WCLC 2025 - Posters & ePosters-EP.04.25 Integrated Biological and Imaging Assessment for Pulmonary Nodule Malignancy: A Real-World Study With Large Cohort Modeling and Validation

EP.04.25 Integrated Biological and Imaging Assessment for Pulmonary Nodule Malignancy: A Real-World Study With Large Cohort Modeling and Validation

Back to course

Pdf Summary

This study by Zihao Li et al. from the First Affiliated Hospital of Zhejiang University presents a real-world investigation into improving early lung cancer diagnosis by integrating biological and imaging data. Traditional CT imaging for pulmonary nodule malignancy shows limited specificity and sensitivity, prompting exploration of combined modalities. Tumor-associated autoantibodies, specifically seven lung cancer-related serum autoantibodies (7-AABs), have shown promise as early detection biomarkers. The authors retrospectively analyzed a cohort of 1,017 patients with pathologically confirmed pulmonary nodules (712 training, 305 validation) collected from 2020 to 2024. Using univariate and multivariate logistic regression, they identified independent predictors including nodule size, nodule type (solid vs. mixed ground-glass), and the integrated CT plus 7-AABs biomarker profile. The combined CT and 7-AABs approach achieved substantially higher diagnostic performance (AUC 0.794) than CT alone (0.667) or 7-AABs alone (0.514). The resulting nomogram model predicting malignancy demonstrated strong discrimination, with AUCs of 0.826 and 0.862 on training and validation sets, respectively. Decision curve analysis indicated greater clinical net benefit for the integrated model compared to single modalities. The findings support that adding serum autoantibody profiles to imaging data significantly improves the accuracy of differentiating benign from malignant pulmonary nodules. This validated multidimensional prediction tool offers clinicians a robust resource to inform personalized diagnostic and therapeutic strategies, potentially enhancing early lung cancer detection and patient outcomes.

Asset Subtitle

Zihao Li

Meta Tag

Speaker Zihao Li

Topic Screening and Early Detection

Keywords

early lung cancer diagnosis

pulmonary nodules

CT imaging

tumor-associated autoantibodies

7 lung cancer-related serum autoantibodies (7-AABs)

multivariate logistic regression

diagnostic performance

nomogram model

decision curve analysis

integrated biomarker model