WCLC 2025 - Posters & ePosters-P2.06.60 Multimodal Diagnostic Model for Intraoperative Diagnosis and Surgery Strategy in Stage Ia LUAD (SuRImage V2): A Prospective, Multicentric Study

P2.06.60 Multimodal Diagnostic Model for Intraoperative Diagnosis and Surgery Strategy in Stage Ia LUAD (SuRImage V2): A Prospective, Multicentric Study

Back to course

Pdf Summary

This prospective multicenter study introduces SuRImage v2, a novel multimodal diagnostic model combining computed tomography (CT) and surgical resection images for intraoperative diagnosis of stage IA lung adenocarcinoma (LUAD). The study addresses limitations of frozen section (FS) examination, which often yields ambiguous results (19.54% of cases) and has an overall accuracy of 86.93%, yet remains time-consuming and suboptimal for intraoperative decision-making. The IASLC grading system guides surgical strategies, emphasizing the need for rapid, reliable diagnostics to choose between segmentectomy and lobectomy.<br /><br />The researchers developed and validated diagnostic models using multicenter datasets: a CT-based model, a surgical resection image model (SuRImage), and the combined SuRImage v2 model. Results show that the CT model achieved AUC values of 0.79 for IAC identification and diagnosis, and 0.77 for grading. The SuRImage model performed slightly better with AUCs of 0.80, 0.82, and 0.80 respectively. Most notably, the integrated SuRImage v2 model combining macro-level surgical image features and radiomic data from CT scans yielded the highest diagnostic accuracy with AUCs of 0.81 in identification, 0.84 in diagnosis, and 0.82 in grading of invasive adenocarcinoma (IAC).<br /><br />The study demonstrates that SuRImage v2 provides prompt, accurate intraoperative assistance for IAC identification, diagnosis, and subtyping, outperforming conventional FS analysis. This multimodal approach supports improved surgical decision-making, potentially guiding segmentectomy versus lobectomy with higher precision. In conclusion, the SuRImage v2 model represents a significant advancement in the intraoperative diagnosis of stage IA LUAD, enhancing diagnostic accuracy and surgical planning through combined radiologic and surgical imaging data.

Asset Subtitle

Lintong Yao

Meta Tag

Speaker Lintong Yao

Topic Pathology and Biomarkers

Keywords

SuRImage v2

multimodal diagnostic model

computed tomography

surgical resection images

intraoperative diagnosis

stage IA lung adenocarcinoma

frozen section examination

IASLC grading system

segmentectomy vs lobectomy

diagnostic accuracy