2023 World Conference on Lung Cancer (Posters)-EP11.03. Quantification of Interstitial Lung Abnormality Using Deep Learning in Cancer Patients Who Underwent Immune Checkpoint Inhibitors

EP11.03. Quantification of Interstitial Lung Abnormality Using Deep Learning in Cancer Patients Who Underwent Immune Checkpoint Inhibitors - PDF(Abstract)

Back to course

Pdf Summary

In this study presented at the WCLC 2023 conference, researchers aimed to quantify interstitial lung abnormalities (ILA) using deep learning in cancer patients who underwent immune checkpoint inhibitor (ICI) treatment. ILA are subtle changes in the lung parenchyma that can indicate the risk of ICI-induced pneumonitis.<br /><br />The study included a total of 291 cancer patients, of which 27.1% had ILA. The researchers used the Masked-attention Mask Transformer for Universal Image Segmentation (Mask2Former) model, developed by Facebook Research, for the segmentation of ILA on baseline chest CT scans. The model consisted of a backbone feature extractor, pixel decoder, and transformer decoder.<br /><br />A 5-fold cross-validation technique was used to develop the ILA segmentation model, and an independent testing cohort of 291 lung cancer and non-lung cancer patients who underwent ICIs was used to evaluate the model's performance. The accuracy and dice similarity coefficient (DSC) of the ILA segmentation were found to be 98.08% and 49.61%, respectively, on the test set.<br /><br />The results suggest that the deep learning model can accurately identify ILA on chest CT scans. This could be a valuable method for early detection of ILA, which can precede the development of interstitial lung disease (ILD) in some cases. ILD patterns are known to be risk factors for ICI-induced pneumonitis in oncologic patients.<br /><br />Overall, this study highlights the potential of using deep learning and AI techniques to segment and quantify ILA in cancer patients undergoing ICI treatment. This could help in the early identification of ILA and the prevention of ICI-induced pneumonitis.

Asset Subtitle

Eun Young Kim

Meta Tag

Speaker Eun Young Kim

Topic Metastatic NSCLC: Immunotherapy - Retrospective

Keywords

WCLC 2023 conference

interstitial lung abnormalities

deep learning

cancer patients

immune checkpoint inhibitor treatment

ILA segmentation

Mask2Former model

chest CT scans

lung parenchyma

ICI-induced pneumonitis