Re: Cancer chemoprevention: progress and promise.

The commentary by Lippman et al. (1) provides a useful review of the topic by organizing a massive amount of information in an organized way. However, because this review will be widely read and referenced, there are two general points made by the authors that could be challenged. The authors define two of the four criteria for identifying a “definitive” chemoprevention trial as “primary end point of cancer incidence” and “large scale (n ù 1000) with the definitive sample size and duration based on anticipated event rates in the intervention arm (treatment effect) and placebo arm.” One major goal of chemoprevention research is to determine the protective effect of an agent while placing as few participants as possible at risk. One straightforward way to accomplish this goal is to understand the process of the disease (carcinogenesis) sufficiently to conduct trials with markers that predispose to or predict the final end point of cancer. In general, most histologically identifiable precancers (e.g., cervical intraepithelial neoplasia, Barrett’s esophagus, adenoma polyps, actinic kerotoses, and dysplastic nevi) evolve to cancer with a sufficiently predictable frequency to conclude that their reversal or suppression can be used to predict cancer development and to assess the value of a chemoprevention agent. There are, in fact, as the authors note, several studies (2–6) that have addressed the effectiveness of chemoprevention in this manner. The medical community accepts hypertension and cholesterol as surrogate “preneoplasias” of cardiovascular disease risk and their modulation as indicative of a favorable or unfavorable drug effect. Modulation of the pathobiology of precancers is as valid a marker of carcinogenesis as is the end point of cancer. Undoubtedly, advances in our understanding of carcinogenesis will allow us to identify and develop new agents by the modulation of a biochemical event earlier in the carcinogenic process and this should be a major goal of chemoprevention research. Large numbers may provide comfort to the investigator that a definitive result has been obtained. However, there are many instances in medicine when large numbers were not required to make the point, i.e., where the underlying cause