2023 World Conference on Lung Cancer (Posters)-EP11.01. Prediction of NSCLC Immunotherapy Outcomes Across Multiple Clinical Indicators

Pdf Summary

This study aimed to develop a classification model using a machine learning algorithm to predict the immunotherapy outcomes of patients with non-small cell lung cancer (NSCLC). The researchers collected data from 538 NSCLC patients, including 15 clinical indicators related to immunotherapy. These indicators included tumor mutational burden, copy number alteration, HLA-I evolutionary divergence, loss of heterozygosity, microsatellite instability, body mass index, sex, blood neutrophil-to-lymphocyte ratio, tumor stage, immunotherapy drug agent, age, albumin levels, platelet levels, and hemoglobin levels. Different predictive models were developed using various clinical factors, and their potential was evaluated using various statistical analyses.<br /><br />The study found that five clinical indicators were significantly different between the responders and non-responders to immunotherapy. These indicators included tumor mutational burden and hemoglobin levels. Most of the clinical indicators were associated with clinical outcomes, except for sex and the specific immunotherapy drug agent. The multivariate Cox regression analysis identified seven clinical indicators (chemotherapy before immunotherapy, age, stage, neutrophil-to-lymphocyte ratio, albumin levels, tumor mutational burden, and copy number alteration) as significant prognostic factors. These indicators had good performance in predicting the 3-year survival rate. The gradient boosting machine analysis also showed that the seven indicators model could predict immunotherapy outcomes well.<br /><br />The study concluded that the established model of seven clinical indicators could help classify immunotherapy responses, guide therapeutic treatment, and improve patient prognosis. This is significant in the field of precision immuno-oncology. The findings are important in screening appropriate patients who can benefit from immunotherapy in precision medicine. This research contributes to improving the sensitivity of biomarkers used in prognostic prediction for NSCLC patients treated with immunotherapy.

Asset Subtitle

Shanshan Xiao

Meta Tag

Speaker Shanshan Xiao

Topic Metastatic NSCLC: Immunotherapy - Biomarker

Keywords

classification model

machine learning algorithm

immunotherapy outcomes

NSCLC

clinical indicators

tumor mutational burden

hemoglobin levels

prognostic factors

precision medicine

sensitivity of biomarkers