2023 World Conference on Lung Cancer (Posters)-P2.13. Pathological Images Machine Learning Predicts Long Term Effects for Immunotherapy in Small-Cell Lung Cancer

Pdf Summary

Researchers from Wakayama Medical University in Japan have developed a machine learning model that predicts the efficacy of immunotherapy in small-cell lung cancer (SCLC) using pathological images. The study suggests that the tumor immune microenvironment (TIME) could be a promising biomarker for immunotherapy in SCLC. However, objectively evaluating the complex interaction of various inflammatory and tumor cells within TIME is challenging for humans. Machine learning has emerged as a potential solution for precise and objective spatial analysis.<br /><br />The predictive model developed in the study is a two-class classifier that predicts 365-day progression-free survival. The primary endpoint of the model is the area under the curve (AUC) of the pathological image model. The secondary endpoints include the AUC of the patient information model and the combined model, as well as progression-free survival and the interpretative importance of variables for determining long-term efficacy.<br /><br />The researchers collected patient information, such as metastatic site and blood test results, and used hematoxylin and eosin staining, programmed death-ligand 1 expression, and multiplex immunofluorescence (CD8 and FoxP3) to analyze the biomarkers in the PATHOLOGICAL IMAGES of patients. They employed a patch-based classifier and a feature extraction module to extract features from the images and predict long-term efficacy.<br /><br />The study demonstrates that a machine learning analysis of TIME in pathological images can accurately predict the long-term efficacy of immunotherapy in SCLC. The researchers also found that certain patient information, such as age and presence of metastasis, played a significant role in determining the efficacy of immunotherapy.<br /><br />These findings provide valuable insights into the potential use of machine learning models in predicting the effectiveness of immunotherapy in SCLC patients. Further research in this area could help improve patient outcomes and personalize treatment approaches. The study was funded by Chugai Pharmaceutical Co., Ltd.<br /><br />In conclusion, a machine learning model using pathological images can predict the efficacy of immunotherapy in small-cell lung cancer. This approach provides an objective and precise analysis of the tumor immune microenvironment, which could aid in determining the effectiveness of immunotherapy in SCLC patients.

Asset Subtitle

Ryota Shibaki

Meta Tag

Speaker Ryota Shibaki

Topic SCLC & Neuroendocrine Tumors: Biomarkers & Radiomics

Keywords

Wakayama Medical University

machine learning model

immunotherapy

small-cell lung cancer

predict

pathological images

tumor immune microenvironment

biomarker

spatial analysis

progression-free survival