Safety and Extrapolation Steven Hirschfeld, MD PhD Office of Cellular, Tissue and Gene Therapy Center for Biologics Evaluation and Research FDA

Clinical Research-Brief History Recorded for at least 2400 years Children were often the first patients for new procedures and interventions Among the founding principles of modern food and drug regulation was protecting children

And yet- Pediatric therapeutic development has never been as thorough or robust as adult therapeutic development Many therapies are administered to children without adequate study Many therapies are not made available for pediatric study until after adult marketing studies are completed

The challenge Assemble sufficient data to establish efficacy and safety in the relevant population The relevant population may be sufficiently rare that confirmatory studies are not feasible Concerns regarding the implications of adverse events in children Establishing and maintaining a framework that would support systematic clinical investigations for the relevant population

Issues regarding safety monitoring in pediatric oncology clinical investigations Acknowledgement that children require special protections Acknowledgement that risk tolerance is higher in oncology therapeutics than in other therapeutic areas No detailed consensus standards on study monitoring

Charge to Committee Suggest ways to incorporate fundamental ethical and scientific principles in protecting patients enrolled in clinical studies for pediatric malignancies while providing clear guidance and minimizing the resource burden

Safety Questions Principles: 1. What are the principles that should be addressed in safety monitoring of clinical studies that enroll children with cancer? If the principles are adequately stated in existing documents, statues, or regulations, please identify the relevant documents and sections.

Safety Questions Practice: 2. Recognizing that particular populations, disease settings, and products may have specific requirements, what general parameters should be monitored for safety in all clinical studies? 3. Based on the response to the previous question, how often should the parameters be monitored?

Safety Questions 4. Based on the response to question 2, who should do the monitoring? Is it adequate to have the personnel involved in the study be responsible for safety monitoring? 5. What circumstances would benefit from a Data Monitoring Committee (Data Safety Review Board) oversight? 6. Are there additional recommendations for safety monitoring?

Extrapolation FDA Working Group on Pediatric Extrapolation* identified four domains that may provide a basis for extrapolation of adult data to pediatric population Non-clinical data Pathophysiology Natural History of the disease or condition Response to Therapy * J. Alexander, D.Birenbaum, S.Hirschfeld, R. Johann-Liang, W.Rodriguez, D. Shetty

Absence of predictive or explanatory non-clinical models in pediatric oncology Safety prediction based on animal studies is estimated at 2/3 for cytotoxic compounds- unknown for other classes of compounds Efficacy prediction is unknown but low Findings in clinical studies, particularly negative findings, remain unexplained Further clinical studies that entail resources and risks are undertaken

Potential Advantages of using Non-Clinical Data Lesser resource burden Ability to answer questions not amenable to available clinical techniques Possibly faster time frame to generate data Dynamic interaction between clinical and non-clinical findings can enhance understanding and confidence in results Avoidance of non-informative and minimization of negative outcome studies Opportunity for new study designs

Charge to Committee Provide advice on: What types of non-clinical data are considered informative to complement or supplement clinical results What the characteristics or properties of non-clinical models and data should be to effectively add to clinical results If no satisfactory models exist, what characteristics should a non-clinical model have to confirm, extend, or substitute for clinical results If there are a set of postulates that can be identified or should be developed

Non-Clinical Questions 1. What types of questions that are of potential clinical relevance but are not feasible or acceptable to answer in a clinical study could be addressed by non-clinical studies? Examples may include the need for repeated tissue sampling, assessment of long term effects of treatment, effects on reproduction, access to critical anatomic structures, exposure to toxic reagents, evaluation of non- monitorable or irreversible toxicities, identification of biomarkers for clinical monitoring

Non-Clinical Questions 2. What type of evidence and data would be recommended in each of the following domains to allow extrapolation from non-clinical data and be informative for a clinical condition? a. Pharmacology and pharmacokinetics b. Safety c. Efficacy d. Behavior e. Long term effects f. Developmental aspects g. Other domains?

Non-Clinical Questions 3. Are there additional recommendations for the effective use of non-clinical data? For example, will open literature reports be generally acceptable? Is documentation of compliance with Good Laboratory Practice (GLP) necessary to evaluate animal data? Should non-clinical data be submitted as an independent report with a presentation of primary data sufficient for verification and review?