Proposal for diagnosis using FLAIR image aimed for pediatric MELAS with recurrent stroke-like episodes on MRI system cannot take ASL imaging

Abstract Background Arterial spin-labeling (ASL) imaging is currently the most useful method for diagnosing mitochondrial encephalomyopathy, lactic acidosis, and stroke-like attack syndrome (MELAS). However, ASL is often an optional feature of standard MRI systems. Therefore, not all MRI systems can perform ASL imaging. In contrast, fluid-attenuated inversion recovery (FLAIR) imaging is one of the common sequences in brain MRI because FLAIR imaging can be performed regardless of the specifications of the equipment. This study aimed to compare the diagnostic performance of quantitative analysis of signal intensity obtained from fluid-attenuated inversion recovery (FLAIR) images with ASL images for MELAS with recurrent stroke-like episodes (SLEs). A total of 68 cases with normal magnetic resonance imaging findings and 25 cases diagnosed MELAS with recurrent SLEs were included. We evaluated the frontal lobe and cuneus as target areas and compared the regional cerebral blood flow (rCBF) values obtained from ASL images with the normalized signal intensity (nSI) obtained from FLAIR images. Results The sensitivity and specificity for diagnosing MELAS from linear discriminant analysis (LDA) obtained from the rCBF values were 0.84 and 0.941, respectively, and those of nSI were 0.8 and 0.897, respectively. The area under the ROC curves (AUC) calculated from the receiver operating characteristic (ROC) curve analysis using rCBF values and nSI were 0.889 and 0.804, respectively. Conclusion Quantitative analysis using the signal intensity of the FLAIR image could have a diagnostic performance equivalent to that of rCBF values obtained from ASL images.