Related Experiment Video
Updated: Nov 5, 2025
Anteromesial Temporal Lobectomy for Medically Intractable Temporal Lobe Epilepsy: An Operative Study
Published on: August 15, 2025
Nomograms to Predict Verbal Memory Decline After Temporal Lobe Resection in Adults With Epilepsy
Robyn M Busch1, Olivia Hogue2, Margaret Miller2
1From the Epilepsy Center (R.M.B., L.F., W.B., I.M.N., L.J.) and Department of Neurology (R.M.B., M.M., I.M.N., L.J.), Neurological Institute, and Department of Quantitative Health Sciences (O.H., M.W.K.), Lerner Research Institute, Cleveland Clinic, OH; Department of Psychology (M.P.M.), University of Toronto; Krembil Brain Institute (M.P.M.), University Health Network, Toronto, Ontario, Canada; Department of Neurology (M.H.), Columbia University, New York, NY; Department of Neurology (M.K., D.L.D.), University of Washington School of Medicine, Seattle; Department of Psychiatry (C.R.M., A.R.), University of California, San Diego; Departments of Neurology and Pediatrics (D.L.D.), Emory University School of Medicine, Atlanta, GA; and Department of Neurology (B.P.H.), University of Wisconsin School of Medicine and Public Health, Madison. buschr@ccf.org.
Predicting verbal memory decline after temporal lobe resection (TLR) for epilepsy is now possible. New models use clinical predictors to estimate the probability of memory loss, aiding surgical decisions.
Area of Science:
- Neurosurgery
- Epilepsy Research
- Cognitive Neuroscience
Background:
- Temporal lobe resection (TLR) is a common epilepsy surgery.
- Predicting postoperative verbal memory decline is crucial for patient counseling.
- Existing prediction methods lack external validation.
Purpose of the Study:
- Develop and validate predictive models for postoperative verbal memory decline.
- Utilize easily accessible preoperative clinical predictors.
- Aid clinicians in patient selection for TLR.
Main Methods:
- Multivariable models developed using Rey Auditory Verbal Learning Test (RAVLT) and Wechsler Memory Scale-Third Edition (WMS-III) subtests.
- Models developed in 359 patients and validated in 290 patients across multiple epilepsy centers.
- Harrell step-down procedure used for variable selection.
Main Results:
- 29% (development) and 26% (validation) of patients showed memory decline.
- Initial models showed good predictive accuracy (c-statistic 0.77-0.80).
- Updated models incorporating both cohorts demonstrated excellent performance (RAVLT c=0.81, LM c=0.76, VPA c=0.78).
Conclusions:
- Clinically applicable nomograms were developed to predict verbal memory decline.
- These tools assist in estimating the probability of memory decline after TLR.
- Class II evidence supports the accuracy of these prediction models.
