Hippocampal Volumetric Integrity in Mesial Temporal Lobe Epilepsy: A Novel Automated MRI Approach
Abstract number :
1.234
Submission category :
5. Neuro Imaging / 5A. Structural Imaging
Year :
2017
Submission ID :
344945
Source :
www.aesnet.org
Presentation date :
12/2/2017 5:02:24 PM
Published date :
Nov 20, 2017, 11:02 AM
Authors :
Mat Hakimi, NYU School of Medicine; Babak Ardekani, Nathan Kline Institute; Christina Pressl, NYU School of Medicine; Ruben Kuzniecky, New York University School of Medicine; and Heath Pardoe, New York University School of Medicine
Rationale: To determine whether a novel automated MRI processing technique for assessing hippocampal volumetric integrity (HVI) can identify hippocampal sclerosis (HS) in mesial temporal lobe epilepsy (mTLE) and determine its performance relative to hippocampal volumetry (HV) and visual inspection. Methods: In this case-control retrospective study, we included individuals who had pre-surgical evaluation at the NYU Comprehensive Epilepsy Center and were referred for research MRI. MRI scans from 177 participants were initially identified for analysis; three were excluded due to prior resections and another was excluded because HVI processing failed. We analyzed HVI in healthy control, mTLE, non-HS temporal lobe epilepsy (TLE), and extratemporal focal epilepsy (EXTLE) subjects imaged using standardized epilepsy research imaging and in non-standardized clinically acquired images from mTLE subjects. HVI was assessed relative to healthy controls, where lower HVI indicates greater hippocampal atrophy. A significantly reduced HVI in mTLE patients suggested that HVI can detect HS. Performance of HVI, HV, and visual inspection was also assessed using receiver operating characteristic (ROC) analysis. Results: There were 173 participants: 35 healthy controls (mean age 33.5 ± 11.5 standard deviation (SD) years, 18 female), 98 patients with research scans (32.5 ± 13.4 years, 48 female), and 40 patients with clinical scans (31.8 ± 15.5 years, 19 female). mTLE patients had significantly reduced ipsilateral HVI relative to controls (effect size: -0.053, 5.62%, d = -1.735, P = 0.002 using research imaging). For lateralizing mTLE, HVI had a sensitivity of 88% while HV had a sensitivity of 92% when specificity was 70%. Conclusions: This novel approach for assessing HVI can detect HS. The fast processing speed suggests this technique could have utility as a quantitative tool to assist with imaging based diagnosis and lateralization of HS in a clinical setting. Funding: Study funded by FACES (Finding a Cure for Epilepsy and Seizures)
Neuroimaging