Hemispherectomy Outcome Prediction Scale: a validity study

Abstract

OBJECTIVE Hemispherectomy is highly effective for patients with medically refractory epilepsy (MRE) arising from a single hemisphere. Recently, the Hemispherectomy Outcome Prediction Scale (HOPS) was developed as a prediction tool for seizure freedom after hemispherectomy. The authors’ goal was to perform a validation study to determine the generalizability of the HOPS score. METHODS The authors present an observational, retrospective, 20-year, single-institution, two-surgeon experience using the lateral peri-insular hemispherectomy approach to validate the HOPS score. Variables used to derive the HOPS score included seizure onset age, semiology, PET hypometabolism, seizure substrate, and history of prior epilepsy resection. Multivariable logistic regression, multiple imputation, and Bayesian analyses were used to determine validity. RESULTS The authors’ cohort comprised 60 patients; 55% of patients were male and 78% were Caucasian. The median age at first hemispherectomy surgery was 72 months. At 1 year postoperatively, 80% of patients had Engel class I outcomes, analogous to most contemporary series. All patients who experienced seizure recurrence after hemispherectomy did so within the first 2 years postoperatively. Sixteen (27%) and 10 (17%) patients had contralateral MRI findings and hypometabolism on PET, respectively. Both a multivariable logistic regression model using HOPS score variables (model p = 0.2588) and a revised model that included contralateral MRI findings (model p = 0.4715) were not statistically significant in this cohort. Bayesian analysis also did not validate the HOPS score. CONCLUSIONS While seizure outcome prediction tools may be helpful for counseling patients about postoperative outcomes, rigorous validity and reliability testing are required. Prospective, standardized, and longitudinal evaluation of patients undergoing hemispherectomy are needed.