INTRODUCTION: Introduction of total mesorectal excision (TME) surgery for rectal cancer decreased local recurrence dramatically. Additional neoadjuvant chemoradiation (nCR) is frequently given in UICC II and III tumors based on TNM staging which is of limited accuracy. We aimed to evaluate determination of circumferential margin by magnetic resonance imaging (mrCRM) as an alternative criterium for nCR. METHODS: Multicenter prospective cohort study which enrolled 642 patients in 13 centers with non-metastasized rectal adenocarcinoma. Patients with T4 tumors or patients with a mrCRM of 1 mm or less were treated by neoadjuvant chemoradiation. All others proceeded directly to surgery when inclusion criteria and no exclusion criteria were met. Quality of TME and accuracy of mrCRM determination were assessed during pathology workup. RESULTS: TME was complete in 381 of 389 patients after surgery without nCR (97.9%) and in 245 of 253 patients (96.8%) after nCR. Negative pathology circumferential margins (pCRM) were seen in 97.4% without nCR and in 89% of patients after nCR. Negative pCRM was predicted by negative mrCRM in 98.3% of rectal cancers. NCR was given to 253 of 642 patients (39.5%). Lymph node count was 23 (range 7-79; median/range) for surgery without nCR and 19 (range 2-56) for surgery after nCR. CONCLUSIONS: Surgical quality determined by pathology workup of specimen was very good in this study. Magnetic resonance imaging guided indication for nCR allows to achieve superb results concerning surrogate parameters for good oncological outcome. Thus, use of neoadjuvant chemoradiation with its potential detrimental side effects may be substantially reduced in selected patients.