Validation of an Online Calculator to Predict Short-term Postoperative Outcomes in Patients with Metastatic Brain Tumors

Abstract

Introduction: Estimating the risk of extended length of stay (LOS) or non-routine discharge disposition is helpful in surgical decision-making for patients with brain metastases (BM). In 2020, an online calculator was introduced by Khalafallah et al. that stratified the risk of patients with brain tumors based on poor surgical outcomes. We applied the calculator to our population of BM patients to determine its generalizability and validity. Methods: We included BM patients who underwent a cranial procedure between 2015 and 2018 at a single academic institution. Patient age, race, marital status, admission status, KPS score, and medical co-morbidities (5-point modified frailty index (mFI-5)) were included in the analysis. We calculated the areas under the Receiver Operating Characteristics (ROC) curves to determine the validity of the model proposed in predicting extended LOS (>7 days) and need for specialty care at discharge (non-routine discharge disposition). Results: We analyzed 244 patients (mean age 61.2 years (SD 11.1), 57.0% female, and 78.1% Caucasian). The areas under the ROC curves were 0.8427 and 0.8422 for extended LOS and non-routine discharge disposition, suggesting high accuracy of the models for these outcomes. However, the (mFI-5) was not a significant predictor of either outcome in our multivariate analyses. Conclusions: We validated Khalafallah et al.’s predictive models of extended LOS and non-routine discharge disposition in our patient population, which included a broader range of surgical procedures. Further investigation of this model could clarify how the type of neurosurgical procedure influences outcomes, the role of the mFI-5, and its overall generalizability.