Construction and validation of a risk score system for diagnosing invasive adenocarcinoma presenting as pulmonary pure ground-glass nodules: a multi-center cohort study in China

• DOI: 10.21037/qims-24-170
• Publication Date: 2024-06-29T09:57:01Z
• Authors: Qingcheng Meng; Changbao Zheng; Lanwei Guo; Pengrui Gao; Wentao Liu; Hong Ge; Tong Liu; Hui Peng; Jie Lu; Xuejun Chen

Background: Anxiety-driven clinical interventions have been queried due to the nondeterminacy of pure ground-glass nodules (pGGNs). Although radiomics and radiogenomics aid diagnosis, standardization reproducibility challenges persist. We aimed assess a risk score system for invasive adenocarcinoma in pGGNs. Methods: In retrospective, multi-center study, 772 pGGNs from 707 individuals The Affiliated Cancer Hospital Zhengzhou University & Henan were grouped into training (509 patients with 558 observations) validation (198 214 sets consecutively January 2017 November 2021. An additional test set 143 observations Hainan was analyzed same period. Computed tomography (CT) signs features manually collected, quantitative parameters achieved by artificial intelligence (AI). positive cutoff ≥3. Risk scores 3 combined carcinoma history, chronic obstructive pulmonary disease (COPD), maximum diameters, nodule volume, mean CT values, type II or III vascular supply signs, other radiographic characteristics. evaluation included area under curves (AUCs), accuracy, sensitivity, specificity, predictive values (PPV), negative (NPV) both systems 1, 2, AI model. Results: [AUC, 0.840; 95% confidence interval (CI): 0.789–0.890] outperformed model (AUC, 0.553; CI: 0.487–0.619), 1 0.802; 0.754–0.851), 2 0.816; 0.766–0.867), 88.0% (0.850–0.904) 95.6% (0.932–0.972) PPV, 0.620 (0.535–0.702) NPV, 89.6% (0.864–0.920) 80.6% (0.717–0.872) specificity sets. sets, performed best AUCs 0.769 (0.678–0.860) 0.801 (0.669–0.933). Conclusions: AI-based scoring using image parameters, features, characteristics effectively predicts