Background. Adjuvant chemotherapy is established routine therapy for colon cancer (CC) patients with radically resected stage III and 'high-risk' stage II disease. The decision on recommending adjuvant chemotherapy, however, is based on data from older patient cohorts not reflecting improvements in pre-operative staging, surgery, and pathological examination. The aim is to review the current risk of recurrence in stage II and III patients and second, to estimate the relative importance of routinely assessed clinico-pathological variables. Methods. The PubMed/MEDLINE and the Cochrane databases were systematically searched for randomized controlled studies and observational studies published after 1 January 2005 with patients included after January 1995 on prognosis in surgically treated stage II and III CC patients. Results. Of 2596 studies identified, 37 met the inclusion criteria and 25 provided data for meta-analysis. The total patient sample size in the 25 studies reporting either disease-free (DFS) or recurrence-free survival was 15 559 in stage II and 18 425 in stage III. Five-year DFS for stage II patients operated without subsequent adjuvant chemotherapy was 81.4% [95% confidence interval (CI) 75.4-87.4; in studies with good/very good quality of reporting 82.7%, (95% CI 80.8-84.6)]. For stage II patients treated with adjuvant chemotherapy, the five-year DFS was 79.3% (95% CI 75.6-83.1). For stage III patients without chemotherapy, five-year DFS was 49.0% (95% CI 23.2-74.8) and for those treated with adjuvant chemotherapy, 63.6% (95% CI 59.3-67.9). The prognostic impact of commonly investigated clinico-pathological parameters, (pT-stage, pN-stage, differentiation, number of lymph nodes studied, MMR-status, and emergency surgery) were confirmed. Conclusions. In this meta-analysis, studies with good quality of reporting show a FIVE-year DFS of 82.7% for stage II CC without adjuvant chemotherapy, whereas the FIVE-year DFS is 63.8% for stage III CC with adjuvant chemotherapy. Due to insufficient reporting on treatment quality the presented DFS is likely an under-estimation of what is achieved at high-quality centers today.
2015. Vol. 54, no 1, 5-16 p.