We constructed histograms to examine distributions of continuous variables. We also computed the number of subjects with ( n ≥ 1) and without ( n = 0) infratentorial or cord lesions on each image. To investigate the potential impact of FLAIR3 on the diagnostic criterion in MS, we calculated the number of lesions detected on either or both image sets. Control ROIs were placed in adjacent normal-appearing white matter, and the lesion conspicuity was assessed by the contrast-to-noise ratio (CNR), calculated as CNR = (S L–S B)/SD N, where S L and S B are the average signal intensities in the lesion and background ROIs, respectively, and SD N is the estimated SD of the image noise, obtained from histogram analysis of local image variance. 13 The ROIs were carefully placed to minimize partial volume effects from surrounding tissue that could affect contrast computations. ROIs were manually drawn on the identified lesions on an axial section showing the largest extent of the lesion using MRIcron software ( ). Lesions were identified as regions with hyperintense signal compared with the background tissue, visible on all 3 planes. Rater 3 repeated the analysis (with a different randomization) after 3 weeks to assess intrarater reliability. Interrater reliability was evaluated using image evaluations by raters 1 and 2, together with the first image evaluation by rater 3. Image evaluation was performed during 1–2 days. For rater 3, 3D T2-FLAIR and FLAIR3 images from all subjects were pooled and randomized for review. In addition, a neuroradiology fellow (O.A., rater 3) who was not familiar with the imaging protocol or the study objective independently evaluated all images to identify infratentorial and cervical cord lesions. Thus, we did not expect a substantial recall bias. There was no specified order for reviewing images in each set. The 3D T2-FLAIR and FLAIR3 images were reviewed in several sessions during 7 weeks (rater 1) and 8.5 weeks (rater 2), with the raters first analyzing all 3D T2-FLAIR images before starting on FLAIR3 images. Due to the large difference in image contrast between T2-FLAIR and FLAIR3, blind evaluation was not feasible. With the extended FOV offered by 3D acquisitions, we also assessed the performance of these two methods for detecting upper cervical cord lesions.Ī board-certified neuroradiologist with 6 years of experience (A.K., rater 1) and an MS neurologist with 15 years of experience in neuroimaging of MS (J.A.L., rater 2) identified all infratentorial and upper cervical lesions seen on 3D T2-FLAIR and FLAIR3 images. The primary aim of this prospective study was to test whether the lesion contrast and detectability of MS infratentorial lesions are improved with FLAIR3 imaging compared with 3D T2-FLAIR alone. 9 A previous study has shown that an algebraic combination of T2-FLAIR and T2-weighted images, 10 referred to as FLAIR3, can yield substantial improvement in lesion contrast, but the performance for infratentorial lesion detection was not specifically addressed in that publication. 8 Combining T2-weighted and proton density–weighted images has also shown improved contrast. 3- 5ģD T2-FLAIR imaging has shown promise in improving the detection of infratentorial lesions, 6, 7 especially with optimized scan parameters. However, the contrast of infratentorial lesions on T2-FLAIR is suboptimal due to partial T1-weighting and different tissue relaxation properties between supratentorial and infratentorial regions. T2-FLAIR MR imaging is currently the most commonly used imaging sequence for identifying brain T2-hyperintense lesions. MR imaging is the primary technique in MS, used for diagnosis, detecting pathology, monitoring the disease course, and patient management. 2 Dissemination in space is established through detection of ≥1 T2-hyperintense lesion characteristic of MS in ≥2 areas in the CNS: infratentorial (including the brain stem and cerebellum), periventricular, and juxtacortical brain regions and the spinal cord. 1 The current McDonald criteria for diagnosing MS require demonstration of the dissemination of CNS lesions in time and space. Infratentorial brain lesions are commonly present in MS, an inflammatory and demyelinating disease of the CNS that affects 2.5 million individuals worldwide. ABBREVIATION: CNR contrast-to-noise ratio
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |