ABSTRACT
PURPOSE: New editions of the TNM staging system for colorectal cancer have been subject to extensive criticism. In the current study, we evaluate each edition of TNM and analyze stage migration caused by the different versions. PATIENTS AND METHODS: Two independent test populations were used: participants derived from a randomized surgical trial from the United Kingdom (n = 455) and patients from a population-based series from Sweden (n = 505). All slides from these patient cases were reviewed with special attention for the presence of tumor deposits. Tumor deposits were classified according to the fifth, sixth, and seventh editions of TNM and correlated with prognosis. RESULTS: Every change in edition of TNM led to a stage migration of between 33% and 64% in patients with tumor deposits. Reproducibility was best in the fifth edition of TNM. The prognostic value of the seventh edition was best only when all tumor deposits irrespective of size or contour were included as lymph nodes. The prognostic value of the fifth edition was better than that of the sixth. CONCLUSION: We demonstrate there is a place for tumor deposits in the staging of patients with colorectal cancer. However, many questions remain about their definition and the reproducibility and use of this category in special situations, such as after neoadjuvant treatment. These should be the subject of additional research before use as a factor in TNM staging. This work demonstrates the necessity of testing modifications before their introduction.