Method for predicting pathological reaction of esophageal squamous carcinoma to neoadjuvant chemotherapy by using multi-mode diffusion model

The invention discloses a method for predicting the pathological reaction of esophageal squamous carcinoma to neoadjuvant chemotherapy by a multi-modal diffusion model. The method comprises the following steps: 1, image acquisition and quality evaluation; step 2, tumor segmentation: segmenting the ESCC lesion on the DWI by using ITK-SNAP; step 3, image conversion and feature extraction; step 4, model construction and evaluation: a mean value model and a radiomics model are constructed by using CTRW, DKI, FROC, IVIM, MEM and SEM respectively; different combinations of three feature selectors and two classifiers are adopted; iteratively retaining the number of features, drawing the AUC according to the number of features, and determining the final number of features by implementing an optimal cross validation AUC model. According to the method, the predicted value of the Delta-NACT model is superior to that of the pre-NACT model, and the Delta-NACT-DKI radiomics model has the best performance in the 24 models.