Background: Although the TNM system is useful in predicting survival in resected colorectal cancer, heterogeneity within the same stages regarding prognosis exists. We are presenting a pooled analysis of prognostic factors from two randomized studies of adjuvant treatment conducted by the Hellenic Cooperative Oncology Group.
Patients and methods: Patients with stage II or III colon (n = 279) or rectal (n = 220) cancer were included in this analysis. Following surgery, patients received: 5-fluorouracil/leucovorin (5-FU/LV) (n = 135), 5-FU/LV and interferon Alfa-2a (IFNA-2a) (n = 138), 5-FU/LV and pelvic chemoradiotherapy (n = 106), and pelvic chemoradiotherapy alone (n = 108).
Results: Median follow up was 92 mo. The number of involved lymph nodes (LNs), tumor differentiation, and the presence of regional implants were independent prognostic factors for both OS and TTP, while nerve invasion was only significant for TTP. Patients were stratified into three prognostic groups (low-risk: no LNs and grade 1/2; high-risk: > 3 LNs and grade 3/4; intermediate-risk: remaining patients) with distinct differences in 5-yr survival (84.7% vs 57.6% vs 32.4%) and 5-yr TTP (81.2% vs 54.5% vs 28.6%).
Conclusion: The combination of clinicopathological prognostic factors can be more informative than the traditional TNM staging system. Such stratification may be necessary in randomized trials and could be useful in deciding the most appropriate adjuvant treatment strategies.