A Robust and Accurate Deep-learning-based Method for the Segmentation of Subcortical Brain: Cross-dataset Evaluation of Generalization Performance

Purpose:To analyze subcortical brain volume more reliably, we propose a deep learning segmentation method of subcortical brain based on magnetic resonance imaging (MRI) having high generalization performance, accuracy, and robustness.Methods:First, local images of three-dimensional (3D) bounding boxes were extracted for seven subcortical structures (thalamus, putamen, caudate, pallidum, hippocampus, amygdala, and accumbens) from a whole brain MR image as inputs to the neural network. Second, dilated convolution layers, which input information of variable scope, were introduced to the blocks that make up the neural network. These blocks were connected in parallel to simultaneously process global and local information obtained by the dilated convolution layers. To evaluate generalization performance, different datasets were used for training and testing sessions (cross-dataset evaluation) because subcortical brain segmentation in clinical analysis is assumed to be applied to unknown datasets.Results:The proposed method showed better generalization performance that can obtain stable accuracy for all structures, whereas the state-of-the-art deep learning method obtained extremely low accuracy for some structures. The proposed method performed segmentation for all samples without failing with significantly higher accuracy (P < 0.005) than conventional methods such as 3D U-Net, FreeSurfer, and Functional Magnetic Resonance Imaging of the Brain’s (FMRIB’s) Integrated Registration and Segmentation Tool in the FMRIB Software Library (FSL-FIRST). Moreover, when applying this proposed method to larger datasets, segmentation was robustly performed for all samples without producing segmentation results on the areas that were apparently different from anatomically relevant areas. On the other hand, FSL-FIRST produced segmentation results on the area that were apparently and largely different from the anatomically relevant area for about one-third to one-fourth of the datasets.Conclusion:The cross-dataset evaluation showed that the proposed method is superior to existing methods in terms of generalization performance, accuracy, and robustness.     

Polymicrogyria with calcification in Pallister-Killian syndrome detected by microarray analysis

BackgroundPallister-Killian syndrome (PKS) is a rare disorder caused by the mosaic tetrasomy of chromosome 12p, and is characterized by facial dysmorphism, developmental delay, hypotonia and seizures.ResultsWe report a patient with PKS showing unique polymicrogyria with calcification. He had delayed development and dysmorphic facial features including frontal bossing, hypertelorism, and high arched palate at 6 months of age. Neuroimaging revealed unilateral polymicrogyria with spot calcifications, which predominantly affected the right perisylvian region. Chromosome G-banding showed the karyotype 46,XY, however, array-based comparative genomic hybridization analysis showed mosaic duplication of chromosome 12p, in which CCND2, which encodes cyclin D2 and is a downstream mediator of PI3K-AKT pathway, is located. Supernumerary chromosome of 12p was detected in 58% of buccal mucosa cells by the interphase fluorescence in situ hybridization analysis using chromosome 12 centromere-specific D12Z3 probe. The diagnosis of PKS was made based on distinctive clinical features of our patient and the results of cytogenetic analyses.ConclusionThis report is, to our knowledge, the first case of a patient with PKS who clearly demonstrates polymicrogyria colocalized with calcifications, as shown by CT scans and MRI, and suggests that a patient with PKS could show structural brain anomalies with calcification. We assume that somatic mosaicism of tetrasomy could cause asymmetrical polymicrogyria in our patient, and speculate that increased dosages of CCND2 at chromosome 12p might be involved in the abnormal neuronal migration in PKS.     

Correlation between number of renal cysts and aortic circumferences measured using autopsy material

Although the presence of renal cysts has been reported to be associated with aortic aneurysm or dissection by imaging studies, an autopsy study has not been performed. Therefore, in our institute, recent consecutive adult autopsy cases (n = 108, 64 males and 44 females) were reviewed. The circumferences and atherosclerosis ratios of both thoracic and abdominal aorta were individually measured and graded. The number of renal cysts was scored and graded. Age of subjects along with histories of smoking, hypertension, and diabetes mellitus were confirmed. Multiple linear regression analyses demonstrated that severity of atherosclerosis and the number of renal cysts were correlated with thoracic aortic circumference, while only the number of renal cysts was correlated with abdominal aortic circumference (p < 0.05), which was more predominant in female subjects (p < 0.05). Microscopically, significantly more dilated renal tubules (by Student's t-test, p < 0.05) along with decreased stainability of basement membrane by Periodic acid-Schiff staining and immunostaining of type IV collagen were noted in background renal tissues in cases with numerous renal cysts than in age- and sex-matched controls without renal cysts (n = 10 vs. 10). The present study suggests that a syndrome that affects both aorta and renal tubules may exist.     

Effects of dextran sulfates on the acute infection and growth stages of Toxoplasma gondii

Toxoplasma gondii is one of the most prevalent parasites, causing toxoplasmosis in various warm-blooded animals, including humans. Because of the broad range of hosts susceptible to T. gondii, it had been postulated that a universal component of the host cell surface, such as glycosaminoglycans (GAGs), may act as a receptor for T. gondii infection. Carruthers et al. (Infect Immun 68:4005–4011, 2000) showed that soluble GAGs have also been shown to disrupt parasite binding to human fibroblasts. Therefore, we investigated the inhibitory effect of GAGs and their analogue dextran sulfate (DS) on T. gondii infection. For up to 24 h of incubation after inoculation of T. gondii, the inhibitory effect of GAGs on T. gondii infection and growth inside the host cell was weak. In contrast, DS markedly inhibited T. gondii infection. Moreover, low molecular weight DS particularly slowed the growth of T. gondii inside host cells. DS10 (dextran sulfate MW 10 kDa) was the most effective agent in these in vitro experiments and was therefore tested for its inhibitory effects in animal experiments; infection inhibition by DS10 was confirmed under these in vivo conditions. In this report, we showed that DSs, especially DS10, have the potential of a new type of drug for toxoplasmosis.