Gene expression classifier capable of predicting lung cancer patient prognosis and construction method of gene expression classifier

Abstract

The invention discloses a construction method of a gene expression classifier capable of predicting lung cancer patient prognosis. The method comprises a data training stage and a verification stage, wherein the training stage comprises a first stage and a second stage; at the first stage, a supervised machine learning method is used to establish a gene expression classifier prototype capable of predicting the lung cancer patient prognosis; and at the second stage, the machine learning method is further used to obtain the gene expression classifier capable of predicting the lung cancer patient prognosis. According to the method, the supervised machine learning method is used to obtain the gene expression classifier, and the prognosis of non-small cell lung cancer patients can be precisely predicted. The gene expression classifier has very high clinical transformation value. By performing gene expression detection of a gene panel, the non-small cell lung cancer patient with a high-risk gene risk score should receive adjuvant therapy, and the non-small cell lung cancer patient with a low-risk gene risk score should receive a low dose or be exempted from adjuvant therapy.

Description

technical field [0001] The present invention relates to a gene expression classifier and its construction method, more specifically to a gene expression classifier for predicting the prognosis of lung cancer patients and its construction method, especially a gene expression classifier for predicting the prognosis of non-small cell lung cancer patients and its construction method. Background technique [0002] The latest cancer epidemiological survey in China shows that there were 730,000 new lung cancer patients and 600,000 deaths in 2015. Lung cancer has become the cancer with the highest incidence and mortality in China (1) . Lung cancer deaths account for 25% of all cancer deaths. Due to the high recurrence risk and low survival rate of lung cancer, the vast majority of lung cancer patients in stage IB-IIIA will undergo postoperative chemotherapy (postoperative chemotherapy, POCT). In addition, based on parameters such as the degree of residual lesions, lymph node meta...

AI Technical Summary

Problems solved by technology

So far, studies have also attempted to develop similar gene expression classifiers in the field of lung cancer to predict the recurrence risk of lung cancer patients. (17-29) , these studies are almost all for non-small cell lung cancer, but there are the following problems: first, the biological heterogeneity of lung cancer in different populations, the gene expression classifiers obtained from this population and that population are very different; the second is the lack of unified standards, such as clinical sample collection, annotation, sample processing, etc.; third, the methods of statistics and machine learning vary greatly, and the number of genes involved in these classifiers varies greatly, and there is no gene overlap, and there is no optimal gene classifier. Consensus developed, therefore no impactful clinical validation studies were conducted

Images