Artificial Intelligence (AI)-Powered H&E Whole-Slide Image Analysis of Tertiary Lymphoid Structure (TLS) Independently Predicts Survival in Patients with Non-Small Cell Lung Cancer (NSCLC) Receiving Immunotherapy
Discuss this preprint
Start a discussion What are Sciety discussions?Listed in
This article is not in any list yet, why not save it to one of your lists.Abstract
Background
Tertiary lymphoid structures (TLSs) within the tumor microenvironment have emerged as potential indicators of treatment response to immune checkpoint inhibitors (ICIs). This study aims to explore the association between artificial intelligence (AI)-powered TLS analysis and treatment outcomes in NSCLC patients treated with ICIs.
Methods
An AI-powered analyzer, Lunit SCOPE, was developed and trained on H&E-stained whole slide images to quantify TLS. An external cohort of 102 advanced-stage NSCLC patients was utilized to assess the predictive value of TLS.
Results
Of 102 patients, 30.3% (31/102) cases were found to contain TLS as determined by the AI analyzer. The AI model achieved an accuracy of 88.2% in identifying the presence of TLS. TLS presence assessed by AI analyzer was associated with significantly favorable progression-free survival (HR 0.49, 95% CI 0.29–0.82, p < 0.01) and overall survival (HR 0.55, 95% CI 0.33–0.92, p = 0.02). The presence of TLS was associated with favorable survival outcomes, independent of PD-L1 status, treatment regimen (ICI monotherapy vs. combination therapy), lines of therapy (first-line vs. second-line and beyond), and tissue harvest site (primary vs. metastatic). The TLS presence enhanced predictive performance when TLS was assessed from WSI obtained from metastatic lesions.
Conclusions
To our knowledge, this study is the first to develop an AI model capable of detecting the presence of TLS in WSI, with clinical validation using real-world data demonstrating its independent predictive value of survival outcomes in advanced-stage NSCLC patients receiving immunotherapy.
