The risk and prognostic factors for lung metastases in oral squamous cell carcinoma: A population-based analysis of the SEER database

Purpose This study aimed to describe the risk and prognostic factors associated with lung metastases among oral squamous cell carcinoma (OSCC) patients, further to establish a nomogram model to predict the risk of lung metastases. Methods Data on OSCC patients was retrieved from the Surveillance, Epidemiology, and End Results (SEER) database between 2010 and 2019. Univariable and multivariable logistic and Cox regression models were used to explore the risk factors for developing lung metastases, further the predictive nomogram was constructed. Results 19, 606 OSCC patients meets the inclusion criteria and were enrolled into this study. Of which, 221 cases have lung metastases at initial diagnosis. Multivariate logistic regression analysis indicated race, T stage, N stage as well as bone metastases, liver metastases were independently associated the development of lung metastases. The diagnostic nomogram for developing lung metastases was constructed, the c-index for this model was 0.830 (0.804-0.856). Both the ROC curve and calibration curves revealed accurate predictability. DCA curve displayed the established nomogram model had good clinical applicability for the prediction of lung metastases. The median OS of OSCC patients with lung metastases was 7.0 months (6.0-9.0), and the 6-months, 12-months, 24-month OS rates were 54.5%, 30.9%, 17.7%, respectively. The multivariate Cox analysis showed that chemotherapy and liver metastases were independently associated with both OS and CSS. Conclusion This study determined the risk and prognostic factors for lung metastases among OSCC patients and the established nomogram had good calibration and discrimination for predicting lung metastases.