Magnetic resonance imaging tumor response score (mrTRS) predicts therapeutic effect and prognosis of locally advanced rectal cancer after neoadjuvant chemoradiotherapy: A prospective, multi-center study

Background and purpose The MRI-assessed tumor regression grade (mrTRG) is limited due to its subjectivity and poor consistency on pathological tumor regression grade (pTRG). A new MRI criterion was established to predict the prognosis of locally advanced rectal cancer (LARC). Materials and methods The new MRI criterion magnetic resonance imaging tumor response score (mrTRS) was based on the retrospective sample of 214 LARC patients (unpublished data). Subsequently, 878 LARC patients were enrolled for a prospective, multicenter study. Baseline and postoperative MRI were obtained, and imaging features were measured by collecting the pathological, clinical and follow-up data. Kaplan-Meier method with log-rank estimate and multivariate cox regression model was used to determine the prognosis of mrTRS in LARC patients with neoadjuvant chemoradiotherapy (NACRT). The predictive capability of 3-year prognosis between mrTRS and mrTRG was determined by time-dependent ROC curves. Results The results demonstrated that mrTRS acted as an independent predictor of survival outcomes. mrTRS stratified by good and moderate responders showed significantly lower risk of death (HR = 0.04, 95%CI 0.01–0.31; HR = 0.35, 95%CI 0.23–0.52), distant metastasis (HR = 0.25, 95%CI 0.13–0.52; HR = 0.42, 95%CI 0.30–0.58), and local recurrence when compared with poor responders(HR = 0.01 95%CI 0.23–0.52;HR = 0.38, 95%CI 0.16–0.90). In contrast, no significant difference was observed among mrTRG stratified groups. Excellent and substantial interobserver agreement for mrTRS and mrTRG evaluation was observed (κ = 0.92 and 0.62), respectively. Conclusion mrTRS can serve as an effective predictor for assessing tumor regression grade in LARC patients with NACRT.