Integrated model of MRI biomarkers and clinical factors for predicting pathologic response after neoadjuvant chemoradiotherapy in locally advanced rectal cancer.

objectiveTo evaluate the utility of multiparametric MRI combined with clinical and laboratory biomarkers in predicting pathologic response to neoadjuvant chemoradiotherapy (nCRT) in patients with locally advanced rectal cancer (LARC).

methodsThis retrospective study included 88 LARC patients who received nCRT followed by total mesorectal excision (TME). Based on pathological tumor regression grade (pTRG), patients were categorized into a good response group (GR, pTRG 0-1, n = 30) and a non-good response group (nGR, pTRG 2-3, n = 58). Independent predictors of GR were identified through univariable and multivariable logistic regression analyses. A combined predictive model was developed, and its predictive performance and calibration were evaluated using receiver operating characteristic curve (ROC) analysis and a calibration curve, with internal validation performed using 1000 bootstrap resamples.

resultsCompared with the nGR group, the GR group had a higher proportion of female patients, a lower rate of carcinoembryonic antigen positivity, a higher proportion of patients with good magnetic resonance tumor regression grade (mrTRG-good), lower ADCₚᵣₑ, higher ADCₚₒₛₜ, and greater ΔADC% (all p < 0.05). Multivariable analysis identified gender, mrTRG, and ΔADC% as independent predictors of GR. The combined model integrating these three independent predictors achieved an apparent area under the curve (AUC) of 0.958, which was significantly higher than that of any individual predictor (all p < 0.05). Bootstrap validation yielded an optimism-corrected AUC of 0.949, and the model showed good calibration.

conclusionThe combined model incorporating gender, mrTRG, and ΔADC% showed promising performance for the prediction of pathologic response after nCRT in patients with LARC and may support individualized treatment decision-making.