Predicting Mortality of Patients with Squamous Cell Carcinoma of the Maxilla: Novel Models with Clinical and Histopathological Predictors

Objective: to develop prediction models that calculate postoperative 2- and 5-year survival probabilities of patients with squamous cell carcinoma of the maxilla (MSCC).

Methods: The TRIPOD-checklist was used as guide for the development of the prediction models. Data was collected from the medical records of patients who had been operated between 2000 and 2015 for MSCC. With potential clinical and histopathological predictors, confounding-(un)adjusted multivariate Cox and logistic regression models were computed with stepwise backward selection. Internal validation was performed to assess calibration and discriminatory ability.

Results: 95 patients with MSCC were included with 2-year follow-up. Ten patients had incomplete 5-year follow-up. The following predictive factors were significantly associated with mortality probability: age, neck treatment, surgical margins, bone invasion, pattern of invasion and vasoinvasion. Models were adjusted for confounding with Charlson Comorbidity Index. The resulting confounding adjusted 2- and 5-year logistic models were internally validated and converted into probability equations. Internal validation of the 2- and 5-year models resulted in C-indexes of .841 and .770, and bootstrapped C-indexes of .838 and .749.

Conclusion: The 2- and 5-year mortality probability can be calculated with newly developed prediction models with MSCC-specific clinical and histopathological predictors. Risk of 2- and 5-year mortality can be assessed with reported odds- and hazard ratios as well.