2023 World Conference on Lung Cancer (Posters)-EP07.02. The Prediction of Pathological Pleural Invasion of Lung Cancer by Artificial Intelligence Image Analysis

EP07.02. The Prediction of Pathological Pleural Invasion of Lung Cancer by Artificial Intelligence Image Analysis - PDF(Slides)

Back to course

Pdf Summary

A study conducted by researchers at Tokyo Medical University in Japan aimed to investigate the use of artificial intelligence (AI) assisted analysis of 3D-CT imaging to predict visceral pleural invasion (VPI) in lung cancer patients. VPI is associated with worse surgical outcomes, so preoperative prediction is crucial. The researchers utilized a training cohort of 408 surgically resected lung cancer cases from 2011 to 2017.5 and a test cohort of 143 cases from 2017.6 to 2018.<br /><br />The AI software used in the study automatically detected lung nodules and reconstructed them into 3D images. It also analyzed 22 features of the nodules to predict VPI. The results showed that the AI prediction model had a sensitivity of 73.9% and a specificity of 65.7% in the test cohort. The significant features identified by the AI software for predicting VPI were solid and pleural attachment.<br /><br />The study concluded that AI assisted analysis of 3D-CT imaging can accurately predict VPI before surgery. This technology has the potential to improve surgical outcomes by enabling early identification of patients at higher risk for VPI.<br /><br />The study was supported in part by a research fund from Fujifilm Corporation. The patient characteristics in the study were similar between the training and test cohorts, except for age and surgical procedure. Overall, the study findings suggest that AI can play a valuable role in assisting radiological findings and improving preoperative prediction in lung cancer patients.

Asset Subtitle

Wakako Nagase

Meta Tag

Speaker Wakako Nagase

Topic Early-Stage NSCLC: New Technology & Innovations

Keywords

artificial intelligence

3D-CT imaging

visceral pleural invasion

lung cancer

surgical outcomes

preoperative prediction

training cohort

test cohort

lung nodules

pleural attachment