Cortical lesions (CL) serve as a valuable biomarker for im- proving the differential diagnosis[1] of multiple sclerosis (MS) and are associated with disease severity[2]. However, detecting CL in vivo remains challenging, particularly in clinical settings, due to their small size and their tendency to affect the more superficial, less myelinated layers of the cortex[3]. While specialized 3T MRI sequences such as dou- ble inversion recovery (DIR) and phase-sensitive inversion recovery (PSIR) have improved CL detection compared to using conventional sequences like fluid-attenuated inversion recovery (FLAIR) or magnetization-prepared rapid gradient echo (MPRAGE)[2], their use in clinical settings is hindered by their long acquisition time. In this study, we aim to propose synthetic DIR (sDIR) and synthetic PSIR (sPSIR) contrasts computed with clinical sequences commonly found in MS protocols to improve CL detection in clinical settings. The use of sDIR and sPSIR allowed to significantly (p < 0.01) increase the number of CL detected (8.237 ± 10.559) when compared to using FLAIR and MPRAGE sequences (4.5±5.658). No statistical difference (p = 0.538) was found between the number of CL detected using sDIR and sPSIR when compared to using advanced DIR and PSIR sequences (respectively, 4.667 ± 6.03 vs. 3.533 ± 3.16). In conclusion, this work proposes a simple and explainable method for generating synthetic contrasts using clinical se- quences commonly found in MS protocols, achieving similar CL detection rates as using advanced research DIR and PSIR sequences. This approach holds promises for expanding the utilization of CL as a valuable biomarker of both MS diagno- sis and prognosis in routine clinical practice.
Vanden Bulcke, Colin ; et. al. SYNTHETIC DIR/PSIR CONTRASTS COMPUTED WITH CLINICAL SEQUENCES FOR CORTICAL LESION ASSESSMENT.IEEE International Symposium on Biomedical Imaging (ISBI 2024) (Athens, du 27/05/2024 au 30/05/2024).