Accuracy and consistency are two key factors in computer-assisted magnetic resonance (MR) image analysis. However, contrast variation from site to site caused by lack of standardization in MR acquisition impedes consistent measurements. In recent yea
rs, image harmonization approaches have been proposed to compensate for contrast variation in MR images. Current harmonization approaches either require cross-site traveling subjects for supervised training or heavily rely on site-specific harmonization models to encourage harmonization accuracy. These requirements potentially limit the application of current harmonization methods in large-scale multi-site studies. In this work, we propose an unsupervised MR harmonization framework, CALAMITI (Contrast Anatomy Learning and Analysis for MR Intensity Translation and Integration), based on information bottleneck theory. CALAMITI learns a disentangled latent space using a unified structure for multi-site harmonization without the need for traveling subjects. Our model is also able to adapt itself to harmonize MR images from a new site with fine tuning solely on images from the new site. Both qualitative and quantitative results show that the proposed method achieves superior performance compared with other unsupervised harmonization approaches.
Deep learning approaches to the segmentation of magnetic resonance images have shown significant promise in automating the quantitative analysis of brain images. However, a continuing challenge has been its sensitivity to the variability of acquisiti
on protocols. Attempting to segment images that have different contrast properties from those within the training data generally leads to significantly reduced performance. Furthermore, heterogeneous data sets cannot be easily evaluated because the quantitative variation due to acquisition differences often dwarfs the variation due to the biological differences that one seeks to measure. In this work, we describe an approach using alternating segmentation and synthesis steps that adapts the contrast properties of the training data to the input image. This allows input images that do not resemble the training data to be more consistently segmented. A notable advantage of this approach is that only a single example of the acquisition protocol is required to adapt to its contrast properties. We demonstrate the efficacy of our approaching using brain images from a set of human subjects scanned with two different T1-weighted volumetric protocols.
Structural magnetic resonance imaging (MRI) can be used to detect lesions in the brains of multiple sclerosis (MS) patients. The formation of these lesions is a complex process involving inflammation, tissue damage, and tissue repair, all of which ar
e visible on MRI. Here we characterize the lesion formation process on longitudinal, multi-sequence structural MRI from 34 MS patients and relate the longitudinal changes we observe within lesions to therapeutic interventions. In this article, we first outline a pipeline to extract voxel level, multi-sequence longitudinal profiles from four MRI sequences within lesion tissue. We then propose two models to relate clinical covariates to the longitudinal profiles. The first model is a principal component analysis (PCA) regression model, which collapses the information from all four profiles into a scalar value. We find that the score on the first PC identifies areas of slow, long-term intensity changes within the lesion at a voxel level, as validated by two experienced clinicians, a neuroradiologist and a neurologist. On a quality scale of 1 to 4 (4 being the highest) the neuroradiologist gave the score on the first PC a median rating of 4 (95% CI: [4,4]), and the neurologist gave it a median rating of 3 (95% CI: [3,3]). In the PCA regression model, we find that treatment with disease modifying therapies (p-value < 0.01), steroids (p-value < 0.01), and being closer to the boundary of abnormal signal intensity (p-value < 0.01) are associated with a return of a voxel to intensity values closer to that of normal-appearing tissue. The second model is a function-on-scalar regression, which allows for assessment of the individual time points at which the covariates are associated with the profiles. In the function-on-scalar regression both age and distance to the boundary were found to have a statistically significant association with the profiles.
