Artificial Intelligence in Cardiovascular Care - Part 2: Applications: JACC Review Topic of the Week.

Recent Artificial Intelligence (AI) advancements in cardiovascular care offer potential enhancements in effective diagnosis, treatment, and outcomes. Over 600 Food and Drug Administration (FDA)-approved clinical AI algorithms now exist, with 10% focusing on cardiovascular applications, highlighting the growing opportunities for AI to augment care. This review discusses the latest advancements in the field of AI, with a particular focus on the utilization of multimodal inputs and the field of generative AI. Further discussions in this review involve an approach to understanding the larger context in which AI-augmented care may exist, and include a discussion of the need for rigorous evaluation, appropriate infrastructure for deployment, ethics and equity assessments, regulatory oversight, and viable business cases for deployment. Embracing this rapidly evolving technology while setting an appropriately high evaluation benchmark with careful and patient-centered implementation will be crucial for cardiology to leverage AI to enhance patient care and the provider experience.

Artificial Intelligence for Cardiovascular Care - Part 1: Advances: JACC Review Topic of the Week.

Recent AI advancements in cardiovascular care offer potential enhancements in diagnosis, treatment, and outcomes. Innovations to date focus on automating measurements, enhancing image quality, and detecting diseases using novel methods. Applications span wearables, electrocardiograms, echocardiography, angiography, genetics, and more. AI models detect diseases from electrocardiograms at accuracy not previously achieved by technology or human experts, including reduced ejection fraction, valvular heart disease, and other cardiomyopathies. However, AI's unique characteristics necessitates rigorous validation by addressing training methods, real-world efficacy, equity concerns, and long-term reliability. Despite an exponentially growing number of studies in cardiovascular AI, trials showing improvement in outcomes remain lacking. A number are currently underway. Embracing this rapidly evolving technology while setting a high evaluation benchmark will be crucial for cardiology to leverage AI to enhance patient care and the provider experience.

Artificial intelligence in neurology: opportunities, challenges, and policy implications.

Neurological conditions are the leading cause of disability and mortality combined, demanding innovative, scalable, and sustainable solutions. Brain health has become a global priority with adoption of the World Health Organization's Intersectoral Global Action Plan in 2022. Simultaneously, rapid advancements in artificial intelligence (AI) are revolutionizing neurological research and practice. This scoping review of 66 original articles explores the value of AI in neurology and brain health, systematizing the landscape for emergent clinical opportunities and future trends across the care trajectory: prevention, risk stratification, early detection, diagnosis, management, and rehabilitation. AI's potential to advance personalized precision neurology and global brain health directives hinges on resolving core challenges across four pillars-models, data, feasibility/equity, and regulation/innovation-through concerted pursuit of targeted recommendations. Paramount actions include swift, ethical, equity-focused integration of novel technologies into clinical workflows, mitigating data-related issues, counteracting digital inequity gaps, and establishing robust governance frameworks balancing safety and innovation.

Development of pathophysiologically relevant models of sickle cell disease and ß-thalassemia for therapeutic studies.

Ex vivo cellular system that accurately replicates sickle cell disease and ß-thalassemia characteristics is a highly sought-after goal in the field of erythroid biology. In this study, we present the generation of erythroid progenitor lines with sickle cell disease and ß-thalassemia mutation using CRISPR/Cas9. The disease cellular models exhibit similar differentiation profiles, globin expression and proteome dynamics as patient-derived hematopoietic stem/progenitor cells. Additionally, these cellular models recapitulate pathological conditions associated with both the diseases. Hydroxyurea and pomalidomide treatment enhanced fetal hemoglobin levels. Notably, we introduce a therapeutic strategy for the above diseases by recapitulating the HPFH3 genotype, which reactivates fetal hemoglobin levels and rescues the disease phenotypes, thus making these lines a valuable platform for studying and developing new therapeutic strategies. Altogether, we demonstrate our disease cellular systems are physiologically relevant and could prove to be indispensable tools for disease modeling, drug screenings and cell and gene therapy-based applications.

An intentional approach to managing bias in general purpose embedding models.

Advances in machine learning for health care have brought concerns about bias from the research community; specifically, the introduction, perpetuation, or exacerbation of care disparities. Reinforcing these concerns is the finding that medical images often reveal signals about sensitive attributes in ways that are hard to pinpoint by both algorithms and people. This finding raises a question about how to best design general purpose pretrained embeddings (GPPEs, defined as embeddings meant to support a broad array of use cases) for building downstream models that are free from particular types of bias. The downstream model should be carefully evaluated for bias, and audited and improved as appropriate. However, in our view, well intentioned attempts to prevent the upstream components-GPPEs-from learning sensitive attributes can have unintended consequences on the downstream models. Despite producing a veneer of technical neutrality, the resultant end-to-end system might still be biased or poorly performing. We present reasons, by building on previously published data, to support the reasoning that GPPEs should ideally contain as much information as the original data contain, and highlight the perils of trying to remove sensitive attributes from a GPPE. We also emphasise that downstream prediction models trained for specific tasks and settings, whether developed using GPPEs or not, should be carefully designed and evaluated to avoid bias that makes models vulnerable to issues such as distributional shift. These evaluations should be done by a diverse team, including social scientists, on a diverse cohort representing the full breadth of the patient population for which the final model is intended.

Genetic Discovery Enabled by A Large Language Model.

Artificial intelligence (AI) has been used in many areas of medicine, and recently large language models (LLMs) have shown potential utility for clinical applications. However, since we do not know if the use of LLMs can accelerate the pace of genetic discovery, we used data generated from mouse genetic models to investigate this possibility. We examined whether a recently developed specialized LLM (Med-PaLM 2) could analyze sets of candidate genes generated from analysis of murine models of biomedical traits. In response to free-text input, Med-PaLM 2 correctly identified the murine genes that contained experimentally verified causative genetic factors for six biomedical traits, which included susceptibility to diabetes and cataracts. Med-PaLM 2 was also able to analyze a list of genes with high impact alleles, which were identified by comparative analysis of murine genomic sequence data, and it identified a causative murine genetic factor for spontaneous hearing loss. Based upon this Med-PaLM 2 finding, a novel bigenic model for susceptibility to spontaneous hearing loss was developed. These results demonstrate Med-PaLM 2 can analyze gene-phenotype relationships and generate novel hypotheses, which can facilitate genetic discovery.

Sustained pigmentation causes DNA damage and invokes translesion polymerase Polκ for repair in melanocytes.

Melanin protects skin cells from ultraviolet radiation-induced DNA damage. However, intermediates of eumelanin are highly reactive quinones that are potentially genotoxic. In this study, we systematically investigate the effect of sustained elevation of melanogenesis and map the consequent cellular repair response of melanocytes. Pigmentation increases γH2AX foci, DNA abasic sites, causes replication stress and invokes translesion polymerase Polκ in primary human melanocytes, as well as mouse melanoma cells. Confirming the causal link, CRISPR-based genetic ablation of tyrosinase results in depigmented cells with low Polκ levels. During pigmentation, Polκ activates replication stress response and keeps a check on uncontrolled proliferation of cells harboring melanin-damaged DNA. The mutational landscape observed in human melanoma could in part explain the error-prone bypass of DNA lesions by Polκ, whose absence would lead to genome instability. Thereby, translesion polymerase Polκ is a critical response of pigmenting melanocytes to combat melanin-induced DNA alterations. Our study illuminates the dark side of melanin and identifies (eu)melanogenesis as a key missing link between tanning response and mutagenesis, mediated via the necessary evil translesion polymerase, Polκ.

Enhancing the reliability and accuracy of AI-enabled diagnosis via complementarity-driven deferral to clinicians.

Predictive artificial intelligence (AI) systems based on deep learning have been shown to achieve expert-level identification of diseases in multiple medical imaging settings, but can make errors in cases accurately diagnosed by clinicians and vice versa. We developed Complementarity-Driven Deferral to Clinical Workflow (CoDoC), a system that can learn to decide between the opinion of a predictive AI model and a clinical workflow. CoDoC enhances accuracy relative to clinician-only or AI-only baselines in clinical workflows that screen for breast cancer or tuberculosis (TB). For breast cancer screening, compared to double reading with arbitration in a screening program in the UK, CoDoC reduced false positives by 25% at the same false-negative rate, while achieving a 66% reduction in clinician workload. For TB triaging, compared to standalone AI and clinical workflows, CoDoC achieved a 5-15% reduction in false positives at the same false-negative rate for three of five commercially available predictive AI systems. To facilitate the deployment of CoDoC in novel futuristic clinical settings, we present results showing that CoDoC's performance gains are sustained across several axes of variation (imaging modality, clinical setting and predictive AI system) and discuss the limitations of our evaluation and where further validation would be needed. We provide an open-source implementation to encourage further research and application.

Large language models encode clinical knowledge.

Large language models (LLMs) have demonstrated impressive capabilities, but the bar for clinical applications is high. Attempts to assess the clinical knowledge of models typically rely on automated evaluations based on limited benchmarks. Here, to address these limitations, we present MultiMedQA, a benchmark combining six existing medical question answering datasets spanning professional medicine, research and consumer queries and a new dataset of medical questions searched online, HealthSearchQA. We propose a human evaluation framework for model answers along multiple axes including factuality, comprehension, reasoning, possible harm and bias. In addition, we evaluate Pathways Language Model1 (PaLM, a 540-billion parameter LLM) and its instruction-tuned variant, Flan-PaLM2 on MultiMedQA. Using a combination of prompting strategies, Flan-PaLM achieves state-of-the-art accuracy on every MultiMedQA multiple-choice dataset (MedQA3, MedMCQA4, PubMedQA5 and Measuring Massive Multitask Language Understanding (MMLU) clinical topics6), including 67.6% accuracy on MedQA (US Medical Licensing Exam-style questions), surpassing the prior state of the art by more than 17%. However, human evaluation reveals key gaps. To resolve this, we introduce instruction prompt tuning, a parameter-efficient approach for aligning LLMs to new domains using a few exemplars. The resulting model, Med-PaLM, performs encouragingly, but remains inferior to clinicians. We show that comprehension, knowledge recall and reasoning improve with model scale and instruction prompt tuning, suggesting the potential utility of LLMs in medicine. Our human evaluations reveal limitations of today's models, reinforcing the importance of both evaluation frameworks and method development in creating safe, helpful LLMs for clinical applications.

Robust and data-efficient generalization of self-supervised machine learning for diagnostic imaging.

Machine-learning models for medical tasks can match or surpass the performance of clinical experts. However, in settings differing from those of the training dataset, the performance of a model can deteriorate substantially. Here we report a representation-learning strategy for machine-learning models applied to medical-imaging tasks that mitigates such 'out of distribution' performance problem and that improves model robustness and training efficiency. The strategy, which we named REMEDIS (for 'Robust and Efficient Medical Imaging with Self-supervision'), combines large-scale supervised transfer learning on natural images and intermediate contrastive self-supervised learning on medical images and requires minimal task-specific customization. We show the utility of REMEDIS in a range of diagnostic-imaging tasks covering six imaging domains and 15 test datasets, and by simulating three realistic out-of-distribution scenarios. REMEDIS improved in-distribution diagnostic accuracies up to 11.5% with respect to strong supervised baseline models, and in out-of-distribution settings required only 1-33% of the data for retraining to match the performance of supervised models retrained using all available data. REMEDIS may accelerate the development lifecycle of machine-learning models for medical imaging.

Respiratory Quinone Switches from Menaquinone to Polyketide Quinone during the Development Cycle in Streptomyces sp. Strain MNU77.

Type III polyketide synthases (PKSs) found across Streptomyces species are primarily known for synthesis of a vast repertoire of clinically and industrially relevant secondary metabolites. However, our understanding of the functional relevance of these bioactive metabolites in Streptomyces physiology is still limited. Recently, a role of type III PKS harboring gene cluster in producing alternate electron carrier, polyketide quinone (PkQ) was established in a related member of the Actinobacteria, Mycobacteria, highlighting the critical role these secondary metabolites play in primary cellular metabolism of the producer organism. Here, we report the developmental stage-specific transcriptional regulation of homologous type III PKS containing gene cluster in freshwater Streptomyces sp. strain MNU77. Gene expression analysis revealed the type III PKS gene cluster to be stringently regulated, with significant upregulation observed during the dormant sporulation stage of Streptomyces sp. MNU77. In contrast, the expression levels of only known electron carrier, menaquinone biosynthetic genes were interestingly found to be downregulated. Our liquid chromatography-high-resolution mass spectrometry (LC-HRMS) analysis of a metabolite extract from the Streptomyces sp. MNU77 spores also showed 10 times more metabolic abundance of PkQs than menaquinones. Furthermore, through heterologous complementation studies, we demonstrate that Streptomyces sp. MNU77 type III PKS rescues a respiratory defect of the Mycobacterium smegmatis type III PKS deletion mutant. Together, our studies reveal that freshwater Streptomyces sp. MNU77 robustly produces novel PkQs during the sporulation stage, suggesting utilization of PkQs as alternate electron carriers across Actinobacteria during dormant hypoxic conditions. IMPORTANCE The complex developmental life cycle of Streptomyces sp. mandates efficient cellular respiratory reconfiguration for a smooth transition from aerated nutrient-rich vegetative hyphal growth to the hypoxic-dormant sporulation stage. Polyketide quinones (PkQs) have recently been identified as a class of alternate electron carriers from a related member of the Actinobacteria, Mycobacteria, that facilitates maintenance of membrane potential in oxygen-deficient niches. Our studies with the newly identified freshwater Streptomyces sp. strain MNU77 show conditional transcriptional upregulation and metabolic abundance of PkQs in the spore state of the Streptomyces life cycle. In parallel, the levels of menaquinones, the only known Streptomyces electron carrier, were downregulated, suggesting deployment of PkQs as universal electron carriers in low-oxygen, unfavorable conditions across the Actinobacteria family.

Temporal analysis of melanogenesis identifies fatty acid metabolism as key skin pigment regulator.

Therapeutic methods to modulate skin pigmentation has important implications for skin cancer prevention and for treating cutaneous hyperpigmentary conditions. Towards defining new potential targets, we followed temporal dynamics of melanogenesis using a cell-autonomous pigmentation model. Our study elucidates 3 dominant phases of synchronized metabolic and transcriptional reprogramming. The melanogenic trigger is associated with high MITF levels along with rapid uptake of glucose. The transition to pigmented state is accompanied by increased glucose channelisation to anabolic pathways that support melanosome biogenesis. SREBF1-mediated up-regulation of fatty acid synthesis results in a transient accumulation of lipid droplets and enhancement of fatty acids oxidation through mitochondrial respiration. While this heightened bioenergetic activity is important to sustain melanogenesis, it impairs mitochondria lately, shifting the metabolism towards glycolysis. This recovery phase is accompanied by activation of the NRF2 detoxication pathway. Finally, we show that inhibitors of lipid metabolism can resolve hyperpigmentary conditions in a guinea pig UV-tanning model. Our study reveals rewiring of the metabolic circuit during melanogenesis, and fatty acid metabolism as a potential therapeutic target in a variety of cutaneous diseases manifesting hyperpigmentary phenotype.

Reverse Genetic Approach to Identify Regulators of Pigmentation using Zebrafish.

Melanocytes are specialized neural crest-derived cells present in the epidermal skin. These cells synthesize melanin pigment that protects the genome from harmful ultraviolet radiations. Perturbations in melanocyte functioning lead to pigmentary disorders such as piebaldism, albinism, vitiligo, melasma, and melanoma. Zebrafish is an excellent model system to understand melanocyte functions. The presence of conspicuous pigmented melanocytes, ease of genetic manipulation, and availability of transgenic fluorescent lines facilitate the study of pigmentation. This study employs the use of wild-type and transgenic zebrafish lines that drive green fluorescent protein (GFP) expression under mitfa and tyrp1 promoters that mark various stages of melanocytes. Morpholino-based silencing of candidate genes is achieved to evaluate the phenotypic outcome on larval pigmentation and is applicable to screen for regulators of pigmentation. This protocol demonstrates the method from microinjection to imaging and fluorescence-activated cell sorting (FACS)-based dissection of phenotypes using two candidate genes, carbonic anhydrase 14 (Ca14) and a histone variant (H2afv), to comprehensively assess the pigmentation outcome. Further, this protocol demonstrates segregating candidate genes into melanocyte specifiers and differentiators that selectively alter melanocyte numbers and melanin content per cell, respectively.

Kupyaphores are zinc homeostatic metallophores required for colonization of Mycobacterium tuberculosis.

Mycobacterium tuberculosis (Mtb) endures a combination of metal scarcity and toxicity throughout the human infection cycle, contributing to complex clinical manifestations. Pathogens counteract this paradoxical dysmetallostasis by producing specialized metal trafficking systems. Capture of extracellular metal by siderophores is a widely accepted mode of iron acquisition, and Mtb iron-chelating siderophores, mycobactin, have been known since 1965. Currently, it is not known whether Mtb produces zinc scavenging molecules. Here, we characterize low-molecular-weight zinc-binding compounds secreted and imported by Mtb for zinc acquisition. These molecules, termed kupyaphores, are produced by a 10.8 kbp biosynthetic cluster and consists of a dipeptide core of ornithine and phenylalaninol, where amino groups are acylated with isonitrile-containing fatty acyl chains. Kupyaphores are stringently regulated and support Mtb survival under both nutritional deprivation and intoxication conditions. A kupyaphore-deficient Mtb strain is unable to mobilize sufficient zinc and shows reduced fitness upon infection. We observed early induction of kupyaphores in Mtb-infected mice lungs after infection, and these metabolites disappeared after 2 wk. Furthermore, we identify an Mtb-encoded isonitrile hydratase, which can possibly mediate intracellular zinc release through covalent modification of the isonitrile group of kupyaphores. Mtb clinical strains also produce kupyaphores during early passages. Our study thus uncovers a previously unknown zinc acquisition strategy of Mtb that could modulate host-pathogen interactions and disease outcome.

Does your dermatology classifier know what it doesn't know? Detecting the long-tail of unseen conditions.

Supervised deep learning models have proven to be highly effective in classification of dermatological conditions. These models rely on the availability of abundant labeled training examples. However, in the real-world, many dermatological conditions are individually too infrequent for per-condition classification with supervised learning. Although individually infrequent, these conditions may collectively be common and therefore are clinically significant in aggregate. To prevent models from generating erroneous outputs on such examples, there remains a considerable unmet need for deep learning systems that can better detect such infrequent conditions. These infrequent 'outlier' conditions are seen very rarely (or not at all) during training. In this paper, we frame this task as an out-of-distribution (OOD) detection problem. We set up a benchmark ensuring that outlier conditions are disjoint between the model training, validation, and test sets. Unlike traditional OOD detection benchmarks where the task is to detect dataset distribution shift, we aim at the more challenging task of detecting subtle differences resulting from a different pathology or condition. We propose a novel hierarchical outlier detection (HOD) loss, which assigns multiple abstention classes corresponding to each training outlier class and jointly performs a coarse classification of inliers vs. outliers, along with fine-grained classification of the individual classes. We demonstrate that the proposed HOD loss based approach outperforms leading methods that leverage outlier data during training. Further, performance is significantly boosted by using recent representation learning methods (BiT, SimCLR, MICLe). Further, we explore ensembling strategies for OOD detection and propose a diverse ensemble selection process for the best result. We also perform a subgroup analysis over conditions of varying risk levels and different skin types to investigate how OOD performance changes over each subgroup and demonstrate the gains of our framework in comparison to baseline. Furthermore, we go beyond traditional performance metrics and introduce a cost matrix for model trust analysis to approximate downstream clinical impact. We use this cost matrix to compare the proposed method against the baseline, thereby making a stronger case for its effectiveness in real-world scenarios.

Baseline cell proliferation rates and response to UV differ in lymphoblastoid cell lines derived from healthy individuals of extreme constitution types.

Differences in human phenotypes and susceptibility to complex diseases are an outcome of genetic and environmental interactions. This is evident in diseases that progress through a common set of intermediate patho-endophenotypes. Precision medicine aims to delineate molecular players for individualized and early interventions. Functional studies of lymphoblastoid cell line (LCL) model of phenotypically well-characterized healthy individuals can help deconvolute and validate these molecular mechanisms. In this study, LCLs are developed from eight healthy individuals belonging to three extreme constitution types, deep phenotyped on the basis of Ayurveda. LCLs were characterized by karyotyping and immunophenotyping. Growth characteristics and response to UV were studied in these LCLs. Significant differences in cell proliferation rates were observed between the contrasting groups such that one type (Kapha) proliferates significantly slower than the other two (Vata, Pitta). In response to UV, one of the fast growing groups (Vata) shows higher cell death but recovers its numbers due to an inherent higher rates of proliferation. This study reveals that baseline differences in cell proliferation could be a key to understanding the survivability of cells under UV stress. Variability in baseline cellular phenotypes not only explains the cellular basis of different constitution types but can also help set priors during the design of an individualized therapy with DNA damaging agents. This is the first study of its kind that shows variability of intermediate patho-phenotypes among healthy individuals with potential implications in precision medicine.

A deep learning system for differential diagnosis of skin diseases.

Skin conditions affect 1.9 billion people. Because of a shortage of dermatologists, most cases are seen instead by general practitioners with lower diagnostic accuracy. We present a deep learning system (DLS) to provide a differential diagnosis of skin conditions using 16,114 de-identified cases (photographs and clinical data) from a teledermatology practice serving 17 sites. The DLS distinguishes between 26 common skin conditions, representing 80% of cases seen in primary care, while also providing a secondary prediction covering 419 skin conditions. On 963 validation cases, where a rotating panel of three board-certified dermatologists defined the reference standard, the DLS was non-inferior to six other dermatologists and superior to six primary care physicians (PCPs) and six nurse practitioners (NPs) (top-1 accuracy: 0.66 DLS, 0.63 dermatologists, 0.44 PCPs and 0.40 NPs). These results highlight the potential of the DLS to assist general practitioners in diagnosing skin conditions.

Histone variant dictates fate biasing of neural crest cells to melanocyte lineage.

In the neural crest lineage, progressive fate restriction and stem cell assignment are crucial for both development and regeneration. Whereas fate commitment events have distinct transcriptional footprints, fate biasing is often transitory and metastable, and is thought to be moulded by epigenetic programmes. Therefore, the molecular basis of specification is difficult to define. In this study, we established a role for a histone variant, H2a.z.2, in specification of the melanocyte lineage from multipotent neural crest cells. H2a.z.2 silencing reduces the number of melanocyte precursors in developing zebrafish embryos and from mouse embryonic stem cells in vitro We demonstrate that this histone variant occupies nucleosomes in the promoter of the key melanocyte determinant mitf, and enhances its induction. CRISPR/Cas9-based targeted mutagenesis of this gene in zebrafish drastically reduces adult melanocytes, as well as their regeneration. Thereby, our study establishes the role of a histone variant upstream of the core gene regulatory network in the neural crest lineage. This epigenetic mark is a key determinant of cell fate and facilitates gene activation by external instructive signals, thereby establishing melanocyte fate identity.