Preoperative prediction of lymphovascular invasion of CRC by artificial neural network

Lymphovascular invasion (LVI) was considered to be important for metastasis of colorectal cancer (CRC). However, there was still no effective method to predict LVI before operation. Our research aimed to construct an artificial neural network (ANN) for the preoperative prediction of LVI. We obtained blood indexes and condition of LVI (confirmed by pathological examination) of 288 cases of CRC patients from a tertiary hospital in China. One hundred and eighty-five CRC patients (training group) were randomly selected to establish neural network and logistic regression models. The remaining 103 cases of CRC patients received the test of ANN and logistic model (validation group). Receiver operating characteristics curve (ROC) and decision curve analysis (DCA) were performed to assess the accuracy of constructed model respectively. All procedures involving human participants were performed by the Shanghai Pudong Hospital ethical committee (2020 No. W2-007) and with the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards. All patients provided their written informed consent. In the training group, the area under the curve (AUC) of ANN was higher than that of the logistic model (0.832 vs 0.692). The ANN correctly predicted 92% cases of LVI, whereas the logistic model only predicted 56% cases. Similar results were also tested in the validation model. Our constructed ANN showed higher accuracy compared with the conventional linear model. The ANN-based on blood indexes may provide value for preoperative prediction of LVI.