Background: Lymph node metastasis (LNM) is difficult to precisely predict before surgery in patients with early-T-stage non-small cell lung cancer (NSCLC). This study aimed to develop machine learning (ML)-based predictive models for LNM. Methods: Clinical characteristics and imaging features were retrospectively collected from 1,102 NSCLC ≤ 2 cm patients. A total of 23 variables were included to develop predictive models for LNM by multiple ML algorithms. The models were evaluated by the receiver operating characteristic (ROC) curve for predictive performance and decision curve analysis (DCA) for clinical values. A feature selection approach was used to identify optimal predictive factors. Results: The areas under the ROC curve (AUCs) of the 8 models ranged from 0.784 to 0.899. Some ML-based models performed better than models using conventional statistical methods in both ROC curves and decision curves. The random forest classifier (RFC) model with 9 variables introduced was identified as the best predictive model. The feature selection indicated the top five predictors were tumor size, imaging density, carcinoembryonic antigen (CEA), maximal standardized uptake value (SUV max), and age. Conclusions: By incorporating clinical characteristics and radiographical features, it is feasible to develop ML-based models for the preoperative prediction of LNM in early-T-stage NSCLC, and the RFC model performed best.

Background Thymic neuroendocrine tumors comprise a heterogeneous group of rare diseases. This study aimed to investigate the real-world clinicopathological features and treatment outcomes of thymic neuroendocrine tumors. Results A total of 104 patients diagnosed with thymic neuroendocrine tumors in a single institution from 1983 to 2021 were eligible. Fourteen (13.46%) and 28 (26.92%) patients diagnosed with thymic neuroendocrine tumors suffered from multiple endocrine neoplasia and ectopic adrenocorticotropic hormone syndrome, respectively. Ninety-seven (93.27%) patients underwent surgical resection, including 79 (81.44%) with radical resection. Except for 5 patients lost during follow-up, the 1-, 3- and 5-year overall survival rates were 91.8%, 70.2% and 54.6%, respectively. The median overall survival was 61.57 months. Multivariate analysis revealed that years at diagnosis (HR 0.559, 95% CI 0.364–0.857, p = 0.008), radical resection (HR 2.860, 95% CI 1.392–5.878, p = 0.004), pathological grade (HR 1.963, 95% CI 1.058–3.644, p = 0.033) and Masaoka–Koga stage (HR 2.250, 95% CI 1.548–3.272, p = 0.000) exerted significant differences in overall survival among 99 patients. In the surgery group, multivariate Cox regression analysis exhibited significant overall survival differences in years at diagnosis (HR 0.563, 95% CI 0.367–0.866, p = 0.009), neoadjuvant therapy (HR 0.248, 95% CI 0.071–0.872, p = 0.030), radical resection (HR 3.674, 95% CI 1.685–8.008, p = 0.001), pathological grade (HR 2.082, 95% CI 1.098–3.947, p = 0.025) and Masaoka–Koga stage (HR 2.445, 95% CI 1.607–3.719, p = 0.000). Conclusions Radical resection and Masaoka–Koga stage were independent prognostic factors for the survival of patients with thymic neuroendocrine tumors. Systemic therapy and integrated management of patients with advanced-stage disease require high-level clinical evidence.

Background: The management of ground-glass opacities (GGOs) depends mainly on personal experience. In clinical practice, benign GGOs are not rare in resected specimens, for which operations may be avoided. We retrospectively compared the clinical features of resected GGOs to identify differential diagnostic characteristics. Methods: Among 1456 patients with suspected malignant GGOs who underwent surgical resection, 105 patients (35 with benign GGOs and 70 matched controls with malignant GGOs) were included. Clinical characteristics, including demographics and radiologic, surgical and pathologic characteristics, were collected. Results: The smoking index (P = 0.044), frequency of coughing (P = 0.026), GGO size (P = 0.003), size change during follow-up (P = 0.011), location (P = 0.022), presence of air bronchogram sign (P = 0.004), distance to the pleura (P = 0.021) and positron emission tomography/computed tomography (PET/CT) appearance (P = 0.003) showed significant differences between the benign and malignant groups. Pathologically, the resected benign GGOs included focal fibrosis (17), inflammation or infection (seven), lymphoproliferative disorder (one), hamartoma (three), inflammatory myofibroblastic tumor (two), hemangioma or vascular malformation (two), endometriosis (two) and pulmonary cyst (one). Conclusions: A higher smoking index, coughing, larger size, similar or increased size during follow-up, location in the upper and middle lobes, air bronchogram sign on CT, lesion margin to pleura distance over 1 cm, and malignant tendency on PET/CT reports were associated with malignant GGOs. Relatively active surgical interventions could be considered for GGOs highly suspected of malignancy.