dEPA
United States
Environmental Protection
Agency
Office of Pesticides Programs
Washington, DC 20460
EPA-540/9-85-019
June 1985
Hazard Evaluation Division
Standard Evaluation Procedure
Oncogenicity Potential (Guidance for
Analysis and Evaluation of Long
Term Rodent Studies)
....
-------
EPA-540/9-85-019
June 1985*
HAZARD EVALUATION DIVISION
STANDARD EVALUATION PROCEDURE
ONCOGENICITY POTENTIAL:
GUIDANCE FOR ANALYSIS AND EVALUATION OF LONG TERM RODENT STUDIES
Prepared by
Orville E. Paynter, Ph.D., D.A.B.T.
Standard Evaluation Procedures Project Manager
Stephen L. Johnson
Hazard Evaluation Division
Office of Pesticide Programs
United States Environmental Protection Agency
Office of Pesticide Programs
Washington, D.C. 20460
* August 9, 1984 (Revised December 1, 1984)
This temporary revision is to be used until public comments
on the EPA Proposed Guidelines for Carcinogen Risk Assessment
(2) are evaluated and are adopted in final form.
-------
STANDARD EVALUATION PROCEDURE
PREAMBLE
This standard Evaluation Procedure (SEP) is one of a set of
guidance documents which explain the procedures used to evaluate
environmental and human health effects data submitted to the
Office of Pesticide Programs. The SEPS are designed to ensure
comprehensive and consistent treatment of major scientific topics
in these reviews and to provide interpretive policy guidance
where appropriate. The Standard Evaluation Procedures will be
used in conjunction with the appropriate pesticide .Assessment
Guidelines and other Agency Guidelines. While the documents were
developed to explain specifically the principles of scientific
evaluation within the Office of pesticide Programs, they may also
be used by other offices in the Agency in the evaluation of
studies and scientific data. The Standard Evaluation Procedures
will also serve as valuable internal reference documents and will
inform the public and regulated community of important consider-
ations in the evaluation of test data for determining chemical
hazards. I believe the SEPs will improve both the quality of
science within EPA and, in conjunction with the pesticide Assess-
ment Guidelines, will lead to more effective use of both public
and private resources.
>hn W. Melone, Director
Hazard Evaluation Division
-------
TABLE OF CONTENTS
Page
PREAMBLE
I. ANALYSIS FOR ONCOGENIC POTENTIAL IN EXPERIMENTAL
ANIMALS
A. Introduction 5
B. Definition of Chemical Oncogenicity 12
C. Documentation and Data Acceptability 14
D. Major Oncogenic Considerations 20
1. Spontaneous Incidence Data 20
2. Rare Neoplasms 26
3. Benign/Malignant Neoplasms 29
4. Degree of Oncogenicity 37
5. Dose-Response Relationships 43
6. Decreased Latency 65
E. Auxiliary Evidence 69
1. Mutagenicity Data 69
2. Metabolic/Pharmacodynamic Data 73
F. Completion of Analysis 76
II. EVALUATION AND CLASSIFICATION OF EVIDENCE
REFERENCES .* 102
-------
TABLES AND FIGURES
No SUBJECT Page
1 INCIDENCE (PERCENT) OF FEMALE CONTROL 42
RATS BEARING THYROID C-CELL TUMORS AMONG
ANIMALS SACRIFICED POST 12-MONTHS
(Same Lab)
2 HISTORICAL CONTROL INCIDENCE OF LUNG TUMORS 42
IN MALE B6C3Fi MICE RECEIVING CORN OIL BY
GAVAGE (Different Labs)
3 EXAMPLES OF NCI USE OF HISTORICAL CONTROL DATA 43
4 SUMMARY OF SPONTANEOUS TUMORS OBSERVED UPON 44
RE-EXAMINATION OF SERIAL SECTIONS OF SELECTED
TISSUES FROM 177 (63 MALES, 114 FEMALES)
SPRAGUE-DAWLEY RATS
5 TUMORS AND ORGANS OF ORIGIN IN 2y082 RATS OF 44
6 SOURCES—CONTINUED
6 PROLIFERATIVE CHANGES AND THEIR SYNONYMS IN 45
RAT LIVER
7 NEOPLASMS (TUMORS, NEW GROWTHS) ONE OF THE 46
DEFINITIONS; NEOPLASM IS AN UNCONTROLLED
GROWTH OF CELLS
8 POTENTIAL DIFFERENCES BETWEEN CHEMICALLY-INDUCED 47
AND TUMORS IN CONTROL RODENTS
9 SAMPLE INCIDENCE TABLE 48
10 SUMMARY INCIDENCE TABLE 49
11 GUIDELINES FOR COMBINING BENIGN AND MALIGNANT 50
NEOPLASMS IN THE FISCHER 344 RAT AND B6C3F1
MOUSE
12 TOTAL NUMBER OF RATS WITH NEOPLASMS (ALL TYPES) 54
13 INCIDENCE OF NEOPLASMS IN MICE KILLED AT END OF 55
TEST
14 MOST COMMONLY INDUCED TUMORS IN THE 98 POSITIVE 56
NCI BIOASSAYS
-------
TABLES AND FIGURES CONT.
NO
15 SELECTED TUMOR INCIDENCE PATTERNS 57
CO
16 TYPICAL STANDARD CARCINOGENS 30
-------
Preamble
For government agencies concerned with public health and/or
environmental effects, one of the most difficult tasks is the
identification and regulation of potential (suspect human)
oncogenic substances, be they food additives, drugs, pesticides
or industrial chemicals. In the Environmental Protection
Agency, the regulatory process is based on scientific evidence,
and socioeconomic considerations. Within EPA, the Office of
Pesticides and Toxic Substances (OPTS) is responsible for the
evaluation of the scientific evidence relative to health risk
assessments for pesticides and industrial chemicals.
The most persuasive evidence of potential oncogenicity
in man comes from competently designed and conducted human
epidemiology studies supported by appropriate animal studies.
However, the most frequently seen evidence is based on long-
term tests in laboratory animals such as mice and rats. In
vivo and in vitro short-term studies (e.g., mutagenicity),
biochemical reactivity information, and metabolic and pharma-
codynamic studies provide additional and sometimes critical
evidence.
This document may not contain anything new or revolutionary
regarding the evaluation of oncogenic potential evidenced by
toxic substances. However, it is the first time, as far as I
know, that the weight of scientific evidence concept, as
being developed by OPTS, and major considerations for analysis
-------
of animal oncogenicity data have been brought together in
one place. This document treats complex issues, incompletely
and in some cases superficially. However, it represents a
base from which OPTS might further develop and strengthen
its hazard identification and risk evaluation abilities in
this complex area.
It is possible that evaluation of the strength of biological
and auxiliary evidence for oncogenicity, prior to the mathematical
calculation of risk, will prevent the untimely polarization
of opinion and interject a higher degree of understanding and
confidence into the hazard identification and risk assessment
process. It is also possible that the standardization of
data organization and evaluation procedures might increase
the efficiency, thoroughness and consistency with which
individual evaluations are prepared and help the evaluators
reach considered and scientifically defensible judgments.
To this end, all references are intended to be an integral
part of the guidance offered by this document and they should
be read.
It must be understood by evaluators that their major
responsibility is the competent analysis, evaluation, and
interpretation of biological and toxicological data according
to sound scientific principles. Therefore, evaluators must not
-------
allow their deliberations to be overly influenced by the
ambiguities of controversial concepts or by their perception
of what potential regulatory decisions or actions their
evaluations may portend. The latter pitfall may be ameliorated
by recognizing that the scientific identification and assessment
of risk is a separate and distinct function from risk management
[i.e., regulatory decision making].1
Three things are certain, no matter how well these
guides are followed: 1) they will not automatically produce
a "carcinogenic" vs. "non-carcinogenic" decision; 2) substitute
for sound scientific judgment; nor 3) prevent criticism or
controversy regarding the judgments made and the conclusions
drawn. However, a review which exhibits internal evidence
of: a) being based on sound scientific principles; b) present-
ing succinct and cogent rationale for judgments and conclusions;
c) presenting quoted material, (i.e., text or tabular) with
proper citation; and d) having been competently and objectively
performed, will require critics to focus their arguments
with equal competence, completeness, and succinctness.
-------
As more experience in use of the weight-of-evidence
procedure is obtained, contents of this document should be
modified and improved. To this end all recommendations will
be most welcome. I am very indebted to my colleagues who
have previously performed this valuable service.
•p g^\-
r&/v3£>t)
Orville E. Paynter, Ph.D., D.A.B.T.
Hazard Evaluation Division
Office of Pesticide Programs
8/9/84
-------
I. Analysis for Oncogenic Potential in Experimental Animals
A. Introduction
The Environmental Protection Agency (EPA) has published
a request for public comment on Proposed Guidelines for
Carcinogen Risk Assessment.2 The purpose of the proposal
is to incorporate the concepts and approaches to oncogenic
risk assessment which have been developed since the 1976
Interim Procedures and Guidelines were issued.3 Although
the_1976 guidelines will be eventually superceded, they are
briefly discussed here because they provide a very clear
picture of the EPA regulatory process as well as an informative
statement of the weight-of-evidence concept.
The 1976 guidelines described the decision process
regarding the regulation of potential oncogens as being two-
phased. The first phase is the determination that a particular
substance constitutes an oncogenic risk. The second phase is
the determination of what regulatory action, if any, should
be taken to reduce that risk. Accordingly, they state: "The
central purpose of the health risk assessment is to provide a
judgment concerning the weight-of-evidence that an agent is a
potential human carcinogen and, if so, how great an impact it
is likely to have on public health" [underlining added].4
-------
In addition the guidelines leave no doubt that an analysis
of health risks must be separate and independent from any
consideration of the socioeconomic consequences of regulatory
action. This is also true of the proposed guidelines.
The preamble of the 1976 guidelines clarifies the meaning
of the "weight-of-evidence" concept.
In considering the risks, it will be necessary
to view the evidence of carcinogenicity in terms
of a warning signal, the strength of which is a
function of many factors including those relating
to the quality and scope of the data, the character
of the toxicological response, and the possible
impact on public health. It is understood that
qualifications relating to the strength of the
evidence for carcinogenicity may be relevant
to this consideration because of the uncertainties
in our knowledge of the qualitative and quantitative
similarities of human and animal responses. In
all events, it is essential in making decisions
about suspect carcinogens that all relevant
information be taken into consideration.5
The weight of biological evidence concept used in this
Evaluation Procedure is that part of the assessment process
which considers and weighs the cumulative observational and
experimental data pertinent to arriving at a level of concern
about a substance's oncogenic potential for humans. it is
-------
composed of a series of judgments concerning the adequacy,
validity, and appropriateness of the observational and
experimental methods used to produce the data base and, those
judgments which bring into causal, complementary, parallel, or
reciprocal relationships all the data considered. Because
our knowledge concerning oncogenic mechanisms is still
developing, because good epidemiological evidence is
seldom available and because animal studies are not always
conclusive, all of the information available at a given time
may provide only "persuasive evidence" (i.e., not clearly
robust or feeble, yet suggestive of a defensible presumption)
one way or the other about the human oncogenic potential of a
given substance. It is for this reason that both guidelines
stress the importance of succinctly articulating the rationale
for judgments and conclusions contained in risk assessment
and the uncertainties pertaining thereto. This becomes
important when new data or new scientific knowledge requires
reevaluation of the data base or a change in a previous risk
assessment or regulatory action.
The 1984 Proposed Guidelines describe in general terms,
the essential aspects of risk assessment and present salient
principles which are the foundation for evaluating biological
and other types of data relating to suspect carcinogens.
They embrace the scientific principles of carcinogenic
assessment developed by the Office of Science and Technology
-------
Policy [OSTP]6 and with modifications, the concept of
risk assessment developed by the National Research Council
[NRC]! and the weight-of-evidence scheme developed by the
the International Agency for Research on Cancer [IARC].
The Proposed Guidelines describe the method to be used
for evaluating studies thus:
Studies are evaluated according to sound biological
and statistical considerations and procedures. These
have been described in several publications [(6) and
(8) through (17)]. Results and conclusions concerning
the agent, derived from different types of information,
whether indicating positive or negative responses,
are melded together into a weight-of-evidence
determination. The strength of the evidence supporting
a potential human carcinogenicity judgment is developed
in a weight-of-evidence stratification scheme.
Pertinent parts of the weight-of-evidence stratification
scheme which is an adaptation of the IARC approach to EPA
needs, is presented in Part II of this document. This scheme
and the guidance provided by the references cited in the
above quote must be used within the Agency for all oncogenicity
study evaluations. This will assure the desirable uniformity
of procedures and conformity to the Agency's prescribed
philosophy for analysis, evaluation, interpretation, and
-------
classification of data generated by such studies. For these
reasons it behooves every evaluator to understand, and strictly
adhere to the guidance presented in the Proposed Guidelines
and the cited references. Any discrepancy between the
general approach presented by the 1984 Proposed Guidelines
and the guidance offered in this Evaluation Procedure must
always be reconciled in favor of the Proposed Guidelines.
The NRC describes risk assessment as containing some or
all of the following steps or components: 1) hazard identifi-
cation; 2) dose-response assessment; 3) exposure assessment;
and 4) risk characterization.^ Guidance offered by this
Evaluation Procedure is confined entirely to the NRC hazard
identification component. This is in keeping with the Proposed
Guidelines which place the other three components within the
dominion and under the aegis of scientists skilled in the
quantitative aspects of health risk assessments. This is
clear from the following Proposed Guideline discussion of the
i v
four components [also see Part III. A. 1. of Reference 2]:
Hazard identification is a qualitative risk assessment,
dealing with the process of determining whether ex-
posure to an agent has the potential to increase the
incidence of cancer. For purposes of these guidelines,
malignant and benign tumors are used in the evaluation
-------
10
of the carcinogenic hazard. The hazard identifi-
cation component qualitatively answers the
question of how likely an agent is to be a human
carcinogen.
The dose-response assessment defines the relationship
between the dose of an agent and the probability of
induction of a carcinogenic effect. This component
usually entails an extrapolation from the generally
high doses administered to experimental animals
or exposures noted in epidemiological studies to the
exposure levels expected from human contact with the
agent in the environment; it also includes consider-
ation of the validity of these extrapolations.
The exposure assessment identifies populations
exposed to the agent, describes their composition
and size, and presents the types, magnitudes,
frequencies, and duration of exposure to the
agent.
In risk characterization the output of the exposure
assessment and the dose-response assessment are
combined to estimate quantitatively some measure
of the carcinogenic risk. As part of risk
characterization, a summary of the strengths and
-------
11
weaknesses in the hazard identification, dose-
response assessment, exposure assessment, and
the public health risk estimates are presented.
Major assumptions, scientific judgments, and, to
the extent possible, estimates of the uncertainties
embodied in the assessment are also presented, dis-
tinguishing clearly between fact, assumption and
science policy.
For discussion of the elements of hazard identification
and the types of data which are relevant to this component
see Parts II, B. 1. through 7. of the Proposed Guidelines.2
In keeping with the Proposed Guidelines and to bring a
sharper focus on the analysis and evaluation of experimental
animal data, OPTS adopts the following definitions and major
considerations as part of its oncogenic hazard identification
procedures.
-------
12
B. Definition of Chemical Oncogenicity.
The International Agency for Research on Cancer (IARC)
has developed the following widely accepted meaning
of the term chemical carcinogenesis.
Chemical carcinogenesis...is [a] the induction
by chemicals of neoplasms that are not usually
observed, [b] the earlier induction by chemicals
of neoplasms that are commonly observed,
and/or [c] the induction by chemicals of more
neoplasms than are usually found--although
fundamentally different mechanisms may be
involved in these three situations.18
OPTS has adopted this meaning as its working definition
with one exception. The term "oncogenicity" is substituted
for the term "carcinogenicity." Carcinogenicity, etymologically,
means induction of malignant neoplasms. The above definition
does not make a distinction between benign and malignant
neoplasms and OPTS should not do so in a working definition.
In the evaluation of health risk, the nature and incidence
of all types of neoplasms and their possible interrelationships
should be considered. Therefore, the term "oncogenicity" is
deemed more appropriate.
-------
13
For the sake of clarity, some implications in this definition
need to be made explicit. All sections of the IARC definition,
imply that the evaluator has knowledge of the types of neoplasms
"usually or commonly observed" in the animals used in the study
being evaluated, including knowledge that goes beyond the
information derived from the concurrent control animals for a
particular study or group of studies ,(i. e. , historical control
data). Part II. B.6. of the Proposed Guidelines also assumes
familiarity with relevant historical control data. Situation
(a) of the IARC definition implies that the reviewer has
knowledge of those neoplasms which are not usually observed
or found, i.e., rare or unusual neoplasms. Situation (b) implies
that OPTS should obtain or derive, and use, data which bears on
the time of tumor appearance (decreased latency or precocity
of onset) and will know when each neoplasm type is usually
expected to appear or be discovered in the animals used.
Finally, the definition implies that OPTS should try to identify
data (e.g., mutagenicity, metabolism, and biochemical reactivity
data) suggestive of mechanisms which may be involved in the
neoplasia produced by the specific chemical.
The Proposed Guidelines relating to hazard identification,
the above working definitipn and its implications, and the
collective experience of OPTS are the foundation for development
of the following considerations.
-------
14
C. Documentation and Data Acceptance
The quality, integrity, and completeness of reporting
observational and experimental data are essential to the
proper analysis and evaluation of study results. In essence,
the "good science" evaluations expected of OPTS have their
foundations in the evaluated evidential documentation.
Therefore, qualitative evaluation of the documentation of a
particular study or group of studies is of special significance
to the acceptability of data.
The following three important considerations address the
acceptability of long-term rodent studies.
1. The adequacy of the experimental design and other
experimental parameters such as: the route of administration;
frequency and duration of exposure; appropriateness of the
species, strain, sex, and age of the animals used; choice of
dosage levels; appropriateness of the observational and
experimental methods; and the conditions under which the
substance was tested.
There are no specific, internationally agreed upon
scientific rules or fixed check lists which make the judgment
regarding the acceptability of data bases a standard routine
procedure. However, there are suggested guidelines concerning
the mechanics of good experimental design, reporting, and
good laboratory practice which are aids to evaluation of
-------
15
report and data acceptability. These may be found in the EPA
suggested guidelines and the EPA and FDA Good Laboratory Prac-
tices Regulations. The Proposal Guidelines2 state that criteria
for the technical adequacy of animal studies can be found in
prescribed publications6' 10' 14' 15' 16' 17' 19' 20, 21
and that these should be used to judge the acceptability
of studies. The evaluator needs to be cautious when using
the above guidance as aids to making an acceptability judgment
for any oncogenicity study. The cardinal question to be
answered is how well does the study, in toto, facilitate the
identification of a potential oncogenic effect or lack thereof
for the substance being evaluated, and not how precisely does
the study fit a prescribed recipe for performance. The
collective experience of OPTS scientists can be very helpful
in resolving difficult questions of acceptability and should
be utilized whenever it is needed.
As the first step in the evaluation process, the evaluator
should carefully read through the report including supporting
data presentations, and make a tentative classification of
acceptability prior to making a detailed evaluation of the
individual data. If there are obvious and significant defi-
ciencies in the report, any further work would be a waste of
resources. The submitter of the report should be notified
of the problem(s) as quickly and as accurately as possible
and any further analysis suspended until these deficiencies
are corrected.
-------
16
Frequently, the subsequent detailed analysis of the data
will bring to light deficiencies which were not obvious during
the initial reading of the report. In this case the deficiencies
should be noted and the evaluation completed as far as possible.
At this time the submitter of the document should be made
aware of the situation along with any scientific questions or
other data needs identified during the detailed data analysis
and evaluation.
2. The competency and completeness with which the study or
studies were conducted and reported.
Doubts on the part of an evaluator regarding completeness
and/or competency with which a study was performed or reported
must be discussed with the evaluator's supervisor. If these
concerns are judged to be reasonable at this level, the study
should be nominated for a laboratory and data base audit.
Any further consideration of the study should be suspended
until the audit is completed and the results evaluated.
3. The effects of modifying factors such as differential
survival, toxicity, or disease which result in major inequalities
between control and treated animals.
This qualitative consideration has more to do with the
evaluation and interpretation of data than with the
-------
17
acceptability of documentation. However, it is placed here
because determination of the various factors influencing
toxicological and oncological data, as may be indicated in
the evidential documentation, needs to be made prior to
application of the major oncogenicity conderations.
There are many factors influencing the response of
experimental animals to chemical substances. Some of these
are discussed by Doull^Z and his presentation of this
subject should be reviewed. Of special interest in oncogenicity
data evaluation are the factors contained in this qualitative
consideration.
Differential survival in any animal group, regardless
of its cause, has an important bearing on the evaluation and
interpretation of oncogenicity studies. An apparent unequal
reduction, real or illusory, of the number of animals at
risk in oncogenicity studies is a complicating influence and
may lead to serious misinterpretation of a substance's oncogenic
potential. Therefore, it is essential to determine, early
in the review period, if this factor has any significance
for the proper application of the major oncogenicity considera-
tions. This determination may not always be a simple routine
[for an example, see reference 79] and the services of a competent
statistician should be obtained in the case of doubt or
controversy. Because of the importance of this factor, time
-------
18
to death or sacrifice, preferably in days, of each animal
should be presented by or obtained from the report submitter.
Such information is useful in certain statistical procedures
(e.g., lifetable method) and may be useful in evaluation of
time to tumor data (see major consideration # 6).
As with differential survival/mortality, the presence
of toxic or pharmacological effects or disease processes can
complicate the evaluation and interpretation of data and,
depending on severity, can cause the study to be of very
limited value for evaluation of the oncogenic potential of a
substance. The effects of these factors can be particularly
troublesome when they are confused with or misinterpreted as
preneoplastic lesions. Examples of these types of problems
can be found in references 80 and 81. Problems related to
the above factors should be resolved prior to application of
the major oncogenicity considerations.
The three qualitative considerations for documentation and
data acceptance discussed above are applicable to all experi-
mental animal studies, no matter what their intended purpose
might be, and essentially establish the acceptability not
only of specific reports but also the acceptability of the
eventual evaluation, interpretation and judgments based upon
them.
-------
19
Resolution of problems related to qualitative or
quantitative considerations is not entirely the responsibility
of the individual evaluator. The submitter of the evidential
documentation may be requested to assist. For particularly
difficult problems, the assistance of consultants and the
Scientific Advisory Panel for pesticides and the Science
Advisory Board for other chemicals may be utilized. Requests
for the latter types of assistance must be through OPTS
management.
The acceptability of reports and other technical information
submitted to OPTS is a scientific judgment and only secondarily
a legal one. Therefore, OPTS bears the burden of defending and
documenting the acceptance or rejection, in part or in whole,
of the evidential documentation and data. The submitters of
information deserve to know the rationale for any rejection
of data. This rationale should be succinctly stated in the
evaluation document.
-------
20
D. Major Considerations for Analysis and Evaluation of Oncogenicit^
1. Spontaneous neoplasm incidence in untreated animals
(concurrent and historical controls).
It is well known among experienced pathologists and
toxicologists that the incidence of spontaneous lesions,
including neoplasms, is unpredictably variable among groups
of concurrent controls in the same study as well as among
control groups of the same strain from different studies and
laboratories.23 Tables 1 and 2 present observed examples
of both situations. Such variation is frequently encountered
and oftentimes complicates the evaluation and interpretation
of toxicity studies in general and oncogenicity studies in
particular. Some of the difficulty in interpretation can
frequently be ameliorated by the judicious consideration of
historical control (spontaneous incidence) data. Such data
should be viewed as an auxiliary aid to interpretation of
study data. It should not be used as a complete substitution
for concurrent control data within a particular study or
group of studies.
The Task Force of Past Presidents of the Society of
Toxicology gives the following examples of how historical
data may be useful.
-------
21
The following propositions may be taken as
scientifically useful in the evaluation of
a chemical carcinogenic response, with
distinctions drawn between the use of con-
current control and historical control
data. (1) If the incidence rate in the
concurrent control group is lower than in
the historical control groups, but the
incidence rates in the treated groups
are within the historical control range,
the differences between treated and control
groups are not biologically significant.
(2) If the incidence rates in the treated
groups are higher than the historical
control range but not statistically signifi-
cantly greater than the concurrent control
incidence> the conclusion would be that there
is no relation to treatment, but with the
reservation that this result could be a
false negative resulting from some flaw.
(3) If the incidence rates in the treated
groups are significantly greater than in
the concurrent controls, and greater than
the historical control range, a treatment
effect may be present which is unlikely to
be a false positive test.24
-------
22
Table 3 presents two actual examples of how the National
Cancer Institute (NCI) has used spontaneous incidence data
(historical control data) as an aid to interpretation of
treatment relationships of particular study lesions. These
examples were chosen because they approximate situations 1)
and 2) presented above by the Task Force of Past Presidents.
Note that historical control data was not substituted for
concurrent control data in either of these examples.
The best historical control data are obtained using the
same species and strain, from the same supplier, maintained
under the same general conditions in the same laboratory
which generated the study data being evaluated. The data
should be from control animals on recent, (5 years but
no later than 10 years) consecutive, long-term oncogenicity
studies. However, even this type of data can be misleading
if not properly organized, evaluated, and interpreted.
It is highly desirable to obtain data from individual
groups of control animals, in order to establish a range of
values, rather than from combined groups of animals yielding
only a single mean value. On this matter, the Task Force of
Past Presidents presented the following examples as a word of
caution:
-------
23
Historical data are often presented as the
incidence of tumor in hundreds of control
animals. Statistical procedures can be used
to relate this overall incidence to the
incidence in a specific study. However, this
leaves much to be desired since the incidence
of tumors can vary considerably between
groups of animals. Thus, in 11 carcinogenic
studies in rats (Charles River Caesarian
Derived) [Sprague Dawley] where there are two
or three concurrent control groups of animals,
the incidence of brain tuitions varied from 1%
to 10% in male rats in three concurrent
control groups. The female rats had no brain
tumors. In other control groups of male
rats, the incidence of brain tumors varied
from 0 to 4%. This type of variation is not
apparent if the incidence in combined control
animals is used.
In another example, the overall incidence of
pheochromocytomas in 1,100 control male rats
was 2% but the variation among groups was 0 to
28%. In one of these studies, the incidence
of pheochromocytoma in male rats was 8% in one
control group, 28% in the concurrent second
-------
24
control group and 14, 18 and 26% in the low,
middle and high dose groups of treated animals,
respectively. If only the control group with
8% incidence of pheochromocytomas had been
used, there would have been a significant
difference between the control group and the
high dose group, and the presence of an upward
trend would have resulted in the conclusion
that the chemical was a carcinogen for male
rats. Obviously, this was not the case.24
It is also necessary to be cautious concerning what is
really represented by tabulated incidence data, spontaneous
incidences or otherwise. Sometimes investigators combine
certain types of neoplasms when presenting tumor incidences in
summary tables. Usually a careful reading of the text accompanying
the summary data table will indicate where tumor combination has
occurred.
To avoid entrapment in these types of pitfalls, evaluators
should specifically request that historical control data be
presented for each neoplasm as discrete control group incidences,
segregated by sex, and updated with each new study submission.
It is also highly desirable that additional information
on each discrete control group be made available. This informa-
tion should include the following:
-------
25
a. Identification of species and strain and name of
the supplier including specific colony identification
if supplier has more than one geographical location.
b. Name of the laboratory in which the study was
performed, and when;
c. Description of general conditions under which the
animals were maintained, including the type or brand of diet,
and type of bedding if possible;
d. The planned duration of the study (e.g., 18 months
or 2 years) and the approximate age, in days, of the control
animals at the beginning of the study and at the time of
killing or death;
e. Description of the control group mortality pattern
observed during or at the end of the study and of any other
pertinent observations (e.g., diseases, infections, etc.);
f. Name of the pathology laboratory and examining
pathologist responsible for gathering and interpreting the
pathological data from the study; and
g. What tumors may have been combined to produce any
of the incidence data.
-------
26
Examples of how historical control data was used in
resolving problems related to a disease incidence may be
found in reference 81. An example of a contrary view of his-
torical incidence (control) data use has been articulated.82
2. Presence and incidence of neoplasms not usually
observed (rare or unusual neoplasms).
The terms "rare" and "unusual" when applied to scientific
observations or events are generally understood to mean
infrequently occurring or not ordinarily or commonly encountered.
Thus the terms are used synonymously for purposes of applying
this major oncogenicity consideration to the data base.
Statistical analysis of lesions observed with low
frequency* in a particular study presents difficult methodolo-
gical and interpretational problems and may be of extremely
limited usefulness as an aid to judging the rarity of such
lesions and their relationships to treatment. Therefore, the
* For purposes of this discussion the criteria used by NCI—
primary tumors occurring in two animals or less in one of the
control or treated groups of 50 animals each and/or where such
tumors are observed in less than 5% of the group,—defines
low frequency.25
-------
27
evaluator must seek another guidepost for the attribution of
the terra "rare" or "unusual" and treatment relationship to the
lesions of low incidence observed in a specific study or
group of studies. Historical control (spontaneous incidence)
data are most useful in this situation. However there is one
particularly important pitfall which must be recognized.
If the standards for applying these attributes to specific
types of toxicologic or oncologic observations or lesions were
static, there would be much less difficulty in the evaluative
and interpretative processes. However, as the opportunity
for observation increases with time or the development and
use of more precise or sensitive methods of observation and
detection, the rarity and unusualness of an event may remain
relatively stable or may slowly or quite rapidly change.
One example of this phenomenon is presented by the 1963
work of Thompson and Hunt.27 The authors decided to re-
examine, "by serial section techniques, representative organs
in which neoplasms, that were not grossly apparent, had been
originally detected upon microscopic examination of randomly
selected single tissue sections." The results of the two
examinations are presented in Table 4.
The most impressive change in tumor incidence occurred in
the thyroid light-cell (C-cell) adenoma. By use of a more
precise technique, the combined observed incidence (all rats)
-------
28
rose from 9/140 (6%) to 55/140 (39%). The incidence for
males and females each rose about six-fold. Since the
pathologists were the same, the increased incidence in C-cell
adenomas can only be explained by the increased opportunity
for observation provided by the serial sectioning technique.
The authors concluded that spontaneous light-cell adenomas
occur with about equal frequency in both sexes of the Sprague-
Dawley strain and, (more importantly for the present illus-
tration) "that this type of tumor is far more common in the
rat than previous reports might suggest."
By contrast, in this study, the brain tumor incidence
did not change because of the increased opportunity for
observation. The incidence for all tumor types and for each
type individually remained 4/126 (3.2%) and 1/126 (0.8%)
respectively. From these data one could conclude that the
observed frequency of brain tumors in the Sprague-Dawley rat
is low and therefore they are not usually or commonly en-
countered (i.e., rare). This conclusion is supported by the
1973 work, a decade later, of MacKenzie and Garner.28 These
pathologists examined various tissues of six rat strains from
different sources. Serial sectioning methods were not used.
The results for brain tumors only are presented in Table 5.
In this study the incidence of brain tumors observed in
the Sprague-Dawley rat was also 0.8% (2/258). The incidence
of all brain tumors among the six strains was 17/2082 (also
0.8%) and ranged from 3/535 (0.6%) for the Charles River-SD
-------
29
rats to 4/217 (1.8%) lor the Diablo-SD rats. See reference 29
for further discussion of brain tumor incidence.
In the case of brain tumors in rats, the incidence has
remained relatively stable with increased time and increased
opportunity for observation. From such types of studies and
their own experience pathologists have reached the consensus
that brain tumors in rats might be considered for the present
rare and unusual neoplasms. For this reason they require
special attention during the evaluation process as do all
other lesions exhibiting this attribute.
The knowledge that shifts in observed spontaneous incidences
for some lesions does occur with increased opportunity for
observation and increased sensitivity of detection should not
be a major impediment to use of historical control data,
especially if such data are continually updated.
References 83 and 84 provide examples of problems
concerning rarity of a tumor or groups of tumors and how
historical incidence (control) data were useful in their
resolution.
3. Increased incidence of benign and/or malignant
neoplasms that are usually found.
-------
30
The pathologist has a unique position in toxicological and
oncological evaluations. Evaluators are usually entirely
dependent on such individuals for descriptions of the variety
of spontaneous and treatment/disease induced lesions present
at any time during a study. This is especially true regarding
the distinction between benign and malignant neoplasms.
Zbinden, in discussing this subject, points out the special
role of the pathologist in providing information on the
morphology of lesions and emphasizes that these data will also
establish the presence or absence of a dose-effect relationship
for some of the lesions. This information is obviously critical
to the establishment of toxic effects produced by the substance and
its oncogenic potential. Zbinden briefly discusses the use of
semi-quantitative methods as well as more accurate morphometric
methods for rating the severity of lesions, but cautions that
even with their use we can not be entirely satisfied with
diagnostic labels for lesions because of the lack of generally
and internationally accepted nomenclature in toxicological
pathology. He gives the following interesting example of
what could happen because the pathologist is permitted to
coin his own diagnostic labels for a mammary gland nodule:
1) it can be labeled "cystic fibromatous hyperplasia" and
make it sound innocent; 2) "ductal carcinoma in situ" to
sound frightening; or 3) being noncommittal -mammary hyperplasia
with squamous metaplasia and a certain potential for malignant
(carcinomatous of sarcomatous) degeneration^
-------
31
To prevent this type of problem, an experienced pathologist
will describe each significant lesion type, at least once,
in such detail that any competent pathologist can perceive a
good mental picture of the lesion and form his own judgment
as to its relevance to the histopathology induced by the
chemical being tested. In spite of improvements in methodology
and descriptive reporting, this area of highly subjective
judgments often times presents special problems of quantification
and reproducibility for toxicologists.
Further examples of potential problems caused by total
reliance on diagnostic labels are provided in Table 6. Note
that the term "hepatoma" has appeared in the scientific
literature as a label for both benign and malignant neoplasms.
Also the term "nodular hyperplasia" has been used as a label
for benign neoplasms and for hyperplasia, in spite of the
fact that the latter is a non-neoplastic lesion.
Most problems with diagnostic terms are encountered in
incidence tables, basically because the tabular information
is meant to summarize descriptive information. For example,
if a table listing liver effects contained only the term
"hepatoma" as the sole designator for tumors, an evaluator
would not know if the incidence data designated benign or
malignant tumors or a combination of both types. Conversely,
if the table listed individual liver effect incidences for
-------
32
nodular hyperplasia, adenomas, and hepatocellular carcinomas,
the evaluator should understand that the pathologist has made a
distinction concerning these different effects and tumor
states. However, if the tabulation only lists nodular
hyperplasia and hepatocellular carcinoma, the evaluator does
not know whether the nodular hyperplasia should be placed in a
hyperplastic or a metaplastic category.
Sometimes incidence tables will contain a collective
diagnostic term as a convenient substitute for more cumbersome
diagnostic terms which do not conveniently fit the tabular
format (e.g., substitution of "adenomatosis", a term which
can be used to label an inflammatory process or a preneoplastic
lesion, in a table for "focal area of alveolar cell
proliferation").
There are only two alternatives for ameliorating this
type of confusion. The first is to rely on the pathologist's
detailed description of the lesion contained in the evidential
documentation. If, however, the submitting pathologist has
not provided a suitable description of the types of lesions
or neoplasms found in the study and/or stated his criteria for
distinguishing between a benign and malignant neoplasm, he
should be requested to do so before the evaluation is completed.
The other alternative is to request the original tissue slides
and have them examined by the OPTS pathologist(s) or a competent
consultant. Either of these requests should be made through
-------
33
the OPTS management.
For an example of interpretational perturbations caused
by the "adenomatosis" substitution cited previously and how
these were resolved by using the pathologist1s description,
see reference 85.
While a competent pathologist must be relied upon for a
final decision regarding the benign or malignant status of a
given neoplasm, a general knowledge of the characteristics of
both types of neoplasms is useful to the evaluator in the
analysis and interpretation of incidence tables. Some of
these are presented in Table 7. These should be perused by
all evaluators from time to time to prevent the possibility
of inappropriately combining benign and malignant neoplasms
during the analysis and evaluation of a study. An example
of what can happen when these characteristics are ignored or
misinterpreted can be found in reference 86.
The evaluator should also be aware of the differences
which may exist between those neoplasms potentially related
to treatment and those which are not so related (spontaneous
neoplasms). Ward and Reznik (32) have discussed some of the
differences, (Table 8). While this concept is not completely
accepted by a majority of pathologists, these differences may
be of aid to evaluation. However, if there are any doubts
on the part of the evaluator about the relationship of neoplas-
-------
34
tic lesions to treatment, an experienced pathologist should
be consulted.
For this major oncogenicity consideration keeping in
mind the above pitfalls, all neoplasms observed in one or
more treated groups which, by inspection, appear to have an
incidence approximately equal to or higher than the concurrent
control incidence should be identified (.low frequency tumors
have been previously discussed). Such data is often
displayed in the evidential documentation in two different
forms approximating Tables 9 and 10. The first of these is
a listing of lesions by individual animals in each group and
is useful in determining the number of animals per group
exhibiting each lesion type. The second is a summary
incidence which presents the number of tissues examined per
group exhibiting each lesion type. The data needed for this
consideration can be obtained, for the most part, from the
Summary Incidence Table (see Table 10) submitted with most
long-term study reports, provided the summary data has been
verified as to its accuracy. However, if appropriate for
completeness of review or for other reasons, the incidence
data may be rearranged and displayed in a more convenient
form. If rearranged by the evaluator or a statistician the
tumor incidences should be segregated by sex, dosage levels,
and tissue or organ site. Part III. B. 1. of Reference 2
states in part:
-------
35
Benign tumors should generally be combined with malignant
tumors for risk estimates unless the benign tumors are not
considered to have the potential to progress to the associated
malignancies of the same morphologic type. However, the con-
tribution of the benign tumors to the total risk should be
indicated.
In order to comply with this latter requirement the
incidence data for benign and maligant neoplasms of the same
histogenic origin found in the same site should be reported
as separate incidences. If the data submitter also wishes to
present combined incidence data, it should be done in a
manner simulating Table 1 and 2'.
The combination of benign and malignant tumors or tumor
sites to evaluate biological and/or statistical significance is
a controversial issue. It is frequently done and may influence
incidence rates and thereby the weight-of-evidence for oncogeni-
city. The basis for the appropriateness or inappropriateness
of combining tumor types and incidences is their histogenesis.
Therefore, when in doubt evaluators should not combine tumor
data without the advice of the OPTS pathologist. When the
combination occurs in the evidential documentation, the
evaluator should expect to find the rationale clearly stated.
If the rationale is absent, it should be requested from the
responsible pathologist or statistician.
-------
36
The National Toxicology Program (NTP) has become involved
with problems attending to the combining procedure and has
proposed draft guidelines33 to its Board of Scientific
Counselors for consideration. These are not yet officially
promulgated by NTP. The reference should be consulted for
the rationale for their proposed use. The following quote
and table are presented here as an illustration of present
thinking on this subject.
Following is a list [Table 11] of organs/tissues
where combining benign and malignant tumors is or
is not appropriate to obtain a clearer understanding
for the evidence of carcinogenicity. This list
comprises those organs/tissues in which neoplasia
is most often observed in Fischer 344 rats and
B6C3F1 mice and may or may not be appropriate for
use in other strains/species. Entities not on
the list would be addressed on a case-by-case
basis; this is a guide only. In addition, as the
depth of knowledge increases in regard to the
biological behavior of neoplasms in a given
organ/tissue, certain combinations in the future
may become inappropriate or appropriate.33
Great care must be exercised when rearranging incidence
data since failure to list all tumors or double listing
-------
37
of tumor types in any animal group may change the biological
and/or statistical significance of the collected data and lead
to a specious conclusion concerning the weight-of-evidence.
Special care must also be taken when the evaluator is confronted
with evidence of widespread systemic disease processes, e.g.,
amyloidosis, arteriosclerosis, or neoplasms of the hematopoietic
system. Evidence of such disease processes is usually present
in multiple organs and tissues within the individual animals.
These lesions should be counted only once in individual
animals. Skin tumors of basal cell origin should be counted
together regardless of their many synonyms. Advice of the
OPTS pathologist should be obtained if there is any doubt on
the part of the reviewer as to how to derive an accurate
incidence for systemic hematopoietic disease and skin tumor
lesions.
Examples of problems created by inappropriate combination
of tumor types have already been presented.8^f 86
4. Degree of Induced Oncogenicity.
For purposes of this consideration, "degree" is defined as
the relative amount of a quality, attribute, or condition and
"induced" means stimulated an occurrence, or caused.
-------
38
For this consideration, the evaluator should first determine
if there is a pattern of potentially induced oncogenicity
discernible in a study or group of studies. Occasionally, a
study will exhibit a high degree of oncogenicity in all
animal groups. The evaluator may obtain an impression of this
situation and its severity by examining the total number of
animals/sex/group exhibiting neoplasms of any kind during the
course of the study. Such data can usually be derived from
study tables presenting individual animal lesions (i.e..
Table 9).
Table 12 presents data recently encountered at termination
of a two-year rat study. This situation is an untidy complica-
tion to say the least. On the surface, the high degree of
oncogenicity in all animal groups obscures any meaningful
pattern of neoplasia or any potential dose response relation-
ships. It also would appear that the degree of neoplasia in
all groups is severe enough to place the study in the ques-
tionable category regarding its usefulness for evaluating
oncogenicity- If a study exhibiting this degree of oncogen-
icity is the only data available or if it represents one-
half of the long-term data base, the problem is a serious one
and the study may have to be repeated.
However, before rejecting the study as unsatisfactory,
further analysis should be done. In the situation above, a
subsequent identification of the predominant types of neoplasms
-------
39
and the degree of malignancy exhibited by all groups revealed
that two tumor types, pituitary and mammary gland tumors,
accounted for the high overall incidence and that increased
incidence of malignancy was not a significant factor in any
group [see reference 87 for details of analysis]. Since the
study represented 1/6 of the total available long-term data
base, it was salvageable. It should be kept in mind that
evaluators do not have a responsibility for salvaging deficient
or defective data bases. However, sufficient analysis must
be done to support a rational judgment regarding the rejection
of or conditional use of a study. In such instances, the
opinion of the OPTS pathologist should be obtained before a
final decision is made.
Another important pattern of neoplasia is the one in
which there are "consistently positive results in two sexes
and in several strains and species and higher incidences at
higher doses".-'- It is generally agreed, at an international
level, that this type of pattern is the best evidence of a
positive oncogenic response obtainable with animal studies.
Obviously, this pattern has major biological significance for
determining the oncogenic potential of the test substance.
If such a pattern is present in appropriately designed and
conducted studies, the substance should be considered
an oncogen for experimental animals and a suspect oncogen
for humans.
-------
40
These two patterns represent extremes, permitting early
decisions on a study, not usually encountered by OPTS evalua-
tors. More frequently encountered are tumor patterns requiring
special attention during data analysis and evaluation.
Among these patterns are: (1) the potential associations of
endocrine tumors; (2) tumors of the hematopoietic system
and; (3) patterns of tumors frequently associated with liver
tumors. There are also patterns of increased or decreased
tumor incidence for some tumor types which are associated
with body weight differences between the treated groups and
their concurrent controls, or which may be related to aging.
It was pointed out in major consideration #3 that special
care must be taken when evaluating incidences for neoplasms
of the hematopoietic system. Haseman34 in investigating
tumor incidence patterns in Fischer 344 rats in 25 National
Toxicology Program (NTP) studies found that leukemia/lymphoma
incidence decreases, in both sexes, were frequently associated
with increased liver tumor incidences in the treated groups.
A clear biological explanation for this association was not
apparent.
Decrease in tumor incidence associated with lower body
weight gains, restricted food consumption, or diet quality
has been frequently reported. Conybeare35 reported
on a study using outbred Swiss mice which received two diets,
-------
41
on an ad lib, or restricted basis each. Table 13 presents
spontaneous tumor incidence differences between the two
feeding regimens. In both sexes on both types of diet,
dietary restriction was associated with slightly better
survival up to 18 months (time of study termination) and with
significant decreases in the incidence of neoplasms of
all types. The excess tumor incidence in mice fed jid lib.
was accounted for by an excess of both lung tumors and liver
tumors. Conybeare discusses these findings in relation to
the interpretation of tests for carcinogenicity. Examination
of food consumption and body weight gain data is helpful in
determining if this phenomenon is present in a particular
study.
Hottendorf and Pachter3^ analyzed the NCI experience
in oncogenesis testing. Table 14 presents the most commonly
found tumors in 98 positive studies. The liver was not only
the most common tumor site, but the mouse liver was a tumor
site about twice as often as the rat liver. In bioassays
where the liver of only one species is considered, the mouse
liver was involved five times as often as the rat liver.
It is this pattern of liver involvement among the NCI
bioassays, other oncogenicity studies, and the Conybeare
study which causes pathologists and toxicologists to be
concerned about the significance of hepatic neoplasms in
-------
42
mouse studies 37' 38 Tomatis e_t a_l. , while not ignoring
mouse data have concluded: "It does not imply that the chemical
which has been tested with negative results in one or more
species should be automatically regarded as having a possible
carcinogenic effect on man solely on the grounds that it
induces liver tumors in the mouse."39
From the data reported by Baseman, Conybeare, Hottendorf
and Pachter, and Tomatis, et aJU34' 35' 36' 39 and
others37' 3^r ^0 it should be obvious that while mouse
liver (and lung) tumor patterns may be very simple to identify,
it may be most difficult to evaluate their significance as
far as a potential oncogenic effect in man is concerned.
Part II. B. 6 of (2), in part, states the following relating
to mouse liver tumors:
These Guidelines take the position that the mouse-liver-
only tumor response, when other conditions for a
classification of "sufficient" evidence in animal
studies are met, should be considered as "sufficient"
evidence of carcinogenicity with the understanding
that this classification could be changed to "limited"
if warranted when a number of factors such as the
following are observed: the occurrence of tumors
only in the highest dose group and/or only at the
end of the study; no substantial dose-related increase
-------
43
in the proportion of tumors that are malignant;
the occurrence of tumors that are predominately
benign, showing no evidence of metastases or
invasion; no dose-related shortening of the time
to the appearance of tumors; negative or inconclusive
results from a spectrum of short-term tests for
mutagenic activity; the occurrence of excess tumors
only in a single sex.
When an increased incidence in mouse liver tumors is observed
it is necessary to examine all other chemical and biological
properties of the test substance in order to arrive at a
final judgment.40 Because mice appear to harbor a significant
population of preexisting initiated or latent tumor cells,40' 4^
some investigators have suggested that the requirement for a
mouse study may be an unnecessary redundancy when a valid rat
study exists.42
5. Dose-response Relationships
The term "dose" can be ambiguous in that its precise
meaning depends, in part, on the route of administration, the
particular interest of an investigator, and the context in
which it is used. In this section, regardless of the complexi-
ities of route of administration, absorption, distribution
'and excretion, dose means that stated quantity or concentration
-------
44
of a substance to which a living organism is experimentally
exposed. Although, the term "response" can be applied to
either beneficial or injurious effects observed at a specific
dose, emphasis here is placed on the latter type of responses
from a multiple dose regimen.
Of all the observations which might be made with respect
to any biological effect, the most fundamental one is that
correlative relationship existing between the dose administered
and the response or spectrum of responses that is obtained.
In essence, this is the classical definition of the term
"dose-response relationship." The concept expressed by this
term is indispensable to the identification, evaluation and
interpretation of most pharmacological and toxicological
responses to chemicals. It is therefore important for an
evaluator to understand the basic assumptions which underlie
and support the concept.
The primary assumption is that a dose-response relationship
is firmly based on knowledge or a defensible presumption that
the response (effect) observed is a result of exposure to a
known substance. Correlative assumptions are: there is a
receptor site(s) with which a substance interacts to produce
the response(s); the observed response(s) and degree of
response are related to the concentration of the substance at
the receptor site(s); and, the concentration at the site(s)
-------
45
is related to the dose received. Therefore the biological
concept of dose-response relationship includes the basic
assumptions that (a) the observed response is a function of
the concentration at a site, (b) the concentration at a site
is a function of the dose, and (c) response and dose are
causally related.43
The essential purpose of long-term animal studies is the
detection of valid biological evidence of the toxic and/or
oncogenic potential of the substance being investigated.
Therefore, protocols should, in an appropriate way, maximize
the sensitivity of the test without significantly altering
the accuracy and interpretability of the biological data ob-
tained. The dose regimen has an extremely important bearing
on these two critical elements. In this regard, two contro-
versial concepts (i.e., maximum tolerated dose (MTD) and
lack of oncogenic thresholds) have had a significant influence
on the selection of doses for long-term oncogenicity studies
and on the interpretation of observed dose responses. The
evaluator should be continually aware that this influence may
have a high probability of interjecting unintended biases
into a data base and the subsequent evaluation. The no oncogenic
threshold concept may also have had an inhibitory influence
on the scientific discussion and development of methods for
assessment of oncogenic potency as well as the development
and use of animal oncogen ranking or classification systems
by regulatory agencies.
-------
46
Part II. B.6. of Reference 2 discusses the dosage
regimen for long-term animal studies and states in
part:
Long-term animal studies at or near the maximum
tolerated dose Level (MTD) are used to ensure an
adequate power for the detection of carcinogenic
activify. Negative long-term animal studies at
exposure levels above the MTD or partial lifetime
exposure at the MTD may not be acceptable because
of toxicity, or if animal survival is so impaired
that the sensitivity of the study is significantly
reduced below that of a conventional chronic
animal study at the MTD. Positive studies at
levels above the MTD should be carefully reviewed to
ensure that the responses are not due to
factors which do not operate at exposure
levels below the MTD. Evidence indicating
that high dose testing produces tumor responses
by indirect mechanisms that may be unrelated
to effects at lower doses should be dealt with
on an individual basis.
Historically, the concept of "maximum tolerated dose"
(MTD) arose from long-term oncogenicity screening studies
which employed very limited dosage regimens and relatively
small numbers of animals. The intent of the studies, under
-------
47
these limited conditions, was to maximize the likelihood of
observing an oncogenic response by administering as high a
dose of chemical as feasible. Little consideration was given
to determining valid dose-response relationships; the major
emphasis was to establish whether or not the chemical had
oncogenic potential in a qualitative sense. To accomplish
this, an extreme condition (i.e., a MTD) was routinely employed
in these studies. Presently the MTD term has almost as many
different connotations as there are individuals who use
it.44, 45, 46 conscientious attempts to accommodate the
concept in long term studies have frequently caused dose
level adjustments in one or more animal groups and these
have frequently introduced interpretational difficulties at
the termination of the study.79
For these reasons and others discussed below, the
characteristics of the highest dose to be administered in
modern long-term animal tests are presently being reconsidered
and more clearly defined by a concerned scientific community.
An Ad Hoc Panel on Chemical Carcinogenesis Testing and
Evaluation has recommended that the following end points from
subchronic studies be used in selecting chronic dose regimens
for NTP long-term studies: a) organ specific and/or systemic
pathology; b) body weight and organ weight data c) clinical
laboratory measurements and; d) pharmacokinetic data.47
-------
48
The developing consensus can be expressed thus. Ideally,
dose selection for long-term oncogenicity studies should
maximize the detection of potential oncogenic dose response
relationships and facilitate the extrapolation of these to
potential risks for other species including humans. There-
fore, the largest administered dose should be at one which
produces signs of minimal toxicity that do not compromise
biological interpretability of the observed responses. For
example, the upper dose should not: (a) alter survival in a
significant manner due to effects other than tumor production;
(b) cause a body weight decrement from the concurrent control
values of greater than 10-12%; (c) exceed 5% of the total
diet; (d) produce toxic, pharmacologic, or physiologic effects
that will shorten duration of the study or otherwise vitiate
the study results.4^
Some of the reasons for this changing attitude toward
MTD are presented here. The potential interpretive difficulties
associated with oncogenic dose-response relationships in
animal groups exhibiting excessive mortality and/or excessive
body weight differences, when compared with their concurrent
controls, have been discussed. It is also known that excessive
stimulation or inhibition of glandular activity through
normal mechanisms or abnormal pharmacological and physiological
effects of excessive dosage can complicate evaluation and
interpretation of oncogenic dose-responses.48' 49' 50' 5^
-------
49
What is not so obvious is the potential problems created by
severe tissue/organ injury produced by excessive dosage
levels in long-term oncogenicity studies.52 Evidence
indicates that 7-methylguanine and O6-methylguanine are
incorporated into liver DNA following administration of
acutely toxic doses of the hepatotoxins hydrazine, carbon
tetrachloride, and ethanol in rats and mice53 and the
male Syrian golden hamster.54 This suggests that aberrant
methylation of DNA may be a response to severe toxic insult
or damage to the rodent liver. If this effect is confirmed
for other substances which induce neoplasms only at or near
severely toxic doses, it will have a significant bearing on
the selection of the dose regimen for long-term oncogenic
studies and the assessment of oncogenic risks for humans.40
It is also known that exaggerated doses can alter, in biologi-
cally significant ways, normal metabolic functions and pharma-
codynamic parameters.55
Although it can be logically argued that responses observed
at exaggerated dose levels (e.g., doses far in excess of
levels experienced under real or potential exposure conditions)
legitimately fall within the classical dose-response concept,
there is a developing suspicion, based on growing scientific
evidence, that such doses are interjecting biases of considerable
importance into the already difficult task of evaluating animal
oncogenic dose responses and the assessment of their relevance
-------
50
to human risk.46' 56' 57' 58 It has been suggested46' 47' 55
that the MTD concept, or at least the term, be abandoned and
that the scientific community rely instead on adverse signs
that are biologically important, but less severe than gross
tissue injury or destruction, in judging the adequacy of the
highest dose administered in long term oncogenicity studies.
A statement as to the adequacy of the dose regimen used
should appear in the evaluation document. The rationale for
this opinion should be concisely stated and should include a
brief presentation of the toxic manifestations observed at
each dose level. Special notation of unusual findings (e.g.,
disease processes unrelated to compound administration,
bladder or kidney stones) and the dose level or levels at
which they were observed should also be made. If a NOAEL is
present for toxic signs, it should be identified in the
evaluation document.
The term "threshold" can be defined as that value at
which a stimulus just produces a sensation, is just appreciable,
or comes just within the limits of perception. In toxicology
and pharmacology the concept of a threshold dose is accepted
as applying to biological responses of nearly all chemical
and some physical stimuli (e.g. one source defines 23 types
of measurable biologic thresholds).59 Generally, the
-------
51
concept is understood by toxicologists to mean that there is
a dose for nearly all chemical substances below which no
response is discernible or detected in the organisms exposed
to it. If this lack of response is exhibited by a reasonable
number of test subjects, the dose is assumed to be a subthreshold
dose. The no oncogenic threshold concept is contrary to the
generally accepted biological threshold dose concept and
requires special consideration because of its potential
impediment to competent scientific evaluation of oncogenicity
data bases and risk assessments.
Gehring and Blau have presented succinct arguments for
both sides of this and the MTD controversy. To paraphrase them
might lessen their impact. For this reason and for the
convenience of the reader, they are quoted here. The reader
should consult the original paper for other important aspects
of oncogenic dose responses and the references cited therein.
Evidence Supporting a Threshold Concept is Substantial.
Some of the Arguments Are:
1. Chemical carcinogenesis is a multistage process
involving :
a. Exposure, absorption, distribution,
activation, deactivation, and elimination
of the chemical per se or products formed
from it.
-------
52
b. Interaction with critical receptor sites
leading to moleculary transmittable
products.
c. Survival and proliferation of transformed
cells to clinical cancer.
Interference with any of these processes may
constitute a threshold. For example, there is a
plethora of data showing that promoters, which
in themselves cannot initiate cancer, can enhance
greatly the incidence of cancer induced by admin-
istration of an initiator. Also, the damaged
receptor site may undergo repair.
2. Alteration of the physiological status may either
augment or inhibit the response to a carcinogen.
For example, age, sex, nutrition, population density,
hormonal state, or concomitant disease may affect
the response to a carcinogen. This suggests that
a precancerous status may exist or may be induced
without development of cancer until the precancerous
status attains some critical level or until the
precancerous status can no longer be held in check
by suppressive mechanisms, whatever they may be.
-------
53
3. As the dose of a carcinogen is decreased, the
latency period for cancer development increases.
This phenomenon was revealed lucidly by Druckrey
(1967), who noted that the dose multiplied by
some power of time was constant, i.e., dtn =
constant in which n = 2 to 4. For all practical
purposes, this relationship implies a threshold
in that multiples of a lifetime will be required
for expression of cancer in response to low
doses. Albert and Altshuler (1973) utilized the
increasing latency with decreasing dose of a car-
cinogen to formulate limits for unavoidable
exposures to carcinogens.
4. Utilizing the relationship of dose to time-of-
appearance of cancer, Jones and Grendon (1975)
postulated that a number of cells in close
proximity require transformation to allow
development of an aberrant clone of cells and
ultimately cancer. This multihit hypothesis, if
true, will result in a marked reduction in the
incidence of cancer as the dose is decreased for
the same reason that trimolecular chemical
reactions become negligible as the concentrations
of the reactants are decreased.
-------
54
5. For many chemical carcinogens, cancer occurs only
when doses are given that exceed those needed to
cause pathological responses, such as grossly and
histologically discernible tissue damage. This
is not surprising, since some cancers develop
clinically in chronically inflamed or scarred
tissue, e.g., colonic cancer in patients with
ulcerative colitis or regional enteritis, squamous
cell carcinomas in ulcers of burn scars, squamous
carcinomas of the bladder in schistosomiasis,
scar carcinomas in lung, carcinomas and sarcomas
arising in osteocutaneous fistulas caused by
chronic osteomyelitis, and carcinoma of the stomach
in autoimmune (atrophic) gastritis (Laroye, 1974).
Perhaps sarcomas induced locally by implants of
inert solid material or local injections of
chemical substances represent an experimental
expression of these phenomena observed clinically.
Even such substances as water, salt, glucose,
and a host of other common nutrients are carcino-
genic when given in this manner (Grasso and
Golberg, 1966). Such evidence of carcinogenicity
is discounted for the most part. Is it any
less reasonable to discount similar evidence
when the administered dose is transported to
another site in the body where it causes chronic
inflammation and subsequently carcinogenesis?
-------
55
6. There is a substantial and growing body of
evidence that carcinogenesis is subject to
immuno-surveillance, particularly cell-mediated
immunity (Roe and Tucker, 1974; Weisburger,
1975) .
7. Stress, such as administration of unrealistically
large doses of chemicals to laboratory animals,
can enhance greatly the response to oncogenic
viruses and perhaps other innate carcinogens
as well. This has been demonstrated eloquently
by Riley (1975) in C3H/He mice infected with
Bittner oncogenic virus, the incidence at 400
days of age was 92% in those under stress and
only 7% in those in a protected environment.
8. Man and animals live in a sea of potential
carcinogens, most of which were not placed
here by man. There is reasonable evidence, in
both humans and animals, that over-nutrition,
particularly excess dietary fat, is a major
cause of cancer (wynder, 1976; Weisburger,
1976). Malonaldehyde, a product of peroxidative
fat metabolism which is also formed spontaneously
in tissues, particularly when the diet is
-------
56
deficient in antioxidants, has been found to
be carcinogenic (Shamberger et_ al. , 1974).
Selenium, an essential micronutrient, calcium
(Krook e_t al. , 1971), and egg whites and
yolks (Szepsenwol, 1963) have all been reported
to cause cancer when given in excess to experi-
mental animals.
Thus, it seems that excesses of many substances
may be expected to induce cancer. Is it not
reasonable to believe that below some threshold,
these naturally occurring environmental carcinogens
will exert no carcinogenic effect? The logical
alternative is to believe that most any substance,
including food, continually gives rise to
small numbers of aberrant cells which eventually
cause cancer if competing causes of death do
not prevail. Adherence to the latter logic
allows acceptance of exposure to levels of man-
made chemicals which do not add measurably to the
background flora of carcinogens, which is likely to
be substantial although not well elucidated.
The authors^O continue:
Arguments that there is no threshold for chemi-
cal carcinogenesis are equally substantive. The
principal argument is, in essence, that cancer is
an expression of a permanent, replicable defect
-------
57
resulting from amplification of a defect initiated in
one cell by reaction of the chemical with a critical
receptor. once such a defect occurs in a cell, the
cell may be dormant for years before expressing a dis-
cernible untoward effect. Unlike classical toxicological
responses, division of a large dose of some carcinogens
into smaller repeated daily doses does not abolish the
response. Indeed, for dimethylaminoazobenzene, 4-
dimethylaminostilbene, and diethylnitrosamine, the
total cumulative dose necessary for carcinogenesis with
small daily doses is smaller than the single dose required
to produce an equivalent response (Druckrey, 1976;
Schramel, 1975; Weisburger, 1975). However, it should
be noted that the size of the repeated doses can be
reduced further, resulting first in an increased latency
for development and, finally, no experimentally discernible
response. It is important to emphasize that these data
were obtained on highly potent, direct-acting carcinogens.
As the doses of such agents are increased, a less than
linear increase in tumors should be anticipated because
their innate reactivity will preclude proportionate
increases in the active agent at the receptor site.
Another frequently referred to piece of evidence is
that exhaustive experiments on radiation-induced cancer
have not revealed a threshold within the realm of statistical
-------
58
reliability. However, the validity of equating chemical
carcinogenesis to radiation-induced carcinogenesis is
questionable. Entry of radiation into a cell and release
of its energy, leading presumably to the local generation
of free radicals, is governed by physical chemical laws;
hence, a particle of radiation is just as likely to do
its dirty deed within the nucleus of a cell as elsewhere.
Such is not the case for chemical carcinogens; all sorts
of deactivating events are feasible and, indeed, likely
to occur before the chemical reaches the critical receptor.
Thus the argument concerning what may occur on the
low end of the dose-response curve continues. Until
recently, the conflict did not have a major impact
because of the philosophy of 'no threshold' was
applied to only a few agents which were very potent
carcinogens; and somewhat more generally to intentional
food additives because of the Delaney Clause. However,
the impact is developing rapidly into a galloping
crisis because the philosophy of 'no threshold' is
being extended to proclaim 'no safe level of exposure'
to any chemical shown to be carcinogenic regardless
of the dose of the chemical needed to elicit a
discernible carcinogenic response. Not unexpectedly,
the chemicals thought heretofore to be safe do increase
cancer when huge doses are administered. In many cases,
-------
59
the doses used have exceeded those required to cause
marked toxicological effects.60
No matter how important or desirable the concepts of MTD
and no oncogenic threshold may appear for prudent regulatory
decisions, it must be kept firmly in mind that presently it
is not known if a substance exhibiting an oncogenic response
as a result of large doses represents a risk when only small
exposure levels are encountered by humans. To treat the no
oncogenic threshold concept as a proven scientific fact for
all substances exhibiting oncogenic potential is contrary to
a growing body of evidence that thresholds may in fact exist
for some such substances,52, 61 an(j ^^ may result in the
evaluator misinterpreting or overlooking important biological
evidence contained in the data base or in auxiliary studies.
The ED01 study has shown that in individual cases, complete
carcinogens may show thresholds of a real or practical nature.
In other cases they may not.61 If in the evaluator's opinion
the data indicates the possibility of a potential threshold
effect, this should be stated and a rationale given.
Table 15 presents four oncogenic incidence patterns
actually encountered in rodent long-term studies. Except
for Figure D, an inhalation study, the route of compound
administration was dietary. Since the data for each figure
was selected to illustrate types of incidence patterns, i.e.,
-------
60
without regard to tumor types and site or toxic, pharmacologic,
or other influences which may have been present, the data
sources and substances are not identified. The fact that all
but one of the selected patterns were evidenced by female
mice is purely accidental.
Figure A represents an incidence pattern which is generally
thought of as the "classical" carcinogenic multidose response
relationship. A malignant tumor, lung adenocarcinoma,
incidence increases with each increase in dose increment and
in such a manner that statistical conformation of the probable
reality of a positive dose response relationship is hardly needed,
Figure B represents data which is in stark contrast to
the data represented by Figure A and is of a more common
occurrence. The incidence data exhibits a random pattern,
statistical analysis does not produce even a borderline value,
and all incidences fall within the expected incidence values
(i.e., historical controls) for the mouse strain used. In
this example, even if strenuous statistical efforts had
produced a borderline p value for the high dose group, there
is no evidence of a dose response relationship and the response
has no biological relationship to treatment, (see Task Force
situation 1). It should be kept in mind that a dose
response relationship should be firmly based on knowledge or
a defensible presumption that the observed response is causally
-------
61
related to the dose. If these conditions are not met the
reality of such a relationship may be illusory. Figures A
and B represent extremes encountered by evaluators. In both
of these cases, all that is required, other aspects of the
study being adequate for evaluation and interpretation, is a
competently performed verification of the data base and the
application of the major oncogenicity considerations.
Unfortunately, the potential incidence patterns which
may be encountered in long-term rodent studies are legion and
often times require considerable analytical skill to identify
a valid dose response relationship, or the lack thereof, for
any particular tumor type or group of tumors. Figures C and
D are only two varieties of the potential problem patterns
encountered between these extremes.
Figure C presents incidence data which are frequently
encountered for many tumor types. The doubling of the
concurrent control incidence at the low and middle dose
levels, on the surface at least, appears to identify a signi-
ficant dose response relationship of biological importance
and the tumor incidence at the high dose appears to be an
artifact. While this evaluation and interpretation is tempting,
it can be misleading unless the artifactual nature of the highest
dose data can be identified. There is nothing in the biological
dose response concept which requires multiple dose regimen
relationships to be unidirectional. Therefore, this type of
-------
62
pattern and its variations must be viewed as a real dose
response in the absence of knowledge or a defensible presump-
tion to the contrary. Such relationships may need considerable
analysis before their biological significance can be interpreted.
Before attempting any complex statistical analysis, the
reviewer should examine the biological data base for a
defensible explanation of the incidence pattern. When accept-
able historical tumor incidence, data for the particular
neoplasm is available, it should be compared with the observed
incidence for each treated group and the concurrent control
group. If all the study incidences are within the historical
tumor incidence range and the expected time of tumor
appearance, the study incidence data may not represent
a treatment relationship and the situation is equivalent to
the Task Force's situations 1 or 2. However, before
this explanation is completely accepted, further analysis of
the data base, including auxiliary data, for corroborative
evidence should be performed and a defensible presumption
for its acceptance presented.
Assuming that the situation just discussed does not
pertain and that the biological data base appears to support
the validity of the data represented in Figure C, the evaluator
should try to discover an explanation for the 4000 ppm incidence,
Significant trends and p values derived for the lower dose
incidences do not prove that the 4000 ppm incidence is arti-
-------
63
factual in nature. It is possible that the high level group
value was heavily influenced by an incidental disease process
exacerbated by severe toxicologic stress, or the 4000 ppm
level was so toxic that it caused significant early mortality
thus reducing the number of animals at oncogenic risk when
compared to control and the other two treatment group survival/
mortality data. The high level may not have caused significant
mortality but caused a severe reduction in body weight by some
mechanism which resulted in the low tumor incidence. It is
also possible that the 4000 ppm dose caused a biologically
significant shift in metabolic pathways or distribution and
elimination patterns.58 Examination of existing metabolism
and or pharmacodynamic data might be helpful in evaluating
this possibility. The reader may recall a similar incidence
pattern and the amount of analytical effort needed to interpret
the results.79 Even after this effort, the interpretation
of the data remained speculative. If no scientifically
defensible explanation can be identified for the response,
this fact should appear in the review.
Figure D represents a tumor incidence pattern encountered
in some long-term rodent studies, although not very frequently
in such a dramatic form. Usually the tumor incidences for
the control and lower dose levels are higher for most tumor
types than the incidences presented in this example. Since,
in this case, the data immediately suggests that a threshold
-------
64
dose has been exceeded, between 5.6 and 14.3 ppm, the first
place to look for a defensible explanation would be any
available metabolic and pharmacodynamic data.58 If this type
of data does not provide reasonable corroberative evidence
for this presumption, the evaluator should proceed, as suggested
in the discussion of Figure C. In this case special attention
to the concurrent control group may be of importance. Occasion-
ally the concurrent control group data, male, female, or
both, do not fall within the "expected normal range" for the
particular strain or species even in the same laboratory.
This phenomenon, sometimes called the "control effect" can be
very troublesome in the interpretive process. In this case
an experienced toxicologist will spend as much time, or
nearly as much time, examining and evaluating the control
data as he or she will in examining data from the groups
receiving the various dose levels (i.e., the treatment groups).
It must be kept in mind that the term "treatment" can have a
specific meaning, as used in this part, or a generic meaning.
In the latter sense, it connotes all the environmental influences,
controlled and uncontrolled, which are inherent in any animal
experiment. Sometimes the uncontrolled influences (e.g.,
diseases,) cause the control group or groups to exhibit
aberrant data bases which may artificially produce statisti-
cally significant differences and false dose response relation-
ships.
-------
65
Because the biological reality of oncogenic dose-response
relationships is so important to risk identification and assess-
ment, it is reiterated that such relationships should be based
on knowledge or a defensible presumption that the response
and dose are causally related. The knowledge or presumption
must be based on the biological, toxicological, metabolic,
pharmacodynamic, and other evidence (i.e., weight-of-evidence)
contained in the submitted documentation. The evaluator must
strenuously resist the temptation to accept a P value(s) as
the sole designator of a biological dose-response relationship
or the sole determinant of an oncogenic effect. The reader
should examine Misconceptions Regarding Significance and P^9
to maintain a balanced perspective regarding this matter.
Evaluators must understand that the weight given to the
level of statistical significance, (i.e., P value) is not an
automatic consequence of some natural law, it is a scientific
judgment. After careful consideration of the data, if an
evaluator chooses to dismiss a statistically* significant
difference between a treatment group and the concurrent
control group, the rationale should be succinctly presented.
6. Decrease in latency (time to tumor discovery) of neoplasms
that are usually observed.
A latent period is generally understood to mean the
interval between the application of a stimulus and the
observation of a response.
-------
66
Some chemicals are known to induce tumors in experimental
animals at very high incidences, much earlier than is usually
expected (precocity), and are for this reason sometimes used
as positive controls (Table 16). As may be seen, 40-51 ppm
diethylnitrosamine, orally, produced a 100% incidence of
liver tumors in 20-35 weeks in two different rat strains. In
the case of 7,12-dimethylbenz(a)anthracene intubation of 15-
20 mg produced a 92-100% incidence of rat mammary tumors in
females in 12-16 weeks and by skin painting in the mouse,
75 mg produced skin neoplasia in 10-25 weeks.
Without adequate serial sacrifices, which are rarely
performed because of cost, tumor latency can only be derived
accurately in the case of visible tumors such as those of the
skin or mammary glands, or the rare tumor types that rapidly
kill the test animal. Therefore, this major consideration
may be less useful for evaluating the oncogenicity of many
substances than other criteria. However, since some chemicals
do shorten the latent period, it is prudent for the evaluator
to perform whatever appropriate analysis the data may allow
and make a statement concerning this potential effect. The
statement might be nothing more than that the study data do
not allow a defensible scientific assessment concerning the
latent period to be made or that the analysis does or does not
suggest a precocity of tumor development.
-------
67
Careful examination of early deaths, those which occur
during the first 15 months of a study, may provide some evidence
of precocious tumor development, but unless the cause of
death can be determined to be directly related to neoplasia
[not easily done in most cases2^], early death may produce
a false impression of decreased latency. This is particularly
true when one or more dosage level groups exhibit a differential
survival rate which results in significant inequalities
between a treatment group survival rate and its concurrent
control group survival rate. An example of this pitfall has
already been cited.79
Historical control incidences and time to tumor discovery
data may be of aid in evaluating latency (see major consideration
#1). These types of data together with analysis of elapsed
time from study initiation to tumor discovery sometimes
allows a qualified statement concerning a latent period to be
made.8** Sometime clinical observations such as those that
relate to palpable tumors or which may be associated with
neoplastic development such as hematuria, abdominal distention,
or impaired respiration might be useful in defining the time
a tumor was first suspected of being present. If the data
allow, it is also sometimes useful to determine the total
cumulative dose and the absolute amount in a single-dose
received by tumor bearing individuals. This approach is
based on the observation by Druckery that in most cases the
-------
68
total tumor yield in any given organ or tissue is generally
proportional to the total cumulative dose received, but the
rate of tumor appearance (latency) is related to the absolute
amount in an individual dose.6^ This is a time consuming
approach and should be used only if there is substantial reason
to be suspicious that a precocity of tumor development is likely
to be present.
-------
69
E. Auxiliary Evidence
If the animal data base gives clear evidence that tumors
are induced at multiple sites in rats, mice, and/or other species
and the tumors are not among those having a high spontaneous
incidence, the problem of assessing the oncogenic potential of
a chemical is diminished and auxiliary evidence may play a minor
supporting role in the evaluation. Seldom, however, is this the
case and evaluators must oftentimes deal with results which are
confined to a single rodent species, or sex, or a tumor type which
has a high background incidence (e.g., mouse lung and liver tumors).
In such situations, auxiliary evidence regarding other toxic
manifestations; metabolic pathways, genotoxicity, biochemical
reactivity; and patterns of absorption, distribution, and elimi-
nation may play a critical role in the weight of evidence approach.40
However, not all of this type of evidence need be given equal
weight and the evaluator should apply prudent judgment, on a
case-by-case basis, when deciding the strength of auxiliary
evidence and its contribution to the evaluative process.
1. Mutagenicity Data
The current use of mutagenicity data as an indication that
a substance may have an oncogenic potential is based on the
hypothesis that an alteration of genetic material in the affected
cells is related directly or indirectly to tumorigenesis. This
-------
70
process is thought to proceed by a series of events. The first
step, initiation, involves damage to DNA resulting in changes
in heritable genetic information. Proliferation of the perm-
anently altered (initiated) cells is thought to result in clone
formation in the tissue of exposed individuals. The progression
of the altered cells to benign or malignant tumors is thought to
be dependent on a series of not well understood mechanisms. Short-
term mutagenicity tests exploit the fact that many oncogenic
substances have the ability to produce DNA damage or chromosomal
anomalies.
It must be kept in mind that tumorigenesis may not always
proceed in this multi-step manner and that some oncogens may be
effective through mechanisms that do not cause genetic effects
[i.e., do not damage DNA.64] In this regard IARC states:
In view of the limitations of current knowledge about
mechanisms of carcinogenesis, certain cautions should
be emphasized: (i) at present, these [mutagenicity]
tests should not be used by themselves to conclude
whether or not an agent is carcinogenic: (ii) even
when positive results are obtained in one or more of
these tests, it is not clear that they can be used
reliably to predict the relative potencies of com-
pounds as carcinogens in intact animals; (iii) since
the currently available tests do not detect all classes
of agents that are active in the carcinogenic process
-------
71
(i.e., hormones, promoters), one must be cautious in
utilizing these tests as the sole criteron
for setting priorities in carcinogenesis research and
in selecting compounds for animal bioassays.18
Furthermore, an international commission65 recently
evaluated the usefulness of mutagenicity studies as an approach
to oncogenesis and concluded:
a) Genotoxic tests for chemical carcinogens are a product
of research conducted during the past 10 years. The
research efforts into these tests are still progressing
rapidly. Therefore it should be anticipated that presently
available test systems may be superceded by new tests with
greater predictive value.
b) The use of an individual test or battery of present tests
for genotoxicity as predictors for the carcinogenicity
of specific chemicals does not give absolutely accurate
results. These tests should therefore be supplemented
by carcinogenesis bioassays in animals if specific
chemicals are expected to enter the environment in apprec-
iable quantities. The genotoxicity tests are of use (1)
in selecting chemicals under development for possible
adverse genetic or carcinogenic effects before costly
product development is attempted; (2) in screening pres-
-------
72
ently available natural or synthetic chemicals for genotoxic
or carcinogenic potential; (3) in screening human body fluids
or excreta for genotoxic agents that may indicate exposure to
noxious agents; and (4) in understanding the mechanisms of
cancer or mutation induction.
(c) Knowledge of the basic mechanisms of carcinogensis in animals
is still in a primitive state. This subject needs increased
research if the present hypothesis, based on correlative
evidence that genotoxic mechanisms are involved in carcinogensis
is to be accepted. Experimental evidence that mutagenicity is
indeed part of the carcinogenic process would greatly increase
confidence in the validity of the tests discussed in this
report.65
Heedful of the cautionary statements, mutagenicity data used
in conjunction with long-term rodent studies, can be useful in
evaluation of oncogenic hazards since they appear to be able to
separate, in some cases, those substances which are genotoxic
(i.e., react with genetic materials) from those substances which
do not appear to do so.
Wright considers genotoxic agents under two main headings:
Precursors Agents - possessing no genetic properties per se but
are converted into ultimate genotoxic agents by metabolism in
-------
73
susceptible organisms; and Ultimate Agents - possessing the
intrinsic properties necessary for interaction with critical
cellular targets, e.g., alkyating agents, thereby initiating
the genotoxic process.6^
2. Metabolic - Pharmacodynamic Data
This rubric covers any data which may be concerned with
the complex of physical and chemical processes involved in
the functioning of any specific substance in, or its actions
on, living systems. It is therefore very broad in scope and
the reader must rely on the cited references for more detailed
discussions of the technical aspects of this type of auxiliary
evidence. Other aids available to the evaluator include (1) the
Office of Pesticide Program Standard Evaluation Procedure (SEP) -
Reviewing Metabolism Studies, and (2) the Chemical Information
System. The latter is a computerized collection of chemical and
regulatory data bases that allow structure, substructure, and
name searching of many thousands of unique substances. It can be
used to obtain lists of structurally related chemicals and also
allows searching of the NIH, EPA, NIOSH, Registry of Toxic Effects
of Chemical Substances and other toxicology data bases to determine
if all or some of the structurally related chemicals have a
common toxicological property.
-------
74
An understanding of the mammalian metabolism of a chemical
agent is basic to the discovery of probable oncogenic mechanisms
(see IARC definition) and an understanding of chemical toxicity
in general.67, 68, 69 ^ prudent investigator would start such
studies prior to initiating long-term rodent studies because, in
addition to identification of major metabolites and metabolic
patterns, it is extremely useful to have information on the
potential effects a long-term dose regimen may have on such
entities and relationships.^5, 67
Consideration of the structures of ultimate carcinogens has
led to the important generalization that such agents are strong
electrophiles, mainly alkylating and arylating agents, although
some carcinogenic acylating agents are known. In certain cases
the instability of the presumed ultimate carcinogen prevents
chemical synthesis. In such instances, e.g., the 2,3-epoxide of
aflatoxin BI, the nature of the ultimate reactant has been inferred
from the structures of adducts generated by reaction of the
formed products with biomacromolecules in situ. There are a few
apparent exceptions to the generalization that ultimate carcinogens
are electrophilic reactants. One such exception, 6-mercaptopurine,
has been reported to cause an increase of certain tumors in the
haemopoietic system of rats and mice.66
-------
75
For further discussion of biochemical reactivity in relation-
ship to oncogenicity see References 70 and 71, and for its importance
to chemical toxicity in general see References 72 and 73.
The usefulness of animal toxicity and oncogenicity data is
also enhanced by knowledge of the absorption, distribution and
elimination patterns of the test substance, i.e., application of
pharmacodynamic principles. Discussions and examples of the
integration of this type of data with chronic toxicity data and
macromolecular events associated with toxicity are available
(e.g., styrene, vinyl chloride, and dioxane .as well as the implica-
tions of this type of data for risk estimation.74' 75' 76' 77
-------
76
F. Completion of Anaylsis
At this point the evaluator should have formulated tentative
judgments and supporting rationale concerning: a) the acceptability
of the evidential documentation and data base; b) the presence
or absence of biologically important toxic and/or oncogenic
effects and the relevancy of any modifying factors; and c)
the likelihood that any of the adverse effects were induced
by the tested substance.
Prior to applying the criteria presented in Part II, an
evaluator should summarize, briefly and cogently, the critical
biological and auxiliary data together with any modifying
factors for all studies under review. Any rationale pertinent to
an evaluation of the oncogenic potential of the substance should
also be included in the summary. The following outline is suggested
It should be modified according to the constraints of the data
base.
1. Acceptability of each study considered.
2. Toxic effects.
3. Increased incidence of one or more histogenetically different
types of neoplasms in multiple a) species, b) strains, c)
sexes, and d) doses.
4. Increased incidence of neoplasms in multiple experiments
with consideration of different routes of administration
and/or dosage levels and relationships).
-------
77
5. Increased incidence of neoplasms to an unusual degree (with
respect to type, site, latency, malignancy and quantitative
considerations).
6- Auxiliary evidence.
-------
78
II. Evaluation and Classification of Evidence of Oncogenic
Potential from Animal Studies
As stated previously, the essential purpose of long-term
animal studies is the detection of valid biological evidence of
the toxic and/or oncogenic potential of the substance being
investigated. Clayson e_t al. discuss four areas of particular
difficulty in the interpretation of oncogenicity tests: a) the
heterogeneous nature of carcinogens in terms of exerting their
effects by a series of differing mechanisms; b) meaning of a
negative animal bioassay; c) significance of tumors induced
against a high spontaneous incidence; and d) transspecies extrapo-
lation. The authors conclude that prevailing evidence clearly
points to the fact that mechanistic considerations taken together
with data on carcinogen potency, dose-response relationships, and
general toxicity will, in the future, lead to an increased ability
to refine risk estimates. Approaching the regulation of carcinogens
within such a conceptual framework makes it possible to exercise
scientific judgment regarding the magnitude of risk. This is
essential if we are to base decisions on sound scientific principles.78
This paper should be read by all reviewers involved in oncogenicity
evaluations.
The strength or weight-of-evidence from animal studies as
well as that of any available auxiliary evidence, should be
evaluated and classified by some agreed upon criteria before
-------
79
mathematical calculation of risk is attempted. Part IV. B. of
Reference 2 presents the following guidance for weighing such
evidence. These assessments are classified into five groups:
1. Sufficient evidence* of carcinogenicity, which indicates
that there is an increased incidence of maligant tumors
or combined maligant and benign tumors**: (a) In multiple
species or strains; or (b) in multiple experiments
(preferably with different routes of administration or
using different dose-levels) or (c) to an unusual
degree with regard to incidence, site or type
of tumor, or age at onset. Additional evidence may be
provided by data on dose-response effects, as well as
information from short-term tests or on chemical structure.
2. Limited evidence of carcinogenicity, which means that the
data suggest a carcinogenic effect but are limited because;
(a) The studies involve a single species, strain, or
experiment; or (b) the experiments are restricted by
inadequate dosage level, inadequate duration of exposure
to the agent, inadequate period of follow-up, poor
* Under specific circumstances, such as the production of neoplasms
that occur with high spontaneous background incidence, the evidence
may be decreased to "limited" if warranted (e.g., there are widely
diverging scientific views regarding the validity of the mouse liver
tumor as an indicator of potential human carcinogenicity when this
is the only response observed, even in replicated experiments in the
absence of short-term or other evidence).
** Benign and malignant tumors will be combined unless the benign
tumors are not considered to have the potential to progress to the
associated malignancies of the same morphologic type.
-------
80
survival, too few animals, or inadequate reporting; or
(c) an increase in the incidence of benign tumors only.
3. Inadequate evidence, which indicates that because of
major qualitative or quantitative limitations, the
studies cannot be interpreted as showing either the
presence or absence of a carcinogenic effect.
4. No evidence, which indicates that there is no increased
incidence of neoplasms in at least two well-designed
and well-conducted animal studies in different species.
5. No data, which indicates that data are not available.
The categories "sufficient evidence" and "limited
evidence" refer only to the strength of the experi-
mental evidence that these agent(s) are carcinogenic
and not to the power of their carcinogenic action.
Part IV. C. of Reference 2 also contains guidance for weighing
of the total evidence (human and animal data) in a stratified
scheme as follows:
Group A - Human Carcinogen
This category is used only when there is
sufficient evidence from epidemiologic studies
to support a causal association between exposure
to the agent(s) and cancer.
-------
81
Group B - Probable Human Carcinogen
This category includes agents for which the
evidence of human carcinogenicity from epidemiologic
studies ranges from almost "sufficient" to "inadequate."
To reflect this range, the category is divided into
higher (Group Bl) and lower (Group B2) degrees of
evidence. Usually, category Bl is reserved for agents
for which there is at least limited evidence of
carcinogenicity to humans from epidemiologic studies.
In the absence of adequate data in humans, it is
reasonable, for practical purposes, to regard agents
for which there is sufficient evidence of carcinogenicity
in animals as if they presented a carcinogenic risk
to humans. Therefore, agents for which there is inadequate
evidence from human studies and sufficient evidence from
animal studies would usually result in a classification
of B2.
In some cases, the known chemical or physical
properties of an agent and the results from short-term
tests allow its transfer from Group B2 to Bl.
Group C - Possible Human Carcinogen
This category is used for agents with limited
evidence of carcinogenicity in animals in the absence
-------
82
of human data. It includes a wide variety of evidence:
(a) definitive malignant tumor response in a single
well-conducted experiment, (b) marginal tumor response
in studies having an inadequate design or reporting
(c) benign but not malignant tumors with an agent
showing no response in a variety of short-term tests
for mutagenicity, and (d) marginal responses in a
tissue known to have a high and variable background
rate.
In some cases; the known physical or chemical
properties of an agent and results from short-term
tests allow a transfer from Group C to B2 or from Group
D to C.
Group D - Not Classified
This category is used for agents(s) with inadequate
animal evidence of carcinogenicity.
Group E - No Evidence of Carcinogenicity for Humans
This category is used for agent(s) that show no
evidence for carcinogenicity in at least two adequate
animal tests in different species or in both epidemio-
logical and animal studies.
-------
83
TABLE 1
INCIDENCE (PERCENT) OF FEMALE CONTROL RATS BEARING THYROID
C-CELL TUMORS AMONG ANIMALS SACRIFICED POST 12-MONTHS**
(Same Lab)
STUDY*
ADENOMA or
CARCINOMA
ADENOMA
CARCINOMA
Group
Group
2
Group
Group
_3
Group
Group
4_
Group
Group
TOTAL
A
B
A
B
A
B
A
B
10/58 (
7/59 (
5/59 (
6/58 (
9/57 (
6/55 (
2/58 (
0/55 (
45/459(
17
11
8.
10
15
10
3.
0.
9.
•
•
5
•
.
•
4
2)
9)
)
3)
8)
9)
)
0)
8
)
10/58 I
6/59 I
5/59 1
6/58 1
6/57 1
5/55 1
2/58 1
0/55 1
40/4591
(17
(10
[8.
[10
[10
[9.
[3.
[0.
(8.
.2)
.2)
5)
.3)
.5)
0)
4)
0)
7)
0/58
1/59
0/59
0/58
3/57
1/55
0/58
0/55
5/459
(
(
(
(
(
(
(
(
(
0)
2)
0)
0)
5)
2)
0)
0)
1.1)
* Each listed study had two control groups, identified as
Group A or B. The rat strain is Sprague-Dawley.
TABLE 2
HISTORICAL CONTROL INCIDENCE OF LUNG TUMORS IN MALE B6C3F] MICE
RECEIVING CORN OIL BY GAVAGE*
(Different Labs)
LABORATORY
Alveolar/
Bronchiolar
Adenoma
Alveolar/
Bronchiolar
Carcinoma
Alveolar/
Bronchiolar
Ademoma or
Carinoma
A
B
C
D
E
F
8/100
12/235
5/120
19/150
4/49
32/248
( 8
( 5
( 4
(12
( 8
(12
.0%)
.1%)
.2%)
.7%)
.2%)
.9%)
6/100 1
17/235 <
3/120 1
4/150 1
3/49 1
11/248 1
: 6
: 7
[ 2
[ 2
[ 6
[ 4
.0%
.2%
.5%
.7%
.1%
.4%
)
)
)
)
)
)
14/100
29/235
8/120
22/150
7/49
43/248
(
(
(
(
(
(
14
12
6
14
14
17
.0%)
.3%)
.7%)
.7%)
.3)
.3%)
TOTAL
80/902 ( 8.9%)
44/902 ( 4.9%) 123/902 (13.6%)
** Nota Bene - This data is for illustrative purposes only.
It must not be used for any other purpose.
-------
84
TABLE 3
EXAMPLES OF NCI USE OF HISTORICAL CONTROL DATA*
EXAMPLE I
EXAMPLE II
REFERENCE:
(TR-160)
(TR-145)
LESION:
Hepatocellular Carcinoma
Endometrial Polyps
SEX/SPECIES: Male B6C3Fi Mice
Female Fisher 344 rats
TUMOR RATES:
Controls: 5/20, 25%
Low-Dose: 26/50, 52%
High-Dose: 27/50- 54%
Controls: 0/19, 0%
Low-Dose: 4/50, 8%
High-Dose: 9/50, 18%
SIGNIFICANT:
Trend: P = 0.039
Low-Dose: P = 0.035
High-Dose: P = 0.025
Trend: P = 0.018
Low-Dose: NS
High-Dose: P = 0.044
(borderline)
INTERPRETATION:
Neoplasm not related
to treatment
Polyps not related
to treatment
COMMENT: Historical control rate
137/422(32%) range up
to 58% compared to
5/20 (25%) in study
control group
Historical control rate
28/284(10%) compared
to 0/19(0%) in study
control group
NB: These examples are for illustrative purposes only.
Consult references (12) and (13) for full data base.
-------
85
TABLE 4*
SUMMARY OF SPONTANEOUS TUMORS OBSERVED UPON RE-EXAMINATION OF
SERIAL SECTIONS OF SELECTED TISSUES FROM 177 (63 MALES, 114
FEMALES) SPRAGUE-DAWLEY RATS
Type of tissue and tumor
Thyriod
light cell adenoma
Adrenal
pheochromocytonia
cortical carcinoma
Hypophysis
adenoma
Ovary
granulosa cell tumor
papillary adenocarcinoma
Uterus
leicmyosarcoma
endome trial polyp
Brain
ependymoma
papillona, choriod plexus
meningioma
pinealoma
Testes
Totals
No. of organs
140
143
50
61
51
126
45
No. of tumors
Single section vs. serial section
Male
4
5
1
3
-
-
-
-
1
0
0
1
0
15
Female
5
2
0
2
1
0
2
1
0
1
1
0
—
15
Male
24
7
1
4
-
-
-
-
1
0
0
1
0
38
Female
31
4
0
4
1
1
2
6
0
1
1
0
—
51
Age in days
300-960
540-930
690
360-900
600
660
480-720
420-690
330
660
510
480
—
*From reference (14) Table 1, p. 834.
Table 5*
Tumors and organs of origin in 2,082 rats of 6 sources-Continued
Tumors
Number of rats
Brain:
Sprague-
Dawley
258
2
Holtz-
man-SD
268
2
Charles
River-SD
535
3
Diablo-
SD
217
3
1
Osborne- Oregon
Mendel
131 . 673
1 4
. 1
Total
2,082
15
2
*Selected form reference (15) Table 2, pp. 1245-46.
-------
86
TABLE 6*
PROLIFERATIVE CHANGES AND
THEIR SYNOMYMS IN RAT LIVER
CHANGES
Hyperplasia (Not Neoplasm)
Benign Neoplasm.
Malignant Neoplasm.
SYNONYMS
Foci and areas of cellular alteration
(clear cell, basophilic-, acidophilic-
Vacuolted-)
Hyperplastic foci and areas
Basophilic hyperplasia
Hyperplastic nodule
Nodular hyperplasia
Adenoma
Neoplastic nodule
Hyperplastic nodule
Nodular hyperplasia
Hepatoma
Hepatic cell adenona
Trabecular carcinoma
Hepatocellular carcinoma
Hepatic cell carcinoma
Hepatoma
Hepatoma Type 1
Hepatoma Type II
Hepatoma malignant
Liver cell carcinoma
*From ref. (18)
-------
87
Table 7*
NEOPLASMS (TUMORS, NEW GROWTHS)
ONE OF THE DEFINITIONS: NEOPLASM IS AN UNCONTROLLED GROWTH OF CELLS
SOME CHARACTERISTICS
BENIGN NEOPLASMS
GROSS CHANGES:
1. Slow growth
2. Expansive type of growth
3. May be capsulated
4. Well defined contoure
5. Focal appearance
6. Not ulcerated
7. Usually not necrotic
HISTOPATHOLOGIC CHANGES;
1. Less anaplastic
2. Not metastatic or infiltrative
3. Moderate cellularity
4. Nuclear chromatin resembles normal
5. Moderate structural differences
from normal tissues
6. Low mitotic index
7. Moderate change in nucleus and
cytoplasm ratio
MALIGNANT NEOPLASMS
GROSS CHANGES
1. Fast growing
2. Infiltrative or metastatic
growth
3. Not capsulated
4. Undefined contoures
5. Diffuse or systematic appearance
6. May be ulcerated
7. May be necrotic
HISTOPATHOLOGIC CHANGES;
1. Highly anaplastic
2. Metastatic or infiltrative
3. Marked cellularity
4. Hyperchromatic nuclei
5. Marked structural differences
from normal tissues
6. Increased mitotic index
7. Marked change in nucleus
and cytoplasmic ratio
*From ref. (1%)
-------
88
Table 8*
Potential differences between chemcially-induced and tumors
in control rodents
Tumors in Induced
control animals tumors
Histogenesis
Hyperplasia not evident present
Preneoplastic lesions not readily present
evident
Precancerous lesions not evident present
Morphology
General characteristic sometimes
of tissue for different
strain of rodent fron usual
control tumor
Histologic tumor types one type several types
Strcmal lymphoid respones usually absent may be present
Biologic behavior often benign more often
malignant
Multiplicity singular multiple, often
involves entire
organ or tissue
*Ref. (19) Table 2 p. 282
-------
TABLE 9
LIVER (Animal No.)
Hepatocellular Carcinoma
Hepatocellular Adenoma
(No. Present)
Malignant Lymphonas
Granulocytic Leukemia
Angiosarcoma
Carcinoma, Metastatic
Sarcoma, Metastatic
Reticulum Cell Sarcoma
Hepatocholang iocarc incma
Multifocal Hepatocellular
Degeneration
Basophilic Foci
Mononuclear Cell Infiltration
Foci of Mononuclear Cells
Angiectasis
Focus of Cellular Change
Multifocal Hepatitis
Multifocal Necrosis
1
1
2
6
2
2
2
4
5
1
3
4
2
2
3
5
3
1
4
3
3
1
5
1
2
2
1
3
2
2
3
2
7
2
1
1
3
N
1
1
2
1
4
P
4
4
2
00
ID
Key
1 = Minimal 2 = Slight 3 = Moderate 4 = Moderately Severe High
5 = Severe/High 1 = Incomplete Section
-------
GROUP I
TABLE 10
GROUP II
GROUP III
GROUP IV
LIVER
(NO. EXAMINED)
Hepatocellular
Carcinoma
Hepatocellular
Adenoma*
Malignant
Lympnoma
Granulocytic
Leukemia
Angiosarccma
Carcinoma,
Metastatic
Sarcoma ,
Metastatic
Reticulum Cell
Sarcoma
Hepatocholang io-
carcinoma
Multifocal
Hepatocellular
Degeneration
Basqphilic Foci
Mononuclear Cell
Infiltration
Foci of Mononu-
clear Cells
Angiectasis
Focus of
Cellular Change
Multifocal
Hepatitis
Multifocal
Necrosis
Scheduled
Sacrifice
(22)
2/2
3
6
1
16
2
Moribund
Sacrifice
& Deaths
(52)
4
1/1
9
1
1
1
9
10
6
Total
(74)
4
3/3
12
1
1
1
15
1
26
8
Scheduled
Sacrifice
(34)
2
1/1
2
1
1
9
1
2
3
23
3
Moribund
Sacrifice
& Deaths
(42)
1
4/3
7
1
1
1
4
8
Total
(76)
3
5/4
9
1
1
1
10
1
3
3
27
11
Scheduled
Sacrifice
(24)
2
30/12
1
1
9
13
1
Moribund
Sacrifice
& Deaths
(52)
1
24/11
7
2
1
2
2
1
7
6
Total
(76)
3
54/23
8
2
1
3
11
1
20
7
Scheduled
Sacrifice
(22)
1
54/17
1
8
2
1
14
2
Moribund
Sacrifice
& Death
(53)
1
21/12
6
1 1
1
1
3
2
8
1
3
9
6
Total
(75)
2
75/29
6
1
1
1
3
3
16
1
5
1
23
8
* Number of neoplasms/number of animals with neoplasms.
-------
Tissue
91
Table 11*
Guidelines for Combining Benign and Malignant Neoplasms
in the Fischer 344 Rat and B6C3F1 Mouse
Tumors Combine
Liver
Mammary Gland
Thyroid
Pituitary
Lung
Hematopoietic System
Neoplastic nodule-rat or
Hepatocellular adenoma-mouse Yes
Hepatocellular carcinoma
Bile duct ademoma
Bile duct carcinoma Yes
Fibroma Yes
Fibroadenona
Carcinoma Yes
Adenocarc inoma
F ibrona/Fibroade nona
Carcincma/Adenocaricinona No
Follicular cell adenoma Yes
Follicular cell carcinoma
C-cell adenoma Yes
C-cell carcinoma
Follicular cell tumors No
C-cell tumors
Adenoma Yes
Carcinoma
Bronchioalveolar adenoma Yes
Bronchioalveolar carcinoma
Rat
Leukemia
mononuclear cell (Fischer rat)
Lymphocytic Yes
Undifferentiated
Myelogenous Leukemia No
Leukemias-other types
Malignant lymphoma (lypnosarcoma)
Lymphocytic Yes
Lymphoblastic
Histiocytic
Reticulum Cell
Mixed cell
*Frcm ref. (20) pp 7-11
-------
92
Pancreas
Gastrointestinal Tract
Leukemias-all types No
Lymphomas-all types
Mouse
Lymphocytic Leukemia Yes
Undifferentiated Leukemia
Myelogenous leukemia No
Leukemia-other types
Malignant lymphoma (lymphosarcona) Yes
Lymphocytic
Lymphoblast ic
Histiocytic
Reticulum cell
Leukemias-all except myelcgenous Yes
Lymphomas-all type
Islet cell adenoma Yes
Islet cell carcinoma
Acinar cell adencma Yes
Acinar cell carcinoma
Islet cell tumors No
Acinar cell tumors
Forestomach papillcmas Yes
Squamous cell carcinomas
Glandular region and intestine
Adencmas/Adenomatous polyps Yes
Adenocarcinomas
Kidney
Leionyanas
Le iomyosarcomas
Fibrcmas
Fibrosarcomas
Squamous cell tumors
Glandular tumor
Mesenchymal tumor
Leiomycmas/leionyosarccmas
Fibronas/f ibrosarconas
Tubular cell adenoma
Tubular cell carcinonas
Transitional cell papillomas
Transitional cell carcincmas
Yes
Yes
No
No
Yes
Yes
-------
93
Urinary Bladder
Skeletal System
Adrenal Gland
Brain
Ovary/Testicle
Lipomas
Liposarcomas
Transitional cell tunors
Tubular cell tunors
Liponatous tunors
Other types of renal tunors
Transitional cell papillcmas
Transitional cell carcinomas
Osteona
Osteosarcoma
Crondrcma
Chrondrosarcoma
Osteoraa/Osteosarcana
Chrondrona/Chrondrosarcana
Cortical adenomas
Cortical carcinomas
Pheochronocytoma
Malignant pheochronocytoma
Cortical tunors
Medullary tumors
All glicmas, i.e.
Oligodendroglioma
Astrocytoma
Granular cell tunors
Gliomas
Nerve cell tumors
Gliomas
Meningionas-all types
Other CNS tumors
Germ cell tumors-all types
Stromal tumors-all types
Germ Cell tumors
Stromal tumors
Yes
No
No
Yes
Yes
Yes
No
Yes
Yes
No
Yes
No
No
No
Yes
Yes
No
-------
94
Uterus
Stronal polyps
Stronal sarconas
Glandular adenonas
Adenocarcinomas
Stronal tumors
Glandular tumors
Yes
Yes
NO
Integument
Subcutis
Preput ial/Clitoral
Gland
Zymbal Gland
Nasal Cavity
Basal cell tumors all types Yes
P ilomatr ixoma
Sebaceous gland tumors-all types Yes
Squamous cell papilloma Yes
Squamous cell carcinoma
Squamous cell tumor No
Adexal tumors
Basal cell tumors
Squamous cell tumors
Keratoacanthoma
Squamous cell carcinoma
Fibrcmas
Fibrosarcomas
Hemangiomas
Hemagiocarccmas
Leionycmas
Le iomy osarcomas
Fibronas/f ibrosarconia
Liecmyonas/le icmyosarccmas
Connective tissue tumors
Endothial tumors
Adenoma
Carcinoma
Adenona
Carcinoma
Adenona
Adenocarcinoma
No
No
Yes
Yes
Yes
No
No
Yes
Yes
Yes
Squamous cell tumors NO
Glandular tumors
Esthestioneuralepithelia tumors No
Other tumors
-------
95
TABLE 12
Total Number of Rats with Neoplasms (all types)
Dosage
Male Group Females
No. _% No. %
46/60 76.7 1* 49/60 81.7
38/58 65.5 II 48/60 80.0
37/57 64.9 III 47/59 79.7
38/58 67.2 IV 47/59 79.7
*Gontrol
-------
TABLK 1.1*
Inr.lilcince of Neoplasm in Mice Kllluil at End of Test
1'i.l.iiy uuimt 1^.)
I' 111) n./ dl. |M)
I'UI) roiritli.il (M)
•11 II ml (id (M)
•4111 renrii.li.il (M)
I'll It .! i.^iulul (1 )
4111 nJ III. (1 )
4lllre,.liLU-l|l)
run <»/ MI i/.t xi i )
I'KI) tc.lilLii.l (M & H
•till ni/ lid |M A 1 )
4111 resiiMcil (M J< II
I'll I)/-*! II u./ Jib. (M)
I'li I>/4III (..iiiKicil (M)
I'lm^lll HI/ /i/i (1 )
I'IU)/-4III loiiiiiul If)
I'KIV-IIU ii./ Jit>. (M i 1 I
I'Un/llll fL'tmcJcri (M A 1 )
I'lsl) in/ /•/. /iciiiiclcil (M A F)
4111 ».' Jifi./ic-.liKlcil (M & !•)
N.I nl in...
lilkil .ii
eml nf K'.ij
IU
SS
'IS
SI
48
U
51
S'J
%
II VI
96
no
9)
106
')<)
12)
192
21")
215
JlV,
N.'ii|ilu>mv
.ii .my sue
,.r..
U
28"*
16
12"'
4
14*
8
31""
12
•12""
24
-17""
21
2t"*
12
73""
36
•4JIII
do
Mahdum
l.v.,i|,l.i;nu
Jl •'">• l"c
4
1
4
1
3
1
t>
3
7"
2
IU"
4
1"
2
9'
4
17"*
6
91
14
1 nng
Itiimniii
IU
i
13
IU
7"
2
9
5
17"
8
22*
15
23
16
16"
7
39"*
23
25
37
Mulli|ile
lung
linnmirs
2
0
1
3
0
0
0
0
j
P)
1
3
3
3
0
0
3
3
2
4
No nl
nine with
I. mil; l.ung
Inilinui of lumniir of Multiple
guile 2
nr mine)
7
4
10'
5
5
2
4
1
12*
6
14"
6
17"
9
9"
3
26"*
12
18
JO
guile 3
or rn.ncj
3
1
j
U
0
0
2
0
j
1
2
0
5
1
2
0
7
1
4
•1
Liver
luimmil
II"'
2
21""
6
2
1
)•
0
13"*
3
24""
6
32""
8
5*
1
37""
9
16(11
30
liver
Illlllllllll
•I'-
ll
7...
0
(l
0
0
0
4"
0
7...
0
11""
u
0
0
II*"*
0
4
7
l.ivcr
liiiiiiMir
guile U,
C1 or O|
7*
2
8
5
U
1
2
0
7
3
"V
5 .
15"
7
2
1
17"
H
10
15
l.iver Nen|>l.iuiit
minimi ,'ii >HLV
BUI
C or
;,
(I
1
U
0
U
1
0
u
0
2
0
1
0
1
0
2
(I
0
2
le ullicr linn
0| lijnj nl liver
3
1
j
1
4'
1
5
1
7"
2-
7
i
5
i
V
4
14"
L
•1
II
M - Mule F - female
| A liil.il nf 19 mice lli.il :il liwluceit wei:l.i 80 anil 8> li.ivc lx.cn eicluvlml
(I. mi); iiiiiiiinu tvxi>: ui.tili:.! nt follows (luieit uu Wjlieik. ' |%6): I - lhciii£n nun-iuvasivc iilcnomu; 2 = tilcnoina (cucmling inln an way i mul/ur inln iiirmurnlini; lunn):
3 - ,nlciiiKMii.iiiiii)i.i wild iin-l.iii.iics in lnl>c uf tni|;iii .it ciilircly rc[ilu^ui|; one lol>c; 4 - aJcnncarcinoriu tilciulinu llirou^li pleuin ui niei.iil.iii/im; lo lulio nllu'i llun Use lode
1. 1 .mi-lit; 5 - lull i.in.ilClil.llll.i Hi>:t.i^l:i>l/lii|! In HIU'I nnlklilc IliC llmrji
|l i.ci luiii.inu \vcic I'.i.iilnl .it liillnwt: A - umtiMi n( uluuiii ndfiii.cl I.M.lin^ liver cclK in alinoil iiornnil iiirnn|;ciiicMi: II *f cunsiili nl icmuiiK.ililc iiaicncliyniiil rclU jiunc-cJ irt
i.nih. C '• iiiiiliiniiicilly iii.iliijii.ini livi'r ttll luniinii ili.ii luii mil mci.uljiiml cltetvltcrc; () >- riieiiiiiusi/ni|: liver fell luriiiuir.
V'.il.ic. in.nU.I uiili ii.icn:.!. » .no ii|;uilir.ii.lly ln^lui ili.iu UK cuiiciii.niJin^ valiK. |;ivcn in llic line below )*(' < (I I . " /' < OtIS; '"(' v Oil) ; ••••/' -' 01X111; Iliuse ni.nl ej uilli
,li(1:..ii JK j^nllK.nill! litvvci III, ill llic iilliliiuiilil'int; v.ilni". Ill llie line liclnn (1C •; 01; III/' < (Mil)
\o
a
a.
I
*!-'r.uin (35) Tal.lt! 3, p. 71
-------
97
TABLE 14*
Most Commonly Induced Tumors in the 98 Positive NCI Bioassays
Site or
Tumor
Type
Liver
Mammary Gland
Lymphoma/Leukemia
Lung
Urinary Bladder
Forestomach
Thyroid
Hemangiosarccma
Uterus
Zymbal Gland
Rat
Male
18
1
5
1
6
7
6
3
—
7
Female
15
13
3
2
10
5
5
1
7
8
No. Rat
Bioassay
Involved
21
13
6
2
11
7
6
3
7
9
Mouse
Male
31
0
4
7
2
5
4
6
-
0
Female
44
3
7
7
2
5
4
5
3
1
No. Mouse
Bioassays
Involved
50
3
8
7
2
5
4
7
3
1
Total No.
Bioassays
Involved
55
16
13
13
11
11
11
10
10
9
-------
SELECTED TUMOR INCIDENCE PATTERNS
60
50
40
30
20
10
0
16/76
0.0
60
60
40
30
20
10
8/72
'44/75
35/75
24/75
20 500 2000 ppm
CD -1 MICE Lung Adenocarcinoma
18/73
J9/73
1/72
0.0 20 500 4000ppm
? CD-1 MICE Retlculo-endothelial Tumor*
60
50
40
30
20
10
0
-11/70
0.0
60
50
40
30
20
10
0/118
B
15/70,
8/70
10/70
250 1000
? SWISS MICE Lung Adenoma
2500 ppm
vo
00
51/117,
0/118
1/119,
0.0 2.0 5.6 14.3ppm
FISHER 344 Rats Squamous-cell carcinoma
-------
TABLE 16*
Typical standard Carcinogens
Carcinogen
Diethylnitrosamine
Diethylnitrosamine
N-2-Fluorenylacetamide
N-2-Fluorenylacetamide
N-2-Fluorenylacetamide
Uracil mustard
Species
(strain) Sex
Rat N or F
(Fischer)
Rat M or F
(CR-SD)
Rat M
(CR-SD)
M
F
Rat F
( Fischer)
Mouse M or F
M or F
Rat M
( Spr ague-Daw ley )
M
F
F
Route
Oral
Oral
Oral
Oral
Oral
Oral
Oral
Oral
i.p.
i.p.
i.p.
i.p.
Main target
Dose Organ and incidence
40 ppm
( in water)
51 ppm
( in water)
223 ppm
( in diet)
80 ppm
(in diet)
80 ppm
(in diet)
2mg
(by gavage,
5 days per
week)
740 ppm
240 ppm
11.5 mg/kg
23 mg/kg
11.5 mg/kg
23 mg/kg
Liver 100%
Liver 100%
Liver 90%
Liver 30%
Breast 50%
Breast 20%
Liver
Liver
Pancreas 4%
Lymphoma 13%
Peritoneum 22%
Pancrease 8%
Lymphoma 25%
Peritoneum 33%
Breast 55%
Lung 10%
Lymphoma 10%
Peritoneum 10%
Breast 53%
Lung 7%
Lymphoma 14%
Ovary 20%
(0.5% controls)
Peritoneum 40%
Latent
period
(weeks) References
20 13, 20
35 17
26-40 17
60-90 Unpublished
60-90 Unpublished
30-40 7
40 Unpublished
90 Unpublished
65 to death 6
50 to death 6
71 6
56 6
vo
vo
*Fron ref. (49) pp 31-33
-------
Carcinogen
Uracil mustard
Uracil mustard
Urethane
N,N-Demethyl-4-
stilbenamine
3-Methyl-4-dimethyl-
aminoazobenzene
Nitrogen mustard
7,12-Dimethalbenz(a)-
anthracene
7,12-Dimethylbenze(a)-
anthracene
3-Aminotriazole
Species
(strain)
Mouse
Mouse
Mouse
(VHe)
Rat
Rat
Mouse
(VJ)
Rat
(SD)
Mouse
Rat
Sex
M or F
M or F
M or F
M
M
F
F
M or F
M or F
M
M
M or F
M or F
M or F
F
M or F
M or F
M
F
Route
i.p.
i.p.
i.p.
i.p.
i.p.
i.p.
i.p.
i.p.
i.p.
Oral
( in feed)
Oral
(in feed)
i.p.
i.p.
i.p.
Oral
Skin
Oral
Oral
Oral
Main target
Dose organ and incidence
0.008 g/kg
0.020 g/kg
0.040 g/kg
9.3 mg/kg
19.3 mg/kg
9.3 mg/kg
19.3 mg/kg
10 mg
20 mg
0.004%
0.05%
0.21 mg/kg
0.87 mg/kg
3.4 mg/kg
15-20 mg
(by tube)
75 mg
300 ppm
300 ppm
300 ppm
Lung 100%
Lung 100%
Lung 100%
Lung 64%
Lung 50%
Lung 60%
Ovary 25%
Lymphoma 50%
Lung 60%
Lymphoma 40%
Ovary 33%
Lung 100%
Lung 100%
Ear duct 63%
Liver 55%
Lung 40%
Lung 69%
Lung 95%
Breast 92-100%
Skin
Thyroid
Liver 65%
Liver 48%
Latent
period
(weeks) I
24
24
24
61 to death
45 to death
58
69
24
24
38
37
39
39
39
12-16
10-25
tefereni
16
16
16
6
6
6
6
16
16
14a
14a
15
15
15
19
3
14
14
14
o
o
-------
Carcinogen
3-Aminotriazole
Safrole
Species
(strain)
Mouse
(C57Bl/6XC3H/Anf)f1
Rat
(Osborne-Mendel)
Sex
M or F
Route
Oral
Mouse M or F
(C57Bl/6XC3H/Anf)fi
Oral
Dose
2,192 ppm
Latent
Main target period
organ and incidence (weeks) Reference
M or F Oral 5,000 ppm
1,112 ppm
(in diet)
Thyroid
Liver
Liver
Liver
78
104
82
10
11
10
-------
102
REFERENCES
(1) National Research Council. (1983). Risk Assessment in the
Federal Government: Managing the Process, National Acad.
Press, Washington, D.C. p. 3.
(2) Ruckelshaus, W.D., (1984). Proposed guidelines for carcinogen
risk assessment. Fed. Reg. 49, pp. 46294-46301.
(3) Train, R., (1976). Health risk and economic impact assess-
ments of suspected carcinogens. Interim procedures and
guidelines. Fed. Reg., Vol. 41, No. 102, Tuesday, May 25, 1976
pp. 21402-21405.
(4) Ibid, p. 21403.
(5) Ibid, p. 21403.
(6) Office of Science and Technology Policy., (1984). Chemical
carcinogens: Review of the science and its associated prin-
ciples, Fed. Reg. 49, pp. 21595-21661.
(7) International Agency for Research on Cancer., (1982).
Monographs on the evaluation of carcinogenic risk of
chemicals to humans. Supplement 4. IARC, Lyon, France.
(8) International Agency for Research on Cancer., (1979). Scien-
tific basis for identification of potential carcinogens and
estimation of risk. J. Natl. Cancer Inst. 63, 245-267.
(9) Peto, R. Pike, M. Day, N., Gray, R., Parish, S., Peto, J.,
Richard, S., and Warrendorf, J., (1980) Guidelines for simple
sensitive, significant tests for carcinogenic effects in long-
term animal experimetns. In Monographs on the Long-Term and
Short- Term Screening Assays for Carcinogens; A Critical
Appraisal. Supplement 2, pp. 311-426., IARC. Lyon, France.
(10) Mantel, N., (1980). Assessing laboratory evidence for neo-
plastic activity. Biometrics, 36, pp. 381-399.
(11) Mantel, N., and Haenszel, W., (1959). Statistical aspects
of the analysis of data from retrospective studies of disease.
J. Natl. Cancer Inst. 22, pp. 719-748.
(12) Interdisciplinary Panel on Carcinogenicity. (1984). Criteria
for evidence of chemical carcinogenicity. Science. 225,
pp. 682-687.
-------
103
(13) National Center for Toxicology Research. (1981), Guidelines
for statistical tests for carcinogenicity in'chronic studies.
NCTR Tech. Rep. 81, p. 100.
(14) National Toxicology Program. (1984). Report of the ad hoc
panel on chemical carcinogensis testing and evaluation of
the NTP Board of Scientific Counselors.
(15) EPA. (1983) Good Laboratory practice standards—toxicology
testing. Fed. Reg. 48, p. 53922.
(16) EPA. (1983) Hazard evaluations: human and domestic animals.
Subdivision F. PB 83-153916.
(17) EPA. (1983). Health effects test guidelines. PB-83-232984.
(18) WHO International Agency for Research on Cancer (1983).
Evaluation of the Carcinogenic Risk of Chemicals to humans.
Vol. 30, p. 14. Lyon.
(19) Food and Drug Administration (1982). Toxicological
principles for the safety assessment of direct food additives
and color additives used in food.
(20) Interagency Regulatory Liasion Group (IRLG). 1979. Scien-
tific Basis for Identification of Potential Carcinogens and
Estimation of Risks, J_. Natl. Cancer Inst. 63; pp. 245-267.
(21) Peron, V.J., Grice, B.C., Griesmer, R., Peto R., Agthe, C.,
Althoff, J., Arnold, D.L., Blumenthal, H., Cabral, J.R.P.,
Delia, Porta, G., Ito, N., Kimmerle, G., Kroes, R., Mohr, U.,
Napalkov, N.P., Odashima, S. , Page, N.P. ,' Schram, T.,
Steinhoff, D., Sugar, J., Tomatis, L., Uehleke, H. and
Vouk, V. 1980. Basic Requirements for Long-Term Assays for
Carcinogenicity. pp 21-83 in Long term and Short-term Screening
Assays for Carcinogens: A critical Appraisal. IARC Monographs,
Supplement 2. International Agency for Research on Cancer,
Lyon, France.
(22) Doull, J., (1980). Factors influencing toxicology. Casarett
and Doull1s Toxicology, 2nd Ed., pp. 70-83, Macmillian Pub.
Co., New York.
(23) Ward, J.M., (1983). Background data and variations in tumor
rates of control rats and mice. Prog. Exp. Tumor Res.,
26, pp. 241-258.
-------
104
(24) Task Force of Past Presidents. (1982). Animal data in
hazard evaluation: Paths and pitfalls. Fundam. and Appl.
Toxicol., pp. 101-107.
(25) NCI. (1979). Bioassay of 2,4,5-Trimethylaniline for possible
carcinogenicity. NCI-CG-TR-160, Table E 2., p. 105.
(26) NCI. (1978). Bioassay of 3-Chloro-p-Toluidine for possible
carcinogenicity. NCI-CG-TR-145, Table 4, p. 27.
(27) Thompson, S.W., and Hunt, R.D., (1963). Spontaneous tumors
in the Sprague-Dawley rat: Incidence rates of some types of
neoplasms as determined by serial section versus single section
technics. Ann. N.Y. Acad. Sci., 108, pp. 832-845.
(28) MacKenzie, W.F., and Garner, F.M., (1973). Comparison of
neoplasms in six sources of rats. J. Nat. Can. Inst., 50,
pp. 1243-1257.
(29) Ward, J.M., and Rice, J.M., (1982). Naturally occurring and
chemically induced brain tumors of rats and mice in carcino-
genesis bioassays. Am. N.Y. Acad. Sci., pp. 304-319.
(30) Zbinden, G. , (1976). The role of pathology in toxicity
testing. Progress in Toxicology 2, pp. 8-18, Springer-Verlag,
New York.
(31) Kasza, L., (1983). Proliferative liver changes. Hazard
Evaluation Division, Toxicology Branch Seminar (unpublished).
(32) Ward, J.M., and Reznik, G., (1983). Refinements of rodent
pathology and the pathologist's contribution to evaluation
of carcinogenesis bioassays. Prog. Exp. Tumor Res., 26,
pp. 266-291.
(33) N.T.P. (1982). Guidelines for combining benign and malignant
neoplasms as an aid in determining evidence of carcinogenicity.
Draft Working Paper, pp. 1-12.
(34) Baseman, J.K., (1983). Patterns of tumor incidence in two-year
cancer bioassay feeding studies in Fisher 344 rats. Fundam.
Appl. Toxicol., 3, pp. 1-9.
(35) Canybeare, G., (1980). Effect of quality and quantity of diet
on survival and tumor incidence in outbred Swiss mice. Food
Cosmet. Toxicol., 18, pp. 65-75.
-------
105
(36) Hottendorf, C.H., and Pachter, I.J., (1982). An analysis
of the carcinogenesis testing experience of the National
Cancer Institute. Tox. Path., 10, pp. 22-26.
(37) Grasso, P., and Crampton, R.F., (1972). The value of
the mouse in carcinogenicity testing. Food Cosmet.
Toxicol., 10, pp. 418-426.
(38) Butler, W.H., and Newberne, P.M., (ed), (1975). Mouse
Hepatic Neoplasia. Elsevier, Amsterdam, pp. 1-5.
(39) Tomatis, L., Partenshy, C., and Montesano, R., (1973).
The predictive value of mouse liver tumour induction in
carcinogenicity testing - a literature survey. Int. j.
Cancer, 12, 1-20.
(40) Doull, J.D., (Chairman) (1983). A report of the international
expert advisory committee to the Nutrition Foundation.
The relevance of mouse liver hepatoma to human carcino-
genic risk. Nutrition Foundation. Wash., D.C.
(41) Vesselinovitch, S.D., and Mihailovich, N., (1983)
Kinetics of diethylnitrosamine hepatocarcinogenesis in the
infant mouse. Cancer Research. 43, pp. 4253-4259.
(42) Schach, von Wittenau, M., and Estes, P.C., (1983). The
redundancy of the mouse carcinogenicity bioassays.
Fundam. Appl. Toxicol. 3, pp. 631-639.
(43) Klaassen, D., and Doull, J., (1980) Evaluation of
safety: Toxicologic evaluation. Casarett and Doull's
Toxicology, 2nd Ed., p. 18, Macmillan Pub. Co., New York.
(44) Roe, F.J.C., (1983). Testing for Carcinogenicity
and the problem of pseudo-carcinogenicity. Science
308, p. 657.
(45) Friedman, L., (1973) Section 4 report of discussion
group No. 3 dose selection and administration. In
Carcinogenesis Testing of Chemicals. L. Golberg,
(ed.): CRC Press, Inc., Boca Raton.
(46) Gori, G.B., (1980). The regulation of carcinogenic
hazards. Science, 208, pp. 256-261.
(47) Doull, J.D., (Chairman) (1984). Ad Hoc Panel on Chemical
Carcinogenesis Testing and Evaluation. Draft report.
2/15/84, pp. 129-157. NTP Board of Scientific Counselors.
(48) Jukes, T.H., and Shaffer, C.B., (1963). Antithroid
effects of Aminotriazole. Science, 132, p. 296.
-------
106
(49) Steinhoff, D., Weber, H., Mohr, U., and Boehme, K. ,
(1983). Evaluation of Amitrol (Aminotriazole) for
potential carcinogenicity in orally dosed rats, mice,
and golden hamsters. Tox. Appl. Pharma., 69, pp. 161-169.
(50) Sevarm, R.L., Roberts, G.K.S., Levy, A.C., and Hines,
L.R., (1973). Observations on the thyroid gland in rats
following the administration of sufamethoxazole and
trimethoprim. Tox. Appl. Pharma., 24, pp. 315-363
(51) Murphy, S.D., (1980). Pesticides. Casarett and Doull's
Toxicology, 2nd Ed., p. 392, Macmillan Pub. Co., New York.
(52) Jensen, R.K., Sleight, S.D., Aust, S.D., Goodman, J.I., and
Troskos, J.E., (1983). Hepatic tumor-promoting ability of
3,3',4,4',5,5'-hexabromobiphenyl: The interrelationship
between toxicity, induction of hepatic microsomal drug
metabolizing enzymes, and tumor-promoting ability. Tox.
Appl. Pharma. 71, pp. 163-176.
(53) Shank, R.C., and Barrows, L.R., (1981). Toxicity -
dependent DNA methylation: Significance to risk
assessment. Health Risk Analyses. C.R., Richmond,
P.J., Walsh, E.M.D., Copenhaven (eds), The Franklin
Inst. Press, pp. 225-233.
(54) Bosan, W.S., and Shank, R.C., (1983). Methylation of
liver DNA guanine in hamster given hydrozine. Tox.
Appl. Pharma., 70, pp. 324-334.
(55) Wolf, F.J., (1980) Effect of overloading pathways on Toxicity,
J. Envir. Path. Tox., 3, pp. 113-134.
(56) Council on Scientific Affairs. (1981) Carcinogen regulation.
JAMA, 246, p. 254.
(57) Tennekes, H.A., Edler, L. and Kunz, H.W., (1982). Dose-
response analysis of the enhancement of liver tumour
formation in CF-1 mice by dieldrin. Carcinogenesis,
3, pp. 941-945.
(58) Gehring, P.J., Watanable, P.G., Young, D.J., and Lebeau,
J.E., (1976). Metabolic Thresholds in assessing carcinogenic
hazard. Collection of DOW Sci., Pap., 2, pp. 56-70.
(59) Dorland's Illustrated Medical Dictionary. (1974). 25
Ed., p. 1605, W.B. Saunders, Philadelphia
(60) Gehring, P.J., and Blau, G.E., (1977) Mechanisms of
Carcinogenesis: Dose response. J. Environ. Path. Tox.
1, pp. 163-179
-------
107
(61) EDQI Task Force (1981) Re-examination of the EDgi Study-
Summary and conclusions. Fundam. Appl. Toxicol. 1: pp. 127-128
(62) Weisburger, J.H., (1974). Report of discussion group No.5
Inclusion of positive control compounds. Carcinogenesis
Testing of Chemicals. pp. 29-34, CRC Press, Inc. Boca
Raton.
(63) Druckery, H., (1967). Potential Carcinogenic Hazards from
Drugs. Trugaut, R. (Ed.): UICC monograph Series 7, pp. 60-76.
Spriner-Verlag, Berlin.
(64) Weisburger, J.H., and Williams, G.M., (1980). Chemical
carcinogens. Casarett and Doull's Toxicology, 2nd. Ed.
Macmillan Pub. Co., New York.
(65) International Commission for Protection Against Environmental
Mutagens and Carcinogens (1982). Committee 2 Final Report,
Mutagenesis testing as an approach to carcinogenesis.
Mutation Res., 99, pp. 73-91.
(66) Wright, A.S., (1980) ICPEMC Working Paper 2/2. The role of
metabolism in chemical mutagenesis and carcinogenesis.
Mutation Res. 75, pp. 215-241; Biol. Zentralbl. 99, pp. 207-304.
(67) Anderson, M.E., (1981). Saturable metabolism and its
relationship to toxicity. CRC Grit. Rev. Tox. pp. 105-150.
(68) Smyth, R.D., and Hottendorf, G.H., (1980) Application of
Pharmacokinetics and biopharmaceutdics in the design of
toxicological studies. Tox. Appl. Pharma. 53, pp. 179-195.
(69) Yacobi, A., Kamath, B.L., and Fai, C., (1982) Pharmacokinetics
in chronic animal toxicity studies. Drug Metabol. Rev.
13(6) , pp. 1021-1051.
(70) Miller, E.G., and Miller, J.A., (1966). Mechanisms of
chemical carcinogens: Nature of proximate carcinogens
and interactions with macromolecules. Pharma. Rev. 18,
pp. 805-838.
(71) Miller, J.A., and Miller E.C., (1976). The Metabolic
activation of chemical carcinogens to reactive
electrophiles. In Biology of Radiation Carcinogenesis.
J.M. Yuhas, R.W., Tennant, and J.D., Regan (eds.), Raven
press, pp. 147-163.
(72) Park, D.V., (1982). Mechanisms of chemical toxicity - A
unifying hypothesis. Reg. Tox. Pharma. 2, pp. 267-286.
-------
108
(73) Mitchell, J.R., Hughes, H., Lauterburg, B.H., and Smith,
C.V. , (1982), Chemical nature of reactive intermediates
as determinant of toxicologic responses. Drug Metabol.
Rev. 13(4) , pp. 539-553.
(74) Ramsey, J.C., and Gehring, P.J., (1980). Application of
pharmacokinetic principles in practice. Federation
Proc. 39, pp. 60-65.
(75) Reltz, R.H., Quast, J.F., Schumann, A.M., Watanabe
P.G., and Gehriing, P.J., (1980). Non-linear
pharmacokinetic parameters need to be considered in
high dose/low dose extrapolations. In Quantitative Aspects
of Risk Assessment in Chemical Carcinogenicis Arch.
Toxicol., Suppl. 3, pp. 79-94.
(76) Dietz, F.K., Stott, W.T., and Ramsey, J.C., (1982).
Nonlinear pharmacokinetics and their impact on toxicology,
illustrated with dioxane. Drug Metabol. Rev. 13(6), pp. 963-981
(77) Hoel, D.G., Kaplan, N.L., and Anderson, M.W., (1983).
Implications of nonlinear kenetics on risk estimation in
carcinogenesis. Science 219, pp. 1032-1037.
(78) Clayson, D.B., Krewski, D., and Munro, I.e., (1983). The
power and interpretation of the carcinogenicity bioassay.
Reg. Tox. Pharma. 3, pp. 329-348.
NB. The following references are from an internal EPA
document which contains confidential data. Removal of such
data will eliminate the instructional values of the document.
Therefore the document must not be released from HED without
the proper authorization and legal clearances.
(79) Paynter, O.E., Budd, E., and Litt, R., (1982). Permethrin:
Responses to a Peer-Review. Unpublished HED document.
Comments 3 and 3a., pp. 3-16.
(80) Ibid., comments 2a., 2b., pp. 35-39 and comments 3a.,
3b., pp.39-48.
(81) Ibid., comments 4 through 5a. pp. 17-22.
(82) Ibid., paragraphs 1 and 2., pp. 134.
(83) Ibid., comments 2a. through 3a., pp. 84-92.
(84) Ibid., comments 4., 4a., pp. 92-97.
(85) Ibid., comments 2a., 2b., pp. 35-38.
-------
109
(86) Ibid., comments 3., 3a., pp. 74-76.
(87) Ibid., comments 4., 4a., pp. 104-109
(88) Ibid., comments 2., 2b., pp. 25-28.
(89) Ibid., comments 7b., pp. 59-60.
------- |