Aaron F Struck1, Michael B Westover2. 1. Massachusetts General Hospital, Harvard Medical School, Division of Epilepsy, Wang 735, 55 Fruit St. Boston, MA 02114, United States. Electronic address: astruck@mgh.harvard.edu. 2. Massachusetts General Hospital, Harvard Medical School, Division of Epilepsy, Wang 735, 55 Fruit St. Boston, MA 02114, United States.
Abstract
PURPOSE: Neuroimaging is critical in deciding candidacy for epilepsy surgery. Currently imaging is primarily assessed qualitatively, which may affect patient selection and outcomes. METHOD: The epilepsy surgery database at MGH was reviewed for temporal lobectomy patients from the last 10 years. Radiology reports for MRI and FDG-PET were compared to the epilepsy conference consensus. First, specific findings of ipsi/contra hippocampal atrophy and T2 signal changes were directly compared. Next the overall impression of presence of hippocampal sclerosis (HS) for MRI and temporal hypometabolism for PET was used for sensitivity/specificity analysis. To assess predictive power of imaging findings logistic regression was used. RESULTS: 104 subjects were identified. 70% of subjects were ILAE class I at 1-year. Radiology reports and the conference consensus differed in 31% of FDG-PET studies and 41% of MRIs. For PET most disagreement (50%) stemmed for discrepancy regarding contralateral temporal hypometabolism. For MRI discrepancy in ipsilateral hippocampal atrophy/T2 signal accounted for 59% of disagreements. When overall impression of the image was used the overall reliability between groups was high with only MRI sensitivity to detect HS (0.75 radiology, 0.91 conference, p=0.02) was significantly different between groups. On logistic regression MRI was a significant predictor of HS, but still 36% of patients with normal MRI as read by both groups had HS on pathology. CONCLUSION: Despite some difference in specific radiologic findings, overall accuracy for MRI and PET is similar in clinical practice between radiology and conference; nonetheless there are still cases of hippocampal pathology not detected by standard imaging methods.
PURPOSE: Neuroimaging is critical in deciding candidacy for epilepsy surgery. Currently imaging is primarily assessed qualitatively, which may affect patient selection and outcomes. METHOD: The epilepsy surgery database at MGH was reviewed for temporal lobectomy patients from the last 10 years. Radiology reports for MRI and FDG-PET were compared to the epilepsy conference consensus. First, specific findings of ipsi/contra hippocampal atrophy and T2 signal changes were directly compared. Next the overall impression of presence of hippocampal sclerosis (HS) for MRI and temporal hypometabolism for PET was used for sensitivity/specificity analysis. To assess predictive power of imaging findings logistic regression was used. RESULTS: 104 subjects were identified. 70% of subjects were ILAE class I at 1-year. Radiology reports and the conference consensus differed in 31% of FDG-PET studies and 41% of MRIs. For PET most disagreement (50%) stemmed for discrepancy regarding contralateral temporal hypometabolism. For MRI discrepancy in ipsilateral hippocampal atrophy/T2 signal accounted for 59% of disagreements. When overall impression of the image was used the overall reliability between groups was high with only MRI sensitivity to detect HS (0.75 radiology, 0.91 conference, p=0.02) was significantly different between groups. On logistic regression MRI was a significant predictor of HS, but still 36% of patients with normal MRI as read by both groups had HS on pathology. CONCLUSION: Despite some difference in specific radiologic findings, overall accuracy for MRI and PET is similar in clinical practice between radiology and conference; nonetheless there are still cases of hippocampal pathology not detected by standard imaging methods.
Authors: Willem M Otte; Pieter van Eijsden; Josemir W Sander; John S Duncan; Rick M Dijkhuizen; Kees P J Braun Journal: Epilepsia Date: 2012-03-01 Impact factor: 5.864
Authors: Andrew J Durnford; William Rodgers; Fenella J Kirkham; Mark A Mullee; Andrea Whitney; Martin Prevett; Lucy Kinton; Matthew Harris; William P Gray Journal: Seizure Date: 2011-09-14 Impact factor: 3.184
Authors: D H Lee; F Q Gao; J M Rogers; I Gulka; I R Mackenzie; A G Parrent; C S Kubu; D G Munoz; R S McLachlan; W T Blume; J P Girvin Journal: AJNR Am J Neuroradiol Date: 1998-01 Impact factor: 3.825
Authors: C R Jack; F W Sharbrough; C K Twomey; G D Cascino; K A Hirschorn; W R Marsh; A R Zinsmeister; B Scheithauer Journal: Radiology Date: 1990-05 Impact factor: 11.105
Authors: B V Savitr Sastri; A Arivazhagan; Sanjib Sinha; Anita Mahadevan; R D Bharath; J Saini; R Jamuna; J Keshav Kumar; S L Rao; B A Chandramouli; S K Shankar; P Satishchandra Journal: J Neurol Sci Date: 2014-03-19 Impact factor: 3.181
Authors: L N Tanenbaum; A J Tsiouris; A N Johnson; T P Naidich; M C DeLano; E R Melhem; P Quarterman; S X Parameswaran; A Shankaranarayanan; M Goyen; A S Field Journal: AJNR Am J Neuroradiol Date: 2017-04-27 Impact factor: 3.825