Purpose: To study the accuracy of different cutoffs for an involved circumferential resection margin (CRM) compared with T and N categories measured by MRI as basis for selective application of neoadjuvant radiochemotherapy (nRCT) in rectal carcinoma.
Materials and methods: In a prospective multicenter observational study involving 153 primarily operated patients, the preoperative results of MRI with pathohistological findings of resected specimens were compared.
Results: For a cutoff of ≤1 mm for involvement of the CRM, the accuracy of preoperative MRI was 90.9% (139/153). The negative predictive value was 98.5% (134/136). The four participating departments did not differ significantly. For a cutoff of >2 mm and >5 mm, the rates of false-positive findings increased significantly from 5% to 12% and 35% with a decrease in accuracy to 82% and 62%, respectively. In contrast, the accuracy in predicting T (69.3%) and N categories (61.4%) was much lower.
Conclusion: The indication for nRCT should be based on the determination of the minimal distance of the tumor from mesorectal fascia with a cutoff point of >1 mm as measured by MRI.
Copyright © 2013 Wiley Periodicals, Inc.