Re: Colon cancer survival rates with the new American Joint Committee on Cancer sixth edition staging.

O'Connell et al. (1) showed that the stage groupings in the newer sixth edition of the tumor-lymph node-metastasis (TNM) staging system more distinctly stratify survival than the fi fth edition (2 , 3). However, survival was worse with T4N0M0 (stage IIB) than with T1-2N1M0 (stage IIIA) disease. Review of 351 000 colon cancer case records in the National Cancer Data Base confi rms this fi nding (A. Stewart, personal communication). In the accompanying editorial , Burke (4) states that this breaks a fundamental rule of the TNM staging system — that survival should decrease with increasing stage grouping. On this basis, he recommends the American Joint Committee on Cancer (AJCC) revert to the fi fth edition of TNM staging for colon cancer staging. In most cases, survival usually decreases with each higher stage group. However, stage group primarily refl ects increasing anatomic extent of disease. For colon cancer, stage II is local disease with negative lymph nodes and stage III denotes positive lymph nodes. The recognition that T4N0 tumors had a worse prognosis than T3N0 was the basis for dividing stage II of the fi fth edition into stages IIA and IIB in the sixth edition. The fi ndings of O'Connell et al. support this division and do not require that we abandon this improved disease classifi cation. Burke also raises more fundamental questions regarding TNM: that it fails to account for the prognostic impact of response to presurgical therapy and does not allow for use of nonanatomic bio-logic factors. Further, he states that the TNM system that assigns cases to stage groups, or " bins, " should be replaced by programs that quantify prognosis for an individual. Grouping patients with similar disease is necessary for assessing the impact of treatments in similar patients and for population surveillance of cancer incidence and outcome. However, this procedure does not preclude the use of any factors that affect cancer prognosis or the future reorganization of groups around such factors. It is everyone's hope that advancing science will fundamentally change understanding of cancer prognosis. The challenge is to identify those factors that are suffi ciently robust that they can be reliably measured and applied to all cases. Toward this goal, the AJCC examined nonanatomic factors and alternate staging schemas for the sixth edition revisions of the TNM staging system. International workshops addressed the use of neural networks and nonanatomic prognostic factors, with concrete …