Finding the missing information in cancer statistics.

Abstract

The current cancer literature is imprecise and confusing. This hampers the ability to draw meaningful conclusions regarding diagnosis and treatment. It is suggested that all future reporting of statistics contain the following features: (1) discussion of case selection criteria; (2) use of TNM system for case classification; (3) listing of minimal treatment details; (4) explanation for length of follow-up used; (5) standardized terms for evaluation (relating these to individual TNM classification); and (6) listing of all failures by standardized terms. It is proposed that editors of journals and organizers of meetings reject all papers that do not meet these criteria.