Dataset Information

Fully automated open-source lesion mapping of T2-FLAIR images with FSL correlates with clinical disability in MS.

ABSTRACT: T2 Lesion Volume (T2LV) has been an important biomarker for multiple sclerosis (MS). Current methods available to quantify lesions from MR images generally require manual adjustments or multiple images with different contrasts. Further, implementations are often not easily or openly accessible.We created a fully unsupervised, single T2 FLAIR image T2LV quantification package based on the popular open-source imaging toolkit FSL.By scripting various processing tools in FSL, we developed an image processing pipeline that distinguishes normal brain tissue from CSF and lesions. We validated our method by hierarchical multiple regression (HMR) with a preliminary study to see if our T2LVs correlate with clinical disability measures in MS when controlled for other variables.Pearson correlations between T2LV and Expanded Disability Status Scale (EDSS: r = 0.344, P = 0.013), Six-Minute Walk (6MW: r = -0.513, P = 0.000), Timed 25-Foot Walk (T25FW: r = -0.438, P = .000), and Symbol Digit Modalities Test (SDMT: r = -0.499, P = 0.000) were all significant. Partial correlations controlling for age were significant between T2LV and 6MW (r = -0.433, P = 0.002), T25FW (r = -0.392, P = 0.004), and SDMT (r = -0.450, P = 0.001). In HMR, T2LV explained significant additional variance in 6MW (R(2) change = 0.082, P = 0.020), after controlling for confounding variables such as age, white matter volume (WMV), and gray matter volume (GMV).Our T2LV quantification software produces T2LVs from a single FLAIR image that correlate with physical disability in MS and is freely available as open-source software.

SUBMITTER: Wetter NC

PROVIDER: S-EPMC4731385 | biostudies-other | 2016 Mar

REPOSITORIES: biostudies-other

ACCESS DATA

Similar Datasets

Project description:IntroductionBladder magnetic resonance imaging (MRI) has been recently integrated in the diagnosis pathway of bladder cancer. However, automatic recognition of suspicious lesions is still challenging. Thus, development of a solution for proper delimitation of the tumor and its separation from the healthy tissue is of primordial importance. As a solution to this unmet medical need, we aimed to develop an artificial intelligence-based decision support system, which automatically segments the bladder wall and the tumor as well as any suspect area from the 3D MRI images.MaterialsWe retrospectively assessed all patients diagnosed with bladder cancer, who underwent MRI at our department (n=33). All examinations were performed using a 1.5 Tesla MRI scanner. All images were reviewed by two radiologists, who performed manual segmentation of the bladder wall and all lesions. First, the performance of our fully automated end-to-end segmentation model based on a 3D U-Net architecture (by considering various depths of 4, 5 or 6 blocks) trained in two data augmentation scenarios (on 5 and 10 augmentation datasets per original data, respectively) was tested. Second, two learning setups were analyzed by training the segmentation algorithm with 7 and 14 MRI original volumes, respectively.ResultsWe obtained a Dice-based performance over 0.878 for automatic segmentation of bladder wall and tumors, as compared to manual segmentation. A larger training dataset using 10 augmentations for 7 patients could further improve the results of the U-Net-5 model (0.902 Dice coefficient at image level). This model performed best in terms of automated segmentation of bladder, as compared to U-Net-4 and U-Net-6. However, in this case increased time for learning was needed as compared to U-Net-4. We observed that an extended dataset for training led to significantly improved segmentation of the bladder wall, but not of the tumor.ConclusionWe developed an intelligent system for bladder tumors automated diagnostic, that uses a deep learning model to segment both the bladder wall and the tumor. As a conclusion, low complexity networks, with less than five-layers U-Net architecture are feasible and show good performance for automatic 3D MRI image segmentation in patients with bladder tumors.

Project description:INTRODUCTION:Progressive brain atrophy, development of T1-hypointense areas, and T2-fluid-attenuated inversion recovery (FLAIR)-hyperintense lesion formation in multiple sclerosis (MS) are popular volumetric data that are often utilized as clinical outcomes. However, the exact clinical interpretation of these volumetric data has not yet been fully established. METHODS:We enrolled 42 consecutive patients with MS who fulfilled the revised McDonald criteria of 2010. They were followed-up for more than 3 years from onset, and cross-sectional brain volumetry was performed. Patients with no brain lesions were excluded in advance from this study. For the brain volumetric data, we evaluated several parameters including age-adjusted gray-matter volume atrophy, age-adjusted white-matter volume atrophy, and T2-FLAIR lesion volume. The numbers of T1-hypointense and T2-FLAIR-hyperintense areas were also measured along the same timeline. The clinical data pertaining to disease duration, expanded disability status scale (EDSS), and MS severity score (MSSS) at the timing of volumetry were collected. RESULTS:Among the 42 patients with MS and brain lesions, the number of T1-hypointensity (rho = 0.51, p<0.001), gray-matter atrophy (rho = 0.40, p<0.01) and white-matter atrophy (rho = 0.49, p<0.001) correlated with the EDSS. T1-hypointensity count divided by FLAIR lesion volume correlated with the MSSS (rho = 0.60, p<0.001). Meanwhile, counts or volumes of FLAIR-hyperintense lesions were associated only with the times of past relapses, and did not correlate with present neurological disability level or ongoing disease activity. These findings were consistent regardless of the presence of spinal cord lesions. CONCLUSION:Numbers of T1-hypointensities and brain atrophy equally indicated the current neurological disability in MS. The number of T1-hypointensities divided by FLAIR lesion volume represented the clinical severity. The size or number of FLAIR lesions reflected earlier relapses but was not a good indicator of neurological disability or clinical severity.

Dataset Information

Fully automated open-source lesion mapping of T2-FLAIR images with FSL correlates with clinical disability in MS.

Similar Datasets

OmicsDI is part of the ELIXIR infrastructure

Tweets