Previous research has demonstrated that after neoadjuvant therapy for rectal cancer, the sensitivity of magnetic resonance complete response (mrCR) for detecting pathologic complete response (pCR) in the surgical specimen ranges from 74 to 94%. Patient and provider interest in nonoperative management of rectal cancer that responds favorably to neoadjuvant therapy has grown, necessitating stronger evidence for how well radiographic complete response truly predicts pCR. We sought to determine the current association between mrCR and pCR in locally advanced rectal cancer. We conducted a retrospective cohort study of patients with rectal adenocarcinoma who underwent neoadjuvant chemoradiation followed by index proctectomy at a single academic referral center from January 2012 to December 2021. Our primary outcomes were mrCR, defined as the absence of residual disease on restaging MRI, and pCR, defined as the absence of residual adenocarcinoma in surgical pathology specimens. Among 523 eligible patients, 157 met the inclusion criteria (38.9% females; 51.0% nonwhite; mean [SD] age, 58.6 [13.2] years). Overall, 8.9% of patients had mrCR and 7.0% had pCR. The sensitivity and positive predictive value of mrCR were 36.4% (95% CI: 10.9 to 69.2) and 28.6% (95% CI: 8.4 to 58.1). Our findings were qualitatively unchanged when only patients in the last 5 years of the study period were included. Study limitations include that neoadjuvant therapy regimens were not standardized and patients who were offered and elected to undergo nonoperative management were not included. The value of mrCR in predicting pathologic response following neoadjuvant therapy in locally advanced rectal cancer is low, and mrCR should be interpreted with caution when counseling patients about nonoperative management. Early, frequent surveillance is critical in patients who elect nonoperative management after mrCR.
Read full abstract