Method for predicting liver cancer feature type, electronic equipment and computer storage medium

Abstract

The invention relates to a method for predicting liver cancer feature types, computing equipment and a storage medium. The method comprises the following steps: based on comparison result data about a tumor sample of a to-be-detected object, generating genome variation data about multiple predetermined genes of the tumor sample of the to-be-detected object; acquiring clinical data about the to-be-detected object; determining tumor mutation load information about the to-be-detected object; obtaining immune checkpoint molecular expression data of the to-be-detected object; generating input data of a prediction model at least based on the genome variation data, the clinical data, the tumor mutation load information and the immune checkpoint molecular expression data; and based on a prediction model trained by multiple samples, the prediction model is constructed based on a neural network model. According to the method, the reliability of predicting the liver cancer feature type can be improved, and good generalization of clinical application is achieved.

Description

technical field [0001] The present disclosure relates generally to biological information processing, and in particular, to methods, electronic devices, and computer storage media for predicting liver cancer feature types. Background technique [0002] Studies have shown that the annual recurrence rate of hepatocellular carcinoma (HCC) patients after surgery is as high as 50%, which is the main factor affecting the long-term survival of early-stage HCC patients. Therefore, it is of great significance to accurately predict the characteristic types of liver cancer to assist in judging the risk of liver cancer recurrence. [0003] Traditional methods for predicting the risk of HCC recurrence include, for example, constructing a cytosine-guanine dinucleotide (CpG) methylation signature to predict the risk of postoperative recurrence of early HCC, or based on radiomics, visual analysis, clinical Pathological and other multi-dimensional information to build early HCC recurrence p...

AI Technical Summary

Problems solved by technology

Although the above-mentioned traditional prediction methods have shown certain predictive ability, because the methylation level detection of the whole genome has not been routinely used in clinical work, there is a lack of clinical evidence support, and the interpretation of image results such as radiomics needs to be enriched. Expert experience and time-consuming and labor-intensive, resulting in lim

Images