Human Carcinogenic Risk Evaluation, Part III: Assessing Cancer Hazard and Risk in Human Drug Development
Abigail Jacobs1 and
David Jacobson-Kram
U.S. Food and Drug Administration, Rockville, Maryland
Received January 13, 2004;
accepted April 7, 2004
 |
ABSTRACT
|
---|
Assessing cancer risk for human pharmaceuticals is important because drugs are taken at pharmacologically active doses and often on a chronic basis. Epidemiologic studies on patient populations have limited value because of the long latency period for most cancers and because these studies lack sensitivity. The Center for Drug Evaluation and Research (CDER) of the U.S. Food and Drug Administration relies on short-term surrogate assays (genetic toxicology studies) to assess risk to patients involved in clinical trials and on rodent carcinogenicity studies to assess cancer risk for drug approval. Unlike some other agencies that typically perform quantitative risk assessments on chemical pollutants or pesticide products, CDER does not perform such quantitative extrapolations. Rather, the evaluation of risk is the result of an integrated assessment of what is known about the drug, and risk is considered in the context of the clinical benefit. Mode of action of carcinogenesis and thresholds for effects are important considerations. The results of carcinogenicity studies of approved products are published in the drug labeling and individual clinicians balance risk and benefit in making prescribing decisions.
Key Words: carcinogenicity; risk; drug development.
 |
INTRODUCTION
|
---|
Assessment of hazard and risk varies throughout drug development as more persons are exposed for longer periods of time and more nonclinical information on the hazard is collected and evaluated. During the drug discovery period, the sponsor may cease development of compounds found to be clearly genotoxic in in vitro screens, depending on possible therapeutic indications. If a compound is negative in the in vitro screen and in the subsequent in vivo study for chromosomal damage, carcinogenicity studies may not need to be reported until submission of the New Drug Application (NDA) to CDER. It is understood that some drugs are nongenotoxic carcinogens and may still present a risk even when results of all genetic toxicology assays are negative. However, the modes of action associated with these types of carcinogens are thought to involve thresholds. Per the International Conference on Harmonization (ICH) Guideline S1A, carcinogenicity studies should be performed for any pharmaceutical whose expected clinical use is continuous for at least 6 months or repeatedly intermittent for chronic conditions. Other possible causes for concern and reasons for conducting carcinogenicity studies include the following:
- Previous demonstration of carcinogenic potential in the product class that is considered relevant to humans
- A structure-activity relationship that suggests a carcinogenic risk
- Evidence of preneoplastic lesions in repeated-dose toxicity studies
- Long-term tissue retention of parent compound or metabolite(s) resulting in local tissue reactions or other pathophysiologic responses
 |
FIRST-IN-HUMAN STUDIES
|
---|
For first-in-human phase 1 studies, a single dose of a genotoxic compound may be acceptable for administration to small numbers of healthy humans. Genotoxicity here is defined as a positive result in an in vitro assay for mutations or chromosomal damage (ICH S2B). A single positive or equivocal result may warrant the conduct of further genotoxicity studies. Exposure of small numbers of healthy humans or patients with nonserious conditions to multiple doses of a nonalkylating genotoxic compound that was negative in in vivo evaluations may also be acceptable for a limited amount of time, based on a weight-of-evidence evaluation of all the genetic toxicology findings. Such a weight-of-evidence evaluation would include mode-of-action considerations. For some life-threatening indications, it may be acceptable to expose larger number of patients for longer durations to genotoxic drugs. When drugs are positive in genotoxicity assays, persons administering or receiving such drugs will be informed in the patient information material or in the investigator's brochure. It is recognized, as per ICH S2A, that the test battery approach of in vitro and in vivo tests is designed to reduce the risk of false negative results for compounds with genotoxic potential. A single positive result in any assay for genotoxicity does not necessarily mean that the test compound poses a genotoxic hazard to humans. In practice, most carcinogenicity findings in rodents for drugs in CDER appear to be attributable to hormonal or immunosuppressive mechanisms or exaggerated pharmacologic actions of the drug, regardless of genotoxicity findings. Neoplasm induction resulting from hormonal or immunosuppressive mechanisms or exaggerated pharmacologic action is thought to involve a threshold. Exposures below that threshold are thought not to increase carcinogenic risk.
 |
PHASE 2 STUDIES
|
---|
For phase 2 studies in patients, the entire ICH battery will have been completed. A positive in one test in the battery may be followed up by other, complementary, short-term assays that look at the potential for genetic damage. For larger phase 2 (or phase 3) studies of proposed drug products for nonlife-threatening conditions, the nature of the in vitro or in vivo findings, threshold considerations (based on probable mechanism), and robustness of the genetic toxicology findings are evaluated. The cancer hazard may be further assessed in the Syrian hamster ovary (SHE) cell transformation assay. A good correlation between results in the SHE assay and the rodent carcinogen bioassay has been reported (Isfort et al., 1996
; Mauthe et al., 2001
). A negative finding in a SHE cell assay or in an alternative carcinogenicity assay, such as the 6-month P53+/ haploinsufficient mouse model, could allow clinical studies to proceed despite positive findings in the ICH genetoxicity battery. If the SHE assay results were positive, an assessment of the carcinogenic potential in an alternative (MacDonald et al., 2004
; Sistare and Jacobs, 2003
) or traditional carcinogenicity study may be necessary before large phase 3 studies are conducted. The P53+/ haploinsufficient mouse model is considered to respond to genotoxic carcinogens (Storer et al., 2001
), so a negative result in that assay may be sufficient to allow clinical development to continue. However, this may depend on the drug indication. It is the drug sponsor's responsibility to propose a way to address the concern, and various options may be accepted by CDER.
 |
PHASE 2 OR 3 STUDIES WITH CHRONIC ADMINISTRATION
|
---|
When carcinogenicity studies are found to have positive findings during clinical development, a sponsor may choose to put the clinical study on hold or the FDA may put the clinical studies on hold while the results are further evaluated. Mechanistic studies may be suggested to support the claim that effects are not relevant or not of great concern to humans under conditions of clinical use. Possible considerations would include the systemic exposure in humans under conditions of maximum use, exposure in animals at which the effects occurred, as well as the mechanistic data. Some positive findings might be considered not relevant to humans under conditions of use. Often positive results in carcinogenicity studies are due to hormonal or immunosuppressive effects and may be secondary to or exaggerations of the pharmacologic effects, for which there may be thresholds. Studies of pharmacologic activity and results of repeat-dose toxicology studies may provide information on the carcinogenicity potential through these modes of action. The presence of clear genotoxic findings, coupled with nonhormonal and nonimmunosuppression-related carcinogenicity findings in a carcinogenicity study, may increase concern. Positive findings in a P53+/ study would warrant a large concern and consideration of whether the benefit of the therapy outweighs the risk (Jacobson-Kram et al., in press).
 |
DURING THE APPROVAL DECISION PROCESS
|
---|
For approval of drugs for chronic conditions, carcinogenicity studies in two species (a traditional rat study and a traditional or alternative mouse study) are generally available. Initially, in the evaluation process of completed studies, statistical and biologic considerations are used in assessing whether a hazard has been identified in the rodents. Even if a hazard has been identified, results may be considered rodent-specific, mouse-specific, or even strain-specific or considered to occur by a mode of action not relevant to humans (Lima and Van der Laan, 2000
; Williams and Iatropoulos, 2001
). Less weight is placed on findings for which there is a very high control incidence or historical incidence, even if statistical significance is achieved. Considerations for assessing the risk include some of the factors mentioned in the paragraph above for such assessments during phase 2 or 3. In making approval decisions, several questions need to be answered:
- Is the drug clearly genotoxic by a mode of action relevant to human use?
- Are there any carcinogenicity findings clearly related to the drug in the rodents? If there are no drug-related neoplastic findings, do the rodents express the pharmacologic activity seen in humans and do the rodents express the main metabolites seen in humans?
- If there are drug-related neoplastic findings, are the findings biologically relevant to humans under conditions of use?
- Are there biomarkers for hormonal or exaggerated pharmacologic effects in humans (e.g., if a neoplasm is prolactin-related, do prolactin levels in humans increase after exposure to the drug)?
- Are there indications that the drug is immunosuppressive? How do the exposures causing suppression compare to clinical exposures?
- Do the effects occur at a high multiple of the human exposure under conditions of use?
- Is there multispecies evidence for carcinogenicity? When studies from two species are available, clear findings in two species/genders may be of greater concern than findings in only one species.
- Are there other drugs in the same pharmacologic class that gave similar rodent findings but no apparent effects on possible biomarkers (e.g., hyperplasia) in humans?
- Are there other drugs for the indication without carcinogenicity findings at relevant exposures?
- Does the benefit of the therapy justify the risk of cancer? If it does not, the product may not be approved. If the benefit is found to outweigh the risks for some patient populations, the findings in the rodent studies are described in the drug product labeling and the multiple of the human systemic exposure at which the risks were found is described. Plasma AUC value comparisons would be used if known and appropriate. Nominal dose per body surface area (mg/m2) comparisons would be used in the absence of appropriate pharmacokinetic data.
 |
POSTAPPROVAL
|
---|
For some products for serious and life-threatening conditions, carcinogenicity studies may be conducted postapproval or not at all if the projected survival is only a few years. Carcinogenic effects may manifest themselves in humans many years after approval of a drug. For some products for which a risk is perceived, data may be collected prospectively. Immunosuppressive compounds or alkylating agents may cause cancer in humans in less than 10 years. However, for most products, it may take 20 years before many cancers would occur. It is very difficult to attribute cancer cases to a particular drug product unless the cancer occurred in a high percentage of persons taking a particular product, the drug relationship to the cancer incidence was being looked for intensively (e.g., the association of PUVA therapy and various skin neoplasms years later), or the drug induced a rare form of cancer, as was the case with DES. Thus, persons making risk/benefit decisions regarding clinical usage must generally rely on nonclinical carcinogenicity studies.
In summary, CDER's paradigm for cancer risk assessment differs from that used by agencies such as the EPA. Unlike with drugs, individuals exposed to environmental contaminants and pesticide residues have only risk and no benefit to be gained from such exposures. Also, because the exposures are generally not voluntary, the EPA must often calculate acceptable risks. Initial estimates of cancer risk by human exposures to a new drug are based on results from short-term, genetic toxicology studies. While it is recognized that nongenotoxic carcinogens may also present a risk, this mechanism of carcinogenesis is thought to have a threshold. Results from chronic carcinogenicity studies are generally not available until the NDA stage of the approval process. With the development of databases associated with expression arrays, better short-term assays for identifying potentially carcinogenic drugs may be available in the near future.
 |
SUPPLEMENTARY DATA
|
---|
See the CDER website (www.fda.gov/cder/guidance/index.htm; or see www.ich.org) for the following ICH guidances: ICH S1A Guideline for Industry The Need for Long-term Rodent Carcinogenicity Studies of Pharmaceuticals; ICH S1B Guidance for Industry S1B Testing for Carcinogenicity of Pharmaceuticals; ICH S2A Guideline for Industry Specific Aspects of Regulatory Genotoxicity Tests for Pharmaceuticals; ICH S2B Guidance for Industry S2B Genotoxicity: A Standard Battery for Genotoxicity Testing of Pharmaceuticals; and ICH M3 Guidance for Industry M3 Nonclinical Safety Studies for the Conduct of Human Clinical Trials for Pharmaceuticals.
 |
ACKNOWLEDGMENTS
|
---|
The opinions expressed in this report are those of the authors and do not necessarily reflect an official FDA opinion.
 |
NOTES
|
---|
1 To whom correspondence should be addressed at HFD-540 the Center for Drug Evaluation and Research, U.S. Food and Drug Administration, Rockville, MD. Fax: (301) 827-2075. E-mail: abigail.jacobs{at}fdahhs.gov.
 |
REFERENCES
|
---|
Isfort, R. J., Kerckaert, G. A., and LeBoeuf, R. A. (1996). Comparison of the standard and reduced pH Syrian hamster embryo (SHE) cell transformation assays in predicting the carcinogenic potential of chemicals. Mutat. Res. 356, 1163.[ISI][Medline]
Jacobson-Kram, D., Jacobs, A., and Sistare, F. (in press). Carcinogenicity testing in transgenic mice. Toxicol. Pathol.
Lima, B. S., and Van der Laan, J. W. (2000). Mechanisms of nongenotoxic carcinogenesis and assessment of the human hazard. Regul. Toxicol. Pharmacol. 32, 135143.[CrossRef][ISI][Medline]
MacDonald, J., French, J. E., Gerson, R. J., Goodman, J., Inoue, T., Jacobs, A., Kasper, P., Keller, D., Lavin, A., Long, G., et al. (2004). The utility of transgenic mouse assays for identifying human carcinogens: A basic understanding and path forward. Toxicol. Sci. 77, 188194.[Abstract/Free Full Text]
Mauthe, R. J., Gibson, D. P., Bunch, R. T., and Custer, L. (2001). The Syrian hamster embryo (SHE) cell transformation assay: Review of the methods and results. Toxicol. Pathol. 29(Suppl.), 138146.[CrossRef][ISI][Medline]
Sistare, F. D., and Jacobs, A. C. (2003). Use of Transgenic Animals in Regulatory Carcinogenicity Evaluations in Alternative Toxicology Methods. CRC Press, Boca Raton, FL.
Storer, R. D., French, J. E., Haseman, J., Hajian, G., LeGrand, E. K., Long, G. G., Mixson, L. A., Ochoa, R., Sagartz, J. E., and Soper, K. A. (2001). P53+/ hemizygous knockout mouse: Overview of available data. Toxicol. Pathol. 29(Suppl.), 3050.[CrossRef][ISI][Medline]
Williams, G. M., and Iatropoulos, M. J. (2001). Chapter 20, pp. 142. Principles of testing for carcinogenic activity. In Principles and Methods of Toxicology (A. W. Hayes, Ed.), 4th ed. Taylor and Francis, Philadelphia