Search

Background: Chronic liver disease diagnoses depend on liver biopsy histopathological assessment. However, due to the limitations associated with biopsy, there is growing interest in the use of quantitative digital pathology to support pathologists. We evaluated the performance of computational algorithms in the assessment of hepatic inflammation in an autoimmune hepatitis in which inflammation is a major component. Methods: Whole-slide digital image analysis was used to quantitatively characterize the area of tissue covered by inflammation [Inflammation Density (ID)] and number of inflammatory foci per unit area [Focal Density (FD)] on tissue obtained from 50 patients with autoimmune hepatitis undergoing routine liver biopsy. Correlations between digital pathology outputs and traditional categorical histology scores, biochemical, and imaging markers were assessed. The ability of ID and FD to stratify between low-moderate (both portal and lobular inflammation ≤1) and moderate-severe disease activity was estimated using the area under the receiver operating characteristic curve (AUC). Results: ID and FD scores increased significantly and linearly with both portal and lobular inflammation grading. Both ID and FD correlated moderately-to-strongly and significantly with histology (portal and lobular inflammation; 0.36≤R≤0.69) and biochemical markers (ALT, AST, GGT, IgG, and gamma globulins; 0.43≤R≤0.57). ID (AUC: 0.85) and FD (AUC: 0.79) had good performance for stratifying between low-moderate and moderate-severe inflammation. Conclusion: Quantitative assessment of liver biopsy using quantitative digital pathology metrics correlates well with traditional pathology scores and key biochemical markers. Whole-slide quantification of disease can support stratification and identification of patients with more advanced inflammatory disease activity.

Background: Preterm infants are at high risk of diffuse white matter injury and adverse neurodevelopmental outcome. The multiple hit hypothesis suggests that the risk of white matter injury increases with cumulative exposure to multiple perinatal risk factors. Our aim was to test this hypothesis in a large cohort of preterm infants using diffusion weighted magnetic resonance imaging (dMRI). Methods: We studied 491 infants (52% male) without focal destructive brain lesions born at

Diffusion-weighted imaging (DWI) is becoming an increasingly important tool for studying brain development. DWI analyses relying on manually-drawn regions of interest and tractography using manually-placed waypoints are considered to provide the most accurate characterisation of the underlying brain structure. However, these methods are labour-intensive and become impractical for studies with large cohorts and numerous white matter (WM) tracts. Tract-specific analysis (TSA) is an alternative WM analysis method applicable to large-scale studies that offers potential benefits. TSA produces a skeleton representation of WM tracts and projects the group's diffusion data onto the skeleton for statistical analysis. In this work we evaluate the performance of TSA in analysing preterm infant data against results obtained from native space tractography and tract-based spatial statistics. We evaluate TSA's registration accuracy of WM tracts and assess the agreement between native space data and template space data projected onto WM skeletons, in 12 tracts across 48 preterm neonates. We show that TSA registration provides better WM tract alignment than a previous protocol optimised for neonatal spatial normalisation, and that TSA projects FA values that match well with values derived from native space tractography. We apply TSA for the first time to a preterm neonatal population to study the effects of age at scan on WM tracts around term equivalent age. We demonstrate the effects of age at scan on DTI metrics in commissural, projection and association fibres. We demonstrate the potential of TSA for WM analysis and its suitability for infant studies involving multiple tracts.

Accurately delineating the brain on magnetic resonance (MR) images of the head is a prerequisite for many neuroimaging methods. Most existing methods exhibit disadvantages in that they are laborious, yield inconsistent results, and/or require training data to closely match the data to be processed. Here, we present pincram, an automatic, versatile method for accurately labelling the adult brain on T1-weighted 3D MR head images. The method uses an iterative refinement approach to propagate labels from multiple atlases to a given target image using image registration. At each refinement level, a consensus label is generated. At the subsequent level, the search for the brain boundary is constrained to the neighbourhood of the boundary of this consensus label. The method achieves high accuracy (Jaccard coefficient > 0.95 on typical data, corresponding to a Dice similarity coefficient of > 0.97) and performs better than many state-of-the-art methods as evidenced by independent evaluation on the Segmentation Validation Engine. Via a novel self-monitoring feature, the program generates the "success index," a scalar metadatum indicative of the accuracy of the output label. Pincram is available as open source software.

Brain images contain information suitable for automatically sorting subjects into categories such as healthy controls and patients. We sought to identify morphometric criteria for distinguishing controls (n = 28) from patients with unilateral temporal lobe epilepsy (TLE), 60 with and 20 without hippocampal atrophy (TLE-HA and TLE-N, respectively), and for determining the presumed side of seizure onset. The framework employs multi-atlas segmentation to estimate the volumes of 83 brain structures. A kernel-based separability criterion was then used to identify structures whose volumes discriminate between the groups. Next, we applied support vector machines (SVM) to the selected set for classification on the basis of volumes. We also computed pairwise similarities between all subjects and used spectral analysis to convert these into per-subject features. SVM was again applied to these feature data. After training on a subgroup, all TLE-HA patients were correctly distinguished from controls, achieving an accuracy of 96 ± 2% in both classification schemes. For TLE-N patients, the accuracy was 86 ± 2% based on structural volumes and 91 ± 3% using spectral analysis. Structures discriminating between patients and controls were mainly localized ipsilaterally to the presumed seizure focus. For the TLE-HA group, they were mainly in the temporal lobe; for the TLE-N group they included orbitofrontal regions, as well as the ipsilateral substantia nigra. Correct lateralization of the presumed seizure onset zone was achieved using hippocampi and parahippocampal gyri in all TLE-HA patients using either classification scheme; in the TLE-N patients, lateralization was accurate based on structural volumes in 86 ± 4%, and in 94 ± 4% with the spectral analysis approach. Unilateral TLE has imaging features that can be identified automatically, even when they are invisible to human experts. Such morphometric image features may serve as classification and lateralization criteria. The technique also detects unsuspected distinguishing features like the substantia nigra, warranting further study.