Approximately 75-80% of patients with stage II colon cancer are cured with surgery alone, and the benefit of adjuvant chemotherapy is controversial . The International Multicentre Pooled Analysis of B2 Colon Cancer Trials (IMPACT B2) has pooled the stage II populations of five similar randomized trials comparing 5-FU and LV vs. observation. This study included 1,016 patients, and failed to demonstrate any effect of chemotherapy . The United Kingdom QUASAR study which included more than 2,000 patients with stage II colon cancer has shown chemotherapy with 5-FU and LV provided a small improvement (~4%) in rates of recurrence and overall survival (OS) compared to patients on observation . Oncologists have frequently used clinical and pathological features such as tumor stage (T3 vs. T4), tumor perforation, inadequately sampled lymph nodes (<12), poor tumor cell differentiation, and extramural venous invasion, to identify patients who may harbor higher risk for recurrence and potentially benefit from adjuvant chemotherapy . Most of these features are not informative for the majority of patients, and have never been validated in perspective studies.
Emerging data have shown that microsatellite instability (MSI) and chromosome 18q loss of heterozygosity (18qLOH) in colon cancer may be useful as molecular prognostic markers in patients with stage II/III colon cancer [56, 64, 65]. The ongoing ECOG 5202 study is a perspective study in stage II colon cancer to identify high-risk patients for adjuvant treatment using molecular marker analysis including MSI and 18qLOH .
Bertagnolli et al presented 18qLOH analysis from Cancer and Leukemia Group B (CALGB) protocol 9581, which randomized 1738 patients with stage II colon cancer to postoperative treatment with monoclonal antibody 17-1A or observation . The result was initially reported in 2004 ASCO annual meeting. There was no difference in 5-year DFS and OS between patients receiving treatment and on observation. Among these patients, 537 tumor samples were obtained for molecular marker analysis, and 23% had MSI. Of the remaining samples, 101 tumor samples had 18qLOH, and 49 had intact 18q. There were significant differences in OS (98 vs. 85 m) and DFS (92 vs. 78 m) between patients with intact 18q and 18qLOH favoring patients with intact 18q. This result is in consistent with prior report that LOH at 18q is prognostic for DFS and OS in patients with early-stage colon cancer did not receive chemotherapy after surgery .
In an effort to develop new clinical tools for risk assessment and treatment decisions in stage II colon cancer, Kerr et al. presented the multi-gene expression assay in patients with stage II colon cancer . This assay is to use real-time RT-PCR to quantitate RNA derived from paraffin-embedded tumor tissue . They have initially identified 761 candidate genes from 1,851 patients' tumor samples in NSABP C-01/C-02/C-04/C-06 and Cleveland Clinic study. After further modeling and analysis, they have prospectively defined the recurrence score based on 7 genes associated with recurrence risk (stromal related genes: FAP, INHBA, BGN; cell-cycle related genes: Ki-67, C-MYC, MYBL-2; and GADD45B). Additionally, 6 genes were chosen and defined as treatment score. The recurrence and treatment scores were validated in 1,436 tumor samples (711 with surgery alone, and 725 with surgery plus adjuvant chemotherapy with 5-FU and LV) from QUASAR study. In 725 patients receiving surgery and adjuvant chemotherapy with 5-FU and LV, treatment score did not predict benefit of adjuvant chemotherapy.
In 711 patients receiving surgery alone, there was significant association between recurrence score and risk of recurrence at 3 years following surgery (P = .004). There were 43.7% in low-risk group (<30 recurrence score), 30.7% in intermediate group and 25.6% in high-risk group (> = 41 recurrence score). Estimated 3-year recurrence risk is 12% (95% CI 9-16%), 18% (95% CI 13-24), and 22% (95% CI 16-29), respectively in low-risk, intermediate and high-risk groups. Multivariate analyses identified three key independent predictors of recurrence in stage II colon cancer after surgery: T4 stage, MSI and recurrence score. In patients with T3 tumor and negative for MSI (~76% of stage II), recurrence score was found to be useful in predicting individual risk of recurrence. This is the first demonstration of a prospectively defined gene expression assay independently predicting risk of recurrence in stage II colon cancer after surgery.
The translational studies of PETACC 3/EORTC 40993/SAKK 60-00 trial were presented in this year's ASCO meeting [70, 71]. This study randomized 3,278 patients with stage II or III colon cancer after surgery to 5-FU/LV or 5-FU/LV/irinotecan . There was no significant difference in 5-yr DFS between these 2 treatment arms. Tumor samples were available from 1,404 patients, and MSI was analyzed in 1,327 samples. There was higher incidence of MSI in stage II (22%) vs. stage III (12%). MSI was a significant prognostic factor for relapse-free survival (HR 0.265, p = 0.0044) and overall survival (median follow up 68 months, HR 0.159, p = 0.011) in stage II colon cancer. There was no significant association between prognosis and MSI in stage III colon cancer, and this may be due to small sample size or possible stage specific biological effects. However, MSI was not predictive for the efficacy of irinotecan/5-FU/LV treatment in this study, which differed from the analysis in CALGB 89803 . Both p53 and the SMAD4 genes had prognostic value for stage III but not for stage II colon cancer. Contradictory to previously published report, 18qLOH failed to demonstrate prognostic value in stage II or III colon cancer. These findings suggest that stage II and III colon cancers may differ biologically.