Large language models approach expert-level clinical knowledge and reasoning in ophthalmology: A head-to-head cross-sectional study

Large language models (LLMs) underlie remarkable recent advanced in natural language processing, and they are beginning to be applied in clinical contexts. We aimed to evaluate the clinical potential of state-of-the-art LLMs in ophthalmology using a more robust benchmark than raw examination scores. We trialled GPT-3.5 and GPT-4 on 347 ophthalmology questions before GPT-3.5, GPT-4, PaLM 2, LLaMA, expert ophthalmologists, and doctors in training were trialled on a mock examination of 87 questions. Performance was analysed with respect to question subject and type (first order recall and higher order reasoning). Masked ophthalmologists graded the accuracy, relevance, and overall preference of GPT-3.5 and GPT-4 responses to the same questions. The performance of GPT-4 (69%) was superior to GPT-3.5 (48%), LLaMA (32%), and PaLM 2 (56%). GPT-4 compared favourably with expert ophthalmologists (median 76%, range 64-90%), ophthalmology trainees (median 59%, range 57-63%), and unspecialised junior doctors (median 43%, range 41-44%). Low agreement between LLMs and doctors reflected idiosyncratic differences in knowledge and reasoning with overall consistency across subjects and types (p>0.05). All ophthalmologists preferred GPT-4 responses over GPT-3.5 and rated the accuracy and relevance of GPT-4 as higher (p<0.05). LLMs are approaching expert-level knowledge and reasoning skills in ophthalmology. In view of the comparable or superior performance to trainee-grade ophthalmologists and unspecialised junior doctors, state-of-the-art LLMs such as GPT-4 may provide useful medical advice and assistance where access to expert ophthalmologists is limited. Clinical benchmarks provide useful assays of LLM capabilities in healthcare before clinical trials can be designed and conducted.

Incorporating participants' welfare into sequential multiple assignment randomized trials

Dynamic treatment regimes (DTRs) are sequences of decision rules that recommend treatments based on patients' time-varying clinical conditions. The sequential, multiple assignment, randomized trial (SMART) is an experimental design that can provide high-quality evidence for constructing optimal DTRs. In a conventional SMART, participants are randomized to available treatments at multiple stages with balanced randomization probabilities. Despite its relative simplicity of implementation and desirable performance in comparing embedded DTRs, the conventional SMART faces inevitable ethical issues, including assigning many participants to the empirically inferior treatment or the treatment they dislike, which might slow down the recruitment procedure and lead to higher attrition rates, ultimately leading to poor internal and external validities of the trial results. In this context, we propose a SMART under the Experiment-as-Market framework (SMART-EXAM), a novel SMART design that holds the potential to improve participants' welfare by incorporating their preferences and predicted treatment effects into the randomization procedure. We describe the steps of conducting a SMART-EXAM and evaluate its performance compared to the conventional SMART. The results indicate that the SMART-EXAM can improve the welfare of the participants enrolled in the trial, while also achieving a desirable ability to construct an optimal DTR when the experimental parameters are suitably specified. We finally illustrate the practical potential of the SMART-EXAM design using data from a SMART for children with attention-deficit/hyperactivity disorder.

Mindfulness-based therapy improves brain functional network reconfiguration efficiency

Mindfulness-based interventions are showing increasing promise as a treatment for psychological disorders, with improvements in cognition and emotion regulation after intervention. Understanding the changes in functional brain activity and neural plasticity that underlie these benefits from mindfulness interventions is thus of interest in current neuroimaging research. Previous studies have found functional brain changes during resting and task states to be associated with mindfulness both cross-sectionally and longitudinally, particularly in the executive control, default mode and salience networks. However, limited research has combined information from rest and task to study mindfulness-related functional changes in the brain, particularly in the context of intervention studies with active controls. Recent work has found that the reconfiguration efficiency of brain activity patterns between rest and task states is behaviorally relevant in healthy young adults. Thus, we applied this measure to investigate how mindfulness intervention changed functional reconfiguration between rest and a breath-counting task in elderly participants with self-reported sleep difficulties. Improving on previous longitudinal designs, we compared the intervention effects of a mindfulness-based therapy to an active control (sleep hygiene) intervention. We found that mindfulness intervention improved self-reported mindfulness measures and brain functional reconfiguration efficiency in the executive control, default mode and salience networks, though the brain and behavioral changes were not associated with each other. Our findings suggest that neuroplasticity may be induced through regular mindfulness practice, thus bringing the intrinsic functional configuration in participants' brains closer to a state required for mindful awareness.

Economic Evaluations of Imaging Biomarker-Driven Companion Diagnostics for Cancer: A Systematic Review

INTRODUCTION

There is a boom in imaging biomarker-driven companion and complementary diagnostics (CDx) for cancer, which brings opportunity for personalized medicine. Whether adoption of these technologies is likely to be cost-effective is a relevant question, and studies on this topic are emerging. Despite the growing number of economic evaluations, no review of the methods used, quality of reporting, and potential risk of bias has been done. We report a systematic review to identify, summarize, and critique the cost-effectiveness evidence for the use of biomarker-driven and imaging-based CDx to inform cancer treatments.

METHODS

The Preferred Reporting Items of Systematic Reviews and Meta-Analyses (PRISMA) guidelines were followed. Systematic literature searches until 30 December 2022 were performed in PubMed, Web of Science, Medline, Embase, and Scopus for economic evaluations of imaging biomarker-based CDx for cancer. The inclusion and exclusion of studies were determined by pre-specified eligibility criteria informed by the 'Patient, Intervention, Comparison, Outcome' (PICO) framework. The Consolidated Health Economic Evaluation Reporting Standards (CHEERS) was used to assess the quality of reporting, and the Bias in Economic Evaluation (ECOBIAS) was used to examine the potential risk of bias of included studies.

RESULTS

A total of 12 papers were included, with eight model-based and four trial-based studies. Implementing biomarker-driven, imaging-based CDx was reported to be cost-effective, cost saving, or dominant (cost saving and more effective) in ten papers. Inconsistent methods were found in the studies, and the quality of reporting was lacking against the CHEERS reporting guideline. Several potential sources of 'risk of bias' were identified. These should be acknowledged and carefully considered by researchers planning future health economic evaluations.

CONCLUSION

Despite favorable results towards the implementation of imaging biomarker-based CDx for cancer, there is room for improvement regarding the quantity and quality of economic evaluations, and that is expected as the awareness of current study limitations increases and more clinical data become available in the future.

Enabling Survival of Transplanted Neural Precursor Cells in the Ischemic Brain

There is no effective therapy for ischemic stroke following the acute stage. Neural transplantation offers a potential option for repairing the ischemic lesion. However, this strategy is hindered by the poor survival of the neural precursor cells (NPCs) that are transplanted into the inflammatory ischemic core. Here, a chemical cocktail consisting of fibrinogen and maraviroc is developed to promote the survival of the transplanted NPCs in the ischemic core of the mouse cerebral cortex. The grafted NPCs survive in the presence of the cocktail but not fibrinogen or maraviroc alone at day 7. The surviving NPCs divide and differentiate to mature neurons by day 30, reconstituting the infarct cortex with vascularization. Molecular analysis in vivo and in vitro shows that blocking the activation of CCR5 on the NPCs protects the NPCs from apoptosis induced by pro-inflammatory factors, revealing the underlying protective effect of the cocktail for NPCs. The findings open an avenue to enable survival of the transplanted NPCs under the inflammatory neurological conditions like stroke.

Mast cell activation in lungs during SARS-CoV-2 infection associated with lung pathology and severe COVID-19

Lung inflammation is a hallmark of Coronavirus disease 2019 (COVID-19) in patients who are severely ill, and the pathophysiology of disease is thought to be immune mediated. Mast cells (MCs) are polyfunctional immune cells present in the airways, where they respond to certain viruses and allergens and often promote inflammation. We observed widespread degranulation of MCs during acute and unresolved airway inflammation in SARS-CoV-2-infected mice and nonhuman primates. Using a mouse model of MC deficiency, MC-dependent interstitial pneumonitis, hemorrhaging, and edema in the lung were observed during SARS-CoV-2 infection. In humans, transcriptional changes in patients requiring oxygen supplementation also implicated cells with a MC phenotype in severe disease. MC activation in humans was confirmed through detection of MC-specific proteases, including chymase, the levels of which were significantly correlated with disease severity and with biomarkers of vascular dysregulation. These results support the involvement of MCs in lung tissue damage during SARS-CoV-2 infection in animal models and the association of MC activation with severe COVID-19 in humans, suggesting potential strategies for intervention.

Systems level identification of a matrisome-associated macrophage polarisation state in multi-organ fibrosis

Tissue fibrosis affects multiple organs and involves a master-regulatory role of macrophages which respond to an initial inflammatory insult common in all forms of fibrosis. The recently unravelled multi-organ heterogeneity of macrophages in healthy and fibrotic human disease suggests that macrophages expressing osteopontin (SPP1) associate with lung and liver fibrosis. However, the conservation of this SPP1+ macrophage population across different tissues and its specificity to fibrotic diseases with different etiologies remain unclear. Integrating 15 single-cell RNA-sequencing datasets to profile 235,930 tissue macrophages from healthy and fibrotic heart, lung, liver, kidney, skin, and endometrium, we extended the association of SPP1+ macrophages with fibrosis to all these tissues. We also identified a subpopulation expressing matrisome-associated genes (e.g., matrix metalloproteinases and their tissue inhibitors), functionally enriched for ECM remodelling and cell metabolism, representative of a matrisome-associated macrophage (MAM) polarisation state within SPP1+ macrophages. Importantly, the MAM polarisation state follows a differentiation trajectory from SPP1+ macrophages and is associated with a core set of regulon activity. SPP1+ macrophages without the MAM polarisation state (SPP1+MAM-) show a positive association with ageing lung in mice and humans. These results suggest an advanced and conserved polarisation state of SPP1+ macrophages in fibrotic tissues resulting from prolonged inflammatory cues within each tissue microenvironment.

A translational perspective towards clinical AI fairness

Artificial intelligence (AI) has demonstrated the ability to extract insights from data, but the fairness of such data-driven insights remains a concern in high-stakes fields. Despite extensive developments, issues of AI fairness in clinical contexts have not been adequately addressed. A fair model is normally expected to perform equally across subgroups defined by sensitive variables (e.g., age, gender/sex, race/ethnicity, socio-economic status, etc.). Various fairness measurements have been developed to detect differences between subgroups as evidence of bias, and bias mitigation methods are designed to reduce the differences detected. This perspective of fairness, however, is misaligned with some key considerations in clinical contexts. The set of sensitive variables used in healthcare applications must be carefully examined for relevance and justified by clear clinical motivations. In addition, clinical AI fairness should closely investigate the ethical implications of fairness measurements (e.g., potential conflicts between group- and individual-level fairness) to select suitable and objective metrics. Generally defining AI fairness as "equality" is not necessarily reasonable in clinical settings, as differences may have clinical justifications and do not indicate biases. Instead, "equity" would be an appropriate objective of clinical AI fairness. Moreover, clinical feedback is essential to developing fair and well-performing AI models, and efforts should be made to actively involve clinicians in the process. The adaptation of AI fairness towards healthcare is not self-evident due to misalignments between technical developments and clinical considerations. Multidisciplinary collaboration between AI researchers, clinicians, and ethicists is necessary to bridge the gap and translate AI fairness into real-life benefits.

A scoping review of the clinical application of machine learning in data-driven population segmentation analysis

OBJECTIVE

Data-driven population segmentation is commonly used in clinical settings to separate the heterogeneous population into multiple relatively homogenous groups with similar healthcare features. In recent years, machine learning (ML) based segmentation algorithms have garnered interest for their potential to speed up and improve algorithm development across many phenotypes and healthcare situations. This study evaluates ML-based segmentation with respect to (1) the populations applied, (2) the segmentation details, and (3) the outcome evaluations.

MATERIALS AND METHODS

MEDLINE, Embase, Web of Science, and Scopus were used following the PRISMA-ScR criteria. Peer-reviewed studies in the English language that used data-driven population segmentation analysis on structured data from January 2000 to October 2022 were included.

RESULTS

We identified 6077 articles and included 79 for the final analysis. Data-driven population segmentation analysis was employed in various clinical settings. K-means clustering is the most prevalent unsupervised ML paradigm. The most common settings were healthcare institutions. The most common targeted population was the general population.

DISCUSSION

Although all the studies did internal validation, only 11 papers (13.9%) did external validation, and 23 papers (29.1%) conducted methods comparison. The existing papers discussed little validating the robustness of ML modeling.

CONCLUSION

Existing ML applications on population segmentation need more evaluations regarding giving tailored, efficient integrated healthcare solutions compared to traditional segmentation analysis. Future ML applications in the field should emphasize methods' comparisons and external validation and investigate approaches to evaluate individual consistency using different methods.

Reducing Falls Among Community-Dwelling Older Adults From Clinicians' Perspectives: A Systems Modeling Approach

Background and Objectives

Falls among older adults are a significant health problem globally. Studies of multicomponent fall prevention programs in randomized controlled trials demonstrate effectiveness in reducing falls; however, the translation of research into the community remains challenging. Although there is an increasing interest to understand the factors contributing to implementation barriers, the dynamic relationships between factors are less well examined. Furthermore, evidence on implementation barriers from Asia is lacking as most of these studies originate from the West. As such, this study aims to engage stakeholders in uncovering the factors that facilitate or inhibit implementing community-based fall prevention programs in Singapore, with a focus on the interrelationship between those factors.

Research Design and Methods

Health care professionals familiar with fall prevention programs were invited to discuss the enablers and challenges to the implementation. This effort was facilitated using a systems modeling methodology of Group Model Building (GMB) to share ideas and create a common conceptual model of the challenges. The GMB employs various engagement techniques to draw on the experiences and perceptions of all stakeholders involved.

Results

This process led to the development of a Causal Loop Diagram (CLD), a qualitative conceptual model of the dynamic relationships between the barriers and facilitators of implementing fall prevention programs. Results from the CLD show that implementation is influenced by two main drivers: health care provider factors that influenced referrals, and patient factors that influenced referral acceptance and long-term adherence. Key leverage points for potential interventions were identified as well.

Discussion and Implications

The overall recommendation emphasized closer coordination and collaboration across providers to ensure sustainable and effective community-based fall prevention programs. This has to be supported by a national effort, involving a multidisciplinary stakeholder advisory group. These findings generated would be promising to guide future approaches to fall prevention.

Synthetic artificial intelligence using generative adversarial network for retinal imaging in detection of age-related macular degeneration

Introduction

Age-related macular degeneration (AMD) is one of the leading causes of vision impairment globally and early detection is crucial to prevent vision loss. However, the screening of AMD is resource dependent and demands experienced healthcare providers. Recently, deep learning (DL) systems have shown the potential for effective detection of various eye diseases from retinal fundus images, but the development of such robust systems requires a large amount of datasets, which could be limited by prevalence of the disease and privacy of patient. As in the case of AMD, the advanced phenotype is often scarce for conducting DL analysis, which may be tackled via generating synthetic images using Generative Adversarial Networks (GANs). This study aims to develop GAN-synthesized fundus photos with AMD lesions, and to assess the realness of these images with an objective scale.

Methods

To build our GAN models, a total of 125,012 fundus photos were used from a real-world non-AMD phenotypical dataset. StyleGAN2 and human-in-the-loop (HITL) method were then applied to synthesize fundus images with AMD features. To objectively assess the quality of the synthesized images, we proposed a novel realness scale based on the frequency of the broken vessels observed in the fundus photos. Four residents conducted two rounds of gradings on 300 images to distinguish real from synthetic images, based on their subjective impression and the objective scale respectively.

Results and discussion

The introduction of HITL training increased the percentage of synthetic images with AMD lesions, despite the limited number of AMD images in the initial training dataset. Qualitatively, the synthesized images have been proven to be robust in that our residents had limited ability to distinguish real from synthetic ones, as evidenced by an overall accuracy of 0.66 (95% CI: 0.61-0.66) and Cohen's kappa of 0.320. For the non-referable AMD classes (no or early AMD), the accuracy was only 0.51. With the objective scale, the overall accuracy improved to 0.72. In conclusion, GAN models built with HITL training are capable of producing realistic-looking fundus images that could fool human experts, while our objective realness scale based on broken vessels can help identifying the synthetic fundus photos.

Retinal Oxygen Metabolism in Patients With Type 2 Diabetes and Different Stages of Diabetic Retinopathy

The aim of this cross-sectional study was to assess retinal oxygen metabolism in patients with type 2 diabetes and different stages of nonproliferative diabetic retinopathy (DR) (n = 67) compared with healthy control subjects (n = 20). Thirty-four patients had no DR, 15 had mild DR, and 18 had moderate to severe DR. Retinal oxygen saturation in arteries and veins was measured using the oxygen module of a retinal vessel analyzer. Total retinal blood flow (TRBF) was measured using a custom-built Doppler optical coherence tomography system. Retinal oxygen extraction was calculated from retinal oxygen saturation and TRBF. Arteriovenous difference in oxygen saturation was highest in healthy subjects (34.9 ± 7.5%), followed by patients with no DR (32.5 ± 6.3%) and moderate to severe DR (30.3 ± 6.5%). The lowest values were found in patients with mild DR (27.3 ± 8.0%, P = 0.010 vs. healthy subjects). TRBF tended to be higher in patients with no DR (40.1 ± 9.2 μL/min) and mild DR (41.8 ± 15.0 μL/min) than in healthy subjects (37.2 ± 5.7 μL/min) and patients with moderate to severe DR (34.6 ± 10.4 μL/min). Retinal oxygen extraction was the highest in healthy subjects (2.24 ± 0.57 μL O2/min), followed by patients with no DR (2.14 ± 0.6 μL O2/min), mild DR (1.90 ± 0.77 μL O2/min), and moderate to severe DR (1.78 ± 0.57 μL O2/min, P = 0.040 vs. healthy subjects). These results indicate that retinal oxygen metabolism is altered in patients with type 2 diabetes. Furthermore, retinal oxygen extraction decreases with increasing severity of DR.

Efficacy, toxicity, and quality-of-life outcomes of ultrahypofractionated radiotherapy in patients with localized prostate cancer: A single-arm phase 2 trial from Asia

AIMS

Ultra-hypofractionated radiotherapy (UHF-RT) is widely utilized in men with localized prostate cancer (PCa). There are limited data in Asian cohorts. We report the outcomes of a single-arm, phase II trial of UHF-RT from an Asian center.

METHODS

We recruited men with histologically confirmed, nonmetastatic localized PCa. UHF-RT regimens were 36.25 Gy (Cohort A) and 37.5 Gy (Cohort B) delivered in five fractions every other day over 1.5-2.5 weeks. Primary endpoint was physician-scored late genitourinary (GU) and gastrointestinal (GI) adverse events (AEs). Quality-of-life (QoL) was assessed by Expanded Prostate Cancer Index Composite (EPIC) at baseline, 1- and 2-year post-UHF-RT.

RESULTS

Between March 2014 and August 2019, 105 men were recruited; four were subsequently excluded from analysis. Median age was 68.0 (Interquartile range (IQR): 63.8-73.0) years. 26 (24.8%) and 68 (64.8%) men had NCCN-defined low-and intermediate-risk PCa, respectively. No late ≥G3 GU or GI toxicities were reported in both cohorts. Peak incidence of acute ≥G2 GU AEs at 14 days post-UHF-RT was 23.6% (17/72) and 24.0% (6/25) in Cohorts A and B, respectively; ≥G2 GI AEs were observed in 9.7% (7/72) and 36.0% (9/25), respectively. Late ≥G2 GU and GI AEs occurred in 4.7% and 3.1% of Cohort A patients, and 5.0% in Cohort B at 12 months, with no AEs at 24 months. EPIC scores changed minimally across all domains. At a median follow-up of 44.9 months, we recorded one (1.3%) biochemical relapse by the Phoenix criteria (Cohort A).

CONCLUSION

UHF-RT is well tolerated in Asian men and can be a recommended fractionation schema for localized PCa.

Stage-dependent differential influence of metabolic and structural networks on memory across Alzheimer's disease continuum

Background

Large-scale neuronal network breakdown underlies memory impairment in Alzheimer's disease (AD). However, the differential trajectories of the relationships between network organisation and memory across pathology and cognitive stages in AD remain elusive. We determined whether and how the influences of individual-level structural and metabolic covariance network integrity on memory varied with amyloid pathology across clinical stages without assuming a constant relationship.

Methods

Seven hundred and eight participants from the Alzheimer's Disease Neuroimaging Initiative were studied. Individual-level structural and metabolic covariance scores in higher-level cognitive and hippocampal networks were derived from magnetic resonance imaging and [18F] fluorodeoxyglucose positron emission tomography using seed-based partial least square analyses. The non-linear associations between network scores and memory across cognitive stages in each pathology group were examined using sparse varying coefficient modelling.

Results

We showed that the associations of memory with structural and metabolic networks in the hippocampal and default mode regions exhibited pathology-dependent differential trajectories across cognitive stages using sparse varying coefficient modelling. In amyloid pathology group, there was an early influence of hippocampal structural network deterioration on memory impairment in the preclinical stage, and a biphasic influence of the angular gyrus-seeded default mode metabolic network on memory in both preclinical and dementia stages. In non-amyloid pathology groups, in contrast, the trajectory of the hippocampus-memory association was opposite and weaker overall, while no metabolism covariance networks were related to memory. Key findings were replicated in a larger cohort of 1280 participants.

Conclusions

Our findings highlight potential windows of early intervention targeting network breakdown at the preclinical AD stage.

Funding

Data collection and sharing for this project was funded by the Alzheimer's Disease Neuroimaging Initiative (ADNI) (National Institutes of Health Grant U01 AG024904) and DOD ADNI (Department of Defense award number W81XWH-12-2-0012). We also acknowledge the funding support from the Duke NUS/Khoo Bridge Funding Award (KBrFA/2019-0020) and NMRC Open Fund Large Collaborative Grant (OFLCG09May0035).

TMEM41B and VMP1 modulate cellular lipid and energy metabolism for facilitating dengue virus infection

Transmembrane Protein 41B (TMEM41B) and Vacuole Membrane Protein 1 (VMP1) are two ER-associated lipid scramblases that play a role in autophagosome formation and cellular lipid metabolism. TMEM41B is also a recently validated host factor required by flaviviruses and coronaviruses. However, the exact underlying mechanism of TMEM41B in promoting viral infections remains an open question. Here, we validated that both TMEM41B and VMP1 are essential host dependency factors for all four serotypes of dengue virus (DENV) and human coronavirus OC43 (HCoV-OC43), but not chikungunya virus (CHIKV). While HCoV-OC43 failed to replicate entirely in both TMEM41B- and VMP1-deficient cells, we detected diminished levels of DENV infections in these cell lines, which were accompanied by upregulation of the innate immune dsRNA sensors, RIG-I and MDA5. Nonetheless, this upregulation did not correspondingly induce the downstream effector TBK1 activation and Interferon-beta expression. Despite low levels of DENV replication, classical DENV replication organelles were undetectable in the infected TMEM41B-deficient cells, suggesting that the upregulation of the dsRNA sensors is likely a consequence of aberrant viral replication rather than a causal factor for reduced DENV infection. Intriguingly, we uncovered that the inhibitory effect of TMEM41B deficiency on DENV replication, but not HCoV-OC43, can be partially reversed using exogenous fatty acid supplements. In contrast, VMP1 deficiency cannot be rescued using the metabolite treatment. In line with the observed phenotypes, we found that both TMEM41B- and VMP1-deficient cells harbor higher levels of compromised mitochondria, especially in VMP1 deficiency which results in severe dysregulations of mitochondrial beta-oxidation. Using a metabolomic profiling approach, we revealed distinctive global dysregulations of the cellular metabolome, particularly lipidome, in TMEM41B- and VMP1-deficient cells. Our findings highlight a central role for TMEM41B and VMP1 in modulating multiple cellular pathways, including lipid mobilization, mitochondrial beta-oxidation, and global metabolic regulations, to facilitate the replication of flaviviruses and coronaviruses.

Given the concerns over potential global health burdens imposed by endless emerging and re-emerging viruses as well as the limited therapeutic options to intervene, host-directed therapeutics can serve as a promising approach to broadly prepare for future pandemics. TMEM41B and VMP1 have been demonstrated as essential host factors for at least two unrelated groups of clinically important RNA viruses with outbreak potential. Therefore these ER membrane proteins could potentially serve as cellular targets for developing host-directed therapeutics. However, the effort must be first supported by a comprehensive understanding of their function in viral infection. Here, we dissected the role of TMEM41B and VMP1 in dengue virus infection, showing that both these proteins are crucial for the normal functionality of mitochondria and the regulation of cellular metabolites. We further provided evidence that these metabolic roles contribute to TMEM41B and VMP1 essentiality in dengue virus infection.

Age-Related Performance in Using a Fully Immersive and Automated Virtual Reality System to Assess Cognitive Function

Introduction

Cognition generally declines gradually over time due to progressive degeneration of the brain, leading to dementia and eventual loss of independent functions. The rate of regression varies among the six cognitive domains (perceptual motor, executive function, complex attention, learning and memory, social cognition and language). Current modality of cognitive assessment using neuropsychological paper-and-pencil screening tools for cognitive impairment such as the Montreal Cognitive Assessment (MoCA) has limitations and is influenced by age. Virtual reality (VR) is considered as a potential alternative tool to assess cognition. A novel, fully immersive automated VR system (Cognitive Assessment using Virtual Reality, CAVIRE) has been developed to assess the six cognitive domains. As cognition is associated with age, VR performance is postulated to vary with age using this system.

Aims

This is a feasibility study to evaluate the VR performance of cognitively healthy adults aged between 35 and 74 years old, based on the performance score and completion time using the CAVIRE system.

Methods

Conducted in a public primary care clinic in Singapore, 25 multi-ethnic Asian adults were recruited in each of the four age groups in years: (1) 35–44; (2) 45–54; (3) 55–64, and (4) 65–74. The eligibility criteria included a MoCA score of 26 or higher to reflect normal cognition and understanding English instructions. They completed common daily activities ranging from brushing teething to shopping, across 13 VR segments. Their performances scores and completion time were automatically computed by the CAVIRE system. These VR performance indices were compared across the four age groups using one-way ANOVA, F-test of the hypothesis, followed by pair-wise comparisons in the event of a significant F-test (p < 0.05).

Results

One participant dropped out from Group 1. The demographic characteristics of 99 participants were similar across the 4 age groups. Overall, younger participants in Groups 1 and 2 attained higher VR performance scores and shorter completion time, compared to those in Groups 3 and 4, in all six cognitive domains (all p < 0.05).

Conclusion

The CAVIRE VR performance scores and completion time significantly differ between the younger and older Asian participants with normal cognition. Enhancements to the system are needed to establish the age-group specific normal performance indices.

SC-MEB: spatial clustering with hidden Markov random field using empirical Bayes

Spatial transcriptomics has been emerging as a powerful technique for resolving gene expression profiles while retaining tissue spatial information. These spatially resolved transcriptomics make it feasible to examine the complex multicellular systems of different microenvironments. To answer scientific questions with spatial transcriptomics and expand our understanding of how cell types and states are regulated by microenvironment, the first step is to identify cell clusters by integrating the available spatial information. Here, we introduce SC-MEB, an empirical Bayes approach for spatial clustering analysis using a hidden Markov random field. We have also derived an efficient expectation-maximization algorithm based on an iterative conditional mode for SC-MEB. In contrast to BayesSpace, a recently developed method, SC-MEB is not only computationally efficient and scalable to large sample sizes but is also capable of choosing the smoothness parameter and the number of clusters. We performed comprehensive simulation studies to demonstrate the superiority of SC-MEB over some existing methods. We applied SC-MEB to analyze the spatial transcriptome of human dorsolateral prefrontal cortex tissues and mouse hypothalamic preoptic region. Our analysis results showed that SC-MEB can achieve a similar or better clustering performance to BayesSpace, which uses the true number of clusters and a fixed smoothness parameter. Moreover, SC-MEB is scalable to large 'sample sizes'. We then employed SC-MEB to analyze a colon dataset from a patient with colorectal cancer (CRC) and COVID-19, and further performed differential expression analysis to identify signature genes related to the clustering results. The heatmap of identified signature genes showed that the clusters identified using SC-MEB were more separable than those obtained with BayesSpace. Using pathway analysis, we identified three immune-related clusters, and in a further comparison, found the mean expression of COVID-19 signature genes was greater in immune than non-immune regions of colon tissue. SC-MEB provides a valuable computational tool for investigating the structural organizations of tissues from spatial transcriptomic data.

Integrative epigenomic and high-throughput functional enhancer profiling reveals determinants of enhancer heterogeneity in gastric cancer

BACKGROUND

Enhancers are distal cis-regulatory elements required for cell-specific gene expression and cell fate determination. In cancer, enhancer variation has been proposed as a major cause of inter-patient heterogeneity-however, most predicted enhancer regions remain to be functionally tested.

METHODS

We analyzed 132 epigenomic histone modification profiles of 18 primary gastric cancer (GC) samples, 18 normal gastric tissues, and 28 GC cell lines using Nano-ChIP-seq technology. We applied Capture-based Self-Transcribing Active Regulatory Region sequencing (CapSTARR-seq) to assess functional enhancer activity. An Activity-by-contact (ABC) model was employed to explore the effects of histone acetylation and CapSTARR-seq levels on enhancer-promoter interactions.

RESULTS

We report a comprehensive catalog of 75,730 recurrent predicted enhancers, the majority of which are GC-associated in vivo (> 50,000) and associated with lower somatic mutation rates inferred by whole-genome sequencing. Applying CapSTARR-seq to the enhancer catalog, we observed significant correlations between CapSTARR-seq functional activity and H3K27ac/H3K4me1 levels. Super-enhancer regions exhibited increased CapSTARR-seq signals compared to regular enhancers, even when decoupled from native chromatin contexture. We show that combining histone modification and CapSTARR-seq functional enhancer data improves the prediction of enhancer-promoter interactions and pinpointing of germline single nucleotide polymorphisms (SNPs), somatic copy number alterations (SCNAs), and trans-acting TFs involved in GC expression. We identified cancer-relevant genes (ING1, ARL4C) whose expression between patients is influenced by enhancer differences in genomic copy number and germline SNPs, and HNF4α as a master trans-acting factor associated with GC enhancer heterogeneity.

CONCLUSIONS

Our results indicate that combining histone modification and functional assay data may provide a more accurate metric to assess enhancer activity than either platform individually, providing insights into the relative contribution of genetic (cis) and regulatory (trans) mechanisms to GC enhancer functional heterogeneity.

Largazole Inhibits Ocular Angiogenesis by Modulating the Expression of VEGFR2 and p21

Ocular angiogenic diseases, characterized by abnormal blood vessel formation in the eye, are the leading cause of blindness. Although Anti-VEGF therapy is the first-line treatment in the market, a substantial number of patients are refractory to it or may develop resistance over time. As uncontrolled proliferation of vascular endothelial cells is one of the characteristic features of pathological neovascularization, we aimed to investigate the role of the class I histone deacetylase (HDAC) inhibitor Largazole, a cyclodepsipeptide from a marine cyanobacterium, in ocular angiogenesis. Our study showed that Largazole strongly inhibits retinal vascular endothelial cell viability, proliferation, and the ability to form tube-like structures. Largazole strongly inhibits the vessel outgrowth from choroidal explants in choroid sprouting assay while it does not affect the quiescent choroidal vasculature. Largazole also inhibits vessel outgrowth from metatarsal bones in metatarsal sprouting assay without affecting pericytes coverage. We further demonstrated a cooperative effect between Largazole and an approved anti-VEGF drug, Alflibercept. Mechanistically, Largazole strongly inhibits the expression of VEGFR2 and leads to an increased expression of cell cycle inhibitor, p21. Taken together, our study provides compelling evidence on the anti-angiogenic role of Largazole that exerts its function through mediating different signaling pathways.

LRG1 Promotes Metastatic Dissemination of Melanoma through Regulating EGFR/STAT3 Signalling

Although less common, melanoma is the deadliest form of skin cancer largely due to its highly metastatic nature. Currently, there are limited treatment options for metastatic melanoma and many of them could cause serious side effects. A better understanding of the molecular mechanisms underlying the complex disease pathophysiology of metastatic melanoma may lead to the identification of novel therapeutic targets and facilitate the development of targeted therapeutics. In this study, we investigated the role of leucine-rich α-2-glycoprotein 1 (LRG1) in melanoma development and progression. We first established the association between LRG1 and melanoma in both human patient biopsies and mouse melanoma cell lines and revealed a significant induction of LRG1 expression in metastatic melanoma cells. We then showed no change in tumour cell growth, proliferation, and angiogenesis in the absence of the host Lrg1. On the other hand, there was reduced melanoma cell metastasis to the lungs in Lrg1-deficient mice. This observation was supported by the promoting effect of LRG1 in melanoma cell migration, invasion, and adhesion. Mechanistically, LRG1 mediates melanoma cell invasiveness in an EGFR/STAT3-dependent manner. Taken together, our studies provided compelling evidence that LRG1 is required for melanoma metastasis but not growth. Targeting LRG1 may offer an alternative strategy to control malignant melanoma.

Clinical implementation of an oncology-specific family health history risk assessment tool

BACKGROUND

The presence of hereditary cancer syndromes in cancer patients can have an impact on current clinical care and post-treatment prevention and surveillance measures. Several barriers inhibit identification of hereditary cancer syndromes in routine practice. This paper describes the impact of using a patient-facing family health history risk assessment platform on the identification and referral of breast cancer patients to genetic counselling services.

METHODS

This was a hybrid implementation-effectiveness study completed in breast cancer clinics. English-literate patients not previously referred for genetic counselling and/or gone through genetic testing were offered enrollment. Consented participants were provided educational materials on family health history collection, entered their family health history into the platform and completed a satisfaction survey. Upon completion, participants and their clinicians were given personalized risk reports. Chart abstraction was done to identify actions taken by patients, providers and genetic counsellors.

RESULTS

Of 195 patients approached, 102 consented and completed the study (mean age 55.7, 100 % women). Sixty-six (65 %) met guideline criteria for genetic counseling of which 24 (36 %) were referred for genetic counseling. Of those referred, 13 (54 %) participants attended and eight (33 %) completed genetic testing. On multivariate logistic regression, referral was not associated with age, cancer stage, or race but was associated with clinical provider (p = 0.041). Most providers (71 %) had higher referral rates during the study compared to prior. The majority of participants found the experience useful (84 %), were more aware of their health risks (83 %), and were likely to recommend using a patient-facing platform to others (69 %).

CONCLUSIONS

65 % of patients attending breast cancer clinics in this study are at-risk for hereditary conditions based on current guidelines. Using a patient-facing risk assessment platform enhances the ability to identify these patients systematically and with widespread acceptability and recognized value by patients. As only a third of at-risk participants received referrals for genetic counseling, further understanding barriers to referral is needed to optimize hereditary risk assessment in oncology practices.

TRIAL REGISTRATION

NIH Clinical Trials registry, NCT04639934 . Registered Nov 23, 2020 -- Retrospectively registered.

A scoping review of studies using observational data to optimise dynamic treatment regimens

BACKGROUND

Dynamic treatment regimens (DTRs) formalise the multi-stage and dynamic decision problems that clinicians often face when treating chronic or progressive medical conditions. Compared to randomised controlled trials, using observational data to optimise DTRs may allow a wider range of treatments to be evaluated at a lower cost. This review aimed to provide an overview of how DTRs are optimised with observational data in practice.

METHODS

Using the PubMed database, a scoping review of studies in which DTRs were optimised using observational data was performed in October 2020. Data extracted from eligible articles included target medical condition, source and type of data, statistical methods, and translational relevance of the included studies.

RESULTS

From 209 PubMed abstracts, 37 full-text articles were identified, and a further 26 were screened from the reference lists, totalling 63 articles for inclusion in a narrative data synthesis. Observational DTR models are a recent development and their application has been concentrated in a few medical areas, primarily HIV/AIDS (27, 43%), followed by cancer (8, 13%), and diabetes (6, 10%). There was substantial variation in the scope, intent, complexity, and quality between the included studies. Statistical methods that were used included inverse-probability weighting (26, 41%), the parametric G-formula (16, 25%), Q-learning (10, 16%), G-estimation (4, 6%), targeted maximum likelihood/minimum loss-based estimation (4, 6%), regret regression (3, 5%), and other less common approaches (10, 16%). Notably, studies that were primarily intended to address real-world clinical questions (18, 29%) tended to use inverse-probability weighting and the parametric G-formula, relatively well-established methods, along with a large amount of data. Studies focused on methodological developments (45, 71%) tended to be more complicated and included a demonstrative real-world application only.

CONCLUSIONS

As chronic and progressive conditions become more common, the need will grow for personalised treatments and methods to estimate the effects of DTRs. Observational DTR studies will be necessary, but so far their use to inform clinical practice has been limited. Focusing on simple DTRs, collecting large and rich clinical datasets, and fostering tight partnerships between content experts and data analysts may result in more clinically relevant observational DTR studies.

Effect of hyperglycaemia and diabetes on acute myocardial ischaemia-reperfusion injury and cardioprotection by ischaemic conditioning protocols

Diabetic patients are at increased risk of developing coronary artery disease and experience worse clinical outcomes following acute myocardial infarction. Novel therapeutic strategies are required to protect the myocardium against the effects of acute ischaemia-reperfusion injury (IRI). These include one or more brief cycles of non-lethal ischaemia and reperfusion prior to the ischaemic event (ischaemic preconditioning [IPC]) or at the onset of reperfusion (ischaemic postconditioning [IPost]) either to the heart or to extracardiac organs (remote ischaemic conditioning [RIC]). Studies suggest that the diabetic heart is resistant to cardioprotective strategies, although clinical evidence is lacking. We overview the available animal models of diabetes, investigating acute myocardial IRI and cardioprotection, experiments investigating the effects of hyperglycaemia on susceptibility to acute myocardial IRI, the response of the diabetic heart to cardioprotective strategies e.g. IPC, IPost and RIC. Finally we highlight the effects of anti-hyperglycaemic agents on susceptibility to acute myocardial IRI and cardioprotection. LINKED ARTICLES: This article is part of a themed issue on Risk factors, comorbidities, and comedications in cardioprotection. To view the other articles in this section visit http://onlinelibrary.wiley.com/doi/10.1111/bph.v177.23/issuetoc.

A tissue-specific collaborative mixed model for jointly analyzing multiple tissues in transcriptome-wide association studies

Transcriptome-wide association studies (TWASs) integrate expression quantitative trait loci (eQTLs) studies with genome-wide association studies (GWASs) to prioritize candidate target genes for complex traits. Several statistical methods have been recently proposed to improve the performance of TWASs in gene prioritization by integrating the expression regulatory information imputed from multiple tissues, and made significant achievements in improving the ability to detect gene-trait associations. Unfortunately, most existing multi-tissue methods focus on prioritization of candidate genes, and cannot directly infer the specific functional effects of candidate genes across different tissues. Here, we propose a tissue-specific collaborative mixed model (TisCoMM) for TWASs, leveraging the co-regulation of genetic variations across different tissues explicitly via a unified probabilistic model. TisCoMM not only performs hypothesis testing to prioritize gene-trait associations, but also detects the tissue-specific role of candidate target genes in complex traits. To make full use of widely available GWASs summary statistics, we extend TisCoMM to use summary-level data, namely, TisCoMM-S2. Using extensive simulation studies, we show that type I error is controlled at the nominal level, the statistical power of identifying associated genes is greatly improved, and the false-positive rate (FPR) for non-causal tissues is well controlled at decent levels. We further illustrate the benefits of our methods in applications to summary-level GWASs data of 33 complex traits. Notably, apart from better identifying potential trait-associated genes, we can elucidate the tissue-specific role of candidate target genes. The follow-up pathway analysis from tissue-specific genes for asthma shows that the immune system plays an essential function for asthma development in both thyroid and lung tissues.

Intra-session repeatability of quantitative metrics using widefield optical coherence tomography angiography (OCTA) in elderly subjects

PURPOSE

To assess the repeatability of retinal vascular metrics using different postprocessing methods as obtained from the swept-source optical coherence tomography angiography (SS-OCTA).

METHODS

Thirty-two participants (63% males; mean [SD] age, 70 [7] years) underwent SS-OCTA imaging (PLEX® Elite 9000, Carl Zeiss Meditec, Inc., Dublin, USA). Each participant underwent 2 repeated scans of 2 scan protocols: a macular-centred 3 × 3-mm2 and a widefield 12 × 12-mm2 for a total of 4 acquisitions. Images of superficial vascular plexuses (SVP) and deep vascular plexuses (DVP) were processed using different filters to generate the perfusion density (PD) and vessel density (VD). Vessel enhancement filters ranged from vessel targeted (Hessian and Gabor filters), classical denoising (Gaussian filter), to a scale-selective adaption (modified Bayesian residual transform [MBRT]). Intra-session repeatability of the different filters and their correlation with the original data set were calculated with the intraclass correlation coefficient (ICC) and Pearson's r.

RESULTS

Of the 32 eyes, 17 and 15 were right and left eyes, respectively. For 3 × 3-mm2 scans, both MBRT and Gabor filters yielded very good repeatable PD and VD (both ICCs > 0.87) values. Gabor filter was the most correlated with the original data set for the OCTA metrics (r = 0.95-0.97). For 12 × 12-mm2 scans, MBRT filter produced good-to-moderate ICC values for SVP (ICC>0.89) and DVP (ICC>0.73) metrics. Both the MBRT and Gabor filters were highly correlated with the original 12 × 12-mm2 scan data set (r = 0.96-0.98). The ICCs for the agreement between 3 × 3-mm2 and cropped 12 × 12-mm2 were high only for the PD values at the SVP layer and were poor for the VD at SVP and DVP measurements (ICC < 0.50).

CONCLUSION

Our findings show that with the proper choice of postimaging processing methods, SS-OCTA metrics can be obtained with high repeatability, which supports its use in various clinical settings.

Optical Coherence Tomography Angiography in Diabetes and Diabetic Retinopathy

Diabetic retinopathy (DR) is a common complication of diabetes mellitus that disrupts the retinal microvasculature and is a leading cause of vision loss globally. Recently, optical coherence tomography angiography (OCTA) has been developed to image the retinal microvasculature, by generating 3-dimensional images based on the motion contrast of circulating blood cells. OCTA offers numerous benefits over traditional fluorescein angiography in visualizing the retinal vasculature in that it is non-invasive and safer; while its depth-resolved ability makes it possible to visualize the finer capillaries of the retinal capillary plexuses and choriocapillaris. High-quality OCTA images have also enabled the visualization of features associated with DR, including microaneurysms and neovascularization and the quantification of alterations in retinal capillary and choriocapillaris, thereby suggesting a promising role for OCTA as an objective technology for accurate DR classification. Of interest is the potential of OCTA to examine the effect of DR on individual retinal layers, and to detect DR even before it is clinically detectable on fundus examination. We will focus the review on the clinical applicability of OCTA derived quantitative metrics that appear to be clinically relevant to the diagnosis, classification, and management of patients with diabetes or DR. Future studies with longitudinal design of multiethnic multicenter populations, as well as the inclusion of pertinent systemic information that may affect vascular changes, will improve our understanding on the benefit of OCTA biomarkers in the detection and progression of DR.

Mitochondria-enriched protrusions are associated with brain and intestinal stem cells in Drosophila

Brain stem cells stop dividing in late Drosophila embryos and begin dividing again in early larvae after feeding induces reactivation. Quiescent neural stem cells (qNSCs) display an unusual cytoplasmic protrusion that is no longer present in reactivated NSCs. The protrusions join the qNSCs to the neuropil, brain regions that are thought to maintain NSCs in an undifferentiated state, but the function of the protrusions is not known. Here we show that qNSC protrusions contain clustered mitochondria that are likely maintained in position by slow forward-and-backward microtubule growth. Larvae treated with a microtubule-stabilizing drug show bundled microtubules and enhanced mitochondrial clustering in NSCs, together with reduced qNSC reactivation. We further show that intestinal stem cells contain mitochondria-enriched protrusions. The qNSC and intestinal stem-cell protrusions differ from previously reported cytoplasmic extensions by forming stem-cell-to-niche mitochondrial bridges that could potentially both silence genes and sense signals from the stem cell niche.

Maternal immunity and antibodies to dengue virus promote infection and Zika virus-induced microcephaly in fetuses

Zika virus (ZIKV), an emergent flaviviral pathogen, has been linked to microcephaly in neonates. Although the risk is greatest during the first trimester of pregnancy in humans, timing alone cannot explain why maternal ZIKV infection leads to severe microcephaly in some fetuses, but not others. The antigenic similarities between ZIKV and dengue virus (DENV), combined with high levels of DENV immunity among ZIKV target populations in recent outbreaks, suggest that anti-DENV maternal antibodies could promote ZIKV-induced microcephaly. We demonstrated maternal-to-fetal ZIKV transmission, fetal infection, and disproportionate microcephaly in immunocompetent mice. We show that DENV-specific antibodies in ZIKV-infected pregnant mice enhance vertical ZIKV transmission and result in a severe microcephaly-like syndrome, which was dependent on the neonatal Fc receptor, FcRN. This novel immune-mediated mechanism of vertical transmission of viral infection is of special concern because ZIKV epidemic regions are also endemic to DENV.

Flavivirus serocomplex cross-reactive immunity is protective by activating heterologous memory CD4 T cells

How previous immunity influences immune memory recall and protection against related flaviviruses is largely unknown, yet encounter with multiple flaviviruses in a lifetime is increasingly likely. Using sequential challenges with dengue virus (DENV), yellow fever virus (YFV), and Japanese encephalitis virus (JEV), we induced cross-reactive cellular and humoral immunity among flaviviruses from differing serocomplexes. Antibodies against JEV enhanced DENV replication; however, JEV immunity was protective in vivo during secondary DENV1 infection, promoting rapid gains in antibody avidity. Mechanistically, JEV immunity activated dendritic cells and effector memory T cells, which developed a T follicular helper cell phenotype in draining lymph nodes upon secondary DENV1 infection. We identified cross-reactive epitopes that promote recall from a pool of flavivirus serocomplex cross-reactive memory CD4 T cells and confirmed that a similar serocomplex cross-reactive immunity occurs in humans. These results show that sequential immunizations for flaviviruses sharing CD4 epitopes should promote protection during a subsequent heterologous infection.