Non-invasive prediction of invasive lung adenocarcinoma and high-risk histopathological characteristics in resectable early-stage adenocarcinoma by [18F]FDG PET/CT radiomics-based machine learning models: a prospective cohort study

BACKGROUND: Precise preoperative discrimination of invasive lung adenocarcinoma (IA) from preinvasive lesions (adenocarcinoma in situ [AIS]/minimally invasive adenocarcinoma [MIA]) and prediction of high-risk histopathological features are critical for optimizing resection strategies in early-stage lung adenocarcinoma (LUAD). METHODS: In this multicenter study, 813 LUAD patients (tumors ≤3 cm) formed the training cohort. A total of 1709 radiomic features were extracted from the PET/CT images. Feature selection was performed using the max-relevance and min-redundancy algorithm and least absolute shrinkage and selection operator. Hybrid machine learning models integrating [18F]FDG PET/CT radiomics and clinical-radiological features were developed using H2O.ai AutoML. Models were validated in a prospective internal cohort ( N = 256, 2021-2022) and external multicenter cohort ( N = 418). Performance was assessed via area under the curve (AUC), calibration, decision curve analysis (DCA), and survival assessment. RESULTS: The hybrid model achieved AUCs of 0.93 (95% CI: 0.90-0.96) for distinguishing IA from AIS/MIA (internal test) and 0.92 (0.90-0.95) in external testing. For predicting high-risk histopathological features (grade-III, lymphatic/pleural/vascular/nerve invasion, and spread through air spaces), AUCs were 0.82 (0.77-0.88) and 0.85 (0.81-0.89) in internal/external sets. DCA confirmed superior net benefit over CT model. The model stratified progression-free ( P = 0.002) and overall survival ( P = 0.017) in the TCIA cohort. CONCLUSION: PET/CT radiomics-based models enable accurate non-invasive prediction of invasiveness and high-risk pathology in early-stage LUAD, guiding optimal surgical resection.