Normal tissue effects: reporting and analysis.
Any effective cancer therapy developed to date is associated with a spectrum of normal tissue effects of varying incidence and severity. With an increasing number of novel therapeutic approaches undergoing clinical testing and an increased effort to optimize the established treatment modalities, methods for reliable quantification of normal tissue effects have become a key element in advancing cancer care. Here, we present a review of many of the issues involved in reporting and analyzing clinical normal tissue effect data. A distinction is introduced between explorative (science-driven) and pragmatic (patient-centered) studies. The desirable properties of criteria for reporting and grading toxicity are discussed from a biological and clinical perspective. Validation of toxicity criteria and the statistical issues involved in analyzing this type of data are presented with special emphasis on descriptors of the time evolution of toxicity. Finally, we discuss surrogate markers for late effects, mechanistic studies, and the design of clinical studies with normal tissue endpoints as a primary outcome. It is concluded that a consensus is required on guidelines for the reporting of normal tissue effects to improve the comparability of published reports on treatment outcome.