet 11 S
values (Snyder et al. 1975) have been used to provide the framework for radia-
tion absorbed dose estimates for patients undergoing diagnostic studies with
radionuclides, more accurate methods are required for radionuclide therapy pro-
cedures. Here, the organ volumes and geometry of standard man models may
not apply. More accurate dose estimates for each patient are also important for
establishing fundamental dose-response relationships for toxicity and efficacy
of Radioimmunotherapy (RIT).
This primer introduces various approaches to internal dose estimation in a ther-
apeutic context. The methodology may also be applied to diagnostic dose esti-
mates. The primer is intended for clinical physicists, radiation oncologists,
nuclear medicine physicians, dosimetrists, nuclear medicine technologists, and
other hospital personnel who are or will be involved in the process of acquiring
and processing dose estimation data.

B. Dose Estimation Relationships

A fundamental equality generally assumed in the estimation of organ doses is
the familiar equation:

 rad  ˜
Dt ← s [rad ] = St ← s   ⋅ As [ µCi − h] (1)
 µCi − h 

where: St←s is a rectangular matrix giving the dose to a target organ t per unit time-
activity in a source organ s (Loevinger and Berman 1976, Loevinger et al. 1991).
Conceptually, eq. (1) separates the analytic process into two segments. Given the
radionuclide, S refers purely to geometric factors. The cumulated activity ÃS is the
area under the curve of activity (AS) versus time. Traditional dimensions, rad/
(µCi-h) for S, are changed to cGy/MBq-h for SI units. Notice that D and à are

1
 Table 1. A Partial List of Radionuclides Used in Internal Emitter Therapy
Radionuclide Half Life Eβ max (MeV) Eγ (MeV)
32
P 14.3 d 1.70 None
64
Cu 12.9 h 0.57 (β− ); 0.66 (β+ ) 0.510 (38%)
67
Cu 61 h 0.57 0.180 (40%)
89
Sr 50.5 d 1.46
90
Y 64.3 h 2.3
117m
Sn 13.6 d 0.13; 0.16 0.158 (87%)
131
I 8.1 d 0.61 0.365 (81%)
153
Sm 1.9 d 0.81 0.103 (29%)
186
Re 3.8 d 1.07 0.137 (9%)
188
Re 17 h 2.12 0.16 (10%)
177
Lu 6.8 d 0.50 0.21 (6%)

Note: All beta emissions are for electrons except in the case of 64Cu, which emits both electrons and
positrons.

column vectors in this formula; their various elements refer to the source and target
organs for a specified radionuclide. If one divides by the injected activity A0, the
resultant quantity is termed the residence time τ. Thus, dose per unit injected activ-
ity is the product of S and τ; this result is often used in reporting estimated doses.
By changing the spatial perspective, elements of the S matrix may be made to rep-
resent other geometric positions such as voxels within an organ or tumor. In such
cases the à vector will represent integrated activities in organ voxels. A short glos-
sary of terms used in internal emitter absorbed dose estimation is included at the
front of this document.
Although an integration over time has been performed in the estimation of dose,
it is important to understand that dose rate estimates may also be made using the
time-derivative of eq. (1). In that case, the left hand side of eq. (1) will be a set of
organ dose rates given as a function of time post-injection of the radiolabeled
agent. Such exposure rates, if referring to a organ at or very near the body sur-
face such as the liver, could be roughly measured using external probes.
We should point out that the traditional calculation using eq. (1) generates the
mean dose to the target organ. The left-hand side may be shown to be the ratio
of total energy absorbed divided by the mass of the target. This need not be the
case in general. Statistical distributions of dose, in the form of dose-volume his-
tograms, have been produced for internal beta emitters (Liu et al. 1998). Such
results are analogous to histograms found in external beam therapy. Although not
yet used extensively in treatment planning, such distributions may eventually
prove to be important in both normal organs and tumors.

2
C. Properties of the S Matrix

1. Non-Penetrating (Charged Particle) Radiation

One may logically separate S values into Snp and Sp parts; these refer to so-called
non-penetrating (np) and penetrating (p) radiation given off by the radionuclide of
interest. Typically, the user assigns charged particle radiation to the former cate-
gory and gamma and x-rays to the latter. Some emitters, such as 90Y, are purely of
the charged-particle type; these are of considerable interest in RIT due to the reduc-
tion in ambient radiation levels during therapy procedures.
If we assume that np radiation cannot escape from the source (and therefore
target) volume, individual terms of Snp can be computed using the relationship:

Snp = (1 / m) ∗ 2.13 ∗ ∑ fE
i = 0, n i i
(2)

where fi refers to the probability of emission of np radiation of energy Ei . A factor

of 2.13 is used when energy is in MeV and the organ mass term (m) in eq. (2) is
in grams (g). In this overview, we have also explicitly neglected Bremsstrahlung
radiation. Thus, Snp will only contain diagonal terms. The investigator can calcu-
late this Snp matrix directly using eq. (2) given emission information and organ or
lesion mass. We will describe below how to augment the computation if edge
effects and/or Bremsstrahlung are to be considered. If these two corrections are
included, the situation, by definition, becomes one in which the radiation is pen-
etrating. Thus, use of the term non-penetrating is only an approximation.

2. Penetrating (Photon) Radiation

For photons, the value of Sp includes an additional term for the absorbed frac-
tion (ϕ) of the ionizing radiation in the target organ of mass m:

S p = (1 / m) ∗ 2.13 ∗ ∑ f Eϕ
i = 0, n i i i
(3)

Computations of Sp can be simplified if we consider an infinite homogeneous

medium (Loevinger et al. 1991). In this case, the reciprocity theorem may be
invoked to prove that Sp is a symmetric matrix (Loevinger et al. 1991). For anthro-
pomorphic phantoms actually used in computing the penetrating part of S, this
simplification can be shown to be only approximately true. The reader may test
this assertion by inspecting off-diagonal elements of S for a photon-emitting
radionuclide of interest (Snyder et al. 1975).
Tumor dose estimates for penetrating radiation represent a challenge since these
tissues are not included in the standard set of source and target organs available to

3
the user. Macey has indicated a graphical method for estimating such S values if
one knows the tumor mass and radionuclide (Meredith et al. 1993). A computer
program for gamma dose estimates to spherical lesions has been developed by
Johnson (1988). While such contributions to tumor dose may be small compared
to the particulate component, it is important to find their magnitude in a given
patient case. This topic is discussed in section VI.C.6.

D. Historical Background
Perhaps because radionuclide therapy with 131I for thyroid cancer has been so suc-
cessful and safe, there has been little historical incentive to provide a rigorous treat-
ment planning approach for clinical radionuclide therapy practice (Hurley and
Becker 1983). In addition, methods and technology, particularly computers, required
to provide more accurate estimates were not available in Nuclear Medicine until the
mid-1970s. The thyroid cancer patient has usually been treated on the basis of an
empirical approach. As is the case for many RIT protocols, the treatment plan
relies on an initial imaging study with a tracer amount of 131I administered to
search for remnant thyroid tissue and metastatic sites in the whole body. A more
complete discussion of radioiodine therapy of the malignant thyroid is presented
in section VII.
Over the past two decades, significant progress in medical technology has been
made and the use of Computed Tomography (CT), Magnetic Resonance Imaging
(MRI), and Single Photon Emission Computed Tomography (SPECT) imaging
can yield most of the data required to provide more accurate patient-specific radi-
ation absorbed dose estimates. Anatomic size, location, and volumes of organs and
tumors can be provided non-invasively from CT, MRI, ultrasound, Positron
Emission Tomography (PET), and SPECT images (Leichner et al. 1981). This
should allow every patient to receive a customized therapy plan involving the
administration of a specified amount of activity (Macey et al. 1991).

E. The Tracer Principle

In the present context, the tracer principle states that a relatively small (tracer)
amount of a radiopharmaceutical can be used to predict the spatial and temporal
distribution of a subsequent larger amount of the same radiopharmaceutical in the
same patient. This principle has been used successfully in radionuclide therapy for
thyroid cancer, although the uptake and distribution of 131I in the thyroid patient
has not been rigorously followed. The concept relies on the assumption that the
response of the biological system is not perturbed by the diagnostic study.
tumor / organ tumor organ
/
 rad  = 
rad 
 mCi  (4)
therapy  mCi  diagnostic

4
Based on the tracer principle, the radiation absorbed dose/administered radioac-
tivity to tumors/organs from a therapy procedure should be equivalent to the values
calculated from a prior diagnostic procedure as given by eq. (4). Information from
the tracer study is used to calculate the amount of activity required for a therapy
procedure. A prescribed radiation absorbed dose to a tumor or dose-limiting organ
usually dictates the amount of activity that can be administered by:

 
 
 Dprescribed therapy dose 
A0 [therapy mCi] = (5)
  rad  
   
  mCi  diagnostic  tumor
A different definition of tracer is often used in Nuclear Medicine. In that context,
the user assumes that a radiolabeled form of a molecule has the same biodistrib-
ution as the native or unlabeled molecule. This may or may not be true in the case
of labeled antibodies due to changes in molecular weight and loss of label (par-
ticularly iodine) in vivo. Thus, our tracer concept is a more restricted one.

F. Treatment Planning for Radioimmunotherapy

Unlike 131I therapy for the thyroid cancer patient, RIT is more toxic because a
large fraction of the administered activity remains in various organs and spaces in
the body. Although the fraction of the administered activity that localizes in tumor
may be comparable to the concentrations of 131I in many thyroid cancer patients
after thyroidectomy, the remainder of the activity administered is not rapidly elim-
inated like 131I. The design of a specific treatment plan for every RIT patient is
important both for the safety of the patient as well as the protection of the staff and
visitors. Radiation absorbed doses could in principle be measured with miniature
dosimeters implanted in selected organ and tumor sites. However in practice this
is impossible and therefore most dosimetry methods depend on acquiring in vivo
pharmacokinetic data and deriving radiation absorbed dose estimates numerically.
Dose estimation plays the most important role in treatment planning. Using the
tracer principle described above, the physicist will estimate absorbed doses to the
important normal organs and to tumor site(s). In the latter case, the site must usually
be visualized in the nuclear image in order that the dose be estimated. If this is not the
case, only normal organ dose limitations will normally enter into the planning phase.

G. Toxicity and Efficacy

Today most clinical research protocols are designed to test and validate a
hypothesis. The practical steps in a clinical trial of a new drug or procedure involve
Phase I, II, and III trials. The Phase I trial is designed to establish the toxicity limit

5
defined as the maximum tolerated dose (MTD) in a small group of patients. Phase
II trials are designed to test the efficacy of the same procedure in a small group of
patients at a dose level somewhat lower than the MTD. Phase III trials usually
involve multi-center studies with patients enrolled at various institutions and dif-
ferent investigators.
Although these steps are required for clinical trials with chemotherapy agents,
radionuclide therapy trials can in principle be designed to combine the Phase I and
Phase II steps. The radionuclide approach is especially valuable since it allows the
investigator to predict to some degree the toxicity and efficacy of a radionuclide
therapy procedure. For example, the biodistribution of the radiopharmaceutical in
organs/tumors in the body can be obtained from serial radionuclide images
obtained at various times after administration. Localization of the radiopharma-
ceuticals in specific organ or tumor sites can be used to determine whether the
patient should proceed to the therapy step. If there is no tumor localization of the
radiopharmaceuticals detected in the diagnostic images, the patient is less likely
to benefit from the subsequent therapy and consequently should not have to be
subjected to the unnecessary risk. All of these data are unique to nuclear tech-
niques and are not found in the classical chemotherapy trial.

II. PROTOCOL DESIGN FOR RADIONUCLIDE

THERAPY

A. What Biological Data are Required

Availability of good biodistribution data for a specific radionuclide therapy pro-
cedure allows the researcher to establish fundamental dose-response relationships.
Simple questions such as what data are required to provide dose estimates can usu-
ally be answered from prior biokinetic information derived from animal—usually
mouse—studies. Frequency of sampling blood, whole body, and normal organs
can be estimated from the animal data. However, judicious design and collection
of data are fundamental to providing reliable absorbed dose estimates in man.
From the patient’s standpoint, one should take into account that all data acquisi-
tion involves inconveniencing the patient and so a compromise must be made
between limiting the amount of biological data collected and reliability of esti-
mates that are calculated from that data.

B. Design of Dose Estimation Protocol

for Radioimmunotherapy
Dose estimation protocols involve serial collection of whole body, blood,
normal organ, and tumor data. The radiopharmaceutical is constantly moving from
blood to various organs and compartments in the body. Schemes to track the

6
 Table 2. Typical Data Acquisition Protocol Designed to Acquire Biokinetic
Data Required to Provide Absorbed Dose Estimates for RIT Patients
Receiving I-131 CC49
Biokinetic Day 0, (End of
Data injection is
Acquired t = 0) Day 1 Day 2 Day 3 Day 4

5 ml Whole 15', 30', 1 h, 6 h 24 h 48 h 72 h 96 h

Blood
Cumulated 0–6 h, 24–48 h 48–72 h 72–96 h 96–120 h
Urine 6–24 h
Tumor/Organ 1h 24 h 48 h 72 h 96 h
Imaging
Whole Body 1h, 6 h 24 h 48 h 72 h 96 h
Probe
Biopsy of 24 h 48 h
Tumor

changing pattern of uptake and clearance in sufficient detail must be devised to

provide good dose estimates. Table 2 illustrates a typical design of a dose estima-
tion protocol for collection of biokinetic data from breast cancer patients enrolled
in a RIT study with 131I CC49. Some of these data can only be acquired via an in
vivo imaging study; for example, by using a gamma-emitting radionuclide labeled
to the CC49 antibody, such as 99mTc or 111In or, alternatively, a low dose of 131I
CC49 prior to therapy. We will describe below a method of using the diagnostic
study with one radionuclide to predict the therapy outcome with a second radionu-
clide attached to the same pharmaceutical.

C. Dose Estimation in Humans from Murine

Biodistribution Data
We first consider a sample computation of human dose estimates based on
murine (obtained from laboratory mice or rats) data. Most of these analyses involve
immune-deficient (nude) mice in the case of RIT. This type of analysis is often used
in the Investigational New Drug Application (IND) application to the Food and
Drug Administration (FDA). Unless such estimates can be supplied, approval for
human use of the radiopharmaceuticals will, in general, not be given.
The murine model used in such biodistribution analyses should be of normal,
tumor-free animals. Normalized counts are obtained from NaI(Tl) counter meas-
urements using an energy window appropriate for the gamma energy of the
radionuclide. Data taken would include percent injected dose (PID) as well as

7
corrected uptake in percent injected dose per gram of the tissue (PID/g). These
measurements are tabulated as decay-corrected biodistributions in various organs
and at a number of time points after injection of the radiolabeled material. Whole
body accumulation may additionally be measured using a single or paired probe
system set at a fixed distance from a constrained animal. As described below,
a mathematical picture of the biodistributions is helpful in determining the total area
under the curve of each source organ. In this example, we will consider using a single
exponential model to represent organ data. The resultant biological half time (T1/2b)
is combined with the physical half life (T1/2p) to obtain the effective half time Teff :

T1 / 2 p ∗ T1 / 2 b
Teff = (6)
T1 / 2 p + T1 / 2 b

The residence time τ for a tumor/organ in the mouse that clears with an effective
half time Teff is given by:

τ tumor / organ = 1.44 ∗ Teff ∗ PIDt = 0 / 100 (7)

where the indicated PID is the percent injected dose at end of injection for a mono-
exponential compartment and Teff is the effective half time in hours. This area
under the curve has the injected activity divided out so as to be in pure time units;
the resultant value is the residence time (τ) of the source organ. A set of such times
is the residence time vector for murine estimates. In the case of animal biodistri-
bution data, there may be 10 or more individual organs or tissue compartments to
consider for this vector.
To estimate the residence time in a human from the biokinetic results in the
mouse, we use a mass correction factor to account for the different ratios of organ
to total body weights in the mouse and in man. To first order, we assume that the
uptake in a given organ, in PID, depends only on the ratio of organ mass divided
by the total body mass. To correct for differences in this apportionment ratio
between murine and human species, we can use the following:

 organ mass 
 total body mass 
τ man   man
CR = = (8)
τ mouse  organ mass 
 total body mass 
  mouse

This process is essentially an organ correction factor (CR) for differences in rela-
tive perfusion between the mouse (or other animal) model and the human subject.
Usually the biokinetic data from the mouse are provided as PID/g, which simpli-
fies this conversion for dose estimation since the murine organ mass is included in
the data set. An example of this type of calculation is given in Appendix A. These

8
correction factors are often sizeable and should be applied when using animal data
to predict human results.

III. CLINICAL ASPECTS OF RADIOIMMUNOTHERAPY

TRIALS

A. Patient Eligibility
Radioimmunotherapy trials are usually fairly demanding for the patient. A
number of additional constraints are placed on the individual and his/her tumor
type beyond those found in radionuclide therapy. In general, RIT patients must
have a good performance status, defined as a Karnofsky scale above 60% and no
other severe systemic disease (Press et al. 1989, Breitz et al. 1992). Only patients
with tumor types known to be reactive with the antibody are to be considered. This
may be done with peroxidase staining of tumor tissue samples obtained from the
patient. Alternatively, an in vivo imaging study may be performed with a gamma-
emitting radionuclide label to demonstrate uptake in one or more known lesion
sites (Breitz et al. 1993, Eary et al. 1990). Additionally, if the antigen is secreted
into the blood, such as carcinoembryonic antigen (CEA), a blood assay may pro-
vide the needed tumor specificity. In the latter case, the test is done with less cost
than that requiring tumor samples. In the event that the patient’s tumor is antigen-
positive, a life expectancy of at least 2 months is necessary so that adequate mon-
itoring of normal organ function and tumor response can be assessed during the
radiotherapy protocol.
On entry into the study, good performance status is important because the
monitoring of the patient for the dose estimation data is demanding. Generally
the protocol will require several sessions of nuclear imaging, of one to two hours
each, depending on the gamma camera system. SPECT studies require that the
patient be able to lie flat with arms raised for a sufficient length of time (perhaps
20 minutes) for an adequate study to be acquired. This may be at one time point
only for a qualitative assessment of tumor uptake. Some investigators will use
SPECT for quantitative estimation of activity in tumor masses and/or normal
organs at one or more times. For older patients or patients with pulmonary com-
plications, this may be an unrealistic expectation.
Good performance status may also be essential because of the high levels of
internally administered activity. Slow excretion of the radiolabeled antibody
metabolites pose radiation hazards similar to those from patients receiving high
dose 131I for treatment of thyroid disease. Exposure to medical personnel will
depend on the radionuclide administered. To reduce such exposure hospitalized
patients should be able to care for themselves with minimal nursing care required.
Patients with lymphoma receiving high dose 131I radiolabeled antibodies have even
been trained to collect their own blood samples (Press et al. 1989). Patients should

9
be capable of collecting their urine and measuring the volume themselves if this is
required in the protocol. Because of radiation hazards, incontinent patients should
not be studied under these conditions.
Baseline blood tests confirming adequacy of bone marrow, renal, and hepatic
function must be obtained. The patient must not have received chemotherapy or
radiation therapy for at least 4 weeks prior to radiolabeled antibody treatment to
allow full recovery from side effects of those treatment modalities. Tumor size must
be documented by conventional diagnostic techniques (physical examination,
planar film, CT image, MRI scan or ultrasound) so that response can be monitored.
At present, trials in patients with unmeasurable disease to detect the value of RIT
as adjunctive therapy are infrequent, although at least one colorectal cancer adju-
vant trial has reported beneficial results using unlabeled antibody (Riethmuller
et al. 1994).
If patients have had prior exposure to a murine or other antibody, circulating
titer against that antibody must be measured. Presence of such a human antibody
to the clinical agent may exclude the patient from participating in the study.
Inclusion of such patients often results in the resultant radioactive antibody-
antigen complexes going directly from the blood to the liver with little activity
seen at the tumor sites.
When the patient has met all the eligibility requirements and signed informed
consent, additional preparation may be necessary before receiving the labeled anti-
body. For example, patients receiving 131I-labeled antibodies must, in addition,
have their thyroid gland appropriately blocked using Lugol’s Solution. Patients
also should be hydrated to reduce kidney and bladder effects due to radiation. If
gastrointestinal excretion is significant, cathartics should be administered to
reduce gastrointestinal mucosa exposure during fecal excretion of the iodine label.
A summary of the required pre-study clinical evaluation is shown in Table 3.
Radiolabeled antibody must not be administered until the quality assurance
studies have been reported. These tests must show adequate assessment of purity,
potency, and safety. These could include an endotoxin assay, Instant Thin Layer
Chromatography, and High Performance Liquid Chromatography (HPLC).
Emergency supplies must be readily available in case of an allergic reaction.
These would include diphenhydramine, epinephrine, and steroids. Patients have
vital signs monitored immediately prior to injection and regularly thereafter.

B. Administration of Radiolabeled Antibody

Radiation precautions must be observed for the infusion of the radiolabeled anti-
body. Gloves and protective glasses must be worn while handling the radiolabeled
antibody. When the radioactivity is to be infused slowly, the syringe is placed in a
Lucite™ and lead-covered infusion pump (Williams et al. 1995a). When a bolus
injection is required, a shielded syringe is used. All couplings and connections in the

10
 Table 3. Pre-Study Patient Evaluation for Radioimmunotherapy

• History and physical exam

• Histologically confirmed diagnosis of antigen presence on tumor by histology or
imaging study. Blood assay may be used if antigen is shed.
• Chest X-Ray, Ultrasound, CT scan or MRI scan for lesion size
• Biochemical Profiles;
Complete blood count (CBC)
Chemistry Screen
Urine Analysis
Prothrombin Time (PT)
• Pregnancy test if applicable
• Signed informed consent
• Baseline Human Anti-Mouse Antibody (HAMA) or other appropriate human
antibody

line between the source and the patient must be tested prior to beginning the injec-
tion procedure. Medical personnel must take appropriate precautions in case of con-
tamination or spills. Absorbent pads are placed under the lines and on the floor below
the bed in case of spills. A more complete discussion of radiation safety aspects of
RIT is included in section IX. Many of the above restrictions apply to the general
case of radionuclide therapy.
Success of this treatment modality depends upon sufficient radioactivity being
carried by the antibody to the tumor site(s). Several routes of administration have
been used. Figure 1 contains a summary of these methods.

1. Intravenous Injections
This is the most commonly used procedure for administration of radionuclide
therapy to patients. Two intravenous lines are established, one for administration of
the radiolabel and one for post-injection blood sampling. The latter may be retained
during the course of the therapy to provide the required samples as described below
in paragraph 4. Duration of injection will vary from several minutes to several hours
depending upon adverse reactions and the amount of protein used in the therapy.

2. Intraperitoneal Injections
Additional factors for intraperitoneal RIT studies must be considered (Jacobs
et al. 1993). Radiological examination is not sensitive enough to define the extent
of the disease. An open laparoscopy or a laparotomy performed within one month
of the study will give the most reliable data, although even this may not provide

11
Figure 1. Routes for administration of radiopharmaceuticals to a therapy patient.

12
complete information. Serum markers may also be used as a guide as to the extent
of disease and response.
Prior to administration of a large dose of radioactivity, the peritoneal cavity
access must be evaluated to be sure that the radiolabel will distribute throughout
the volume. A 99mTc-sulfur colloid study may be performed to assess this; activ-
ity must be seen in the pelvis and paracolic gutters. When severe ascites is pres-
ent, as much ascites fluid as possible must be drained prior to the infusion. The
radiolabeled antibody may be administered through a porta-cath or Tenckhoff
catheter. The volume of fluid administered with the antibody should be sufficient
to allow distribution throughout the peritoneal cavity. This volume can range
from 300 ml to 1000 ml of normal saline or Ringer’s lactate. The patient is
instructed to roll 360 degrees three times prior to the gamma camera imaging.
Later, they are encouraged to move about freely to promote uniform distribution.
To estimate clearance of radioactivity from the peritoneal cavity, and to estimate
absorbed dose to the peritoneal surface, aspiration of peritoneal fluid samples is
desirable.

3. Intra-Arterial Injections
Known metastases must be confined to one region that can be perfused by intra-
arterial injection. On the day of the radiolabeled antibody injection, a catheter must
be placed under fluoroscopy into the artery perfusing the area of interest. The
catheter is removed following the injection. Additional monitoring is required for
adverse events resulting from catheter placement.

4. Intralesional Injections
This is a more invasive and time-consuming technique and can require hospi-
talization for several weeks. For example, in the treatment of gliomas, a catheter
is placed in the lesion at craniotomy, or by stereotaxy some days later. This
catheter is placed in the residual cavity after as much tumor as possible has been
resected. Removable or indwelling catheters may be used. A reservoir may be
connected to the catheter and placed under the skin for repeated injections (Riva
et al. 1994).
We should point out that this approach has been followed in the case of radionu-
clide therapy of pancreatic carcinomas. Here, a local anesthetic is used and a 22-
gauge needle inserted into the lesion under CT guidance. Aggregated albumin
(MAA) is given initially with a follow-on dose of 32P to yield the desired absorbed
radiation dose estimates (Order et al. 1994). Imaging of the lesion may be per-
formed using the Bremsstrahlung radiation from the single concentrated source
of radiophosphorous.

13
C. Patient Monitoring After Radiolabeled Antibody
Administration
Patient monitoring must be done under the supervision of a physician. Patients
are observed for acute allergic reactions such as flushing and hives. Adverse reac-
tions must be treated and reported as required in the protocol. An assessment for
allergy may be done using a small test dose of the protein to be injected. A 5 micro-
gram amount can be injected subdermally and the site observed for 25 minutes
to determine any sensitivity. If swelling and other reactions are seen, injection of
the larger imaging or therapy dose (typically 5 milligrams of protein) can then
be prevented. Vital signs must be monitored for at least one hour post antibody
administration.
To obtain data for dosimetry estimation, imaging should begin immediately fol-
lowing injection and prior to voiding to assess the baseline activity. When high
dose 131I is administered, imaging may begin only after the activity level has
decreased to values on the order of 30 mCi or less. This restriction is based on cur-
rent state and federal regulations that are designed to reduce radiation exposure to
hospital personnel and also because the Anger camera is only designed to detect
relatively low amounts of radioactivity in patients. High count rates may lead to
dead-time effects including spatial location errors and count losses so as to make
quantitative imaging difficult.
Biochemical and hematological parameters are monitored regularly for 3 months
or longer following the study. Typically, biochemical changes are not significant.
With increased levels of radioactivity administered, bone marrow suppression is
observed as a reduction in the number of white blood cell and/or platelet counts from
baseline values. With increased radioactivity, these counts will drop earlier and to a
lower level. The platelet count decreases approximately a week earlier than the white
cell count—usually at week four.
Levels of patient antibody to the injected protein are assessed for up to 6 months.
In the case of a murine antibody, the human response is termed a Human Anti-
Mouse Antibody or HAMA. Similar acronyms such as HACA and HAHA refer
to human anti-chimeric and human anti-human antibodies respectively. It should
be noted that the patient may also develop antibodies against the chelator used to
attach radiometal to an injected protein. In any event, these titers should be deter-
mined weekly for 6 weeks and then monthly. Presence of elevated levels of human
antibody may preclude repeat treatments.
At a specified time, usually 6 to 8 weeks, a follow-up radiological exam is per-
formed to assess tumor response. Generally, this would be a CT or MRI image so
that the tumor volume(s) post-therapy may be compared directly with the baseline
study. Nuclear bone scan images may be used to evaluate the progress of therapy
on bone metastatic sites. Such evaluations would also be appropriate for general
radionuclide therapy treatments.

14
D. Repeat Injections of the Same Antibody
Prior to repeat administration of antibody, serum human antibody levels must
be measured. In patients with lymphoma, cycles of radioimmunotherapy can often
be administered. In patients with other tumors who have received murine mono-
clonal antibodies, elevated HAMA levels generally preclude a subsequent admin-
istration. In Phase I studies, human anti-antibody levels must be monitored until
the characteristics of the patient response are known.

E. Protein Preloading
Early studies with radiolabeled antibodies established that a minimum mass of
a foreign protein was required for administration into the blood. This followed
from the fact that various receptor sites in the liver, lung, and other normal tissues
must be saturated before the concentration of the radiopharmaceutical in the blood
was high enough so that antibody would localize in the tumor. For lymphoma, a
few mg of protein were found to be adequate. Literature studies have pointed out
that there may (Patt et al. 1988) or may not be (Wong et al. 1998) an advantage in
increasing the amount of injected protein above this minimal level. Some proto-
cols, because of significant levels of circulating antigen, require relatively large
amounts of unlabeled protein prior to the injection of the labeled antibody (Breitz
et al. 1992, DeNardo et al. 1990).
It should be mentioned that increasing the injected amount of protein has two
possible negative results. First, the likelihood of patient antibody response will
increase with the amount of injected foreign antibody. In addition, the cost of large
amounts of engineered protein may preclude such procedures. For example, in
mice, about 700 µg of antibody were required to significantly reduce an anti-CEA
antibody’s hepatic uptake (Beatty et al. 1989). Scaling these results to man implies
clinical injected protein doses approaching one gram. No one has yet attempted
such injections in solid tumor therapy although values on the order of several hun-
dred milligrams have been required in treating hematological malignancies due to
circulating antigen in the blood or high levels of B-cell antigens on lymphoid
tissue (DeNardo et al. 1990).
Most RIT patients receive their radiopharmaceutical as an intravenous injection
that is either infused over a planned time or injected as a bolus. Localization of the
radiolabeled antibody in the target tumor tissue and normal organs is a continuous
process of uptake and clearance and the pattern observed is the net result of inflow
and outflow from sites that are visualized on a radionuclide image. Intravenous and
intra-arterial routes depend on the assumption that the radiopharmaceutical can
leave the blood pool and enter the various volumes or partitions in a tumor. In the
case of intraperitoneal injection, radioactivity enters the lesion sites directly from
the peritoneal fluid as well as from circulation. In the latter case, there is a signif-
icant time delay before the material enters into the blood circulation.

15
IV. ACQUISITION OF PHARMACOKINETIC
PATIENT DATA

A. Biological Data of Interest

The activity and concentration of the radiopharmaceutical in all major source
organs and spaces in the body must be determined for the dose calculation step.
Usually the activities in whole blood and whole body are simple to acquire while
tumor and organ uptake measurements require quantitative organ imaging proce-
dures.

1. Blood Activity
Changing activities in whole blood and plasma are determined from serial
blood samples acquired at various times after administration of the radiopharma-
ceutical. Usually 3 to 5 ml of blood are collected at 1, 4, 24, 48, and 72 hours from
a vein in the limb opposite to that in which the radiopharmaceutical was admin-
istered. Other schedules may be used with animal data providing some predictions
as to human blood kinetics. In addition, the pharmacokinetic modeling program
(see below) can be used to predict optimal times for blood sampling. These times,
however, may not be convenient for patient or clinical logistics.

2. Whole Body Retention

This information can be measured most conveniently with a probe placed at 2 to
6 meters from the surface of the patient as shown in Figure 2. Here, one obtains a
reading for the count rate from the patient about 10 to 20 minutes after the activ-
ity is administered and prior to any excretion via urine or feces. This initial value
is used to normalize all subsequent measurements made in the same geometry. The
assumption of this approach is that the sensitivity of the probe or dosimeter used
to measure the various readings is reasonably independent of where the activity
is located within the patient.
Because of the redistribution of the activity over time, it is best that a pulse
height analyzer be used in conjunction with the probe system. Exclusion of scat-
tered radiation from the patient is the primary motivation for this strategy. Thus,
as we describe below for gamma cameras, an energy window should be set up over
the photoelectric peak of the detected gamma radiation.
A second method for monitoring whole body retention is a sequence of whole
body images taken with a gamma camera in scanning mode. Generally, these
images will include regions from the top of the head to the knees with the arms
being visible at the sides of smaller patients. As in the probe case, positioning of
the patient for repeat scans will require a localization method. One technique is
the use of wall-mounted lasers to reposition the patient serially. This method has

16
Figure 2. Measurement of whole body retention of activity with probe detector.

the advantage that distributions within patient organs can be acquired at the same
time as the whole body image(s).

3. Other Normal Organ and Tumor Activity

Activity in normal organs and tumor is most conveniently determined from a
quantitative Anger camera image. Special acquisition protocols must be used to
calculate activity from the number of counts detected in a specified organ or tumor
site that is visualized in the radionuclide image. Magnitude of activity in a given
organ is often referred to as uptake and is usually measured in percent injected
activity (PIA).
Uptake in a tumor or organ is most readily estimated if the number of counts/
pixel (count density) detected from that region in the Anger camera image exceeds
the count density in the immediate vicinity. A tumor near high count density
regions such as the liver and cardiac volumes requires a much larger ratio of uptake

17
compared with tumors in the limbs and skull. Most region of interest (ROI) meth-
ods used for scoring uptake from these images rely on drawing a background region
in the immediate vicinity of an uptake site or a contralateral region of the body if
the background surrounding the uptake region is not uniform. These background
subtraction techniques are designed to remove the contribution of counts to a tumor/
organ site from radioactivity in overlying and underlying tissue.

4. Minimal Levels of Uptake for Visualization

Simple statistical analysis methods can be used to determine the minimum
number of counts that are required to visualize the typical level of uptake of radi-
olabeled antibodies in tumors. Usually the concentration in a tumor must be sev-
eral times greater than the concentration in the immediate vicinity to make the
tumor visible in a planar Anger camera image. Biokinetic data indicate the typ-
ical range for human tumors is about 0.001 to 0.02 PID/g (20% ID/kg). For a
20 g tumor in a patient, these levels of uptake would be translated to 0.020 to
0.40 PID. If the patient receives 10 mCi 131I labeled antibody for the diagnos-
tic study, the 20 g tumor would be expected to have an uptake of 2 to 40 µCi
(1.5 MBq). Whether this range of uptake is visualized in a whole body scan or
a planar image depends upon multiple factors such as the background count den-
sity, physical decay, sensitivity of the camera-collimator system, and observer
training.
Planar image acquisition times of 10 to 20 minutes are required to visualize
tumor sites in RIT patients. These long times and the requirement for repeated
images make it important to have a dual-head Anger camera with a large field of
view. Generally, a whole body scan image is acquired to characterize the distri-
bution in the whole body followed by a series of conjugate image sets that include
expanded views of chest, abdomen, and pelvis. If available, laser methods should
be used to register the patient for a set of serial images.

B. Uptake Measurements in Vivo

1. Attenuation Corrections
All uptake measurements depend upon correction of recorded data for back-
ground and attenuation in the patient. Several strategies exist for these corrections.
In the single projection image method, which can be used for superficial uptake
sites, the attenuation correction factor can be set to unity. Here the uptake is given
simply by:

Nt [counts]
Atumor / organ [Ci] = (9)
ε [counts / Ci s] ∗ ∆t[s]

18
where ε is the efficiency of the camera and collimator system and Nt is the number
of counts recorded in time ∆t.
For the conjugate view approach, the user must obtain, essentially simultane-
ously, two diametrically opposed views of the same object (Thomas et al. 1976,
Sorenson 1974). Figure 3 contains a geometric description of the process. The
number of counts detected from the anterior (Nap) and posterior (Npa ) projection
images of an organ of thickness  cm in the body, with overlying and underlying
tissue thicknesses a and b cm is:

Ae − µa Ae − µb
[ ] [ ]
N ap N pa
=ε∗ 1 − e−µ l ; =ε∗ 1 − e−µ l (10)
∆t µl ∆t µl

The geometric mean of these values becomes:

sinh( µl / 2)
GM = N ap N pa = ε ∗ ∆t ∗ Ae − µ ( a + b + l ) / 2 (11)
( µ l / 2)
For organ thicknesses that are sufficiently small, the source thickness factor
(µ/2)/sinh(µ/2) can be set to one. This simplification is often the basis for using
the geometric mean of counts to estimate organ and tumor uptake from conjugate
view Anger camera images. The geometric mean of anterior and posterior image
counts can provide an estimate for tumor/organ uptake very simply from a meas-
urement of the total thickness (a + b + ) of the body region in which the organ is
imaged. We refer to this as the total thickness T. The linear attenuation coefficient
is defined as µ in the above equalities.
The attenuation factor in the above equation for the GM can also be determined
from a transmission source imaging technique. If the transmission and emission
sources have similar µ values, the activity in an organ can be calculated using:

N ap N pa N ap N pa N (0)
Atumor / organ [ µCi] = e + µ ( a + b + l) / 2 = (12)
ε ∗ ∆t ε ∗ ∆t N (T )

where N(T ) is the number of counts with the patient between source and gamma
camera and N(0) is the number of counts without the patient in place. Source
thickness factor has been set to unity in eq. (12). The transmission emission
approach can be invoked to estimate uptake of most gamma-emitting radionu-
clides currently used in nuclear medicine to an accuracy of about 10% to 20%
(Macey and Marshall 1982). If the radionuclide used to determine the attenuation
correction factor emits a different energy than the therapy radionuclide (e.g., if
57
Co is used for the transmission measurement in a patient who received 131I for
RIT), special correction factors are required to account for the differences in tissue
linear attenuation coefficient values.

19
Figure 3. Acquisition of transmission and emission images. ACF is the attenuation
correction factor.

20
2. Scatter Correction Techniques
In the above analyses, the photons were assumed to travel in such a manner
that scattered radiation did not enter the camera or other detector system. Scatter
correction is usually necessary because the recorded counts in a given camera
pixel will include primary as well as scattered photons.
Typical scatter corrections have included using adjacent areas on the planar image
to subtract such counts or the use of variable windowing on the gamma camera photon
spectra to exclude lower energy photons from the imaging. For a single photon emit-
ter such as 99mTc, this would entail setting two energy windows: one over the direct
photons at 140 keV and one over the scattered radiation at 100 to 120 keV. A sub-
traction of the latter from the former gives the net direct (unscattered) counts for
that region of the patient. This is discussed in greater detail in the section on quan-
titative SPECT imaging. Use of a triple-energy window (TEW) with satellite win-
dows set both above and below the photopeak has been described (Ichihara et al.
1993). A weighted average of the count rates in these two satellites is then subtracted
from the count rate in the center of the spectrum to produce a primary count rate.
If the radionuclide used in the imaging procedure emits more than one photon,
these strategies for spectroscopic scatter correction become more complicated. For
example with 111In, the dual-window program would require a total of four energy
windows to be set. Yet the lowest of these, at 140 to 150 keV, would contain both
scattered photons from the higher energy peak at 247 keV as well as scattered pho-
tons from the lower peak at 174 keV.
It is important to understand that because a pulse is recorded with a lower
energy than the photoelectric peak in the detector does not necessarily indicate
scattering. This ambiguity occurs due to the finite energy resolution of the NaI(Tl)
crystal in the gamma camera. Direct photon energies appear to spread around a
mean value with a full width at half maximum (FWHM) of approximately 10% or
higher. Thus, correction using such secondary windows is uncertain and necessar-
ily leads to loss of counts attributable to unscattered (direct) photons. Scattering
within the crystal of the camera is also ignored in this analysis. A sample gamma
camera photon spectrum for 111In is shown in Figure 4.
Comparisons of the various types of spectroscopic scatter-correction tech-
niques have shown (Ljungberg et al. 1994, Buvat et al. 1995) that no single method
is best under all possible circumstances. One can conclude that some correction,
however, must be applied to the acquired camera data.
Another method to include the effects of scattered radiation is to insert a
buildup factor (B) into the counting equation. Thus, eq. (10) would have a revised
anterior projection given by:

A ∗ B − µa
[ ]
N ap
=ε e 1 − e−µ l (10a)
∆t µl

21
Figure 4. Photon spectrum from a gamma camera for 111In.

An analogous change would occur for the posterior projection. In the original deri-
vation by Thomas et al. (1976), the B factor was set equal to unity. Siegel and co-
workers have determined such B functions for a variety of radionuclides by doing
single source attenuation experiments with a set of tissue-equivalent phantoms (Wu
and Siegel 1984). One may use either a depth-dependent or depth-independent value
for B given the radionuclide (van Rensberg et al. 1988).
An artificial neural network to compensate for scatter and attenuation has been
proposed as an additional method of correction (Maksud et al. 1998). The network
learning time was found to be reduced if a geometric mean was used as the input
data to the system.

3. Use of Geometric Data from Anatomic Imaging

As indicated, geometric thicknesses specified in the above equations may be
determined using transmission nuclear imaging. There are at least two arguments,
however, why a hybrid analysis, with both anatomic and physiological images,
could be used instead. First, distances, organ sizes, and positions can now be rig-
orously established. Such data, typically obtained via CT scan, are required for
the assessment of patient disease in the case of malignancy. Second, there is now
no need of external transmission sources with their added patient exposure and
cost (van Rensberg et al. 1988). In the case of anatomic information, the attenua-
tion equation is expressed in the integral form:

Ni
∆t ∫
= ε ⋅ ai ( x ) B ( x ) e − µx dx (13)

22
where Ni are the counts recorded along the ith ray through the patient and ai(x) is
the unknown linear activity density (µCi/cm or MBq/cm) at the depth x along that
ray. Other factors in the equation have been defined previously. Liu et al. (1996)
have solved the equation set (13) with a CT-assisted matrix inversion method
(CAMI). By imaging a phantom with both CT and nuclear camera, these authors
reported errors in three pseudo organ uptakes of between 5% and 15%. Using the
geometric mean (GM) approach and the same phantom, comparable errors were
on the order of 30%. Background activity was included in the phantom as a fourth
radioactive source.
Other observers, using a humanoid phantom, have reported geometric mean
activity estimates with errors on the order of 30% (van Rensberg et al. 1988). This
uncertainty is one of the primary reasons why GM analyses must be carefully uti-
lized. A second reason is the possible lack of a geometric mean value if the source
can only be viewed from a single projection. It is important to recognize that the GM
analysis also depends upon observing the organ clearly in both projections—with-
out overlap of other tissues. For example, the right kidney and liver interfere with
each other in almost any projection so as to make geometric mean activity measure-
ments for these two organs difficult if both show accumulation of the radiopharma-
ceutical. Thomas and co-workers have indicated methods, involving hyperbolic
functions, to account for other organs in the field of view (Thomas et al. 1988). These
methods require use of lateral images to determine thicknesses of the tissues
involved. A general review of uptake measurements, involving both non-scatter and
scatter situations, is found in MIRD Pamphlet 16 (Siegel et al. 1999).
Finally, it should be emphasized that because a hot spot appears in a compara-
ble location in two opposed projections there is no guarantee that it is the same
source. For example, a tumor may appear in the lower left quadrant in the ante-
rior view, but the spleen may be the strong uptake source seen in the correspon-
ding position in the posterior projection. By using CT data and merging that image
with nuclear camera information, this last kind of confusion may be reduced.
Table 4 contains a summary of the methods to evaluate organ uptake of radioac-
tivity. Three logical cases are possible for the nuclear projections. The observer

Table 4. Non-SPECT Methods to Determine Organ Uptake

in Nuclear Medicine

Method
Available Projection(s) GM Single Image CAMI
Diametric Opposed Yes Yes Yes
Single Projection No Yes Yes
None No No Yes

Note: CAMI means CT-assisted matrix inversion and requires the fusion of coronal CT (or MRI)
projections and coronal nuclear images.

23
may have diametrically opposed projections, a single projection, or no nuclear
image at all (!). This last case occurs when a CT or MRI image reveals a site that
does not show enough contrast to be imaged by the nuclear camera. Those doing
uptake measurements will have to deal with each of these three cases.

4. Quantitative SPECT Imaging

If a lesion or organ can be encompassed by a single set of SPECT images, these
cross-sectional counts may be used to estimate uptake (Chang 1980). In this analy-
sis, scatter corrections are applied to the reconstructed section or to each projection
obtained over the angular interval (typically 360 degrees). As Ljungberg and Strand
(1990) have indicated, the scatter corrections are generally done in one of two ways.
Either the observer sets a secondary energy window over the Compton energy peak
in the spectrum (as described above), or a convolution analysis is employed to deter-
mine the amount of scatter within the photoelectric energy window.
In the former case, two images are acquired at each angle over the range of
detector rotation. Explicitly, we find a corrected projection count value via:

N pr = Ntotal − kN sc (14)

where Npr represents primary counts and Nsc are counts recorded in a scatter
window. The constant k, called the scatter multiplier, is the ratio of scatter counts
detected in the photopeak and scatter windows respectively.
This simple dual energy window scatter subtraction method was originally
developed for SPECT images (Jaszczak et al. 1984). The method was found to be
accurate to within 10% for 99mTc. With a photopeak window set at 127 to 153 keV
(±10%) and scatter window at 92 to 125 keV, a k value of 0.5 was empirically deter-
mined. Both line sources and cold spheres were used in the measurements.
Koral and co-workers (1990) have extended the k-factor analysis to individual
projection images. Here, somewhat different values of k were determined using
only radioactive sources of 99mTc. In their study, k varied between 0.73 to 1.29
depending upon the geometry of the source and the size of the ROI selected by
the investigator. Thus, it is clear that the user must determine a correction constant
appropriate for the local camera and collimator system.
In the convolution technique, which requires only a single photopeak image set,
a similar equation has been applied.

N pr = Ntotal − Q ⊗ N pr (15)

where Q is a scatter function and ⊗ the convolution symbol (Axelsson et al. 1984).
Generally, Q is represented by a monoexponential function of counts versus dis-
tance off axis in a 22-cm diameter cylinder. Subsequently, Q has been found to be

24
non-stationary and, consequently, the method cannot provide the same accuracy
as the dual energy window approach. A general review of quantitative SPECT
imaging should be consulted for more details of the various scatter-correction
methods (Rosenthal et al. 1995).

C. Organ and Tumor Volumes

The accuracy of determining tumor/organ volumes is important for dosimetry,
since dose is a measure of energy density. One usually requires the anatomical
volume of an organ or tumor site. Because this is a challenge for most nuclear
medicine techniques, we must use volumes furnished by CT and MRI. Such vol-
umes are also used clinically to determine the progress of the malignancy in the
patient. While several imaging methods can be used, we should point out that pal-
pation may be sufficient for external (skin) lesions.

1. Palpation
Palpation has been a classical approach to estimating volumes of surface
lesions/organs in the clinic. The physician usually records the length (l), height
(h), and width (w) of an abnormal lump (e.g., a lymph node), usually indicated by
the patient, and the volume is calculated simply from the relationship:

4 (l ∗ w ∗ h )
Vtumor = π (16)
3 8

This subjective approach is simple and easy to apply, and can be valuable for
assessing the “hardness” and response of superficial nodes in various regions of
the body. An example of this sort of volume estimation using CT data is given in
Appendix B.

2. Imaging Methods to Determine Organ Mass

More objective volumetric methods rely on CT, MRI, and ultrasound to pro-
vide size estimates of organs and tumors in the body. These data can be used to
provide organ/tumor-specific dose estimates (cf. section VI) and serve as a crite-
ria for evaluating response based on serial imaging. Organs and tumors that have
diseased/necrotic volumes may not localize a radiopharmaceutical equally in all
regions. Sometimes this variation is due to poor perfusion. This can often be
demonstrated in large tumor volumes where the necrotic segment can represent a
significant fraction of the measured volume. Radiation dose estimation requires
determining the functioning volumes inside these organs/tumors.

25
 Radionuclide tomography can provide information for larger organs and
tumors, and PET and SPECT can be used to measure the functioning volumes for
organs and tumor sites larger than 10 ml. Biopsy with autoradiography can pro-
vide information on the microscopic distribution/fraction of a tumor/organ in
which a radiopharmaceutical is distributed. This method is usually impractical for
dose estimation because it is invasive and very localized. Biopsies in animal stud-
ies have demonstrated the distribution of protein agents to be heterogeneous. The
longer range of energetic beta emitters tends to compensate for the heterogeneity
and can be used to justify their use in clinical therapy.

V. DATA ANALYSIS METHODS

A. Activity and Percent Injected Activity (PIA)

Methods used to convert count data to activity, A (in mCi or MBq) or percent
injected activity (PIA) or fractional injected activity (FIA = PIA/100) have been
considered. One practical aspect of those methods is the use of a standard source
(usually less than 1% of the injected activity in 10 ml) that can be placed in the
camera field of view for every image. Test source count rates can be used to estab-
lish camera reproducibility over the course of the clinical trial. Total counts from
the standard can also be used to assess possible dead-time correction factors for
whole body and planar imaging.
Use of a well counter to measure blood, urine, and biopsy samples also requires
standards that are a fixed fraction of the total radioactivity administered. Usually
these standards are made up with a dilution factor of 10,000. In almost all cases, the
well counter information is in concentration form: mCi/ml or MBq/ml. Estimation
of the total amount of radioactivity in the blood or urine from well counter samples
requires additional information. Frequently, plasma samples are collected and hema-
tocrit values are required to extrapolate from plasma counts to whole blood con-
centration. Total urine volume would have to be recorded in the case of evaluating
excreted activity.

B. Justifications for Data Modeling

A number of reasons exist for the generation of a mathematical biodistribution
model. To first order in radiological physics, a model of animal or human data is
required in estimation of absorbed radiation doses. This follows from the need to
perform integration of normal organ and tumor activity curves out to many physi-
cal half-lives. By using only physical decay or some crude form of extrapolation
from the last data point(s), the investigator is not effectively understanding the long-
term course of the data. The same difficulty holds for estimation of activity values
between data points as taken (i.e., the model also provides better interpolation).

26
 A secondary reason for modeling, of more conceptual interest, is determina-
tion of rate constants and volumes descriptive of the tracer biodistributions. These
parameters can be compared across different proteins and other agents to help in
understanding the handling of the radiopharmaceutical in vivo. One of the most
important of these parameters is the volume of distribution. Generally, these vol-
umes are inverse functions of the molecular weight (MW) of a protein. Generally,
as the MW decreases, the volume of distribution will increase.
Finally, there is the possibility that the model may point out that some data
values are incorrect. Most such errors are due to mistakes made in the handling of
information and not because of camera or other detector malfunction. These erro-
neous values will be far away from any model-generated functions and thereby
noted by the investigator. In a data base analysis, such values must be screened out
before the integrations to determine areas under the curve.

C. Types of Models
Two types of physiological model are in general use: simple curve fitting and
compartmental analysis. We will describe the advantages and disadvantages of
each approach. Both require the use of a computer-based algorithm to produce the
best fit, usually via least-squares analyses, of a given set of model parameters.
Goodness of fit may be judged by use of standard statistical parameters such as
the correlation coefficient or R2.

1. Individual Organ Curve Fitting via Multi-Exponential

Functions
As the simpler alternative, one may consider a separate multi-exponential rep-
resentation for each of the various blood, normal organ, and tumor curves. In this
open model strategy, there is no explicit relationship between the various sets of
exponentials from one organ to another as these fitting processes go on in isola-
tion. This type of solution has certain justification since Laplace Transform analy-
sis shows that linear compartmental modeling, as described below, leads to a set
of linear differential equations whose solutions are indeed combinations of vari-
ous exponential functions (Wagner 1975). In practice it may be difficult, however,
to fit more than two exponential functions to any organ curve. For one thing, only
a single exponential solution is unique. Solutions involving multiple exponential
functions have different results depending upon starting conditions for the mod-
eling algorithm used. This was the reason that single exponential fits were shown
in the murine data analyses of Appendix A. Thus, complicated uptakes are not
readily represented via this strategy. The PLOT software package, allowing up to
three exponential functions, is available with MIRDOSE2 software from Oak
Ridge Associated Universities (Watson et al. 1984).

27
2. Compartmental Modeling of the Entire
Physiological System
The general motivation for compartmental modeling lies in the ability to fit all
of the biodistribution data simultaneously. Given the model, one then has a repre-
sentation that gives distribution volumes, rate constants, and other parameters.
Some advantage in the understanding of the biodistribution may result from this
more cohesive set of fitted parameters. Two general programs are presently avail-
able: SAAM and CONSAAM (Foster and Boston 1983) and ADAPT II (D’Argenio
and Schumitzky 1979). Both are operable on a standard PC. SIMPLE, available
from UCLA, runs on the Macintosh platform (Gambir et al. 1991).
Two general types of local relationships are found in closed compartmental analy-
ses. These are the mammillary and catenary forms as given in Figure 5. Note that,
for most agents, the injection of the labeled material into the blood implies that
the vasculature will act as the source compartment for all other organs (i.e., the
mammillary picture). This holds even for cases of intraperitoneal, intra-arterial
or intralesional injection or even inhalation of radioactivity due to eventual appear-
ance of the label in the blood compartment. Catenary (chain-like) features are
found in the excretion or processing of the labeled material. Renal and gut excre-
tion might be expected to exhibit this type of behavior. A given agent would
probably show some of each of these patterns. Because much of the input data
were acquired via imaging, blood may be used as part of any organ compart-
ment. This assumption may be not be necessary in the case of animal biodistri-
butions where blood has been removed from normal tissue and tumor samples
before counting.
Two catenary subroutines are built into both MIRDOSE2 and MIRDOSE3 pro-
grams. One is used to describe the chain from blood to kidney to urinary bladder
(Watson et al. 1984). The other describes excretion via the gut; small, upper large,
and lower large intestines are included in that catenary picture. Readers should not,
however, believe that their radiopharmaceutical will necessarily follow either of
these two schema. Instead one needs to confirm that such simple processes actu-
ally describe elimination of proteins or their metabolic products. It may be more
likely that a patients’s gut excretion rather than renal excretion would be similar
to that available in the MIRDOSE programs (Breitz et al. 1993). This follows from
the possible interaction of fragmented antibodies within the renal tubules.

D. Which Data to Analyze

Modeling is done using organ biodistribution data presented as a function of
time. Generally, a model’s mathematical form would deal with an organ’s uptake,
not the organ’s uptake per gram. This means A(activity/organ), as given as a series
of time points, is the appropriate input information. Motion of material and the

28
Figure 5. Mammillary and catenary model subsystems.

29
conservation of matter (activity) are readily accomplished with these variables.
Data in the form of u(activity/gram) are usually not immediately appropriate since
such specific uptakes are not directly relatable to the model variables. It may be
possible to take the mathematical picture as derived for the A variables and recast
it into the u variables by respectively dividing through by organ masses. In the case
of clinical results, the last conversion may not be achievable, however, due to
uncertainties in patient organ size.
This conversion is required, however, in the case of blood samples whereby
the total volume of distribution is unknown—particularly for a novel protein or
new tracer. In this case, the A variable for the blood is explicitly modeled as
u(blood) ∗ V(blood) where the latter volume is determined by the fitting algorithm.
Note that u is the measured quantity in this case, not A.

E. Physical Decay as a Clearance Mechanism in the Model

Most literature biodistribution data have been corrected for physical decay so
that the given percent injected dose/organ (PID/organ) or PID/g value is that of an
associated hypothetical pharmaceutical—not that of the decaying radiopharma-
ceutical. If the analyst is interested only in pharmacokinetics results and the model
is of the linear type, such data may be taken directly into the modeling equations.
Rate constants, volumes, and other parameters will then be appropriate for the
decay-independent model.
In the case of radiological physics, however, such decay-corrected data are
not appropriate for analyses prior to Medical Internal Radiation Dose (MIRD)-
type absorbed dose estimations (Snyder et al. 1975). This follows from the fact
that the user is interested in finding an area under the organ’s curve of activity
versus time. These integrals are the total number of decays in the various source
organs within the patient. Thus, the user may prefer to go back to the original data
to find the appropriate inputs into the modeling program. Alternatively, one may
simply multiply decay-corrected data by exp(−λt) in order to take out the decay
corrections the author has placed in published reports. Here, λ is the physical
decay constant in units of inverse time. Some statistical impact is expected in such
cases as measured data will have fewer counts at long times as compared to the
decay-corrected published data. Notice that the model must, in this case of deal-
ing with decay-containing data, include the physical decay as an output from each
compartment. If the model is non-linear and/or non-homogeneous, however, the
modeler must be careful to substitute the true data (uncorrected) into the equa-
tions. The latter must themselves contain physical decay as an explicit route out
of each compartment (Williams et al. 1995b). In this case, one cannot simply cor-
rect the data for decay; the decay must be included in the model a priori or else
incorrect results may appear in the solutions.

30
F. Integration of Area Under the Curve (AUC)

1. Mathematical Form
Some users prefer to estimate activity integrals under PIA versus time curves
with a simple trapezoidal or other rule (Chaney and Kinkaid 1985). Usually, the
region of the curve beyond the last measured point is represented as a single expo-
nential based on the effective half time calculated from the last 2 or 3 points. This
is still a model—albeit one without a physical basis. The accuracy of the AUC will
depend mostly on the magnitude of the last point relative to the maximum values
for each uptake site.
A purely mathematical form of integration has another application in dose esti-
mation in a modeling context. It may be the case that one organ system, such as the
kidneys, is only imaged in a relatively small subset of the patients. Thus, due to a
lack of renal data, the model cannot contain a kidney compartment. For those
patients whose renal system is imaged, the user may prefer to perform that particu-
lar integration via a trapezoidal or other mathematical rule as a separate procedure.

2. Multiple Exponential Functions

Integration of the open model or multiple exponential picture of organ activity
is performed analytically using the fact that the integral of a i exp(−k it) from t = 0
to ∞ is simply a i/ki. Thus the sum of such exponentials becomes a sum over the
amplitudes (ai) of each term divided by the relevant rate constant (ki).

∑  k 
a
AUC = (17)
i =1, n i

It is useful to recall that the physical decay constant is generally implicitly part of
the k parameters. Any given k can then be represented as kb + λ with kb being the
biological clearance constant.

3. Compartmental Model
In either the ADAPT (D’Argenio and Schumitzky 1979) or SAAM (Foster and
Boston 1983) programs, integration of the time-activity curves is accomplished
by defining a mathematical compartment as the integral of the organ compartment
of interest. For example, if we set the activity in compartment one equal to A(1),
then its integral is:
∞
à (1) = ∫
t =0
A(1) dt (18)

31
Figure 6. Five-compartment model for the intact antibody cT84.66.

Here, Ã(1) is formally the time integral of compartment one. Notice that Ã(1) is
not sampled, and can only be computed using the fitted organ one activity curve.
With ADAPT, this computation may require extensive processor time due to the
step-by-step integration necessary in a modeling program. Convergence may also
be questioned and is best answered by looking at the graph of Ã(1) as a function
of time. If the integral converges, this function approaches a constant value. An
integration compartment must be defined for each of the modeled physiological
systems. Thus, if 5 organs are modeled, 5 equations of the form given above will
need to be added to the analysis. In this case, a total of 10 differential equations
will have to be solved simultaneously. One may consider this complexity to be a
disadvantage of the modeling approach as compared to a purely mathematical
integration. This last consideration is a reason for multi-exponential representa-
tion of an organ’s activity data. Figure 6 contains the five compartment model used
to represent the antibody cT84.66.
Uncertainties in the integrals are generally not discussed by investigators. One
method to estimate the errors in the AUCs is to use the covariance matrix avail-
able from the fitting procedure. This matrix, with Monte Carlo methods, will allow
simultaneous prediction of the variation of areas for each of the organ compart-
ments (Kaplan et al. 1997). While this type of analysis may be performed sepa-
rately on each of the organs in an open compartment analysis, that result will tend
to misrepresent the variation of AUC since the cross-correlations are not included
in the process.
Given the Ã, the user must substitute this set of values into the fundamental
MIRD formula of eq. (1). Generally, one retains the units as taken, µCi − h or
decays, for this computation. The resultant dose is then in rads or cGy if the S

32
matrix value is in appropriate units. If one wishes to compare various patients with
each other or with animal studies, the computation is in rad/µCi or cGy/MBq
injected activity so that residence times (τ) are used in lieu of à in eq. (1).

4. Variation of the Radionuclide

It may occur that the imaging radionuclide must, of necessity, be different from
the therapy radionuclide. One example of this sort is the use of the 111In-labeled
protein as imaging tracer for the 90Y-labeled protein. Organ activity information,
obtained via the photons emitted by the imaging label, must then be recast into
activity appropriate for the therapy agent. This is conventionally done by simply
changing the physical decay constant in the modeling equations. Whether the user
has an open model or compartmental picture, the decay constant of the therapy
radionuclide is entered in lieu of that of the imaging radionuclide.

VI. DOSE CALCULATION METHODS

AND PROGRAMS

A. Normal Organ Dose Estimates in Standard Man

1. MIRD 11 Pamphlet S Values

As a first choice in normal organ dose estimation, the user may elect to multiply
the cumulated activity à value set by a standard tabulated S matrix. The MIRD 11
Pamphlet, published by the Society of Nuclear Medicine is probably the best-known
source of these values (Snyder et al. 1975). Radionuclides available include 131I, 90Y,
and 111In. In the pamphlet, both penetrating and non-penetrating radiation S values
are summed to provide a net S matrix. It will be found that so-called pure beta emit-
ters such as 90Y and 32P are diagonal matrices (i.e., Bremsstrahlung radiation is not
included in these Monte Carlo S values). A method to include Bremsstrahlung radi-
ation dose is described in section VI.C.5.
The Monte Carlo S matrix computation done in the MIRD Pamphlet 11 tabu-
lation was performed using a computer-based set of humanoid phantoms. Uniform
distribution of radioactivity in the source organs was also assumed in this
approach. It is important to understand that these phantoms do not exist as phys-
ical objects; instead they are mathematical intersections of various geometric
solids which simulate shape and position of a given organ system. For example,
the thyroid is a set of tangent ellipsoids and the bladder is a sphere. Absolute size
of these intersections is determined by standard organ masses available from phys-
iological data sets. Because of these standard geometric shapes and locations, the
MIRD phantom set is most appropriate for a dose estimate involving legal ques-
tions and theoretical consequences. An example is an Investigational New Drug

33
(IND) Application sent to the FDA. We have previously discussed using animal
biodistribution data as an input to this type of computation.
Reviewers of the IND application are generally not interested in a specific
patient’s absorbed dose estimate. Instead, they want to find out how the applicant’s
dose estimation for a novel radiopharmaceutical compares to other, similar
agents. By using a standard set of S values, the applicant allows the FDA to follow
its computations more readily than if a patient-specific dose estimate were made.
We will describe individualized absorbed dose estimates in section VI.B.

2. MIRDOSE2 and MIRDOSE3

A more efficient method for selecting S values is the use of either of the two
standard programs available from Oak Ridge Associated Universities (ORAU).
These S values are generated in real time using a parameterization of the MIRD
Pamphlet 11 values for a given radionuclide. The older of these, MIRDOSE2, is
available as a BASIC program which runs using the BASIC interpreter on a PC
system (Watson et al. 1984). Source code is available which the user may exam-
ine and change as needed. For example, one could add a subroutine to supply an
organ system model to describe the clearance of the radiopharmaceutical. PLOT,
a three-exponential fitting algorithm, in BASIC, comes with the program. Catenary
models of renal and gut excretion are part of this package. Figure 7 contains a sum-
mary of the steps required in the estimation of normal organ doses.
In 1996, an updated version, termed MIRDOSE3, was made available from
ORAU (Stabin 1996). MIRDOSE3 runs under WINDOWS 3.1 with an icon
system of choices. These include radionuclide and phantom selection. One can
also print the S matrix of interest to accompany the dose estimation report.
Unfortunately, the PLOT software is no longer part of the package so that the mod-
eling aspects must be done externally. The catenary models for renal and gut excre-
tion are included however.
One difficulty in using MIRD Pamphlet 11 is the lack of a “remainder of the
body” as a possible source organ in the tabulated S matrices. One cannot use the
total body term as that Monte Carlo estimate assumed that the activity was uni-
formly distributed in the total body of the phantom (Loevinger et al. 1991). The
remainder of the body is the difference between the total body and the organs that
show individual uptake. Thus, it is a variable source organ. The MIRDOSE3 pro-
gram allows direct input of the remainder of the body residence time so as to elim-
inate this problem.
A total of six phantoms are available with MIRDOSE3. These range from fetus
to adult male- and female-sized individuals. Since a typical RIT protocol will
only involve patients over the age of 18, it is unlikely that this selection range
will be required in a treatment plan. There is, of course, the possibility that if a
patient is unusually small, then the 15-year old or even younger phantom might

34
Figure 7. Outline of the general dose estimation process.

be of interest. A sample absorbed dose estimate for the adult male phantom using
the geometric mean method for uptake measurement is given in Appendix C.
Integration of the activity curves was done with the five-compartment model of
Figure 6.

3. Bone Marrow Dose Estimation

Because of the enhanced radiosensitivity of the blood-forming cells, the red
marrow is frequently the dose-limiting organ in RIT. The sensitivity of the red
marrow to ionizing radiation is limited to about 500 rads for external beam radi-
ation and about 200 rads for internal emitters. Most RIT treatment plans conse-
quently rely on providing the best estimates for red marrow dose based on a range
of assumptions and uncertainties. Frequently, patients enrolled in RIT protocols
have undergone other therapies which may have damaged a large fraction of their
marrow reservoirs. The marrow mass and distribution in these patients may be
uncertain and the mass correction factor will be unreliable—albeit greater than
one since the residual marrow mass is likely to be less than the value assumed in
MIRD Pamphlet 11.

35
 Three source organs are generally assumed for estimating marrow dose in RIT.
These include activities in the marrow, bone, and residual body. This can be ex-
pressed as:

Drm = Drm ← rm + Drm ← bone + Drm ← RB (19)

Radionuclide organ imaging methods often cannot separate these three source
organs and sometimes a biopsy sample may be useful. Counts from an iliac crest
or other bone marrow sample may be used to normalize to imaging data taken
from the gamma camera (Macey et al. 1995). Such samples, however, only rep-
resent a single determination (snapshot) of a radioactivity distribution that is
changing and repeated samples are not easy to procure because of the trauma to
the patient.
If the radiolabeled antibody is assumed to be localized uniformly in the red
marrow, an aliquot can be used to provide estimates for total marrow uptake based
on imaging a small region of the marrow. A first order approach to providing esti-
mates of specific marrow uptake has been attempted based on the ROI counts
detected in the sacrum or 3 lumbar vertebrae (Macey et al. 1988). Integration of
the uptake curves obtained over any of these areas is used to provide an estimate
for specific uptake in total marrow. If little or no net activity can be visualized in
these regions, an alternative strategy must be used. Generally, this involves the
blood curve of the patient.
It has become more generally accepted that, in the case of marrow dose esti-
mation, one can use the blood concentration curve as a surrogate for the marrow
concentration curve (Siegel et al. 1990). Instead of a direct correspondence, the
marrow à value is set equal to a fraction ( f ) of the blood curves Ã:

A˜ rm ← rm = f A˜ blood 
1500 
(20)
 5000 

where 5000 (g) refers to the whole blood mass and 1500 (g) the normal red
marrow mass in the adult. Physically, this relationship is an attempt to correct, to
lowest order, for the mass difference between the whole blood and red marrow.
Values for the f factor have been estimated to lie between 0.2 and 0.4; a most
probable value of 0.34 to 0.36 has been determined by Sgouros (1993). Note that
this is only one term in the estimation of Ã. Two other terms arise from specific
uptake in the bone itself and from radioactivity in the remainder of the body.
Figure 8 summarizes the steps required to estimate marrow dose. Appendix D
demonstrates a bone marrow dose estimation using the technique of eq. (20) and
a residual body source term.

36
Figure 8. Schematic of red marrow dose estimation.

B. Patient-Specific Normal Organ Doses

1. Target Organ Mass Correction

One cannot assume that a given patient is exactly the size prescribed by one
of the MIRDOSE3 phantoms. As was noted in section III, a CT scan is generally
done before RIT in order that the medical oncologist has a baseline measurement
of tumor sizes and locations. As a side benefit, this scan, or perhaps a compara-
ble set of MRI images, can be spatially integrated to provide true organ sizes.
Given this size, one may correct the MIRDOSE3 or other normal organ S value
by simply scaling a mass correction factor for the non-penetrating (np) contribu-
tion. Let us consider a sample patient-specific computation using MIRD Pamphlet
11 tables of S matrices and breaking the value down into its penetrating (p) and
non-penetrating (np) components:

S p = SMIRD11 − Snp (21)

37
As an illustration of this simple mass correction approach, suppose the mass of
the spleen in a patient is known to be enlarged by a factor of 2 compared with the
MIRD Pamphlet 11 phantom model of 173.6 g. For 131I in this spleen, the Sp value
will be given by:

0.408
S pspleen = 0.0026 − = 0.00025 (22)
173.6

where the np component has been calculated using eq. (2) and the emission data
for 131I. We have written this term purposely as a ratio of the numerator of eq. (2)
divided by the MIRD phantom splenic mass. For the enlarged spleen the first order
corrected S value will be:

0.408
347 g ) =
S(spleen + 0.00025 = 0.00143 (23)
347

For the 347 g spleen, the Sp value for 131I could be more accurately derived by
interpolating the change in absorbed fraction with mass for the range of photons
emitted, but since this changes slowly with mass, a first order estimate is accurate
to better than 5% of the final value that would be returned. In our example, this
simple correction for the S value reduced the absorbed dose to the spleen by 45%.
The primary correction factor was the inverse dependence upon mass given in
eq. (2). Anatomic information is essential to the method so that CT or MRI data
will need to be available for the computation to proceed. A more complete descrip-
tion, including discussion of the gamma ray contribution which has been neglected
here, is given by Shen et al. (1997).
Other methods may be used in lieu of the MIRD Committee and MIRDOSE
approach. These techniques for individual doses are not based on Monte Carlo
methods and mathematical phantoms. Instead, one uses the actual patient geom-
etry given from the anatomical image set and explicitly calculates the absorbed
dose to a voxel in the target organ space given a set of voxels of activity distrib-
uted in the patient’s source organs. We will discuss two of these methods.

2. Convolution Dose Estimates

The most exact method for normal organ (or tumor) dose estimation follows
from the use of a point source kernel or function (PSF) for a selected radionuclide
in an infinite, unit density medium. These functions provide a look-up table for
the dose rate as an isotropic function of distance between source and target.
Tabulations are available for electron and photon energies commonly found in RIT
(Leichner 1992). Previously, PSFs have been generated for monoenergetic elec-
trons (Berger 1971). The PSF for a radionuclide is the sum of the dose rate from
each particle and photon emitted. The differential dose at a position y with a point

38
source at x is given the convolution:

∆D( y) = ∆A˜ ( x ) ⊗ PSF( y − x ) (24)

where ∆Ã(x) is the differential cumulated activity at point x. This technique is

entirely independent of S matrices but requires exact knowledge of patient anatomy
and activity distribution.
The execution time required for the integration of this equation is the greatest
obstacle to the method. For beta radiation, the only target usually considered is the
source organ itself so that the integral is carried out rapidly. For photon doses, how-
ever, all possible sources must be included in the integration step. Consequently,
implementation of the equation also requires information on the distribution of Ã
in each voxel; this may be difficult to determine in a patient. Such constraints force
many users to assume uniform uptake as was done in providing MIRD Pamphlet
11 S results and the MIRDOSE programs. In spite of this, the actual spatial volume
of the organ is being considered in the estimation so that the process is inherently
superior to use of a precalculated S matrix.
One limitation of the point source method is that the medium is assumed to be
uniform and infinite in extent. Usually a soft tissue framework is assumed. Hence,
if there is significant variation in tissue type between source and target, the point
source method may be significantly in error. This would not be the case within a
soft tissue such as liver or spleen, but could occur near transition points such as
bone structures or estimating dose to lung from sources in the liver.

3. Voxel Source Kernel Method

One strategy that is useful in patient-specific dose estimation is the generation
of a small scale S value (i.e., one computed on a voxel basis). Here, the voxel would
be one appropriate for the nuclear image being used for the uptake measurements
(Liu et al. 1998). Any smaller scale S cannot be justified since the nuclear data do
not have the requisite spatial resolution. Monte Carlo methods are used in the
development of such voxel source kernels (VSKs). Dose estimates are then com-
puted using a convolution approach, as in eq. (24), using the actual anatomic
volume of the patient organ(s). These computations may also be performed on
tumor sites and lymph nodes. A recent MIRD publication gives several VSK S
values for a number of radionuclides of interest (Bolch et al. 1999).

C. Tumor Dose Estimations in Radioimmunotherapy

and Radionuclide Therapy
In principle, there are three types of emission that may be considered in tumor
dose estimates. Of primary interest are alpha and beta rays, which have a relatively
short range and hence are localized in the lesion(s). Photon doses are usually of

39
secondary significance. In a practical sense, the most important non-penetrating
radiation is the beta ray.

1. Alpha and Auger Emissions

Application of alpha particles has been difficult due to their extremely short
range (10 to 100 µm) in soft tissue. Because antibody deposition is rarely within
this distance of solid tumor cell nuclei, one has the possibility of ineffectiveness of
such therapy. Exceptions can take place because of the movement of the antibody-
antigen complex within the tumor cell. Additionally, there have been applications
to leukemia patients wherein malignant cells are freely circulating in the blood.
In either of those cases, the estimates must be done on a microscopic basis. In fact,
alpha dose estimation is generally a microdosimetric method that requires exqui-
site geometric knowledge of radionuclide deposition. Such knowledge probably
requires autoradiographic methods, which are difficult to use clinically.
A similar argument holds for Auger electrons emitted due to vacancies in atomic
shells. These particles have ranges on the order of 10 µm at 20 keV. If the antibody
radiolabel will be taken into the cell, preferably into the nucleus, such emissions may
be considered for RIT. Otherwise, their application to therapy is limited. For these
reasons, we will not consider alpha-emitters or Auger electrons in the following.

2. Beta Emissions
a. Uniform Uptake

Because of its essentially random location and size, tumor cannot be a source or
target organ in the standard MIRD phantom format. There is, within MIRDOSE3,
a separate algorithm which estimates beta doses to spherical tumors of a given size
and having uniform uptake of the radioagent. Here, one supplies the tumor mass,
the radionuclide, and an à value for this computation. Edge effects are explicitly
taken into account (Buras et al. 1994). Doses due to other organs, however, are not
included in the computation. Such effects are probably not significant for relatively
large tumors. There may be, however, significant cross talk of high-energy betas
from a nearby normal organ to the tumor (e.g., from the liver to a liver metastasis).
A VSK may be used in this analysis as indicated in section VI.B.3.

b. Non-Uniform Uptake

Let us next consider the most general case of a non-uniform distribution of beta
activity in the tumor and energy deposition that is non-local. The latter assump-
tion follows from the extensive range of beta rays from radionuclides such as 90Y
(1.1 cm in soft tissue). Howell et al. (1989) have given S factors for various spher-
ical tumor sizes and different non-uniform radionuclide distributions. In this case

40
one could directly substitute the resultant S into the general MIRD formula if the
à is known. Thus, an S value does not necessarily imply uniform uptake within
the lesion. The VSK approach, as described above, will also provide dose esti-
mates for non-uniform uptakes by generating a voxel-sized S matrix. In any event,
estimation will be separated into its two classical aspects: the evaluation of à and
the determination of an S value or equivalent for the tumor geometry (Loevinger
and Berman 1976, Loevinger et al. 1991).

3. The Tumor à Algorithm

Activity information may be obtained from an associated gamma emitter that
has an identical or similar biodistribution in the patient. For example, if 131I is the
antibody radiolabel, the gamma rays from radioiodine point out the position of the
radioactive atoms that are also a source of the beta radiation. Alternatively, one may
use the 111In-antibody in lieu of the 90Y-antibody to establish the biodistribution if
the labeling type does not affect the actual targeting in vivo. In practice, this method
is less certain since no chelator is equally effective as one changes the radius of the
radioactive ion. Finally, if a pure high-energy beta-labeled antibody is injected, one
might use the Bremsstrahlung radiation images to quantify activity as a function of
time in the tumor. In this case, however, the image quality is often poor unless one
injects directly into the lesion (Siegel et al. 1994).
Using any of the above three strategies of following tumor radionuclide depo-
sition, serial nuclear medicine images provide the data to be modeled for à evalu-
ation. If one wants some idea of the heterogeneity of activity in the lesion, SPECT
imaging may be the only possible modality. Otherwise, a ROI is drawn around a
planar image of the tumor to delineate the pixels of interest and the activity as a
function of time. Notice that the gamma camera image must be corrected for
attenuation and scatter and be calibrated so that counts per minute or per image
can be translated into µCi or MBq in the lesion. These methods are described in
section IV.

4. Tumor Mass Determination

Tumor mass must generally be determined by some imaging method since exter-
nal measurement or palpation is unlikely to be accurate for lesions within the patient.
One exception to this rule may be Na131I therapy of the thyroid, which is discussed
below. Thus, CT, MRI, ultrasound or even planar X-ray images are preferable. If
several dimensions are available, the physicist may represent the lesion as an ellip-
soid with minor axes equal to half of each of the respective three dimensions. A
lesion with two dimensions may be idealized as an ellipsoid of rotation. Likewise,
if only one length is observed, the tumor may be represented as a sphere whose diam-
eter is equal to that distance. Unit density is generally assumed in any case. Eq. (16)
may be used in these computations.

41
 Because of a lack of other information, it could be necessary to use the nuclear
image to provide tumor geometry. If this is done, the resultant volume may be either
larger or smaller than that found with anatomical imaging or at surgery. The former
case of increased size occurs due to the scatter of the photons in the patient before
reaching the detector. Some of this scatter is in the collimation system. As
described in section IV, corrections to the nuclear image can be made by merging
the CT and nuclear images so as to better define the actual lesion volume.
Additionally, one may draw a satellite ROI near the source organ. Using count den-
sity from this background region, one may subtract counts from the source organ
so as to get a better idea of the actual decay events therein. To a major degree, this
is a correction for over- and under-lying activities.
The nuclear image may be smaller than the anatomical due to lack of perfusion
of certain regions of the neoplasm. In this case, a dose estimate based on nuclear
volumes may be too high and not appropriate for the tumor in question. If avail-
able, a comparison of nuclear image and non-nuclear image lesion dimensions
may be informative. Overlaying images from the two modalities can help estab-
lish zones of poor perfusion and/or incomplete targeting. This last feature may also
occur due to the heterogeneous expression of the antigen.

5. Computational Algorithms
If the activity distribution is heterogeneous and known, a direct Monte Carlo
approach is one way to estimate absorbed dose in different regions of the tumor.
Several programs are available including EGS4 (Ford and Nelson 1978), MCNP
(Briesmeister 1993), and ETRAN (Berger and Seltzer 1973). Some of these codes,
including MCNP, are available on PC computers. Geometric volumes allowed by
these programs are limited, however, so that the true geometry of the lesion and
its surroundings may not be acceptable as an input set of tally spaces. Computer
time can be quite extensive and this sort of solution may not, therefore, be prac-
tical on a routine basis.
An additional approach to this general problem is to use a convolution method
based on point source functions for the beta emitter of interest (Loevinger et al.
1956). Here, the estimated dose at the position y is given by the result of eq. (24).
By means of Fast Fourier Transform (FFT) algorithms, this computation can be
done relatively quickly for a simple tumor shape (Brigham 1974). The Fourier
transform maps the convolution into a simple product of functions in the fre-
quency (cm−1) domain. Roberson and co-workers (1994) have performed exten-
sive FFT calculations of this type in variously shaped human xenografts in a nude
mouse model using autoradiographic data for the activity distribution Ã(x).
Alternatively, one may perform the integral of eq. (24) directly in the spatial
domain. In this case, the time required may be prohibitive when the lesion has a
random shape.

42
 Finally, for high-energy beta emitters, there is the Bremsstrahlung radiation
contribution (Williams et al. 1989, Stabin et al. 1994). This effect is rather small,
on the order of one percent or less, unless the lesion lies within the normal organ
(e.g., the gut for a colorectal cancer). A point source function is available for this
effect in the case of 90Y (Stabin et al. 1994). Given this function, convolution tech-
niques as outlined above may prove effective in rapid computations for simple
geometries. With a voxel source kernel, the Bremsstrahlung radiation can be
explicitly included within the S matrix (Liu et al. 1998).

6. Photon Contributions to Tumor Dose Estimates

Both à and organ mass may be determined as described above. If the activity
distribution and geometry are relatively simple (e.g., uniform uptake in a spheri-
cal tumor), one can find tabulated absorbed fractions appropriate to the photon
energy. Such S values allow direct, rapid estimation of tumor-to-tumor doses.
Even in such a simple case, the estimates of photon doses to tumors from other
organs is difficult due to the random location of the lesion within the standard
human MIRD geometry. One solution to this problem is to use a normal organ in
lieu of the tumor to provide an approximate S value for the photon(s) of interest.
A similar replacement strategy can be used to provide an estimate of normal organ
target dose from a source within the tumor. Such replacements are sometimes
called analogue source or target organs. If no normal organs of appropriate size
and location are available, the whole body could be used in lieu of the tumor as a
target organ; this is clearly an overestimate of the tumor-subtended solid angle.
Resultant estimates will probably be too large and can only serve as an upper
bound to the gamma-induced absorbed dose.
If we consider the most general case of heterogeneous activity distribution that
is known, we can estimate absorbed fractions most generally with Monte-Carlo
methods. Limitations are, as mentioned, due to the time required. Parallel pro-
cessing has been advocated to permit more rapid estimates for a given number of
photon histories (Johnson 1988). Convolution or direct integration over the source
volume(s) may also be performed. Leichner (1992) has published point source
functions for various monoenergetic photons. These two methods explicitly permit
dose estimates to normal organs just as in the case of the beta radiation. As in that
case, however, the computational time required may not permit practical clinical
operation.
A probable better approach to the above situation is to use the MABDOSE pro-
gram developed by Johnson and associates (Johnson 1988). Here, the gamma-
derived absorbed dose estimates can be made for spherical tumors at various
locations within the MIRD standard phantom geometries. Three Cartesian coor-
dinates of the center of this sphere must be specified to begin the Monte Carlo
based estimation process. Photon-derived doses to the tumor from activity both

43
within and without the lesion are included in the estimation. This program, run-
ning under Windows 3.1, is available on a PC.

7. Clinical Computations (Beta and Photon

Contributions Combined)
All radionuclides, in practice, produce both particulate and photon emissions.
Internal conversion and Auger electrons are seen, for example, in purely photon
transitions. Likewise, pure beta emitters give rise to Bremsstrahlung radiation
because of the presence of local tissue surrounding the emitter. Thus, the physi-
cist will need to sum over both types of emissions to produce a general absorbed
dose estimate.

8. Doses to Tumors That Are Not Imaged

One of the reasons for antibody or other systemic therapy is that undetected
tumors may also be irradiated during the course of the treatment. Small metasta-
tic lesions, lymph node localizations, and even microscopic clusters of tumor cells
are possible targets for a radiolabeled antibody. It has been observed in human
xenografts in mice and in a limited number of human results that an inverse rela-
tionship usually exists between tumor uptake (u) in PID/g and the mass (m) of the
lesion (Williams et al. 1988). This relationship is in the form of a power-law:

u( m) = a ∗ m b (25)
where a and b are constants independent of m. The exponent b is typically in the
range −1.0 < b < −0.1.
For example, for primary colorectal lesions in humans, b = −0.362 was meas-
ured (Williams et al. 1993). Thus, if the beta radiation does not escape the tumor
volume, one anticipates improved radiation dose delivery as the lesion mass
decreases. Eq. (25) can be used to predict uptake for one or more lesions seen only
on CT or MRI scan. Here, we suppose that the a and b coefficients have been deter-
mined using animal or human data. Human uptakes of antibody tracers have been
measured by taking surgical samples post systemic injection. Values of u, cor-
rected for decay of the radionuclide, are generally in the range of 2 to 20% ID/kg
at periods of time between 5 and 15 days post-injection (Williams et al. 1993).
These magnitudes are consistent with the results with human xenografts in nude
mice provided that one takes into account the species mass difference and the vari-
ation of antigen concentration between the xenograft and human wild-type
tumors. Thus, analysis of a murine model does provide a good prediction of the
human tumor uptake in a clinical setting. A summary of tumor dose estimation
methods is given in Figure 9.

44
Figure 9. Steps in tumor dose estimation. ACF refers to attenuation correction factor.

9. Treatment Strategy and Results

It is often the case that red marrow toxicity is the activity-limiting step in
RIT treatment planning. Although, for example, only weak correlations may be
found between estimated marrow dose and the depletion of platelets, many
patients exhibit a reduction of platelets with a nadir at approximately 4 weeks
post-treatment (Figure 10). The likelihood that marrow effects will occur in a
given patient depends upon the uncertain history of that individual’s marrow
depletion as a result of earlier therapies. Marrow storage is often used to preclude
potentially fatal results. Figure 11 gives the overall strategy for treatment plan-
ning with a limiting estimated marrow dose of 200 rads or 2.00 Gy. If the pro-
jected dose exceeds this value, a reduction of injected activity or rejection of the
RIT protocol are possible choices for the patient. Repeat RIT activities are gen-
erally the same as those used initially.

45
Figure 10. Typical platelet data from four representative 131I-CC49 RIT patients.

VII. THYROID CANCER THERAPY - THYROID

THERAPY WITH 131I
The prototype for RIT is treatment of malignant thyroid disease with radioiodine.
Here, there are three methods for estimating the amount of radioactivity to be given
to treat the patient’s remnant thyroid (Hurley and Becker 1983, Harbert 1987). These
include a fixed small amount generally 30 mCi, a larger activity 75–150 mCi, or a
calculated absorbed dose on the order of 3 × 104 rads (300 Gy) to the residual malig-
nant tissue. Reasons for the 30 mCi protocol turn on the issue of patient isolation
due to historical state and other regulations on radiation safety. These are discussed
in section IX.
Some disagreement occurs as to the efficacy of a 30 mCi treatment strategy
however, and that method would probably not be recommended generally
(McCowen et al. 1976, Siddiqui et al. 1981). The most common treatment regi-
men is using a fixed activity of 75 to 150 mCi of 131I with some authors claiming
85% success in achieving complete ablation—that is, no subsequent uptake of
radioiodine in the neck area (Siddiqui et al. 1981).
If one uses a calculated absorbed dose value, calibration of the probe system
usually relies on the counting of a known source in the same geometry as the
patient. A neck phantom may be useful in this regard to simulate attenuation.
Absolute uptake quantification has been developed using coincidence decays of
gamma and X-rays from 123I to evaluate the depth of the emitter in the neck (Siegel
1981).
If the physician elects to treat the patient by estimating absorbed dose, the
activity-limiting organ has usually been the blood. Seen from the perspective of

46
Figure 11. Treatment strategy for RIT based on limited dose (200 rad) to marrow.

47
RIT, this is very similar to the limitation due to red marrow dose. Generally, one
attempts to keep estimated blood dose to 200 rads (2.0 Gy) or less. Since this treat-
ment will follow surgical removal of the majority of the thyroid, the actual dose
to remnant tissue remains uncertain, a difficulty that is also found in treating a
small tumor of unknown size within the body of a patient using RIT.
As in the case of non-thyroid tumor therapy, the mass of residual thyroid tissue
remains the primary uncertainty in the target organ absorbed dose estimation
process. Both penetrating and non-penetrating radiation are included in the esti-
mation. Possible radionuclides are 131I and 125I with advantages seen in the use of
the former due to its greater beta ray penetration in soft tissue. It should be noted
that use of 131I may allow imaging of the biodistribution of activity during the
course of treatment. Here, we assume rapid excretion of most of the activity after
the first 24 hours so that the gamma camera is not overwhelmed with the count
rate during therapy. Additional lesions may also be visualized using this high level
of 131I activity prescribed for the radiotherapy treatment. Thus, for most clinicians,
131
I is the radionuclide of choice in thyroid therapy. It is administered as the sodium
salt—usually by mouth.

VIII. BONE PAIN PALLIATION WITH

RADIOPHARMACEUTICALS

A. Clinical Situation
Prostate, breast, and other cancers may lead to multiple bone metastatic sites.
While single sites may be treated with external beam therapy, multiple sites imply
the use of a radioactive agent that can distribute as a bone seeker. In general, the
treatment is not intended to reduce the metastatic site in size, but rather to induce
a reduction in the patient’s pain symptoms. The anatomic target of this radiation
is not known. In several clinical studies, such palliation has been achieved at levels
between 50% and 90% of the population. Pain may recur with subsequent treat-
ment becoming necessary (Lewington 1996). Absorbed dose estimates have been
made in only a minority of the clinical trials. Thus, it has not been possible to cor-
relate the clinical outcome (i.e., the reduction of symptoms) with the estimated
dose.

B. Radionuclides
A number of beta emitters, including 32P, 89Sr, 153Sm, and 186Re, have been
investigated in these applications. Particulars regarding their beta (and possible
gamma) emissions are described in Table 1. More recent trials have involved the
last two radionuclides in the reduction of breast and prostate cancer bone pain. In

48
particular, 186Re has chemical properties similar to 99mTc so that the latter tracer
may be used to indicate the eventual biodistribution of rhenium. In this way, the
physician may correlate the conventional bone scan image with that found during
therapy (de Klerk et al. 1992). Presence of a gamma emission in the decay spectra
of 153Sm and 186Re allows imaging of the distribution activity after injection of these
radionuclides. In that way, the clinician can document uptake and, possibly, the
reduction of a lesion’s size.

C. Dose Estimation
Determination of activity within bone is probably the greatest obstacle to dose
estimation using bone agents. Projection images may show accumulations that
actually reflect uptake in hard bone, as well as trabecular bone and marrow. Thus,
the clinician is usually limited to bone biopsy information. These data, however,
are generally only obtained once for a given patient due to trauma and associated
pain. It may be possible to correlate the biopsy with the image data, but this can
be difficult.
Stabin and co-workers have used MIRDOSE2 and its ICRP 30 (International
Commission on Radiation Protection and Measurements Report 30) subroutine
to estimate marrow dose with 153Sm bound to the chelator EDTMP (Eary et al.
1993). In that case, values of 5.7 rad/mCi administered activity were calculated.
Since patients were treated with up to 3.0 mCi/kg, this led to red marrow
absorbed doses of up to 2250 rad or 22.50 Gy. Yet only 2 of 4 patients experi-
enced even mild hemotoxicity at this level. The authors comment that these
doses were almost certainly overestimated due to the ICRP 30 assumption that
all bone surfaces are in contact with marrow. Instead, the marrow is heteroge-
neously distributed within the bony matrix and has, particularly in the older
patient, large amounts of yellow (fatty) marrow admixed within its assumed
space inside the bone. In addition, many of these patients have been pretreated
with various chemotherapy agents that may also have led to a change in their
marrow mass.
A group at the University of Cincinnati has attempted to improve such bone
marrow absorbed dose estimates using pathology samples and a Monte-Carlo
approach (Samaratunga et al. 1995). They applied their method to 186Re(Sn)-
HEDP. In this case, S values were found to be considerably higher in the Monte-
Carlo method than that found using a homogeneous model of the soft-tissue
(lesion) distribution within the bone matrix. This work, entirely microscopic, did
not attempt to compare its results with the traditional MIRD organ-sized analyses
described in this report.
We would conclude that bone marrow dose estimation for bone-seeking
radiopharmaceuticals is not a simple exercise (Bayouth et al. 1995). As indicated
above, while new information is available on S matrix values, bone and marrow

49
biodistribution data are difficult to obtain. The latter result will probably not
improve in the near term. It is therefore tempting to validate instead the compu-
tations (if any) of marrow absorbed dose with clinical blood chemistry post-
therapy. This method will, at least, demonstrate a practical method for estimating
red marrow absorbed dose. It may also be compared to external beam results.

IX. RADIATION PROTECTION CONSIDERATIONS

We have previously discussed protection of the patient by use of dose estima-
tion techniques. We will now describe protection of personnel involved in prepar-
ing radiopharmaceuticals, caring for the patient, patient relatives, and the general
public. In particular, nursing staff should receive training on caring for patients
who have received radiolabeled antibodies and how to manage a spill of radioac-
tive material. Specific precautions required during administration of the radio-
pharmaceutical have been described in section IIIC.

A. General Radiation Safety

The following guidelines, which are similar to those for patients receiving 131I
therapy, should be observed. Visitors must be at least 18 years old. Pregnant
women are not permitted in the room. Patients should be confined to their rooms
except for special medical procedures. Patient meals must be ordered on isolation
trays and trays must remain in the room until checked by radiation safety person-
nel. The floor around the toilet must be covered with absorbent pads. A summary
of these guidelines is given in Table 5.
To reduce exposure, nursing staff should spend a minimal amount of time in
the room, i.e., only that required to administer satisfactory care to the patients. The
radiation safety officer (RSO) or staff physicist should post the maximum time on
the door. A personal dosimeter may be provided for hospital staff. In that case, a
logbook is placed at the entrance to the patient area and dosimeter readings are
recorded at the time of entrance and exit by each person entering. Shoe covers
must be worn and placed in a designated container on leaving the room. When in
the room, assume radioactivity is present in all bodily fluids and excretions.
Observe universal precautions and wash hands well when leaving the room.
Gloves are mandatory when drawing blood, handling urinals, bedpans, emesis
basins or other containers having any materials from the patients. Gloves and
sharps are to be disposed of in containers in the patient’s room. The toilet should
be flushed three times after use. If urine is to be collected, all urine collections
should be handled over disposable pads. Any contamination on container surface
should be wiped off. If urine is to be transferred, the container must be rinsed sev-
eral times with hot soapy water.

50
 Table 5. Radiation Safety Guidelines for Patients Receiving
Radionuclide Therapy

• Admit patient to single room

• Cover floor near toilet with absorbent pads; use pads for any procedure where con-
tamination is likely.
• Cover traffic areas with absorbent pads
• Cover door knobs, bed controls, and telephone handset with plastic wrap
• Confine patient to room except for special medical procedures
• Radiation placard on door
• Principles of time, distance, and shielding applied
• No visitors under 18 years of age.
• No pregnant visitors
• Universal precautions
• Assume all body fluids are radioactive
• Gloves and sharps disposed in container in patient’s room
• Flush toilet three times after each usage
• Isolation trays used
• Contaminated non-disposable materials stored in isolation until decay (10 half-lives).
• Specimens leaving room monitored for contamination.
• No housekeeping until patient discharged and room cleared by RSO or Staff
Physicist

B. Discharge of Patients with Radioactivity from Hospital

The RSO or staff physicist will determine when the patient is to be discharged.
Regulations depend upon the local license stipulations, but will be within general
guidelines specified by the controlling agency. For the Nuclear Regulatory
Commission (NRC), these rules have changed recently and now may be related
to the actual total effective dose (TEDE) to individuals in the patient’s home envi-
ronment (USNRC 1997). For the 20 states covered by the NRC, three alternative
release criteria are possible. These are:
(1) documented activity in the patient,
(2) exposure rate at one meter from the patient or
(3) estimated TEDE value to patient family members.
In the last case, the limit is 0.5 rem and the estimation must be made on a patient-
specific basis prior to release. In the case of the (1) or (2) criterion, the magnitude
of the maximum activity or exposure rate value at time of release depends upon
the radionuclide (USNRC 1997).

51
 For agreement states, criteria may differ from the NRC values. For example,
the previous federal rules permitted a maximum of 30 mCi of activity or 5 mR/h
at one meter from the patient at the time of release. These values are still in force
at many institutions. Zanzonico (1997) has shown that if hyperthyroid patients
were released with a mean activity of 29 mCi, their family members received an
average of 500 mrem. Both direct irradiation and radioiodine uptake were con-
sidered in these measurements which justified the older, more restrictive, regula-
tions for non-malignant disease.
Patients with thyroid malignancies, given that they will initially undergo thy-
roidectomy, will demonstrate much more rapid clearance of the radioiodine and
hence justify more liberal release criteria than that used for hyperthyroid patients.
It is unlikely that RIT patients will show such rapid clearance. Dehalogenation of
the radioiodine in a RIT protocol may allow more rapid excretion than with a
radiometal label on the antibody, however. Whole body and various organ reten-
tions of the antibody LYM-1 are given for a typical patient in Figure 12.

Figure 12. Whole body and organ retention data for a 131I-LYM-1 patient.

52
 When the patient is discharged, the room must be surveyed for contamination
before it is released for use by others. This process involves both room surveys
and the swabbing of surfaces to detect removable activity. Pure beta sources may
be counted using either Bremsstrahlung or, if energetic enough, Cerenkov radia-
tion. Calibration of the counter is provided by an aliquot of the injected activity.

X. SUMMARY OF RADIOIMMUNOTHERAPY
AND RADIONUCLIDE THERAPY
IN CLINICAL PRACTICE

A. Treatment Planning
RIT requires extensive staff involvement and equipment preparation. Nuclear
Medicine personnel must acquire uptake information from the major organ sys-
tems and perform integrations out to times on the order of 10 physical half-lives.
As we have seen, activity values in patients may be determined by one of several
methods with uncertainties between ±10% and ±30%. Generally, the GM strategy
has been the most commonly used with the requirement that two opposed images
be determined at each time point in the study. Alternatively, quantitative SPECT
or the CAMI method may be invoked if this condition is not met. Integration of
activity is best done with a physiological model that includes all the imaged com-
partments as well as fluid samples. Errors in this process are probably on the order
of ±2% to ±10%. Lastly, the provision of S values in the standard MIRD formula
requires that the target organ or tumor mass be determined with anatomic imag-
ing. If this is not possible, errors on the order of factors of two- or three-fold are
not unexpected due to the differences between true organ sizes and the MIRD
phantom values assumed (Liu et al. 1998).
Overall, one anticipates that at best the estimated absorbed dose in RIT trials
is correct to within ±20%. This result is much worse than the value usually given
for external beam therapy (±5%). Such comparisons are probably better done,
however, by looking at dose-volume histograms whereby the differences may not
be as striking. As the use of Monte Carlo methods increases, such histograms will
become common and allow us to better visualize the estimated dose distributions
provided by each modality.

B. Clinical RIT Results

At present, only two commercial RIT trials, both involving treatment of B-cell
lymphoma, have been initiated. A large number of local Phase I and Phase I/II pro-
tocols are underway for the treatment of blood diseases and solid tumors. While lym-
phomas have shown good response to RIT, little progress has been seen with solid
lesions. This result is viewed by some as being consistent with the ineffectiveness
seen with external beam therapy of the same diseases (e.g., colon cancer). Use of

53
RIT in the adjuvant setting is not yet in any widespread use. As we noted in section
VI.C.8, the smaller tumors are better targets for radiation therapy using internal emit-
ters. One can expect more RIT protocols following surgical removal of the primary
disease site(s). This application may eventually become their primary clinical use.
In the near term, one can anticipate that the number of trials will continue to
grow as more antibodies and their fragments are engineered. Certain applications,
such as the purging of diseased marrow using labeled antibodies, will increase and
may eventually supplant external beam treatments. It is possible that RIT will
evolve as a preferred modality for lymphoma therapy. The sensitivity of this dis-
ease to ionizing radiation and the localization of the absorbed dose derived from
RIT are the primary reasons for this prediction.
Patient-specific absorbed dose estimates will become necessary to understand
the effectiveness of all trials. While MIRD-type estimates will still be made for
the FDA, one expects the specific dose estimate, or treatment plan, will become
much more common. Such absorbed dose values will then be compared with the
tumor effects and normal organ toxicities observed in the clinic. Here, one may
find that the rad or grey dose estimated via a patient-specific calculation is not
directly comparable to external beam doses delivered over much shorter times
with extended rest intervals between treatments.
We also expect that there will be some effort to add radiation sensitizers to the
treatment regimen. Presumably given prior to RIT, these agents could lead to
greater cancer cell sensitivity to the beta (or alpha) radiation involved in the ther-
apy. In such work, the importance of physiological models will become much more
significant than we have indicated. Using such models and knowledge of the cell
cycle, one could hope to optimize the treatment of a given solid tumor.

C. Limitations of the RIT Method

Since the cancer patient can never be assumed permanently cured, ongoing
therapy is a requisite for any treatment system. At present, the appearance of
patient-derived antibodies to any administered protein may be the greatest limi-
tation to the RIT strategy. Engineering work on the design of these proteins is con-
tinuing, however, and will result in improved agents that elicit lower responses
from the patient’s immune system. In this work, the human backbone structure will
be included with initially an animal-derived recognition site. Eventually, the proteins
may become entirely human with consequently little chance of eliciting a human
antibody response.

D. Other Internal Emitter Therapies

The therapy provided by radioiodine for non-medullary thyroid cancer can be
anticipated to continue in the indefinite future. In addition to this technique, we may
anticipate other, rarer applications to increase. Among these will be the use of

54
MIBG (metaiodobenzylguanidine) in the treatment of neuroendocrine tumors
(Shapiro et al. 1992) and the various bone pain agents described above. In the former
case, diseases such as pheochromocytoma may be containable with sufficient
activity levels of 131I-MIBG. In the latter example, one imagines at least a con-
tinuing therapy strategy to alleviate the patient’s discomfort. No other modality
seems to be able to provide this service which implies a continuing increase in the
number of such clinical therapies.

REFERENCES
Axelsson, B., P. Msaki, and A. Israelsson (1984). “Subtraction of Compton-
scattered photons in single photon emission computed tomography.” J. Nucl.
Med. 25: 490–494.

Bayouth, J. E., D. J. Macey, A. L. Boyer, and R. E. Champlin (1995). “Radiation

dose distribution within the bone marrow of patients receiving Ho-166 labeled
phosphonate for marrow ablation.” Med. Phys. 22: 743–753.

Beatty, J. D., B. G. Beatty, L. E. Williams, R. J. Paxton, J. E. Shively, M.

O’Connor-Tressel (1989). “Effect of specific antibody pretreatment of liver
uptake of In-111-labeled anticarcinoembryonic antigen monoclonal antibody in
nude mice bearing human colon cancer xenografts.” Cancer Res. 49: 1587–1594.

Berger, M. J. “Distribution of Absorbed Dose Around Point Sources of Electrons

and Beta Particles in Water and Other Media” in Journal of Nuclear Medicine
Pamphlet No. 7. New York: Society of Nuclear Medicine, pp. 5–23, 1971.

Berger, M. J., and S. M. Seltzer. “ETRAN, Monte Carlo Code System for
Electron and Photon Transport Through Extended Media” in Oak Ridge
National Laboratory Documentation for RSIC Computer Code Package
CCC-107. Oak Ridge, TN: ORNL, 1973.

Bolch, W. E., L. G. Bouchet, J. S. Robertson, B. W. Wessels, J. A. Siegel, R. W.

Howell, A. K. Erdi, B. Aydogan, S. Costes, and E. E. Watson (1999). “MIRD
pamphlet no. 17: The dosimetry of nonuniform activity distributions—
radionuclide S values at the voxel level.” J. Nucl. Med. 40: 11S–36S.

Breitz, H. B., D. R. Fisher, P. L. Weiden, J. S. Durham, B. A. Ratliff, M. J.

Bjorn, P. L. Baumier, and P. G. Abrams (1993). “Dosimetry of 186-rhenium
labeled monoclonal antibodies: Methods, prediction from 99m-Tc-labeled
antibodies and results of phase I trials.” J. Nucl. Med. 34: 908–917.

55
Breitz, H. B., P. L. Welden, J. L. Vanderheyden, J. W. Appelbaum, M. J. Bjorn,
M. F. Fer, S. B. Wolf, B. A. Ratliff, C. A. Seiler, D. C. Foisie, D. R. Fisher, R.
W. Schroff, A. R. Fritzberg, and P. G. Abrams (1992). “Clinical experience with
186-rhenium labeled monoclonal antibodies for radioimmunotherapy: Results
of phase I trials.” J. Nucl. Med. 33: 1099–1112.

Briesmeister, J. (ed.). “MCNP- A General Monte Carlo N-particle Transport

Code Version 4A.” Los Alamos Report LA-12625. Los Alamos, NM: Los
Alamos National Lab (LANL), 1993.

Brigham, E. O. The Fast Fourier Transform. Englewood Cliffs, NJ: Prentice-

Hall, 1974.

Buras, R. R., L. E. Williams, B. G. Beatty, J. Y. C. Wong, and J. D. Beatty

(1994). “A method including edge effects for the estimation of radioimmuno-
therapy absorbed doses in the tumor xenograft model.” Med. Phys. 21:
287–292.

Buvat, I., M. Rodriguez-Villafuerte, A. Todd-Pokropek, H. Banali, and

R. Di Paola (1995). “Comparative assessment of nine scatter correction
methods based on spectral analysis using Monte Carlo simulations.” J. Nucl.
Med. 36: 1476–1488.

Chaney, E. W., and D. Kinkaid. Numerical Mathematics and Computing.

Monterey, CA: Brooks/Cole, pp. 268–286, 1985.

Chang, L. T. (1980). “A method for attenuation correction in radionuclide

computed tomography.” IEEE Trans. Nuc. Science NS 25: 363.

D’Argenio, D. Z., and A. Schumitzky (1979). “A program package for

simulation and parameter estimation in pharmacokinetic systems.” Comput.
Prog. Biomed. 9: 115–134.

deKlerk, J. M. H., A. van Dijk, A. D. van het Schip, B. A. Zonnenberg, and P. P.

van Rijk (1992). “Pharmacokinetics of rhenium-186 after administration of
rhenium-186-HEDP to patients with bone metastases.” J. Nucl. Med. 33:
646–651.

DeNardo, G. L., S. J. DeNardo, L. F. O’Grady, N. B. Levy, G. P. Adams, and

S. L. Mills (1990). “Fractionated radioimmunotherapy of B-cell malignancies
with I-131 Lym-1.” Cancer Res. 50 (Suppl): 1014s–1016s.

56
Eary, J. F., C. Collins, M. Stabin, C. Vernon, S. Petersdorf, M. Baker, S.
Hartnett, S. Ferency, S. J. Addison, F. Appelbaum, and E. E. Gordon (1993).
“Samarium-153-EDTMP biodistribution and dosimetry estimation.” J. Nucl.
Med. 34: 1031–1036.

Eary, J. F., O. W. Press, C. C. Badger, L. D. Durack, K. Y. Richter, S. J. Addison,

K. A. Krohn, D. R. Fisher, B. A. Porter, D. L. Willliams, P. J. Martin, F. R.
Appelbaum, S. L. Brown, R. A. Miller, W. B. Nelp, and I. W. Bernstein (1990).
“Imaging and treatment of B-cell lymphoma.” J. Nucl. Med. 31: 1257–1268.

Ford, R. L., and W. R. Nelson. “The EGS Code System.” Stanford Linear
Accelerator Report SLAC-210. Stanford, CA: Stanford University, 1978.

Foster, D. M., and R. C. Boston. “The Use of Computers in Compartmental

Analysis: The SAAM and CONSAM Programs” in Compartmental
Distribution of Radiotracers. J. S. Robertson (ed.). Boca Raton, FL: CRC
Press, pp. 73–81, 1983.

Gambir, S. V., D. K. Mahoney, G. Rosenquest, M. S. Turner, and A. Wong.

SIMPLE (Symbolic Interactive Modeling Package and Learning Environment).
UCLA, 1991–1994.

Harbert, J. C. Nuclear Medicine Therapy. New York: Thieme Medical,

pp. 37–89, 1987.

Howell, R. W., D. V. Rao, and K. S. R. Sastry (1989). “Macroscopic dosimetry

for radioimmunotherapy: Nonuniform activity distributions in solid tumors.”
Med. Phys. 16: 66–74.

Hurley, H. R., and D. V. Becker. “The Use of Radioiodine in the Management

of Thyroid Cancer” in Nuclear Medicine Annual. L. M. Friedman and H. S.
Weiss (eds.). New York: Raven Press, pp. 329–384, 1983.

Ichihara, T., K. Ogawa, N. Motomura, A. Kubo, and S. Hashimoto (1993).

“Compton scatter compensation using the triple-energy window method for
single- and dual-isotope SPECT.” J. Nucl. Med. 34: 2216–2221.

Jacobs, A. J., M. Fer, F. M. Su, H. Breitz, J. Thompson, H. Goodgold, J. Cain,

J. Heap, and P. Weiden (1993). “A phase I trial of a rhenium-186-labeled
monoclonal antibody administered intraperitoneally in ovarian carcinoma:
toxicity and clinical response.” Obstetrics Gynecology 82: 586–593.

57
Jaszczak, R. J., K. L. Greer, C. E. Floyd, C. C. Harris, and R. E. Coleman
(1984). “Improved SPECT quantification using compensation for scattered
photons.” J. Nucl. Med. 25: 893–900.

Johnson, T. K. (1988). “MABDOS: A generalized program for internal

radionuclide dosimetry.” Comput. Meth. Progr. Biomed. 27: 159–167.

Kaplan, D. D., L. E. Williams, K. G. Clarke, T. L. Odom-Maryon, A. Liu, G.

Lopatin, A. A. Raubitschek, and J. Y. C. Wong (1997). “Estimating residence
times and their associated errors in patient absorbed dose calculation.” J. Nucl.
Med. Tech. 25: 264–268.

Koral, K. F., F. M. Swailem, S. Buchbinder, N. H. Clinthorne, W. L. Rogers,

and B. M. W. Tsui (1990). “SPECT dual-energy-window Compton
correction: scatter multiplier required for quantification.” J. Nucl. Med. 31:
90–98.

Lewington, V. J. (1996). “Cancer therapy using bone-seeking isotopes.” Phys.

Med. Biol. 41: 2027–2042.

Leichner, P. K. (1992). “A unified approach to photon and beta particle

dosimetry.” J. Nucl. Med. 33: 1721–1729.

Leichner, P. K., J. L. Klein, J. B. Garrison, R. E. Jenkins, E. L. Nickoloff, D. S.

Ettinger, and S. E. Order (1981). “Dosimetry of I-131 labeled anti-ferritin in
hepatoma. A model for radioimmunoglobulin dosimetry.” Int. J. Radiat.
Oncol. Biol. Phys. 7: 323–333.

Liu, A., L. E. Williams, and A. A. Raubitschek (1996). “A CT assisted method

for absolute quantitation of internal radioactivity.” Med. Phys. 23:
1919–1928.

Liu, A., L. E. Williams, J. Y. C. Wong, and A. A. Raubitschek (1998). “Monte

Carlo assisted voxel source kernel method (MAVSK) for internal beta
dosimetry.” Nucl. Med. Biol. 25: 423–433.

Ljungberg, M., M. A. King, G. J. Hademenos, and S.-E. Strand (1994).

“Comparison of four scatter correction methods using Monte Carlo simulated
source distributions.” J. Nucl. Med. 35: 143–151.

Ljungberg, M., and S.-E. Strand (1990). “Scatter and attenuation correction in
SPECT using density maps and Monte Carlo simulated scatter functions.”
J. Nucl. Med. 31: 1560–1567.

58
Loevinger, R., and M. A. Berman. “A Revised Schema for Calculating the
Absorbed Dose from Biologically Distributed Radionuclides” in MIRD
Pamphlet No. 1 Revised. New York: Society of Nuclear Medicine, 1976.

Loevinger, R., E. M. Japha, and G. L. Brownell. “Discrete Radioisotope

Sources” in Radiation Dosimetry. G. J. Hine and G. L. Brownell (eds.). New
York: Academic, pp. 823–824, 1956.

Loevinger, R., T. F. Budinger, and E. E. Watson. MIRD Primer for Absorbed

Dose Calculations, Revised Edition. New York: Society of Nuclear Medicine,
1991.

Macey, D. J., and R. Marshall (1982). “Absolute quantitation of radiotracer

uptake in the lungs using a gamma camera.” J. Nucl. Med. 23: 731–734.

Macey, D. J., G. L. DeNardo, and S. J. DeNardo (1991). “A treatment planning

program for radioimmunotherapy.” Front. Radiat. Ther. Oncol. 24: 123–131.

Macey, D. J., S. J. DeNardo, and G. L. DeNardo (1995). “Estimation of

radiation absorbed dose for the red marrow from I-131 radiolabeled antibodies.”
Clin. Nuc. Med. 20: 117–125.

Macey, D. J., S. J. DeNardo, G. L. DeNardo, J. L. Goodnight, and M. W. Unger

(1988). “Uptake of In-111 labeled monoclonal antibody ZME018 as a function
of tumor size in a patient with melanoma.” Am. J. Physiol. Imaging 3: 1–6.

Maksud, P., B. Fertil, C. Rica, G. El Fakhri, and A. Aurengo (1998). “Artificial

neural network as a tool to compensate for scatter and attenuation in
radionuclide imaging.” J. Nucl. Med. 39: 735–745.

McCowen, K. D., R. A. Adler, N. Ghaed, T. Verdon, and F. D. Hofeldt (1976).

“Low dose radioiodide thyroid ablation in postsurgical patients with thyroid
cancer.” Am. J. Med. 61: 52–58.

Meredith, R. F., T. K. Johnson, G. Plott, D. J. Macey, R. L. Vessella, L. A.

Wilson, H. B. Breitz, and L. E. Williams (1993). “Dosimetry of solid tumors.”
Med. Phys. 20: 583–592.

Order, S. E., J. A. Siegel, R. A. Lustig, R. Principato, L. S. Zeiger, E. Johnson,

H. Zhang, P. Lang, and P. E. Waller (1994). “Infusional brachytherapy in the
treatment of non-resectable pancreatic cancer: A new radiation modality
(Preliminary report of the phase I study).” Antibod. Immunoconj.
Radiopharm. 7: 11–27.

59
Patt, Y. Z., L. M. Lamki, T. P. Haynie, M. W. Unger, M. G. Rosenblum,
A. Shirkhada, and J. L. Murray (1988). “Improved tumor localization with
increasing dose of indium-111-labeled anti-carcinoembryonic antigen
monoclonal antibody ZCE-025 in metastatic colorectal cancer.” J. Clin. Oncol.
6: 1220–1230.

Press, O., J. Eary, C. C. Badger, P. J. Martin, F. R. Appelbaum, R. Levy, R.

Miller, S. Brown, W. B. Nelp, K. A. Krohn, D. Fisher, K. DeSantes, B. Porter,
P. Kidd, E. D. Thomas, and I. D. Bernstein (1989). “Treatment of refractory
non-Hodgkin’s lymphoma with radiolabeled MB-1 (Anti-CD37) antibody.”
J. Clin. Oncol. 7: 1027–1038.

Riethmuller, G., E. Schneider-Gadicke, G. Schlimok, W. Schmiegel, R. Raab,

K. Hoffken, R. Gruber, H. Pichlmaier, H. Hirche, R. Pichlmayr, P. Buggisch,
J. Witte (1994). “Randomised trial of monoclonal antibody for adjuvant therapy
of resected Dukes’C colorectal carcinoma.” Lancet 343: 1177–1183.

Riva, P., A. Arista, V. Tison, C. Sturiale, G. Francheschi, A. Spinelli, N. Riva,

M. Casi, G. Moscatelli, and M. Frattarelli (1994). “Intralesional radioimmuno-
therapy of malignant gliomas: An effective treatment in recurrent tumors.”
Cancer 73 Suppl.: 1076–1082.

Roberson, P. L., D. B. Heidorn, M. L. Kessler, R. K. Ten Haken, and D. J.

Buchsbaum (1994). “Three-dimensional reconstruction of monoclonal antibody
uptake in tumor and calculation of beta dose-rate nonuniformity.” Cancer 73:
912–918.

Rosenthal, M. S., J. Cullom, W. Hawkins, S. C. Moore, B. M. W. Tsui, and

M. Yester (1995). “Quantitative SPECT imaging: A review and
recommendations by the focus committee of the society of nuclear medicine
computer and instrumentation council.” J. Nucl. Med. 36: 1489–1513.

Samaratunga, R. C., S. R. Thomas, J. D. Hinnefeld, L. C. Von Kuster, D. M.

Hyams, J. S. Moulton, M. I. Sperling, and H. R. Maxon (1995). “A Monte Carlo
simulation model for radiation dose to metastatic skeletal tumor from Rhenium-
186(Sn)-HEDP.” J. Nucl. Med. 36: 336–350.

Sgouros, G. (1993). “Bone marrow dosimetry for radioimmunotherapy:

theoretical considerations.” J. Nucl. Med. 34: 689–694.

Shapiro, B., J. C. Sisson, D. M. Wieland, T. J. Mangner, S. M. Zempel,

E. Mudgett, M. D. Gross, J. E. Carey, K. R. Zasadny, W. H. Beierwaltes (1992).

60
“Radiopharmaceutical therapy of malignant pheochromocytoma with 131-I-
metaiodobenzylguanidine: Results from ten years of experience.” J. Nucl. Biol.
Med. 35: 269–276.

Shen, S., G. L. DeNardo, D. J. Macey, R. T. O’Donnell, A. Yuan, D. A.

DeNardo, S. J. DeNardo (1997). “Practical determination of organ S values for
individual patients for therapy.” Nucl. Med. Biol. 24: 447–449.

Siddiqui, A. R., J. Edmondson, H. N. Wellman, R. C. Hamaker, R. E. Lingeman,

H.-M. Park, and C. C. Johnston (1981). “Feasibility of low doses of 131-I for
thyroid ablation in postsurgical patients with thyroid carcinoma.” Clin. Nucl.
Med. 6: 158–161.

Siegel, J. A. Thyroid Handling of Iodine-123 and Technetium-99m-

pertechnetate. Ph.D. Thesis, UCLA, 1981.

Siegel, J. A., S. Whyte-Ellis, L. S. Zeigler, S. E. Order, and P. E. Walner (1994).

“Bremsstrahlung SPECT imaging and volume quantitation with 32Phosphorus.”
Antibody Immunoconj. Radiopharm. 7: 1–10.

Siegel, J. A., B. W. Wessels, E. E. Watson, M. G. Stabin, H. M. Vriesendorp,

E. W. Bradley, C. C. Badger, A. B. Brill, C. S. Kwok, D. R. Stickney, K. F.
Eckerman, D. R. Fisher, D. J. Buchsbaum, and S. E. Order (1990). “Bone
marrow dosimetry and toxicity in radioimmunotherapy.” Antibod.
Immunoconj. Radiopharm. 3: 213–233.

Siegel, J. A., S. R. Thomas, J. B. Stubbs, M. G. Stabin, M. T. Hays, K. F.

Koral, J. S. Robertson, R. W. Howell, B. W. Wessels, D. R. Fisher, D. A.
Weber, and A. B. Brill (1999). “MIRD Pamphlet No. 16: Techniques for
quantitative radiopharmaceutical biodistribution data acquisition and analysis
for use in human radiation dose estimates.” J. Nucl. Med. 40: 37S–61S.

Snyder, W. S., W. R. Ford, G. G. Warner, and S. B. Watson. “S Absorbed Dose

Per Unit Cumulated Activity for Selected Radionuclides and Organs.” MIRD
Pamphlet No. 11. New York: Society of Nuclear Medicine, 1975.

Sorenson, J. A. “Quantitative Measurement of Radioactivity In-Vivo by Whole

Body Counting” in Instrumentation in Nuclear Medicine. Vol. 2. G. J. Hine
and J. A. Sorenson (eds.). New York: Academic Press, 1974.

Stabin, M. G. (1996). “MIRDOSE: Personal computer software for internal

dose assessment in nuclear medicine.” J. Nucl. Med. 37: 538–546.

61
Stabin, M. G., K. F. Eckerman, J. C. Ryman, and L. E. Williams (1994).
“Bremsstrahlung radiation dose in yttrium-90 therapy applications.” J. Nucl.
Med. 35: 1377–1380.

Thomas, S. R., H. R. Maxon, and J. G. Kereiakes (1976). “In-vivo quantitation

of lesion radioactivity using external counting methods.” Med. Phys. 3:
253–255.

Thomas, S. R., R. H. Maxon, and J. G. Kereiakes. “Techniques for Quantitation

of In Vivo Radioactivity” in Effective Use of Computers in Nuclear
Medicine. M. J. Gelfand and S. R. Thomas (eds.). New York: McGraw-Hill,
pp. 468–484, 1988.

United States Nuclear Regulatory Commission (USNRC). Release of Patients

Administered Radioactive Materials. Washington D.C.: USNRC, 1997.

van Rensberg, A. J., M. G. Lotter, A. du P. Heyns, and P. C. Minnaar (1988).

“An evaluation of four methods of In-111 planar image quantification.” Med.
Phys. 15: 853–861.

Wagner, J. G. Fundamentals of Clinical Pharmacology. Hamilton, IL: Drug

Intelligence Publications Inc., pp. 231–246, 1975.

Watson, E. E., M. G. Stabin, and W. E. Bolch. Documentation Package for

MIRDOSE (Version 2). Oak Ridge, TN: Oak Ridge Associated Universities,
1984.

Williams, L. E., J. Y. C. Wong, D. O. Findley, and B. W. Forell (1989).

“Measurement and estimation of organ Bremsstrahlung radiation dose.”
J. Nucl. Med. 30: 1373–1377.

Williams, L. E., L. M. Ettinger, E. Cofresi, A. A. Raubitschek, and J. Y. C.

Wong (1995a) “A shielded, automated injector for energetic beta-emitting
radionuclides.” J. Nucl. Med. Tech. 23: 29–32.

Williams, L. E., R. B. Bares, J. Fass, S. Hauptmann, V. Schumpelick, and

U. Buell (1993). “Uptake of radiolabeled anti-CEA antibodies in human
colorectal primary tumors as a function of tumor mass.” Eur. J. Nucl. Med. 20:
345–347.

Williams, L. E., T. L. Odom-Maryon, A. Liu, A. Chai, A. A. Raubitschek, J. Y.

C. Wong, and D. Z. D’Argenio (1995b). “On the correction for radioactive
decay in pharmacokinetic modeling.” Med. Phys. 22: 1619–1626.

62
Williams, L. E., R. B. Duda, R. T. Proffitt, B. G. Beatty, J. D. Beatty, J. Y. C.
Wong, J. E. Shively, and R. J. Paxton (1988). “Tumor uptake as a function of
tumor mass: A mathematic model.” J. Nucl. Med. 29: 103–109.

Wong, J. Y. C., D. Z. J. Chu, D. Yamauchi, T. L. Odom-Maryon, L. E. Williams,

A. Liu, J. M. Esteban, A. M. Wu, F. J. Primus, J. D. Beatty, J. E. Shively, and
A. A. Raubitschek (1998). “Dose escalation trial of indium-111-labeled
anti-carcinoembryonic antigen chimeric monoclonal antibody (cT84.66) in
pre-surgical colorectal cancer patients.” J. Nucl. Med. 39: 2097–2104.

Wu, R. K., and J. A. Siegel (1984). “Absolute quantitation of radioactivity using

the buildup factor.” Med. Phys. 11: 189–192.

Zanzonico, P. (1997). “Radiation dose to patients and families incident to I131

therapy.” Thyroidology 7: 193–198.

63
 APPENDIX A

Human Residence Time Estimates Made with Murine Data

The CR factor is defined in eq. (8). It is based on a relative perfusion argument.
If the model animal’s organ mass is the same fraction of the total body mass as is
the case in the human, CR is unity. We consider a sample liver calculation using
mouse biodistribution data for the antibody mT84.66.
Using the PLOT routine from MIRDOSE2, mouse liver data (in fractional
injected dose per gram or FID/g) was fitted with a single exponential function.
The amplitude was 0.0778/g with a T1/2b of 995 h. Notice that we are analyzing
data corrected for the decay of the radiolabel. Using eq. (6), the effective half
life became 63.56 h for an 111In label. Given that value, specific murine residence
time for the liver was 7.14 h/g. Since the mouse average body mass was 21.4 gram,
the modified CR factor of eq. (8) became 0.56 g when a human liver was assumed
to be 1833 g and a total body human mass was taken as 70 kg. Notice that we have
included the animal organ mass in the uptake value so that CR is modified accord-
ingly. Multiplying the modified CR value times the specific murine residence time
of 7.14 h/g yielded a human residence time estimate of 4.00 h [τliver (human)].
Table A1 gives a summary of these results for other normal organs.
Taking this τliver value and using the MIRD Pamphlet11 table of S values
resulted in a human liver-to-liver dose estimate of 0.52 rads/mCi of administered
111
In-labeled mT84.66 antibody. Other estimates (not shown) would follow
accordingly.

Table A1. Sample Human Residence Time Estimates Made

with Murine Data for 111In-mT84.66
Organ Tb Tef FID/g τ (mouse)/g CR(g) τ (human)
Blood 142 h 46 h 0.309 20.50 h 1.65 33.83 h
Liver 996 63.6 0.078 7.14 0.56 4.00
Spleen 1280 64.5 0.102 9.49 0.05 0.51
Kidneys 343 56.7 0.098 8.00 0.09 0.70
Lungs 191 50.1 0.151 10.90 0.30 3.27
Stomach 426 58.6 0.031 2.62 0.12 0.32
Bone 1429 64.8 0.033 3.09 3.05 9.42
Muscle Infinite 67.9 0.011 1.07 9.17 9.79

64
 APPENDIX B

Tumor Dose Estimation

Tumor FIA (fractional injected activity) was determined over a sequence of time
points. Since tumor was not made part of the Figure 6 model (relatively few patients
showed tumor uptake), integration of the FIA was carried out using a trapezoidal
method. Size of the lesion (5.4 × 4.0 × 3.0 cm) was taken directly from the radiol-
ogist’s reading of the CT scan. The resulting mass was determined with the approx-
imation [eq. (16)] that each lateral dimension was an ellipsoidal diameter. The
resultant volume was 33.9 ml or 33.9 gram if a unit density object.
For 90Y, the absorbed fraction of the beta radiation for this spherical size was
determined via Monte Carlo simulation as 0.91. Here, because of a lack of distri-
bution information, uniform uptake in the metastatic site was assumed. Given the
0.43 h residence time (cf. Table C3 in Appendix C) calculated using a trapezoidal
integration, the resultant dose estimate was 22.1 rad/mCi or 6.0 mGy/MBq.

65
 APPENDIX C

Organ Data Acquired and Dose Estimates

We consider the chimeric anti-CEA antibody cT84.66 in a clinical trial (#91169).
The patient (#14) had primary colon cancer with several metastatic lesions. Anterior
and posterior counts were obtained at several time points using 111In-cT84.66.
Background was taken from an area of muscle (thigh). Table C1 contains the raw
data for liver and thigh. We subtract background on the basis of counts per pixel. At
0.71 h post-injection (elapsed time), anterior and posterior images were acquired
over 12 min using a dual-headed gamma camera.

Table C1. Partial List of Anterior and Posterior Counts Obtained with a
Whole Body Gamma Camera

Elapsed Time Anterior Posterior

Organ (h) Pixels Counts Counts
Liver (and blood) 0.71 4018 301355 254356
Soft Tissue 0.71 703 7286 6773

Using CT data, it was seen that the patient thickness (T) at the liver was 26 cm
while the liver thickness () was 16 cm. By previous attenuation measurements with
an 111In source, it was determined that the linear attenuation coefficient (µ) was
0.116 cm−1. We subtract the background (cf. Figure 3) via:

Liver (net counts/pixel) = Liver (raw counts/pixel) – Background (counts/pixel)

For the anterior projection, this difference amounted to 64.6 counts/pixel; the pos-
terior net count value was 53.7 counts/pixel. Upon multiplying by 4018 pixels for the
liver, we obtain totals of 259 × 103 and 216 × 103 counts for anterior and posterior
projections in 12 min. Their geometric mean is 237 × 103 counts in 12 minutes. The
two geometric factors exp(µT/2) and (µ/2)/sinh(µ/2) are 4.39 and 0.869 respec-
tively. Substituting into eq. (11) with a sensitivity of 7063 × 103 counts/(12 min IA),
we obtain a fractional injected activity (FIA) of 0.128 for the liver. We should note
that since this organ is imaged in vivo, the liver contains blood as well as hepatic
tissue. Thus, the model contains both tissues for this image data.
Table C2 contains a summary of FIA for organ systems that were imaged or
sampled via well counter assay. In the latter case, an aliquot of the injected dose
was used to determine the FIA.

66
 Table C2. Fractional Injected Activities (FIAs) for 111In-cT84.66
Elapsed Time Whole Body Liver Spleen Heart Tumor

0.71 h 0.993 0.128 0.027 0.168 0.0017

19.86 0.820 0.128 0.020 0.114 0.0031
42.86 0.616 0.113 0.016 0.076 0.0033
143.32 0.170 0.047 0.003 0.015 0.0010

A five-compartment model (Figure 6) was fitted to the above data set using
SAAM II. The resultant residence times are given in Table C3. Notice that the
model, since it contains decay as an explicit output channel from each compart-
ment, predicts values for both 111In and 90Y labels on cT84.66.

Table C3. Residence Times for 111In-cT84.66 and 90Y-cT84.66

Whole Red Residual
Body Liver Spleen Heart Marrow Tumor Body
111
In- cT84.66 83.80 h 16.70 h 2.15 h 10.94 h 4.49 h 0.43 h 46.70 h
90
Y- cT84.66 80.70 15.90 2.08 10.62 4.33 0.41 45.10

We then substitute these residence times into MIRDOSE3 and obtain the values
shown in Table C4.

Table C4. Estimated Absorbed Doses for cT84.66

Whole Heart Red
Body Liver Spleen Wall Marrow
111
In-cT84.66 0.47 rad/mCi 2.39 2.27 2.73 0.62
90
Y-cT84.66 2.00 16.60 22.60 24.00 4.11
111
In-cT84.66 0.13 mGy/MBq 0.65 0.62 0.74 0.17
90
Y-cT84.66 0.54 4.48 6.10 6.48 1.11

67
 APPENDIX D

Red Marrow Data and Dose Estimation

Red marrow dose was estimated using the blood time-activity curve, meas-
ured with iv sampling, as a surrogate for the unmeasurable red marrow activity
vs. time curve via eq. (20). We assumed an f factor of 0.3 in these computations
with the MIRD assumptions that the plasma mass is 5.0 kg and the red marrow
mass is 1.5 kg. Since we did not observe uptake in bone, we had only two source
organs for 90Y-cT84.66: red marrow (as represented by blood) and residual body.
The corresponding residence times (4.33 h and 45.10 h) are listed in Table C3. In
the case of the 111In-cT84.66, all organs having measurable uptake were included
as source organs.
Resulting red marrow dose estimates were then 0.62 and 4.11 rad/mCi (0.17
and 1.11 mGy/MBq) for 111In-cT84.66 and 90Y-cT84.66 respectively as shown in
Table C4. These were based, as mentioned above, on the standard human whole
blood and marrow masses. If a patient-specific red marrow dose estimate is
desired, individualized mass values should be entered for these two parameters.
Because of earlier therapies the red marrow mass is, however, difficult to predict
for an individual RIT patient.

68
AAPM REPORT SERIES ($10.00 Each)
No. 1 “Phantoms for Performance Evaluation and Quality Assurance of CT
Scanners” (1977)
No. 3 “Optical Radiations in Medicine: A Survey of Uses, Measurement and
Sources” (1977)
No. 4 “Basic Quality Control in Diagnostic Radiology,” AAPM Task Force
on Quality Assurance Protocol (1977)
No. 5 “AAPM Survey of Medical Physics Training Programs,” Committee
on the Training of Medical Physicists (1980)
No. 6 “Scintillation Camera Acceptance Testing & Performance Evaluation,”
AAPM Nuclear Medicine Committee (1980)
No. 7 “Protocol for Neutron Beam Dosimetry,” AAPM Task Group #18
(1980) (FREE)
No. 8 “Pulse Echo Ultrasound Imaging Systems: Performance Tests &
Criteria,” P. Carson & J. Zagzebski (1980)
No. 9 “Computer-Aided Scintillation Camera Acceptance Testing,” AAPM
Task Group of the Nuclear Medicine Committee (1982)
No. 10 “A Standard Format for Digital Image Exchange,” Baxter et al. (1982)
No. 11 “A Guide to the Teaching of Clinical Radiological Physics to Residents
in Radiology,” AAPM Committee on the Training of Radiologists
(1982)
No. 12 “Evaluation of Radiation Exposure Levels in Cine Cardiac
Catherization Laboratories,” AAPM Cine Task Force of the Diagnostic
Radiology Committee (1984)
No. 13 “Physical Aspects of Quality Assurance in Radiation Therapy,” AAPM
Radiation Therapy Committee Task Group #24, with contribution by
Task Group #22 (1984)
No. 14 “Performance Specifications and Acceptance Testing for X-Ray
Generators and Automatic Exposure Control Devices” (1985)
No. 15 “Performance Evaluation and Quality Assurance in Digital Subtraction
Angiography,” AAPM Digital Radiology/ Fluorography Task Group
(1985)
No. 16 “Protocol for Heavy Charged-Particle Therapy Beam Dosimetry,”
AAPM Task Group #20 of the Radiation Therapy Committee (1986)

69
No. 17 “The Physical Aspects of Total and Half Body Photon Irradiation,”
AAPM Task Group #29 of the Radiation Therapy Committee (1986)
No. 18 “A Primer on Low-Level Ionizing Radiation and its Biological
Effects,” AAPM Biological Effects Committee (1986)
No. 19 “Neutron Measurements Around High Energy X-Ray Radiotherapy
Machines,” AAPM Radiation Therapy Task Group #27 (1987)
No. 20 “Site Planning for Magnetic Resonance Imaging Systems,” AAPM
NMR Task Group #2 (1987)
No. 21 “Specification of Brachytherapy Source Strength,” AAPM Radiation
Therapy Task Group #32 (1987)
No. 22 “Rotation Scintillation Camera Spect Acceptance Testing and Quality
Control,” Task Group of Nuclear Medicine Committee (1987)
No. 23 “Total Skin Electron Therapy: Technique and Dosimetry,” AAPM
Radiation Therapy Task Group #30 (1988)
No. 24 “Radiotherapy Portal Imaging Quality,” AAPM Radiation Therapy
Task Group #28 (1988)
No. 25 “Protocols for the Radiation Safety Surveys of Diagnostic Radiological
Equipment,” AAPM Diagnostic X-Ray Imaging Committee Task
Group #1 (1988)
No. 26 “Performance Evaluation of Hyperthermia Equipment,” AAPM
Hyperthermia Task Group #1 (1989)
No. 27 “Hyperthermia Treatment Planning,” AAPM Hyperthermia Committee
Task Group #2 (1989)
No. 28 “Quality Assurance Methods and Phantoms for Magnetic Resonance
Imaging,” AAPM Nuclear Magnetic Resonance Committee Task
Group #1 (1990)
No. 29 “Equipment Requirements and Quality Control for Mammography,”
AAPM Diagnostic X-Ray Imaging Committee Task Group #7 (1990)
No. 30 “E-Mail and Academic Computer Networks,” AAPM Computer
Committee Task Group #1 (1990)
No. 31 “Standardized Methods for Measuring Diagnostic X-Ray Exposures,”
AAPM Diagnostic X-Ray Imaging Committee Task Group #8 (1991)
No. 32 “Clinical Electron-Beam Dosimetry,” AAPM Radiation Therapy
Committee Task Group #25 (1991)

70
No. 33 “Staffing Levels and Responsibilities in Diagnostic Radiology,”
AAPM Diagnostic X-Ray Imaging Committee Task Group #5 (1991)
No. 34 “Acceptance Testing of Magnetic Resonance Imaging Systems,”
AAPM Nuclear Magnetic Resonance Task Group #6 (1992)
No. 35 “Recommendations on Performance Characteristics of Diagnostic
Exposure Meters,” AAPM Diagnostic X-Ray Imaging Task Group #6
(1992)
No. 36 “Essentials and Guidelines for Hospital Based Medical Physics
Residency Training Programs,” AAPM Presidential AD Hoc
Committee (1992)
No. 37 “Auger Electron Dosimetry,” AAPM Nuclear Medicine Committee
Task Group #6 (1993)
No. 38 “The Role of the Physicist in Radiation Oncology,” Professional
Information and Clinical Relations Committee Task Group #1 (1993)
No. 39 “Specification and Acceptance Testing of Computed Tomography
Scanners,” Diagnostic X-Ray Imaging Committee Task Group #2
(1993)
No. 40 “Radiolabeled Antibody Tumor Dosimetry,” AAPM Nuclear Medicine
Committee Task Group #2 (1993)
No. 41 “Remote Afterloading Technology,” Remote Afterloading Technology
Task Group #41 (1993)
No. 42 “The Role of the Clinical Medical Physicist in Diagnostic Radiology,”
Professional Information and Clinical Relations Committee Task
Group #2 (1993)
No. 43 “Quality Assessment and Improvement of Dose Response Models,”
$25, (1993).
No. 44 “Academic Program for Master of Science Degree in Medical
Physics,” AAPM Education and Training of Medical Physicists
Committee (1993)
No. 45 “Management of Radiation Oncology Patients with Implanted Cardiac
Pacemakers,” AAPM Task Group #4 (1994)
No. 46 “Comprehensive QA for Radiation Oncology,” AAPM Radiation
Therapy Committee Task Group #40 (1994)
No. 47 “AAPM Code of Practice for Radiotherapy Accelerators,” AAPM
Radiation Therapy Task Group #45 (1994)

71
No. 48 “The Calibration and Use of Plane-Parallel Ionization Chambers for
Dosimetry of Electron Beams,” AAPM Radiation Therapy Committee
Task Group #39 (1994)
No. 49 “Dosimetry of Auger-Electron-Emitting Radionuclides,” AAPM
Nuclear Medicine Task Group #6 (1995)
No. 50 “Fetal Dose from Radiotherapy with Photon Beams,” AAPM Radiation
Therapy Committee Task Group #36 (1995)
No. 51 “Dosimetry of Interstitial Brachytherapy Sources,” AAPM Radiation
Therapy Committee Task Group #43 (1995)
No. 52 “Quantitation of SPECT Performance,” AAPM Nuclear Medicine
Committee Task Group #4 (1995)
No. 53 “Radiation Information for Hospital Personnel,” AAPM Radiation
Safety Committee (1995)
No. 54 “Stereotactic Radiosurgery,” AAPM Radiation Therapy Committee
Task Group #42 (1995)
No. 55 “Radiation Treatment Planning Dosimetry Verification,” AAPM
Radiation Therapy Committee Task Group #23 (1995), $48, (Includes
2 disks, ASCII format). Mail, fax, or phone orders to: AAPM
Headquarters, One Physics Ellipse, College Park, MD 20740-3846,
Phone: (301) 209-3350, Fax: (301) 209-0862
No. 56 “Medical Accelerator Safety Considerations,” AAPM Radiation
Therapy Committee Task Group #35, Reprinted from Medical Physics,
Vol. 20, Issue 4, July/August 1993
No. 57 “Recommended Nomenclature for Physical Quantities in Medical
Applications of Light,” AAPM General Medical Physics Committee
Task Group #2 (1996)
No. 58 “Managing the Use of Fluoroscopy in Medical Institutions,” AAPM
Radiation Protection Committee Task Group #6 (1998)
No. 59 “Code of Practice for Brachytherapy Physics,” AAPM Radiation
Therapy Committee Task Group #56, Reprinted from Medical Physics,
Vol. 24, Issue 10, October 1997
No. 60 “Instrumentation Requirements of Diagnostic Radiological Physicists,”
AAPM Diagnostic X-Ray Committee Task Group #4 (1998)
No. 61 “High Dose Brachytherapy Treatment Delivery,” AAPM Radiation
Therapy Committee Task Group #59, Reprinted from Medical Physics,
Vol. 25, Issue 4, April 1998

72
No. 62 “Quality Assurance for Clinical Radiotherapy Treatment Planning,”
AAPM Radiation Therapy Committee Task Group #53, Reprinted from
Medical Physics, Vol. 25, Issue 10, October 1998
No. 63 “Radiochromic Film Dosimetry,” AAPM Radiation Therapy
Committee Task Group #55, Reprinted from Medical Physics, Vol. 25,
Issue 11, November 1998
No. 64 “A Guide to the Teaching of Clinical Radiological Physics To
Residents in Diagnostic and Therapeutic Radiology,” Revision of
AAPM Report #11, AAPM Committee on the Training of Radiologists,
January 1999
No. 65 “Real-Time B-Mode Ultrasound Quality Control Test Procedures,”
AAPM Ultrasound Task Group #1, Reprinted from Medical Physics,
Vol. 25, Issue 8, August 1998
No. 66 “Intravascular Brachytherapy Physics,” AAPM Radiation Therapy
Committee Task Group #60, Reprinted from Medical Physics, Vol. 26,
Issue 2, February 1999
No. 67 “Protocol for Clinical Reference Dosimetry of High-Energy Photon
and Electron Beams,” AAPM Task Group #51, Reprinted from
Medical Physics, Vol. 26, Issue 9, September 1999
No. 68 “Permanent Prostate Seed Implant Brachytherapy,” AAPM Medicine
Task Group #64, Reprinted from Medical Physics, Vol. 26, Issue 10,
October 1999
No. 69 “Recommendations of the AAPM on 103Pd Interstitial Source
Calibration and Dosimetry: Implications for Dose Specification and
Prescription,” Report of the Low Energy Interstitial Brachytherapy
Dosimetry Subcommittee of the Radiation Therapy Committee, In
progress (2001)
No. 70 “Cardiac Catherization Equipment Performance,” Task Group #17
Diagnostic X-ray Imaging Committee, February 2001
No. 71 “A Primer for Radioimmunotherapy and Radionuclide Therapy,” Task
Group #7 Nuclear Medicine Committee, April 2001

73
Return to: Medical Physics Publishing
4513 Vernon Blvd.
Madison, WI 53705-4964
(800) 442-5778 phone
(608) 265-2121 FAX

Name: __________________________________________________________

Address: _________________________________________________________

_________________________________________________________________

________ copies of_____________________________________ @_________

________ copies of_____________________________________ @_________

________ copies of_____________________________________ @_________

Prepayment required
U.S. check or International Money Order
Amount enclosed $_____