Zhen-Ming Wang1,2, Peng-Hu Wei3, Chunxiu Wang4, Yaqin Hou1, Kun Guo1, Bixiao Cui1, Yongzhi Shan3, Guo-Guang Zhao3,5, Jie Lu6,7. 1. Department of Radiology and Nuclear Medicine, Xuanwu Hospital, Capital Medical University, No. 45 Changchun Street, Xicheng District, Beijing, 100053, China. 2. Beijing Key Laboratory of Magnetic Resonance Imaging and Brain Informatics, Beijing, China. 3. Department of Neurosurgery, Xuanwu Hospital, Capital Medical University, No. 45 Changchun Street, Xicheng District, Beijing, 100053, China. 4. Department of Evidence-based Medicine, Xuanwu Hospital, Capital Medical University, Beijing, China. 5. Clinical Research Center for Epilepsy, Capital Medical University, Beijing, 100053, China. 6. Department of Radiology and Nuclear Medicine, Xuanwu Hospital, Capital Medical University, No. 45 Changchun Street, Xicheng District, Beijing, 100053, China. imaginglu@hotmail.com. 7. Beijing Key Laboratory of Magnetic Resonance Imaging and Brain Informatics, Beijing, China. imaginglu@hotmail.com.
Abstract
OBJECTIVES: To integrate the glucose metabolism measured using [18F]FDG PET/CT and anatomical features measured using MRI to forecast the post-surgical seizure outcomes of intractable temporal lobe epilepsy. METHODS: This retrospective study enrolled 63 patients with drug-resistant temporal lobe epilepsy. Z-transform of the patients' PET images based on comparison with a database of healthy controls, cortical thickness, and quantitative anisotropy (QA) of the diffusion spectrum imaging concordant/non-concordant with cortical resection was adopted to quantify their predictive values for the post-surgical seizure outcomes. RESULTS: The PET hypometabolism region was concordant with the surgical field in 47 of the 63 patients. Forty-two patients were seizure-free post-surgery. The sensitivity and specificity of PET in predicting seizure freedom were 89.4% and 68.8%, respectively. Complete resection of foci with overlapped PET, cortical thickness, and QA abnormalities resulted in Engel I in 27 patients, which was a good predictor of seizure freedom with an odds ratio (OR) of 19.57 (95% CI 2.38-161.25, p = 0.006). Hypometabolism involved in multiple lobes (OR = 7.18, 95% CI 1.02-50.75, p = 0.048) and foci of hypometabolism with QA/cortical thickness abnormalities outside surgical field (OR = 14.72, 95% CI 2.13-101.56, p = 0.006) were two major predictors of Engel III/IV outcomes. ORs of QA to predict Engel I and seizure recurrence were 14.64 (95% CI 2.90-73.80, p = 0.001) and 12.01 (95% CI 2.91-49.65, p = 0.001), respectively. CONCLUSION: Combined PET and structural pattern is helpful to predict the post-surgical seizure outcomes and worse outcomes of Engel III/IV. This might decrease unnecessary surgical injuries to patients who are potentially not amenable to surgery. KEY POINTS: • A combined metabolic and structural pattern is helpful to predict the post-surgical seizure outcomes. • Favorable post-surgical seizure outcome was most likely reached in patients whose hypometabolism overlapped with the structural changes. • Hypometabolism in multiple lobes and QA or cortical thickness abnormalities outside the surgical field were predictors of worse seizure outcomes of Engel III/IV.
OBJECTIVES: To integrate the glucose metabolism measured using [18F]FDG PET/CT and anatomical features measured using MRI to forecast the post-surgical seizure outcomes of intractable temporal lobe epilepsy. METHODS: This retrospective study enrolled 63 patients with drug-resistant temporal lobe epilepsy. Z-transform of the patients' PET images based on comparison with a database of healthy controls, cortical thickness, and quantitative anisotropy (QA) of the diffusion spectrum imaging concordant/non-concordant with cortical resection was adopted to quantify their predictive values for the post-surgical seizure outcomes. RESULTS: The PET hypometabolism region was concordant with the surgical field in 47 of the 63 patients. Forty-two patients were seizure-free post-surgery. The sensitivity and specificity of PET in predicting seizure freedom were 89.4% and 68.8%, respectively. Complete resection of foci with overlapped PET, cortical thickness, and QA abnormalities resulted in Engel I in 27 patients, which was a good predictor of seizure freedom with an odds ratio (OR) of 19.57 (95% CI 2.38-161.25, p = 0.006). Hypometabolism involved in multiple lobes (OR = 7.18, 95% CI 1.02-50.75, p = 0.048) and foci of hypometabolism with QA/cortical thickness abnormalities outside surgical field (OR = 14.72, 95% CI 2.13-101.56, p = 0.006) were two major predictors of Engel III/IV outcomes. ORs of QA to predict Engel I and seizure recurrence were 14.64 (95% CI 2.90-73.80, p = 0.001) and 12.01 (95% CI 2.91-49.65, p = 0.001), respectively. CONCLUSION: Combined PET and structural pattern is helpful to predict the post-surgical seizure outcomes and worse outcomes of Engel III/IV. This might decrease unnecessary surgical injuries to patients who are potentially not amenable to surgery. KEY POINTS: • A combined metabolic and structural pattern is helpful to predict the post-surgical seizure outcomes. • Favorable post-surgical seizure outcome was most likely reached in patients whose hypometabolism overlapped with the structural changes. • Hypometabolism in multiple lobes and QA or cortical thickness abnormalities outside the surgical field were predictors of worse seizure outcomes of Engel III/IV.