The Society of Toxicologic Pathology's Recommendations on Statistical Analysis of Rodent Carcinogenicity Studies

TheSociety forToxicologicPathology(STP)isanonproe t organization dedicated to improving the discipline of toxicologic pathology through education and professional interactions. The Society’s membership includes over 800 pathologists and toxicologists involvedworldwide in the nonclinical assessment of toxicity and carcinogenicity of chemicals, pharmaceutical candidates, and medical devices. Many of the study pathologists who interpret rodent carcinogenicity studies are members of the STP. The STP, in conjunction with representatives of the U.S. Food and Drug Administration, recognized that there were issues in the application of statistical methodstocarcinogenicitystudies (2).While most pharmaceutical companies used the Peto test (5), some did not categorize neoplasms as Fatal or Incidental (no contexts ofobservationswereassigned).Somecategorizedneoplasms as Fatal or Incidental based solely on the type of neoplasm (diagnostic terminology) rather thanon an animal-by-animal basis. Others categorized neoplasms as Fatal or Incidental based on the gross and microscopic e ndings for each animal. There were no guidelines for consistent assignment of Fatal and Incidental categories. The STP created an ad hoc Peto Working Group to review statistical methods for rodent carcinogenicity studies and make recommendations on how microscopic observations from these studies could be consistently and reliably categorized for analysis using existing statistical methods. The STP Peto Working Group has reviewed the FDA draftGuidanceforIndustry‐StatisticalAspectsoftheDesign, Analysis, and Interpretation of Chronic Rodent Carcinogenicity Studies of Pharmaceuticals (4) and has provided written comments and suggestions to FDA regarding this draft guidance. At the time these comments were provided, the STP was still gathering information and evaluating the use of pathology data in statistical analysis of rodent carcinogenicity studies. The STP has since solicited and published explanations and recommendations of best statistical practices from biostatisticians familiar with rodent carcinogenicity studies, and some positions provided by the STP to the FDA have been modie ed. The Peto Working Group also sponsored a discussion of the Peto test and the Poly-3 test with the STP membership duringthe 2001 STP annual meeting.Theopinionsofabroadsegmentofpathologistsinvolved with rodent carcinogenicity studies varied and included: 1. Many pathologists believe that experienced pathologists can accurately and reliably determine if a neoplasm contributed to death or euthanasia of an individual animal, based on a review of clinical, necropsy, and microscopic e ndings for that individual. While most pathologists are