Application Artificial Intelligence to Identifying and Personalize Lung Cancer Treatments

Lung cancer stands as the primary cause of cancer-related fatalities on a global scale. Personalized targeted therapy has emerged as the preferred treatment approach for lung cancer, particularly yielding advantages for patients with genetic mutations. The primary aim of this research is to identify genetic variations and forecast mutant genes with the potential to be targeted for treatment in non-small cell lung cancer (NSCLC). We built an algorithm model based on data collected from PubMed, HGMD, and COSMIC. Research, design, and development of software for gene mutation identification in patient samples and personalized targeted therapy for lung cancer treatment. The software utilizes machine learning models and automated processes to identify gene mutations in patient samples and support personalized targeted therapy. The collected dataset includes 225,149 articles, of which 124,850 belong to the positive class (with 4079 articles from HGMD as the rest are similarly cited articles from PubMed) and 100,299 articles belong to the negative class. The PubMed BERT model was currently achieving the highest results for two modules, NER and RE. In this study, we developed a deep learning model for variant naming based on a combination of the CNN model, GATK, and XGBoost algorithm. The model results give accuracy with non-somatic were 94%. The application of machine learning models in the field of medicine has been increasingly developed in recent years, especially in the fields of diagnosis, treatment prognosis, and prediction of targeted drug treatment responses.