Volumetric histogram analysis of amide proton transfer-weighted imaging for predicting complete tumor response to neoadjuvant chemoradiotherapy in locally advanced rectal adenocarcinoma.

Abstract

To investigate the potential of histogram analysis applied to pre-treatment amide proton transfer-weighted (APTw) imaging in predicting complete pathological regression (pCR) in patients with locally advanced rectal cancer (LARC) undergoing neoadjuvant chemoradiotherapy (nCRT). This retrospective study enrolled LARC patients who underwent preoperative rectal magnetic resonance imaging (MRI). Based on histologic assessment, the patients were divided into a pathological complete response (pCR) group or a non-pCR group. APTw histogram features, apparent diffusion coefficient (ADC), and clinical parameters were analyzed. Mann-Whitney U-test, Spearman rank correlation, and univariate and multivariate logistic regression were used for statistical analysis. The predictive performances of different models were evaluated by the receiver operating characteristic curve (ROC). One-hundred forty-five patients were included (mean age, 61.6 years ± 11.8 [SD]; 87 men). pCR patients exhibited lower pre-treatment ADC value, higher pre-treatment APTw-10%, APTw-90%, minimum, maximum, median, mean, range, and root mean square (RMS) of the primary tumor compared to non-pCR patients (all p < 0.05). APTw-10%, APTw-90%, maximum, mean, median, minimum, range, and RMS showed negative correlations with the tumor regression grade (TRG) category (r ranged between -0.457 and -0.173; all p < 0.005). Skewness, kurtosis, and entropy exhibited positive correlations with the TRG category (r = 0.278, 0.319, and 0.324, respectively; all p < 0.05). The combined model had a higher AUC of 0.930, with 93.9% sensitivity and 83.9% specificity. Histogram analysis of pre-treatment APTw may hold promise as a novel approach for predicting the response of LARC patients to nCRT. Question Predicting response to nCRT is crucial for early stratified management of LARC patients; however, current radiological studies remain inconclusive. Finding LARC patients with pCR is correlated with higher pre-treatment APTw intensity-related and lower shape-related histogram features. Clinical relevance The APTw-histogram model and the APTw-clinical combined model demonstrated strong diagnostic efficacy and clinical practicality in predicting LARC patients' responsiveness to nCRT, offering new insights for early clinical decision-making.