Using Supervised Machine Learning to Predict 30-day Readmission and Reoperation Following Resection of Supratentorial High-grade Gliomas: An NSQIP Study Involving 9,418 Patients

Introduction: High-grade gliomas (HGG) are among the rarest yet most aggressive tumor types in neurosurgical practice. In the current literature, few studies have assessed the drivers of early outcomes following surgical resection of these tumors and investigated their association with quality of care. We aimed to identify the clinical predictors for 30-day readmission and reoperation following HGG surgery using the NSQIP database and seek to create web-based applications predicting each outcome.

Methods: Our study is a retrospective, multicenter cohort analysis of patients who underwent resection of supratentorial HGG between January, 2016, and December, 2020, using the NSQIP database. Demographics and comorbidities were extracted. The primary outcomes were 30-day unplanned readmission and reoperation. A stratified 80:20 split of the available data was carried out. Supervised machine learning algorithms were trained to predict 30-day outcomes.

Results: A total of 9,418 patients were included in our cohort. The observed rate of unplanned readmission within 30-days of surgery was 14.9%(N=1221). In terms of predictors, weight, chronic steroid use, preoperative BUN level, and WBC count were associated with a higher risk of readmission. The observed rate of unplanned reoperation within 30-days of surgery was 5.47%(N=489). In terms of predictors, increased weight, higher operative time, and a higher number of days between hospital admission and the operation were associated with an increased risk of early reoperation. Our Random Forest(RF) algorithm showed the highest predictive performance for early readmission (AUC=0.967),while the XGBoost(XGB) algorithm showed the highest predictive performance for early reoperation (AUC=0.985). Web-based tools for both outcomes were deployed respectively:https://glioma-readmission.herokuapp.com/ ;https://glioma-reoperation.herokuapp.com/.

Conclusion : In this study, we provide the first nationwide analysis for short-term outcomes in patients undergoing resection of supratentorial HGG. Multiple patient, hospital, and admission factors are associated with readmission and reoperation,confirmed by machine learning predicting patients’ prognosis leading to better planning preoperatively and subsequently improved personalized patient care.