Mathematical Models for Intraoperative Prediction of Metastasis to Lymph Nodes in the Hilar-intrapulmonary or the Mediastinal Region in Patients With Clinical Stage I Non-small Cell Lung Cancer: a Retrospective Cohort Study

Abstract Background: It remains challenging to determine the regions of metastasis to lymph nodes during operation for clinical stage I non-small cell lung cancer (NSCLC). This study aimed to establish intraoperative mathematical models with nomograms for predicting the hilar-intrapulmonary node metastasis (HNM) and the mediastinal node metastasis (MNM) in patients with clinical stage I NSCLC.Methods: The clinicopathological variables of 585 patients in a derivation cohort who underwent thoracoscopic lobectomy with complete lymph node dissection were retrospectively analysed for their association with the HNM or the MNM. After analysing the variables, we developed multivariable logistic models with nomograms to estimate the risk of lymph node metastasis in different regions. The predictive efficacy was then validated in a validation cohort of 418 patients.Results: It was confirmed that CEA (> 5.75 ng/ml), CYFRA211 (> 2.85 ng/ml), the maximum diameter of tumour (> 2.75 cm), tumour differentiation (grade III), bronchial mucosa and cartilage invasion, and vascular invasion were predictors of HNM, and CEA (>8.25 ng/ml), CYFRA211 (> 2.95 ng/ml), the maximum diameter of tumour (> 2.75 cm), tumour differentiation (grade III), bronchial mucosa and cartilage invasion, vascular invasion, and visceral pleural invasion were predictors of MNM. The validation of the prediction models based on the above results demonstrated good discriminatory power.Conclusions: Our predictive models are helpful in the decision‑making process of specific therapeutic strategies for the regional lymph node metastasis in patients with clinical stage I NSCLC.