Weakly supervised perivascular spaces segmentation with salient guidance of Frangi filter.

PURPOSE: To develop a weakly supervised 3D perivascular spaces (PVS) segmentation model that combines the filter-based image processing algorithm and the convolutional neural network. METHODS: We present a weakly supervised learning method for PVS segmentation by combing a rule-based image processing approach Frangi filter with a canonical deep learning algorithm Unet using conditional random field theory. The weighted cross entropy loss function and the training patch selection were implemented for the optimization and to alleviate the class imbalance issue. The performance of the model was evaluated on the Human Connectome Project data. RESULTS: The proposed method increases the true positive rate compared to the rule-based method and reduces the false positive rate by 36% in the weakly supervised training experiment and 39.4% in the supervised training experiment compared to Unet, which results in superior overall performance. In addition, by training the model on manually quality controlled and annotated data which includes the subjects with the presence of white matter hyperintensities, the proposed method differentiates between PVS and white matter hyperintensities, which reduces the false positive rate by 78.5% compared to weakly supervised trained model. CONCLUSIONS: Combing the filter-based image processing algorithm and the convolutional neural network algorithm could improve the model's segmentation accuracy, while reducing the training dependence on the large scale annotated PVS mask data by the trained physician. Compared to the filter-based image processing algorithm, the data driven PVS segmentation model using quality-controlled data as the training target could differentiate the white matter hyperintensity from PVS resulting low false positive rate.

Life After Mild Traumatic Brain Injury: Widespread Structural Brain Changes Associated With Psychological Distress Revealed With Multimodal Magnetic Resonance Imaging.

Background: Traumatic brain injury (TBI) can alter brain structure and lead to onset of persistent neuropsychological symptoms. This study investigates the relationship between brain injury and psychological distress after mild TBI using multimodal magnetic resonance imaging. Methods: A total of 89 patients with mild TBI from the TRACK-TBI (Transforming Research and Clinical Knowledge in Traumatic Brain Injury) pilot study were included. Subscales of the Brief Symptoms Inventory 18 for depression, anxiety, and somatization were used as outcome measures of psychological distress approximately 6 months after the traumatic event. Glasgow Coma Scale scores were used to evaluate recovery. Magnetic resonance imaging data were acquired within 2 weeks after injury. Perivascular spaces (PVSs) were segmented using an enhanced PVS segmentation method, and the volume fraction was calculated for the whole brain and white matter regions. Cortical thickness and gray matter structures volumes were calculated in FreeSurfer; diffusion imaging indices and multifiber tracts were extracted using the Quantitative Imaging Toolkit. The analysis was performed considering age, sex, intracranial volume, educational attainment, and improvement level upon discharge as covariates. Results: PVS fractions in the posterior cingulate, fusiform, and postcentral areas were found to be associated with somatization symptoms. Depression, anxiety, and somatization symptoms were associated with the cortical thickness of the frontal-opercularis and occipital pole, putamen and amygdala volumes, and corticospinal tract and superior thalamic radiation. Analyses were also performed on the two hemispheres separately to explore lateralization. Conclusions: This study shows how PVS, cortical, and microstructural changes can predict the onset of depression, anxiety, and somatization symptoms in patients with mild TBI.

Perivascular Space Imaging at Ultrahigh Field MR Imaging.

The recent Food and Drug Administration approval of 7 T MR imaging scanners for clinical use has introduced the possibility to study the brain not only in physiologic but also in pathologic conditions at ultrahigh field (UHF). Because UHF MR imaging offers higher signal-to-noise ratio and spatial resolution compared with lower field clinical scanners, the benefits of UHF MR imaging are particularly evident for imaging small anatomic structures, such as the cerebral perivascular spaces (PVS). In this article, the authors describe the application of UHF MR imaging for the investigation of PVS.