Explore the vast range of titles available.

Posterior fossa ependymomas (EPN_PF) in children comprise two morphologically identical, but biologically distinct tumor entities. Group-A (EPN_PFA) tumors have a poor prognosis and require intensive therapy. In contrast, group-B tumors (EPN_PFB) exhibit excellent prognosis and the current consensus opinion recommends future clinical trials to test the possibility of treatment de-escalation in these patients. Therefore, distinguishing these two tumor subtypes is critical. EPN_PFA and EPN_PFB can be distinguished based on DNA methylation signatures, but these assays are not routinely available. We have previously shown that a subset of poorly prognostic childhood EPN_PF exhibits global reduction in H3K27me3. Therefore, we set out to determine whether a simple immunohistochemical assay for H3K27me3 could be used to segregate EPN_PFA from EPN_PFB tumors. We assembled a cohort of 230 childhood ependymomas and H3K27me3 immunohistochemistry was assessed as positive or negative in a blinded manner. H3K27me3 staining results were compared with DNA methylation-based subgroup information available in 112 samples [EPN_PFA ( n  = 72) and EPN_PFB tumors ( n  = 40)]. H3K27me3 staining was globally reduced in EPN_PFA tumors and immunohistochemistry showed 99% sensitivity and 100% specificity in segregating EPN_PFA from EPN_PFB tumors. Moreover, H3K27me3 immunostaining was sufficient to delineate patients with worse prognosis in two independent, non-overlapping cohorts ( n  = 133 and n  = 97). In conclusion, immunohistochemical evaluation of H3K27me3 global reduction is an economic, easily available and readily adaptable method for defining high-risk EPN_PFA from low-risk posterior fossa EPN_PFB tumors to inform prognosis and to enable the design of future clinical trials.

Posterior fossa ependymoma comprise three distinct molecular variants, termed PF-EPN-A (PFA), PF-EPN-B (PFB), and PF-EPN-SE (subependymoma). Clinically, they are very disparate and PFB tumors are currently being considered for a trial of radiation avoidance. However, to move forward, unraveling the heterogeneity within PFB would be highly desirable. To discern the molecular heterogeneity within PFB, we performed an integrated analysis consisting of DNA methylation profiling, copy-number profiling, gene expression profiling, and clinical correlation across a cohort of 212 primary posterior fossa PFB tumors. Unsupervised spectral clustering and t-SNE analysis of genome-wide methylation data revealed five distinct subtypes of PFB tumors, termed PFB1-5, with distinct demographics, copy-number alterations, and gene expression profiles. All PFB subtypes were distinct from PFA and posterior fossa subependymomas. Of the five subtypes, PFB4 and PFB5 are more discrete, consisting of younger and older patients, respectively, with a strong female-gender enrichment in PFB5 (age: p  = 0.011, gender: p  = 0.04). Broad copy-number aberrations were common; however, many events such as chromosome 2 loss, 5 gain, and 17 loss were enriched in specific subtypes and 1q gain was enriched in PFB1. Late relapses were common across all five subtypes, but deaths were uncommon and present in only two subtypes (PFB1 and PFB3). Unlike the case in PFA ependymoma, 1q gain was not a robust marker of poor progression-free survival; however, chromosome 13q loss may represent a novel marker for risk stratification across the spectrum of PFB subtypes. Similar to PFA ependymoma, there exists a significant intertumoral heterogeneity within PFB, with distinct molecular subtypes identified. Even when accounting for this heterogeneity, extent of resection remains the strongest predictor of poor outcome. However, this biological heterogeneity must be accounted for in future preclinical modeling and personalized therapies.

Tumors of the pineal region comprise several different entities with distinct clinical and histopathological features. Whereas some entities predominantly affect adults, pineoblastoma (PB) constitutes a highly aggressive malignancy of childhood with a poor outcome. PBs mainly arise sporadically, but may also occur in the context of cancer predisposition syndromes including DICER1 and RB1 germline mutation. With this study, we investigate clinico-pathological subgroups of pineal tumors and further characterize their biological features. We performed genome-wide DNA methylation analysis in 195 tumors of the pineal region and 20 normal pineal gland controls. Copy-number profiles were obtained from DNA methylation data; gene panel sequencing was added for 93 tumors and analysis was further complemented by miRNA sequencing for 22 tumor samples. Unsupervised clustering based on DNA methylation profiling separated known subgroups, like pineocytoma, pineal parenchymal tumor of intermediate differentiation, papillary tumor of the pineal region and PB, and further distinct subtypes within these groups, including three subtypes within the core PB subgroup. The novel molecular subgroup Pin-RB includes cases of trilateral retinoblastoma as well as sporadic pineal tumors with RB1 alterations, and displays similarities with retinoblastoma. Distinct clinical associations discriminate the second novel molecular subgroup PB-MYC from other PB cases. Alterations within the miRNA processing pathway (affecting DROSHA , DGCR8 or DICER1 ) are found in about two thirds of cases in the three core PB subtypes. Methylation profiling revealed biologically distinct groups of pineal tumors with specific clinical and molecular features. Our findings provide a foundation for further clinical as well as molecular and functional characterization of PB and other pineal tumors, including the role of miRNA processing defects in oncogenesis.

Multimodal learning often outperforms its unimodal counterparts by exploiting unimodal contributions and cross-modal interactions. However, focusing only on integrating multimodal features into a unified comprehensive representation overlooks the unimodal characteristics. In real data, the contributions of modalities can vary from instance to instance, and they often reinforce or conflict with each other. In this study, we introduce a novel MultiModal loss paradigm for multimodal learning, which subgroups instances according to their unimodal contributions. MultiModal loss can prevent inefficient learning caused by overfitting and efficiently optimize multimodal models. On synthetic data, MultiModal loss demonstrates improved classification performance by subgrouping difficult instances within certain modalities. On four real multimodal datasets, our loss is empirically shown to improve the performance of recent models. Ablation studies verify the effectiveness of our loss. Additionally, we show that our loss generates a reliable prediction score for each modality, which is essential for subgrouping. Our MultiModal loss is a novel loss function to subgroup instances according to the contribution of modalities in multimodal learning and is applicable to a variety of multimodal models with unimodal decisions. Our code is available at https://github.com/DMCB-GIST/MultiModalLoss

Many animal groups exhibit signatures of persistent internal modular structure, whereby individuals consistently interact with certain groupmates more than others. In such groups, information relevant to a collective decision may spread unevenly through the group, but how this impacts the quality of the resulting decision is not well understood. Here, we explicitly model modularity within animal groups and examine how it affects the amount of information represented in collective decisions, as well as the accuracy of those decisions. We find that modular structure necessarily causes a loss of information, effectively silencing the input from a fraction of the group. However, the effect of this information loss on collective accuracy depends on the informational environment in which the decision is made. In simple environments, the information loss is detrimental to collective accuracy. By contrast, in complex environments, modularity tends to improve accuracy. This is because small group sizes typically maximize collective accuracy in such environments, and modular structure allows a large group to behave like a smaller group (in terms of its decision-making). These results suggest that in naturalistic environments containing correlated information, large animal groups may be able to exploit modular structure to improve decision accuracy while retaining other benefits of large group size. This article is part of the theme issue ‘Liquid brains, solid brains: How distributed cognitive architectures process information’.

Abstract The diagnosis of medulloblastoma incorporates the histologic and molecular subclassification of clinical medulloblastoma samples into wingless (WNT)-activated, sonic hedgehog (SHH)-activated, group 3 and group 4 subgroups. Accurate medulloblastoma subclassification has important prognostic and treatment implications. Immunohistochemistry (IHC)-based and nanoString-based subgrouping methodologies have been independently described as options for medulloblastoma subgrouping, however have not previously been directly compared. We describe our experience with nanoString-based subgrouping in a clinical setting and compare this with our IHC-based results. Study materials included FFPE tissue from 160 medulloblastomas. Clinical data and tumor histology were reviewed. Immunohistochemical-based subgrouping using β-catenin, filamin A and p53 antibodies and nanoString-based gene expression profiling were performed. The sensitivity and specificity of IHC-based subgrouping of WNT and SHH-activated medulloblastomas was 91.5% and 99.54%, respectively. Filamin A immunopositivity highly correlated with SHH/WNT-activated subgroups (sensitivity 100%, specificity 92.7%, p < 0.001). Nuclear β-catenin immunopositivity had a sensitivity of 76.2% and specificity of 99.23% for detection of WNT-activated tumors. Approximately 23.8% of WNT cases would have been missed using an IHC-based subgrouping method alone. nanoString could confidently predict medulloblastoma subgroup in 93% of cases and could distinguish group 3/4 subgroups in 96.3% of cases. nanoString-based subgrouping allows for a more prognostically useful classification of clinical medulloblastoma samples.

Autism spectrum disorder (ASD) presents significant challenges in diagnosis and intervention due to its diverse clinical manifestations and underlying biological complexity. This study explored machine learning approaches to enhance ASD screening accuracy and identify meaningful subtypes using clinical assessments from AGRE database integrated with molecular data from GSE15402. Analysis of ADI-R scores from a large cohort of 2794 individuals demonstrated that deep learning models could achieve exceptional screening accuracy of 95.23% (CI 94.32–95.99%). Notably, comparable performance was maintained using a streamlined set of just 27 ADI-R sub-items, suggesting potential for more efficient diagnostic tools. Clustering analyses revealed three distinct subgroups identifiable through both clinical symptoms and gene expression patterns. When ASD were grouped based on clinical features, stronger associations emerged between symptoms and underlying molecular profiles compared to grouping based on gene expression alone. These findings suggest that starting with detailed clinical observations may be more effective for identifying biologically meaningful ASD subtypes than beginning with molecular data. This integrated approach combining clinical and molecular data through machine learning offers promising directions for developing more precise screening methods and personalized intervention strategies for individuals with ASD.

Amyotrophic lateral sclerosis (ALS) is a multisystem disease with marked pathophysiological and clinical heterogeneity, making individual and objective characterization of the degree of disease progression and disease‐related subtrajectories challenging. Here, we use in vivo multimodal neuroimaging data and computational models to generate personalized indices of ALS progression and subtrajectory. We used structural and diffusion weighted imaging of 691 participants (58% ALS) from two independent ALS data sets (North American and Utrecht cohorts) to extract regional values of grey matter (DM) density and white matter (WM) microstructural integrity. Contrastive trajectory inference (cTI) allowed us to identify and separate latent, multivariate patterns in neuroimaging features highlighting ALS‐associated pathological processes, which were used to generate subject‐specific indices of disease progression and subtrajectory. Disease subtrajectories were based on distinct patterns of alterations in neuroimaging data considering subjects at different disease progression levels. The neuroimaging‐based, personalized index of disease progression is indicative of clinical symptom severity (North American: p < 0.01 and Utrecht: p < 0.01) and displays alignment with the King's College staging system (p = 0.001 and p = 0.002). Three ALS subtrajectories were identified that displayed distinct alterations in the motor, limbic system, and widespread cortical and subcortical changes that also differed in clinical symptom manifestation. Our analysis has shown that neuroimaging data encodes subject‐specific, disease‐related patterns that can be leveraged to obtain an in vivo proxy of disease progression and putative disease subtype. Amyotrophic lateral sclerosis (ALS) is recognized as a multifaceted disease with varying symptom profiles calling for a more comprehensive understanding of ALS. We demonstrate the feasibility of extracting personalized disease progression markers as well as distinct disease subtrajectories from in vivo multimodal neuroimaging data in a heterogeneous ALS population.

Abstract Current approaches to automated sleep staging from the electroencephalogram (EEG) rely on constructing a large labeled training and test corpora by aggregating data from different individuals. However, many of the subjects in the training set may exhibit changes in the EEG that are very different from the subjects in the test set. Training an algorithm on such data without accounting for this diversity can cause underperformance. Moreover, test data may have unexpected sensor misplacement or different instrument noise and spectral responses. This work proposes a novel method to learn relevant individuals based on their similarities effectively. The proposed method embeds all training patients into a shared and robust feature space. Individuals who share strong statistical relationships and are similar based on their EEG signals are clustered in this feature space before being passed to a deep learning framework for classification. Using 994 patient EEGs from the 2018 Physionet Challenge (≈6,561 h of recording), we demonstrate that the clustering approach significantly boosts performance compared to state-of-the-art deep learning approaches. The proposed method improves, on average, a precision score from 0.72 to 0.81, a sensitivity score from 0.74 to 0.82, and a Cohen’s Kappa coefficient from 0.64 to 0.75 under 10-fold cross-validation.

Background There are large variations in symptoms and prognostic factors among patients sharing the same musculoskeletal (MSK) diagnosis, making traditional diagnostic labelling not very helpful in informing treatment or prognosis. Recently, we identified five MSK phenotypes across common MSK pain locations through latent class analysis (LCA). The aim of this study was to explore the one-year recovery trajectories for pain and functional limitations in the phenotypes and describe these in relation to the course of traditional diagnostic MSK groups. Methods We conducted a longitudinal observational study of 147 patients with neck, back, shoulder or complex pain in primary health care physiotherapy. Data on pain intensity and function were collected at baseline (week 0) and 1, 2, 3, 4, 6, 8, 12, 26 and 52 weeks of follow up using web-based questionnaires and mobile text messages. Recovery trajectories were described separately for the traditional diagnostic MSK groups based on pain location and the same patients categorized in phenotype groups based on prognostic factors shared among the MSK diagnostic groups. Results There was a general improvement in function throughout the year of follow-up for the MSK groups, while there was a more modest decrease for pain intensity. The MSK diagnoses were dispersed across all five phenotypes, where the phenotypes showed clearly different trajectories for recovery and course of symptoms over 12 months follow-up. This variation was not captured by the single trajectory for site specific MSK diagnoses. Conclusion Prognostic subgrouping revealed more diverse patterns in pain and function recovery over 1 year than observed in the same patients classified by traditional diagnostic groups and may better reflect the diversity in recovery of common MSK disorders.