An audit of performance, interpretation, and influence of pretherapeutic MRI in rectal cancer : a Swedish population-based cohort study

Background: The performance of magnetic resonance imaging (MRI) interpretation and communication of findings and its implication on treatment decisions has not fully been explored in rectal cancer. Purpose: To investigate in a region the adherence to MRI protocol standards and the relation between MRI interpretation and selection to preoperative therapy in rectal cancer. Material and Methods: Data on consecutive patients who underwent elective rectal cancer surgery in the region from January to June 2010 were obtained from the National Colorectal Cancer Registry. Pretherapeutic MRI images were re-evaluated. Agreement between the original reports and the re-evaluation was compared using Cohen's kappa coefficient. Results: Among the 94 patients included, 81 (86%) had pretherapeutic MRI in accordance with defined imaging guidelines. In 34% of the original MR reports, data on extramural venous invasion (mrEMVI) and mrT category were not reported. Complete tumor staging was not possible because of missing data in 33 (35%) of the patients. The agreement between the original MR reports and the re-evaluation regarding tumor stage was moderate (kappa = 0.48). For decided treatment compared to recommended preoperative treatment according to the re-evaluation, the agreement was fair (kappa = 0.33). Conclusion: Established MRI protocol standards were not universally applied. Missing data and inadequacies in original MRI reports resulted in moderate agreement between the original report and the re-evaluation indicating a risk of inappropriate treatment selection. The results call for further educational efforts in rectal cancer MRI acquisition and repeated audits of image protocol adherence and interpretation quality.