2023 North America Conference on Lung Cancer (NACLC) - Posters and Abstracts-PP01.44 (Poster) A CT IMAGING BIOMARKER FOR CD8+ LYMPHOCYTES INFILTRATION STRATIFICATION IN PATIENTS WITH NON-SMALL CELL LUNG CANCER

PP01.44 (Poster) A CT IMAGING BIOMARKER FOR CD8+ LYMPHOCYTES INFILTRATION STRATIFICATION IN PATIENTS WITH NON-SMALL CELL LUNG CANCER

Back to course

Pdf Summary

A study was conducted to investigate the ability of radiomics features to stratify patients with non-small cell lung cancer (NSCLC) based on CD8 lymphocyte infiltration levels. Four publicly available datasets were collected, and radiomics features were extracted from each NSCLC lesion using segmentation validated by an expert radiologist. Features with near zero variance, highly correlated features, and linear combinations were eliminated. The patients were split into two groups (CD8 high/ CD8 low), and statistical tests were used to compare the groups. A logistic regression model with Elastic Net regularization was performed to classify CD8 cell infiltration levels.<br /><br />The efficacy of immunotherapy is influenced by the presence of CD8 T cell infiltration in the tumor immune microenvironment. CT image-based radiomics was used as a non-invasive approach to analyze the entire tumoral tissue and identify relevant radiomics features associated with CD8 lymphocyte infiltration levels.<br /><br />The study concluded that CT texture biomarkers can differentiate patients with high and low CD8 lymphocyte infiltration levels. These features have the potential to serve as surrogate predictors for patient responses to immunotherapy and aid in clinical decision-making regarding treatment.<br /><br />The model trained on discriminative features achieved a mean area under the curve (AUC-ROC) of 0.73 and an AUC-ROC of 0.67 on the test set. Overall, the study suggests that radiomics features can be used to stratify patients with NSCLC based on CD8 lymphocyte infiltration levels, which can help in identifying individuals who are more likely to benefit from immunotherapy.<br /><br />This research was presented as a poster at the IASLC 2023 NACLC conference in Chicago, IL. For further information, the contact email is provided.

Asset Subtitle

Fadila Zerka

Keywords

radiomics features

non-small cell lung cancer

CD8 lymphocyte infiltration

publicly available datasets

segmentation

expert radiologist

logistic regression model

Elastic Net regularization

immunotherapy

CT texture biomarkers