Resting-state functional magnetic resonance imaging with independent component analysis for presurgical seizure onset zone localization: A systematic review and meta-analysis.

Abstract

One of the greatest challenges of achieving successful surgical outcomes in patients with epilepsy is the ability to properly localize the seizure onset zone (SOZ). Many techniques exist for localizing the SOZ, including intracranial electroencephalography, magnetoencephalography, and stereoelectroencephalography. Recently, resting-state functional magnetic resonance imaging (rs-fMRI) in conjunction with independent component analysis (ICA) has been utilized for presurgical planning of SOZ resection, with varying results. In this meta-analysis, we analyze the current role of rs-fMRI in identifying the SOZ for presurgical planning for patients with drug-resistant epilepsy. Specifically, we seek to demonstrate its current effectiveness compared to other methods of SOZ localization. A literature review was conducted using the PubMed, MEDLINE, and Embase databases up to May of 2020. A total of 253 articles were screened, and seven studies were chosen for analysis. Each study was analyzed for SOZ localization by ground truth, SOZ localization by rs-fMRI with ICA, principal component analysis, or intrinsic connectivity contrast, and outcomes of surgery. A meta-analysis was performed to analyze how ground truth compares to rs-fMRI in SOZ localization. The odds ratio comparing ground truth to rs-fMRI was 2.63 (95% confidence interval = 0.66-10.56). Average concordance of rs-fMRI SOZ localization compared with ground truth localization across studies was 71.3%. In the hunt for less invasive presurgical planning for epilepsy surgery, rs-fMRI with ICA provides a promising avenue for future standard practice. Our preliminary results show no significant difference in surgical outcomes between traditional standards of SOZ localization and rs-fMRI with ICA. We believe that rs-fMRI could be a step forward in this search. Further investigation comparing rs-fMRI to traditional methods of SOZ localization should be conducted, with the hope of moving toward relying solely on noninvasive screening methods.