Clinical FDG-PET Findings in Patients with Temporal Lobe Epilepsy: Concordance with EEG and MRI Zeenat Jaisani , Robert S. Miletich, Murali Ramanathan, Arie L. Weinstock From the Department of Neurology, University of Alabama at Birmingham (UAB), Birmingham, AL (ZJ); Department of Neurology, University at Buffalo, Jacobs School of Medicine and Biomedical Sciences, Buffalo, NY (ALW); Department of Pharmacological Sciences, University at Buffalo, Buffalo, NY (MR); and Department of Nuclear Medicine, University at Buffalo, Jacobs School of Medicine and Biomedical Sciences, Buffalo, NY (RSM). ABSTRACT BACKGROUND AND PURPOSE: We aimed to study the yield of PET in temporal lobe epilepsy (TLE) by analyzing the correlation of PET findings with MRI, and interictal and ictal EEG findings, in a single-center cohort of patients with TLE. Predictors of PET thalamic changes and its role in predicting postsurgical outcome were also studied. METHODS: This was a retrospective study of 39 patients with TLE who underwent MRI, PET, and scalp video EEG monitoring at the University at Buffalo, New York from 2001 to 2011 during presurgical evaluation. PET-defined metabolism of the temporal lobes was evaluated using a 4-point ordinal rating scale. RESULTS: PET hypometabolism was associated with a variation in ictal (P = .034) and interictal (P < .001) foci in both lesional (by MRI) and nonlesional patients. Nonlesional MRI scans were associated with none to mild temporal PET hypometabolism (71% of patients) while lesional MRI scans were associated with moderate to severe hypometabolism (82% of patients) (P = .006). The odds of thalamic hypometabolism were 5.36 times higher when there was moderate to severe temporal hypometabolism (P = .039). CONCLUSION: This study underscores the utility of PET in localizing ictal foci in TLE patients even in those with normal MRI. The degree of PET hypometabolism corresponds to presence of MRI pathology. Coexistent thalamic hypometabolism with temporal hypometabolism suggests a secondary effect of distant temporal network disruption. Extratemporal metabolism is a predictor of poor postsurgical seizure outcome in TLE patients. Keywords: FDG-PET, temporal lobe epilepsy, thalamic hypometabolism, temporal hypometabolism. Acceptance: Received April 29, 2019, and in revised form October 8, 2019. Accepted for publication October 9, 2019. Correspondence: Address correspondence to Zeenat Jaisani. E-mail: zjaisani@uabmc.edu or zjaisani@yahoo.com. Acknowledgments: We thank Dr. Jerzy Szaflarski (University of Alabama at Birmingham) and Ganna Chaitanya, MD, PhD for reviewing the manuscript. The Grammarly tool (https://www.grammarly.com) was used for language editing. Conflict of Interest: The authors declare no conflict of interest. J Neuroimaging 2020;30:119-125. DOI: 10.1111/jon.12671 Introduction Temporal lobe epilepsy (TLE) is the most common subtype of medically refractory focal epilepsy. It is estimated that about 30% of TLE patients have identifiable MRI lesions such as tu- mor, cortical dysplasia (CD), or cavernous hemangioma. Mesial temporal lobe epilepsy (MTLE) characterizes the subgroup that has hippocampal sclerosis (HS positive) on MRI and constitutes about 50% of TLE. The remaining TLE patients, however, have no definite pathology on MRI. 1–3 Several studies have been undertaken to evaluate the role of PET in studying TLE in the presence or absence of positive MRI findings. Fluorodeoxyglucose (FDG)-PET is highly sensitive in localizing the epileptogenic region in combination with MRI and EEG. 4–6 FDG-PET detects interictal hypometabolism in as much as 80% of patients with MTLE. 5,7 Previous studies have also shown that PET hypometabolism corresponds to the ictal EEG region in patients without evidence of HS on MRI. 8–13 It has been, however, noted that the hypometabolic zone of- ten extends to involve the ipsilateral temporal neocortex, the ipsilateral basal ganglia and thalamus and, occasionally, the contralateral temporal lobe. 6,14–21 Recent studies also show that FDG-PET can provide prognostic data based on the extent of initial hypometabolism present; where a more extensive region of hypometabolism may favor a poor prognosis and a longer du- ration of epilepsy. 22–24 Despite the completion of several studies, our understanding of the relevance of metabolic abnormalities in subcortical regions, such as thalamus, is not well-defined to date. Variable outcomes linked to thalamic metabolic abnor- malities in predicting postsurgical seizure outcome have been reported. 22,25,26 Our study aimed to analyze certain PET patterns and their correlation with MRI lesions, ictal and interictal EEG, epilepsy duration and type, and age, in a single-center cohort of TLE patients. Past studies have analyzed these aspects individually and in patients who underwent surgery, whereas, in our study, we analyzed patients who had TLE irrespective of their surgical status. We included patients with a full range of MRI findings including normal MRI, mesial temporal lobe sclerosis (MTS), and other miscellaneous abnormalities such as cavernoma, tumor, or CD. The primary objectives of this research were to: (1) assess the yield of PET in the presurgical evaluation of patients ◦ C 2019 by the American Society of Neuroimaging 119