2023 World Conference on Lung Cancer (Posters)-P1.20. Computed Tomography-based Deep Learning Model for Spread Through Air Spaces Prediction in Ground-Glass Predominant Lung Adenocarcinoma

P1.20. Computed Tomography-based Deep Learning Model for Spread Through Air Spaces Prediction in Ground-Glass Predominant Lung Adenocarcinoma - PDF(Abstract)

Back to course

Pdf Summary

In this study presented at the World Conference on Lung Cancer 2023, the authors aimed to develop a deep learning model for predicting tumor spread through air spaces (STAS) in ground-glass-predominant early-stage lung adenocarcinoma. Currently, sublobar resection is associated with poor prognosis in these cases, so accurate prediction of STAS is crucial for surgical planning.<br /><br />The authors retrospectively analyzed data from 581 patients from two institutes between 2015 and 2019. They developed a deep learning model (STAS-DL) to extract features of solid components through the use of solid components gated (SCG) for predicting STAS. The performance of the STAS-DL model was compared with other models, including the deep learning model without SCG (STAS-DLwoSCG), radiomics-based model, consolidation-to-tumor (C/T) ratio, and physicians.<br /><br />The proposed STAS-DL model outperformed the other methods on the testing set, achieving an area under the curve (AUC) of 0.82 and an accuracy of 74%. This was significantly better than the STAS-DLwoSCG, which achieved an accuracy of 70%, as well as the physicians with an AUC of 0.68. The STAS-DL model also achieved the highest standardized net benefit compared to the radiomics-based model, C/T ratio, and physicians.<br /><br />The authors concluded that the STAS-DL model they developed has great potential for accurately predicting STAS preoperatively in patients with ground-glass-predominant early-stage lung adenocarcinoma. The model could assist clinicians in making informed decisions regarding surgical planning for these patients.<br /><br />Keywords for this study include lung adenocarcinoma, tumor spread through air spaces, and deep learning model. This research falls under the Pathology and Biomarkers track of the conference.

Asset Subtitle

Mong-Wei Lin

Meta Tag

Speaker Mong-Wei Lin

Topic Pathology & Biomarkers: Artificial Intelligence in Pathology

Keywords

World Conference on Lung Cancer 2023

tumor spread through air spaces

STAS

ground-glass-predominant

lung adenocarcinoma

deep learning model

sublobar resection

surgical planning

radiomics-based model

consolidation-to-tumor ratio