The Achilles heel of MRI imaging for assessing the anterior CRM of Low Anterior Rectal Cancer: Do we need to do better?

Abstract Aim To evaluate the accuracy of MRI in predicting circumferential resection margin (CRM) involvement in low anterior rectal cancer (LARC) in males. Methods MRI scans of 288 male patients with LARC from 2012 to 2023 were retrospectively evaluated for their CRM status on post-neoadjuvant therapy MRI (ymrCRM), using final histopathology as a gold standard. CRM involvement is defined as tumor < = 1mm from the mesorectal fascia. Results Post-neoadjuvant treatment, the sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of MRI (ymrCRM) were 84%, 69.9%, 32.8%, and 96.1%, respectively. Prostate infiltration was observed in 15.1% (n = 25) of patients post-neoadjuvant therapy. The diagnostic accuracy of MRI in detecting prostate infiltration showed sensitivity, specificity, PPV, NPV, and accuracy at 69.2%, 77.1%, 58.7%, 84.2%, and 74.6%, respectively. Signet ring adenocarcinomas showed higher rates of CRM positivity (24.3%) and local recurrence rates (17%). In patients younger than 30 years, higher CRM positivity (37%) and local recurrence rates (22.2%) were observed, with CRM positivity showing a statistically significant difference (p = 0.002). Conclusion While post-treatment MRI has an excellent negative predictive value (96%) for anterior CRM in low rectal tumours, the positive predictive value is rather poor, at only 32.8%. This means a significant proportion of patients possibly undergo a more morbid surgery, like total pelvic exenteration (TPE), to achieve negative margins because of the inability to distinguish post-treatment fibrosis on an MRI from residual tumour in an anatomically limited space, i.e, the recto-prostatic groove.