A flexible 2.5D medical image segmentation approach with in-slice and cross-slice attention.

Deep learning has become the de facto method for medical image segmentation, with 3D segmentation models excelling in capturing complex 3D structures and 2D models offering high computational efficiency. However, segmenting 2.5D images, characterized by high in-plane resolution but lower through-plane resolution, presents significant challenges. While applying 2D models to individual slices of a 2.5D image is feasible, it fails to capture the spatial relationships between slices. On the other hand, 3D models face challenges such as resolution inconsistencies in 2.5D images, along with computational complexity and susceptibility to overfitting when trained with limited data. In this context, 2.5D models, which capture inter-slice correlations using only 2D neural networks, emerge as a promising solution due to their reduced computational demand and simplicity in implementation. In this paper, we introduce CSA-Net, a flexible 2.5D segmentation model capable of processing 2.5D images with an arbitrary number of slices. CSA-Net features an innovative Cross-Slice Attention (CSA) module that effectively captures 3D spatial information by learning long-range dependencies between the center slice (for segmentation) and its neighboring slices. Moreover, CSA-Net utilizes the self-attention mechanism to learn correlations among pixels within the center slice. We evaluated CSA-Net on three 2.5D segmentation tasks: (1) multi-class brain MR image segmentation, (2) binary prostate MR image segmentation, and (3) multi-class prostate MR image segmentation. CSA-Net outperformed leading 2D, 2.5D, and 3D segmentation methods across all three tasks, achieving average Dice coefficients and HD95 values of 0.897 and 1.40 mm for the brain dataset, 0.921 and 1.06 mm for the prostate dataset, and 0.659 and 2.70 mm for the ProstateX dataset, demonstrating its efficacy and superiority. Our code is publicly available at: https://github.com/mirthAI/CSA-Net.

Long-term safety and efficacy of zilucoplan in patients with generalized myasthenia gravis: interim analysis of the RAISE-XT open-label extension study.

Background: Generalized myasthenia gravis (gMG) is a chronic, unpredictable disease associated with high treatment and disease burdens, with a need for more effective and well-tolerated treatments. Objectives: To evaluate the long-term safety, tolerability, and efficacy of zilucoplan in a mild-to-severe, acetylcholine receptor autoantibody-positive (AChR+) gMG population. Design: Ongoing, multicenter, phase III open-label extension (OLE) study. Methods: Eligible patients had completed a qualifying randomized, placebo-controlled phase II or phase III zilucoplan study and received daily, self-administered subcutaneous 0.3 mg/kg zilucoplan. The primary endpoint was incidence of treatment-emergent adverse events (TEAEs). Secondary efficacy endpoints included change from baseline in Myasthenia Gravis Activities of Daily Living (MG-ADL) score. Results: In total, 200 patients enrolled. At the cut-off date (8 September 2022), median (range) exposure to zilucoplan in RAISE-XT was 1.2 (0.11-4.45) years. Mean age at OLE baseline was 53.3 years. A total of 188 (94%) patients experienced a TEAE, with the most common being MG worsening (n = 52, 26%) and COVID-19 (n = 49, 25%). In patients who received zilucoplan 0.3 mg/kg in the parent study, further improvements in MG-ADL score continued through to Week 24 (least squares mean change [95% confidence interval] from double-blind baseline -6.06 [-7.09, -5.03]) and were sustained through to Week 60 (-6.04 [-7.21, -4.87]). In patients who switched from placebo in the parent study, rapid improvements in MG-ADL score were observed at the first week after switching to zilucoplan; further improvements were observed at Week 24, 12 weeks after switching (-6.46 [-8.19, -4.72]), and were sustained through to Week 60 (-6.51 [-8.37, -4.65]). Consistent results were observed in other efficacy endpoints. Conclusion: Zilucoplan demonstrated a favorable long-term safety profile, good tolerability, and sustained efficacy through to Week 60 with consistent benefits in a broad AChR+ gMG population. Additional long-term data will be available in future analyses. Trial registration: ClinicalTrials.gov identifier: NCT04225871 (https://clinicaltrials.gov/ct2/show/NCT04225871).

MicroRNAs as biomarkers of brain injury in neonatal encephalopathy: an observational cohort study.

Neonatal Encephalopathy (NE) is a major cause of lifelong disability and neurological complications in affected infants. Identifying novel diagnostic biomarkers in this population may assist in predicting MRI injury and differentiate neonates with NE from those with low-cord pH or healthy neonates and may help clinicians make real-time decisions. To compare the microRNA (miRNA) profiles between neonates with NE, healthy controls, and neonates with low cord pH. Moreover, miRNA concentrations were compared to brain injury severity in neonates with NE. This is a retrospective analysis of miRNA profiles from select samples in the biorepository and data registry at the University of Florida Health Gainesville. The Firefly miRNA assay was used to screen a total of 65 neurological miRNA targets in neonates with NE (n = 36), low cord pH (n = 18) and healthy controls (n = 37). Multivariate statistical techniques, including principal component analysis and orthogonal partial least squares discriminant analysis, and miRNA Enrichment Analysis and Annotation were used to identify miRNA markers and their pathobiological relevance. A set of 10 highly influential miRNAs were identified, which were significantly upregulated in the NE group compared to healthy controls. Of these, miR-323a-3p and mir-30e-5p displayed the highest fold change in expression levels. Moreover, miR-34c-5p, miR-491-5p, and miR-346 were significantly higher in the NE group compared to the low cord pH group. Furthermore, several miRNAs were identified that can differentiate between no/mild and moderate/severe injury in the NE group as measured by MRI. MiRNAs represent promising diagnostic and prognostic tools for improving the management of NE.

Improvement of fatigue in generalised myasthenia gravis with zilucoplan.

BACKGROUND: Fatigue is a debilitating symptom of myasthenia gravis (MG). The impact of fatigue on MG can be assessed by Quality of Life in Neurological Disorders (Neuro-QoL) Short Form Fatigue scale. Transformation of raw Neuro-QoL fatigue scores to T-scores is a known approach for facilitating clinical interpretation of clinically meaningful and fatigue severity thresholds. METHODS: In the Phase 3, double-blind, placebo-controlled RAISE study (NCT04115293), adults with acetylcholine receptor autoantibody-positive generalised MG (MG Foundation of America Disease Class II-IV) were randomised 1:1 to daily subcutaneous zilucoplan 0.3 mg/kg or placebo for 12 weeks. Patients completing RAISE could opt to receive zilucoplan 0.3 mg/kg in an ongoing, open-label extension study, RAISE-XT (NCT04225871). In this post-hoc analysis, we evaluated the long-term effect of zilucoplan on fatigue in RAISE patients who entered RAISE-XT. We report change in Neuro-QoL Short Form Fatigue T-scores and fatigue severity levels from RAISE baseline to Week 60. RESULTS: Mean Neuro-QoL Short Form Fatigue T-scores improved from baseline to Week 12 in the zilucoplan group (n = 86) with a clinically meaningful difference versus placebo (n = 88; least squares mean difference: - 3.61 (nominal p-value = 0.0060]), and these improvements continued further to Week 60. At Week 12, more patients on zilucoplan (n = 34, 47.2%) experienced improvements in ≥ 1 fatigue severity level from baseline versus placebo (n = 23, 28.4%; p = 0.017). At Week 60, most (n = 55, 65.5%) patients had mild fatigue or none. CONCLUSION: Treatment with zilucoplan demonstrated statistical and clinically meaningful improvements in fatigue scores and severity versus placebo during RAISE, which were sustained to Week 60 in RAISE-XT.

NRBC concentrations over time in neonates with moderate to severe neonatal encephalopathy with and without sentinel events.

OBJECTIVE: To study the serum concentrations of nucleated red blood cells (NRBC) over time in neonates with moderate to severe neonatal encephalopathy (NE). STUDY DESIGN: A retrospective cohort study with subjects subdivided into three groups: definite sentinel events (n = 52), probable sentinel events (n = 20) and no history of sentinel events (n = 63). Peak absolute NRBC and NRBC/100 WBC were compared between groups and with MRI Injury score, cord and admission pH/base deficit. RESULTS: Absolute NRBC peaked at 24.05 h after birth (CI: 15.30-32.79), 17.56 h after birth (CI: 7.35-27.77), and 39.81 h after birth (CI: 28.73-50.89) in each respective group. The peak in absolute NRBC correlated with the severity of injury in the grey matter in group 2 and white matter in groups 1 and 2. Higher peak absolute NRBC value correlated to a lower admission ABG pH. CONCLUSION: NRBC peak at 24 h after birth in neonates with sentinel events.

MiR-126 and miR-146a as Melatonin-Responsive Biomarkers for Neonatal Brain Ischemia.

Despite advances in obstetric and neonatal care, challenges remain in early identification of neonates with encephalopathy due to hypoxia-ischemia who are undergoing therapeutic hypothermia. Therefore, there is a deep search for biomarkers that can identify brain injury. The aims of this study were to investigate the serum and brain expressions of two potential biomarkers, miR-126/miR-146a, in a preclinical model of hypoxia-ischemia (HI)-induced brain injury, and to explore their modulation during melatonin treatment. Seven-day-old rats were subjected to permanent ligation of the right carotid artery followed by 2.5 h hypoxia (HI). Melatonin (15 mg/kg) was administered 5 min after HI. Serum and brain samples were collected 1, 6 and 24 h after HI. Results show that HI caused a significant increase in the circulating levels of both miR-126 and miR-146a during the early phase of ischemic brain damage development (i.e. 1 h), with a parallel and opposite pattern in the ischemic cerebral cortex. These effects are not observed 24 h later. Treatment with melatonin restored the HI-induced effects on miR-126/miR-146a expressions, both in the cerebral cortex and in serum. We conclude that miR-126/miR-146a are promising biomarkers of HI injury and demonstrate an associated change in concentration following melatonin treatment.

Therapeutic targeting of ALS pathways: Refocusing an incomplete picture.

Numerous potential amyotrophic lateral sclerosis (ALS)-relevant pathways have been hypothesized and studied preclinically, with subsequent translation to clinical trial. However, few successes have been observed with only modest effects. Along with an improved but incomplete understanding of ALS as a neurodegenerative disease is the evolution of more sophisticated and diverse in vitro and in vivo preclinical modeling platforms, as well as clinical trial designs. We highlight proposed pathological pathways that have been major therapeutic targets for investigational compounds. It is likely that the failures of so many of these therapeutic compounds may not have occurred because of lack of efficacy but rather because of a lack of preclinical modeling that would help define an appropriate disease pathway, as well as a failure to establish target engagement. These challenges are compounded by shortcomings in clinical trial design, including lack of biomarkers that could predict clinical success and studies that are underpowered. Although research investments have provided abundant insights into new ALS-relevant pathways, most have not yet been developed more fully to result in clinical study. In this review, we detail some of the important, well-established pathways, the therapeutics targeting them, and the subsequent clinical design. With an understanding of some of the shortcomings in translational efforts over the last three decades of ALS investigation, we propose that scientists and clinicians may choose to revisit some of these therapeutic pathways reviewed here with an eye toward improving preclinical modeling, biomarker development, and the investment in more sophisticated clinical trial designs.

Genetic analysis and natural history of Charcot-Marie-Tooth disease CMTX1 due to GJB1 variants.

Charcot-Marie-Tooth disease (CMT) due to GJB1 variants (CMTX1) is the second most common form of CMT. It is an X-linked disorder characterized by progressive sensory and motor neuropathy with males affected more severely than females. Many reported GJB1 variants remain classified as variants of uncertain significance (VUS). In this large, international, multicentre study we prospectively collected demographic, clinical and genetic data on patients with CMT associated with GJB1 variants. Pathogenicity for each variant was defined using adapted American College of Medical Genetics criteria. Baseline and longitudinal analyses were conducted to study genotype-phenotype correlations, to calculate longitudinal change using the CMT Examination Score (CMTES), to compare males versus females, and pathogenic/likely pathogenic (P/LP) variants versus VUS. We present 387 patients from 295 families harbouring 154 variants in GJB1. Of these, 319 patients (82.4%) were deemed to have P/LP variants, 65 had VUS (16.8%) and three benign variants (0.8%; excluded from analysis); an increased proportion of patients with P/LP variants compared with using ClinVar's classification (74.6%). Male patients (166/319, 52.0%, P/LP only) were more severely affected at baseline. Baseline measures in patients with P/LP variants and VUS showed no significant differences, and regression analysis suggested the disease groups were near identical at baseline. Genotype-phenotype analysis suggested c.-17G>A produces the most severe phenotype of the five most common variants, and missense variants in the intracellular domain are less severe than other domains. Progression of disease was seen with increasing CMTES over time up to 8 years follow-up. Standard response mean (SRM), a measure of outcome responsiveness, peaked at 3 years with moderate responsiveness [change in CMTES (ΔCMTES) = 1.3 ± 2.6, P = 0.00016, SRM = 0.50]. Males and females progressed similarly up to 8 years, but baseline regression analysis suggested that over a longer period, females progress more slowly. Progression was most pronounced for mild phenotypes (CMTES = 0-7; 3-year ΔCMTES = 2.3 ± 2.5, P = 0.001, SRM = 0.90). Enhanced variant interpretation has yielded an increased proportion of GJB1 variants classified as P/LP and will aid future variant interpretation in this gene. Baseline and longitudinal analysis of this large cohort of CMTX1 patients describes the natural history of the disease including the rate of progression; CMTES showed moderate responsiveness for the whole group at 3 years and higher responsiveness for the mild group at 3, 4 and 5 years. These results have implications for patient selection for upcoming clinical trials.

Safety and efficacy of zilucoplan in patients with generalised myasthenia gravis (RAISE): a randomised, double-blind, placebo-controlled, phase 3 study.

BACKGROUND: Generalised myasthenia gravis is a chronic, unpredictable, and debilitating rare disease, often accompanied by high treatment burden and with an unmet need for more efficacious and well tolerated treatments. Zilucoplan is a subcutaneous, self-administered macrocyclic peptide complement C5 inhibitor. We aimed to assess safety, efficacy, and tolerability of zilucoplan in patients with acetylcholine receptor autoantibody (AChR)-positive generalised myasthenia gravis. METHODS: RAISE was a randomised, double-blind, placebo-controlled, phase 3 trial that was done at 75 sites in Europe, Japan, and North America. We enrolled patients (aged 18-74 years) with AChR-positive generalised myasthenia gravis (Myasthenia Gravis Foundation of America disease class II-IV), a myasthenia gravis activities of daily living (MG-ADL) score of least 6, and a quantitative myasthenia gravis score of at least 12. Participants were randomly assigned (1:1) to receive subcutaneous zilucoplan 0·3 mg/kg once daily by self-injection, or matched placebo, for 12 weeks. The primary efficacy endpoint was change from baseline to week 12 in MG-ADL score in the modified intention-to-treat population (all randomly assigned patients who received at least one dose of study drug and had at least one post-dosing MG-ADL score). Safety was mainly assessed by the incidence of treatment-emergent adverse events (TEAEs) in all patients who had received at least one dose of zilucoplan or placebo. This trial is registered at ClinicalTrials.gov, NCT04115293. An open-label extension study is ongoing (NCT04225871). FINDINGS: Between Sept 17, 2019, and Sept 10, 2021, 239 patients were screened for the study, of whom 174 (73%) were eligible. 86 (49%) patients were randomly assigned to zilucoplan 0·3 mg/kg and 88 (51%) were assigned to placebo. Patients assigned to zilucoplan showed a greater reduction in MG-ADL score from baseline to week 12, compared with those assigned to placebo (least squares mean change -4·39 [95% CI -5·28 to -3·50] vs -2·30 [-3·17 to -1·43]; least squares mean difference -2·09 [-3·24 to -0·95]; p=0·0004). TEAEs occurred in 66 (77%) patients in the zilucoplan group and in 62 (70%) patients in the placebo group. The most common TEAE was injection-site bruising (n=14 [16%] in the zilucoplan group and n=8 [9%] in the placebo group). Incidences of serious TEAEs and serious infections were similar in both groups. One patient died in each group; neither death (COVID-19 [zilucoplan] and cerebral haemorrhage [placebo]) was considered related to the study drug. INTERPRETATION: Zilucoplan treatment showed rapid and clinically meaningful improvements in myasthenia gravis-specific efficacy outcomes, had a favourable safety profile, and was well tolerated, with no major safety findings. Zilucoplan is a new potential treatment option for a broad population of patients with AChR-positive generalised myasthenia gravis. The long-term safety and efficacy of zilucoplan is being assessed in an ongoing open-label extension study. FUNDING: UCB Pharma.