The invention belongs to the technical field of
medical imaging, and relates to a method for predicting complete
pathologic remission of
head and neck squamous
cell carcinoma and related equipment. The method comprises the following steps: respectively inputting a T1W sequence image before treatment, a T1W sequence image
after treatment, a T2W sequence image before treatment and a T2W sequence image
after treatment into an MR I multi-
time sequence fusion module for decoupling
feature extraction operation; inputting the
time sequence fused T1W sequence MR I
time sequence decoupling
feature data and the time sequence fused T2W sequence MR I time sequence decoupling
feature data into an MR I multi-sequence fusion module to carry out multi-sequence fusion operation; inputting the WSI image into a WSI
feature extraction module to carry out WSI
feature extraction operation; the MRI multi-sequence fusion
feature data and the WSI feature data are input to an MRI-WSI multi-
modal fusion module for cross fusion operation; and performing pCR prediction operation on the MRI-WSI cross fusion feature data according to a
convolutional neural network and a multi-layer
perceptron to obtain a pCR prediction result. The method can be independently competent for a multi-factor prediction task of pCR.