Stage II/III rectal cancer with intermediate response to preoperative radiochemotherapy: Do we have indications for individual risk stratification?

BACKGROUND: Response to preoperative radiochemotherapy (RCT) in patients with locally advanced rectal cancer is very heterogeneous. Pathologic complete response (pCR) is accompanied by a favorable outcome. However, most patients show incomplete response. The aim of this investigation was to find indications for risk stratification in the group of intermediate responders to RCT. METHODS: From a prospective database of 496 patients with rectal adenocarcinoma, 107 patients with stage II/III cancers and intermediate response to preoperative 5-FU based RCT (ypT2/3 and TRG 2/3), treated within the German Rectal Cancer Trials were studied. Surgical treatment comprised curative (R0) total mesorectal excision (TME) in all cases. In 95 patients available for statistical analyses, residual transmural infiltration of the mesorectal compartment, nodal involvement and histolologic tumor grading were investigated for their prognostic impact on disease-free (DFS) and overall survival (OS). RESULTS: Residual tumor transgression into the mesorectal compartment (ypT3) did not influence DFS and OS rates (p = 0.619, p = 0.602, respectively). Nodal involvement after preoperative RCT (ypN1/2) turned out to be a valid prognostic factor with decreased DFS and OS (p = 0.0463, p = 0.0236, respectively). Persistent tumor infiltration of the mesorectum (ypT3) and histologic tumor grading of residual tumor cell clusters were strongly correlated with lymph node metastases after neoadjuvant treatment (p < 0.001). CONCLUSIONS: Advanced transmural tumor invasion after RCT does not affect prognosis when curative (R0) resection is achievable. Residual nodal status is the most important predictor of individual outcome in intermediate responders to preoperative RCT. Furthermore, ypT stage and tumor grading turn out to be additional auxiliary factors. Future clinical trials for risk-adapted adjuvant therapy should be based on a synopsis of clinicopathologic parameters.