New York University School Of Medicine

NIH Research Projects · FY 2025 · 2025-09

Project Summary/Abstract We will conduct a detailed analysis of the long-term health outcomes of living kidney donors (LKDs) utilizing the comprehensive Cosmos dataset, which includes electronic health records (EHR) from over 245 million individuals, among them 90,000 LKDs. Our goal is to understand how reduced kidney function after donation impacts the health of LKDs compared to the general population. We aim to develop new clinical practices and care guidelines specifically for LKDs, addressing this critical gap in care. We will examine the natural history of kidney function, albuminuria, and systolic blood pressure (SBP) in LKDs, adjusting for variables such as age, sex, and medication use to accurately model these health indicators, thus providing an in-depth view of the health trajectory of LKDs following their donation. Furthermore, we will validate and possibly recalibrate the CKD-PC and PREVENT equations within our LKD cohort. The CKD-PC (Chronic Kidney Disease Prognosis Consortium) equations are used to predict renal outcomes based on factors such as age, sex, GFR, and albuminuria levels, originally validated in the general CKD population. The PREVENT equations are used to predict cardiovascular outcomes by integrating kidney function measures and have informed AHA guidelines on statin use for persons with CKD. We will adjust these well-established equations to more accurately reflect the unique post-donation physiological changes that LKDs experience, which will improve their utility in predicting risks such as cardiovascular diseases and end-stage kidney disease for this specific population. In addition, we plan to develop LKD-specific adaptations of the KFRE (Kidney Failure Risk Equation) predictive models. KFRE is used to estimate the risk of progression to kidney failure based on clinical inputs, including kidney function measures. We will integrate LKD status into the models, modifying parameters to cater specifically to the nuances of kidney donor profiles to enable more precise and individualized risk assessments for LKDs, which could improve post-donation care and monitoring. The culmination of our project will be a virtual consensus conference, where we will engage with key stakeholders from major nephrology and transplantation organizations. At this conference, we will present our research findings, discuss potential updates to clinical practice guidelines for LKDs, and strategize on the most effective methods to disseminate our research outcomes to healthcare providers and the broader LKD community. By leveraging the largest national EHR data, which includes detailed information on over 90,000 racially representative kidney donors, we strive to enhance medical care for LKDs. We aim to equip healthcare providers with tools and knowledge to manage LKDs' health trajectories effectively, ensuring better outcomes for those who've altruistically donated kidneys. We anticipate establishing a new standard of care for living kidney donors across their lifespan.

NIH Research Projects · FY 2025 · 2025-09

Improving upstream components of health is critical for assuring improvement in longer term health and wellbeing. This is particularly relevant for pre-kindergarten education. While evidence exists on smaller scale programs that are not universal in nature, we have a paucity of information on the role of true universal expansions, at the level of a large city, on health or education outcomes for children at the population level. In 2014, New York City (NYC) launched a universal pre-kindergarten (UPK) expansion, more than tripling the number of slots available in 2 years. We will examine the role that his program had in both health and education outcomes for children in NYC through the school years, through 12th grade. UPK could influence a wide variety of health (e.g., mental health, asthma, vision/hearing, obesity) and education (e.g., test scores, attendance, retention) outcomes. The research leverages the NYC Student Population Health Registry (SPHR), a uniquely inclusive, longitudinal database of all NYC public school students, created jointly by the NYC Department of Health and Mental Hygiene and NYC Department of Education to examine these and other outcomes. SPHR links multiple data sources at the child-level, allowing us to examine the influence of UPK on myriad outcomes. The specific aims are: Aim 1: Determine the influence of NYC’s citywide UPK expansion on child health outcomes. H1: Relative to children who entered public school before UPK expansion, children who entered school after UPK expansion will have improved health outcomes. Aim 2: Determine the influence of NYC’s citywide UPK expansion on child education outcomes. H2: Relative to children who entered public school before UPK expansion, children who entered school after UPK expansion will have improved education outcomes. Aim 3: Determine the influence of NYC’s citywide UPK expansion on health and education for all children. H3: Different groups of children will see differential impacts from UPK on health and education outcomes.

NIH Research Projects · FY 2025 · 2025-09

PROJECT SUMMARY Epilepsy affects over 50 million people worldwide, with developmental and epileptic encephalopathies (DEEs) representing the most severe and treatment-resistant forms. DEEs are characterized by early-onset seizures and developmental delays, often caused by pathogenic variants in genes critical for brain function. Understanding these genetic underpinnings has revolutionized epilepsy research, paving the way for precision medicine—therapeutic strategies tailored to specific genetic and molecular mechanisms. Among these disorders, EEF1A2-related neurodevelopmental disorder (EEF1A2 syndrome) stands out as a rare yet compelling example, highlighting the need for targeted research to understand and treat epilepsy. Mutations in the EEF1A2 gene, which encodes the eukaryotic elongation factor 1A2 (eEF1A2), have been identified as a cause of severe neurodevelopmental disorders, including DEEs. eEF1A2 facilitates protein synthesis by delivering aminoacyl-tRNAs to ribosomes in a GTP-dependent manner. Approximately 50 distinct missense mutations have been reported, with recurrent variants such as G70S, E122K, and D252H associated with severe intellectual disability. These mutations map to key functional domains of eEF1A2—GTP binding (G70S), tRNA binding (E122K), and actin binding (D252H)—and exhibit distinct clinical phenotypes. While G70S and E122K are linked to early-onset epilepsy, D252H is associated with milder or no epilepsy. The role of eEF1A2 in actin dynamics through its third functional domain is also unclear, suggesting potential additional roles in neurons. This project aims to address these gaps through two specific aims: 1. Determining how EEF1A2 mutations alter protein synthesis in human neurons. 2. Determining how EEF1A2 mutations alter actin dynamics, neuronal morphology, and development. Previous studies have been limited by the species-specific isoform switch between eEF1A1 and eEF1A2 that occur in humans but are not fully replicated in mouse models. By utilizing human-induced pluripotent stem cells (hiPSCs) differentiated into cortical neurons, this research will overcome these limitations by determining the alterations due to these mutations on protein synthesis, elongation rate, translatome profile, and translational efficiency (Aim 1). Additionally, alterations in actin mobility and subcellular structure, as well as neuronal morphology, synapse number, and neuronal electrophysiological properties will be determined (Aim 2). The findings will advance our understanding of eEF1A2's neuron-specific roles and inform precision medicine strategies for EEF1A2 syndrome treatment.

NIH Research Projects · FY 2026 · 2025-09

PROJECT SUMMARY/ABSTRACT Disparities in weight outcomes begin in infancy and persist across the lifespan, fueling obesity rates and its comorbid diseases. Most current prevention strategies are narrow in scope, coaching parents to invest substantial effort into managing their infant’s feeding and activity practices. These approaches face limitations in families with financial challenges, as they compete directly with efforts to manage co-existing food, housing, or utility needs. Social risks account for up to 27.4% of obesity outcomes, yet only 1.6% of strategies target them. The Medicaid Section 1115 Demonstration Waiver enables each US state to customize its Medicaid program to tailor disease prevention for its specific population. Researchers have shown that screening for social needs — adverse social conditions unevenly distributed across racial and ethnic communities — may be more effective than genetics screening in predicting health outcomes. Subsequently, New York state was awarded a temporary Waiver to reimburse health and community systems for social needs screening and resource connections, with targeted eligibility for early maternal-child health. This proposal represents a rare and scientifically significant opportunity to identify policy strategies that may prevent racial, ethnic, and socioeconomic disparities in obesity. We propose a strategic study design that capitalizes on the time-sensitive window to evaluate New York’s Medicaid 1115 Waiver, set to expire on March 31, 2027. This proposal’s overall goal is to evaluate impacts of New York’s Waiver Demonstration on early obesity risk in New York City Health + Hospitals, the largest US public hospital system. Our specific aims are: 1) Evaluate whether a greater degree of Waiver exposure is associated with healthier weight outcomes, as well as healthier infant feeding and activity practices; 2) Describe stakeholder attitudes and perspectives on facilitators and barriers to Waiver implementation. AIM 2’s qualitative exploration will identify contextual factors that influence real-world implementation, informing the predictive framework in AIM 3; 3) Assess trends in infant weight outcomes before and after Waiver implementation and evaluate how stakeholder-identified facilitators and barriers from AIM 2 may influence Waiver impact. This proposal draws upon world-class clinical, research, and teaching resources available at NYU Grossman School of Medicine, New York City Health + Hospitals, and Stanford University School of Medicine. The PI has assembled a scientific team comprised of leading experts in pediatrics, health services research, social determinants of health, advanced biostatistics, epidemiology, early childhood obesity disparities, and policy implementation. This urgent research will provide actionable insights for other states considering similar Waivers and support integration of social needs into obesity prevention.

NIH Research Projects · FY 2025 · 2025-09

PROJECT SUMMARY/ABSTRACT Follicular lymphoma (FL) is an indolent, virtually incurable B-cell lymphoma and the most common small B-cell lymphoma in the United States. Most FL patients experience long-term survival rates similar to the general population; however, approximately 20% experience early disease progression within 24 months (POD24) of frontline chemoimmunotherapy (CIT)), often with histologic transformation (HT) to high-grade lymphoma and face a poor prognosis. When initiated early, intensified frontline treatment can improve outcomes in these high- risk patients, but currently, there are no reliable predictive biomarkers to guide frontline risk stratification and assignment to a risk-adapted treatment protocol at the time of diagnosis. The challenge of predictive biomarker discovery in FL is rooted, in significant part, in its heterogeneity both genetically and in the composition of the tumor microenvironment (TME). Consistent with these observations, the identification of recurrent genetic alterations in FL has led to only modest improvements in the predictive value of established clinical risk models. The m7-FLIPI, for example, currently represents one of the most robust clinicogenetic FL risk models to date for the prediction of POD24, yet performs with only 77% accuracy and a positive predictive value under 50%, suggesting that the determinants of early progression in FL extend beyond the clinicogenetic parameters that inform the current risk-classification paradigm. In addition, multiple in-depth genomic analyses of FL have failed to identify a single unifying genetic driver of transformation, supporting the concept that it more likely represents the culmination of multiple, progressively acquired genetic alterations. Several studies have identified components of the TME that contribute to disease pathogenesis or are correlated with clinical outcomes, including tumor-associated macrophage, T cells, and even components of the stroma, but these findings have not been translated to routine clinical practice. We have shown that the FL TME is enriched for transcriptional and functional subtypes of neoplastic B cells and non-B-cells and that TME alterations in cell composition and/or function can predict FL patient outcomes following CIT. To characterize the TME changes that are predictive of clinical outcomes, we will utilize bulk transcriptome, single-cell resolution protein, transcriptional profiling, and flow cytometry (FC) approaches to characterize pretreatment tissues and peripheral blood from well-annotated FL patient samples to identify biomarkers of CIT response. We will also elucidate the role of neutrophilic myeloid-derived suppressor cells (PMN-MDSCs) in FL development and therapy response using a novel mouse model of FL. In addition to identifying biomarkers of high-risk FL that can serve as the basis of novel predictive transcriptomic, immunohistochemical (IHC), or FC assays for use in the clinical setting, we expect these studies to improve our understanding of the biological mechanisms driving FL development and treatment outcomes.

NIH Research Projects · FY 2025 · 2025-09

PROJECT SUMMARY Cutaneous T cell lymphoma (CTCL) involves dermal expansion and accumulation of malignant CD4+ T cells. While putative triggers ranging from chemical exposures to microbial elements have been suggested, the etiology of CTCL is unknown. However, studies have observed increased skin colonization with S. aureus in CTCL patients compared to healthy controls, and a small trial documented disease regression in response to aggressive antibiotic treatment that eradicated S. aureus colonization. These findings suggest that S. aureus colonization may play a role in the progression of CTCL, although no unifying mechanism for this relationship has been uncovered to date. There are major gaps in our knowledge, which will be addressed by the experiments in this proposal: 1) Is S. aureus indeed enriched in CTCL patient skin microbiomes, and what strain(s) and virulence factor(s) are found in patient samples? And how does the presence of S. aureus impact malignant cell features in CTCL patient samples? 2) Is S. aureus contributing to CTCL progression in vivo? 3) And is in vivo CTCL progression mediated by an antigen-specific response in the skin? Prior studies have been limited by low patient numbers, as CTCL is a rare disease, and a lack of comprehensive integration of 16S sequencing, microbial molecular genetics, and malignant cell characteristics. This study will overcome these limitations through sample collection of over 200 CTCL patients for analysis, including 102 previously collected samples from collaborating clinicians, and an additional 100 CTCL samples to be collected through our active, IRB-approved human studies spanning multiple institutions. Microbial analysis will reveal the proportions of, and genetic characteristics of, S. aureus in the CTCL microbiome, and the relationship between the skin microbiome and human malignant T cell activation, as determined by single- cell surface expression and transcriptomics (Aim 1). In vivo mouse models of CTCL subjected to S. aureus colonization will allow study of skin-resident immune responses to S. aureus, and its mechanistic relationship to disease progression (Aim 2). And finally, sophisticated in vivo models of CTCL will be used to determine if topical antigenic stimulation is sufficient and necessary to drive disease progression in the skin (Aim 3).

NIH Research Projects · FY 2025 · 2025-09

Project Summary Prostate Cancer (PCA) is the second most prevalent malignancy and a leading cause of mortality among men in the United States. The urgency for early identification of Clinically Significant Prostate Cancer (CSPCA) within the at-risk population cannot be overstated, as it holds the key to significantly enhancing clinical outcomes. Yet, the current standard-of-care for CSPCA surveillance, the Prostate Specific Antigen (PSA) test, is plagued by low specificity, resulting in a cascade of unnecessary advanced diagnostic imaging and invasive biopsy procedures, inflicting a substantial wasteful cost burden on the health system and subjecting patients to avoidable trauma. Despite the established efficacy of Magnetic Resonance Imaging (MRI) in diagnosing CSPCA, it is not used as a first-in-line tool to identify patients with CSPCA at the population level. The reluctance to embrace MRI is rooted in the requirement to generate high-fidelity images, imposing inherent challenges – the need for expensive scanners installed in specialized imaging centers, executing complex and slow protocols tailored to acquire copious high- quality k-space data, and the dependence on sub-specialized radiologists for image interpretation. Driven by our mission to democratize MR diagnostics for widespread surveillance of critical illnesses like CSPCA on a population scale, we challenge the conventional wisdom that high-fidelity images are necessary for accurate disease inference. Instead, we advocate for inferring disease presence directly from a meticulously curated subset of degraded k-space data. We hypothesize that modern Machine Learning (ML) models can accurately infer the presence of the disease (a binary decision - which is all that may be required for disease surveillance) using only a minimal amount of carefully selected degraded k-space data. This carefully chosen degraded data, that would otherwise be insufficient to generate a diagnostic-quality image, can be swiftly acquired using inexpensive and accessible scanning devices, thereby paving the way for making MR technology accessible for disease identification at the Point-of-Care (POC). In this proposal we will first establish the feasibility of ML models to accurately detect CSPCA directly from a fixed under-sampled k-space data, without images (Aim 1). Subsequently, we will determine the minimum quantity of k-space data required for accurate disease detection, irrespective of image quality, by proposing an innovative end-to-end ML methodology that identifies a diminutive subset of k-space containing ample information for precise CSPCA inference and uses it to draw conclusions about CSPCA presence (Aim 2). Finally, we will aim to show that CSPCA can be accurately inferred from the degraded k-space data acquired using low-field scanners, without images (Aim 3). Findings of this project will establish the groundwork for affordable and accessible MR devices that maintain accuracy in detecting CSPCA. These devices, capable of operating closer to the POC for disease surveillance, will promote equitable access to MR diagnostics beyond specialized environments, revolutionizing disease surveillance on a broader scale.

NIH Research Projects · FY 2025 · 2025-09

SUMMARY Breast cancer is the most commonly diagnosed cancer among US women, accounting for ~30% of all new cancer cases in women. Identifying biomarkers associated with breast cancer risk may reveal new pathophysiological mechanisms and potentially lead to improvements in risk stratification. The immune system is known to play a critical role in cancer control, and experimental studies suggest that common autoantibodies (AAbs) are involved in the earliest phase of immunosurveillance, i.e. elimination of transforming cells. Common AAbs, which include both natural and acquired AAbs, are abundant in healthy individuals and have been implicated in maintaining immune homeostasis and clearing apoptotic and cancerous cells. Therefore, varying levels of certain common AAbs could indicate active immunosurveillance mechanisms that prevent cancer development through antibody-mediated mechanisms. Levels of AAbs are now measurable at the omics-scale in small volumes of serum. However, no prospective studies have searched for common AAbs associated with risk of breast cancer. The New York University Women's Health Study (NYUWHS) is a prospective cohort study that enrolled 14,274 women aged 35–65 years between 1985 and 1991. All women donated blood at the time of enrollment. The cohort is well suited to discover new biomarkers of breast cancer risk because of its biorepository, long follow-up, excellent cancer case ascertainment, and detailed data on breast cancer risk factors. In a study involving NYUWHS healthy participants and two serum samples collected 1 year apart from each woman, we tested for common AAbs using the HuProt human proteome microarray with >20,000 AAb probes. We found that common AAbs are abundant in healthy women and that levels of most of the AAbs are stable over time. Our goal is to identify a panel of circulating common AAbs that are protective against breast cancer development. We will continue cancer case ascertainment through linkages to the Virtual Pooled Registry Cancer Linkage System and electronic health records (EHRs), medical records review and data harmonization to expand our data on tumor subtype and increase the representation of non-white cases in the proposed study (Aim 1). We will conduct a prospective case–control study nested in the NYUWHS including 800 cases of invasive breast cancer and 1-to-1 matched controls, with 300 case–control pairs in a discovery phase and 500 pairs in a validation phase (Aim 2). We will use cases of invasive breast cancer who had serum samples collected >2 years before diagnosis. We will also identify correlates of AAbs (Aim 3). This study has the potential to advance our understanding of the pathophysiological processes of breast cancer.

NIH Research Projects · FY 2025 · 2025-09

PROJECT SUMMARY Hippocampal theta is one of the most prominent and well-studied neural oscillations in rodent exploration and memory. In rodents, sustained hippocampal theta (4–7 Hz) and coupled gamma (30–100 Hz) rhythms dominate the exploratory state and spatial memory, organizing and sequencing place information for later use. Yet whether mechanisms supporting rodent navigation apply to human episodic memory has been debated for decades. Unlike in rodents, human “theta” occurs in bouts, ranging between 1–13 Hz. The discrepancy between rodent and human theta may result from differences in recording practices, interspecies differences in hippocampal organization, and exclusion of exploratory visuomotor behavior in previous human experiments. Our goal is to define and test the relationship between human exploratory eye and body movements and hippocampal activity along the long axis, at the single unit and population level. We will examine three possible scenarios: (1) hippocampal theta oscillations occur independently of saccades and saccades reset the phase of ongoing theta, potentially to a different degree in the anterior and posterior segments; (2) hippocampal LFP and associated spiking activity is mainly evoked by saccades; and (3) bouts of internally generated theta waves are induced by saccades under cognitive demands. To test these competing models, we will leverage: 1) high- resolution hippocampal intracranial EEG (iEEG) in surgical patients as they participate in visual exploration while measuring single unit and population-level activity; and 2) chronic hippocampal iEEG recordings (Responsive Neurostimulation System, RNS, NeuroPace Inc) in ambulatory epilepsy patients. Chronic iEEG will be integrated with high-density scalp EEG (hd-EEG), and peripheral accessories that track eye and body movements. We propose the following: Aim 1. Define the neurophysiological mechanisms along the hippocampal long axis during visual exploration in surgical patients with high-resolution iEEG. Aim 2. Define the neurophysiological mechanisms along the long axis during ambulatory exploration in patients with chronic iEEG (RNS, NeuroPace). Our experiments will determine if and how hippocampal “theta” or other mechanisms are influenced by exploratory visuomotor behavior, thus referencing decades of rodent literature. Our proposal is innovative in concept and method because we: 1) measure spontaneous eye and body movements during naturalistic exploration, to resemble rodent navigation; 2) use micro/macro contacts to track hippocampal activity at the single unit and population level along the long axis; and 3) integrate chronic iEEG with scalp hd-EEG, eye- tracking, and peripheral accessories in a mobile setting. Our interdisciplinary team of clinicians, neuroscientists, and engineers in the fields of epilepsy, neuroscience, and vision rehabilitation medicine will advance a mechanistic explanation of human hippocampal neurophysiology relevant to a future U01 proposal investigating the hippocampal-neocortical mechanisms of naturalistic memory behaviors.

NIH Research Projects · FY 2025 · 2025-09

PROJECT SUMMARY Developmentally-defined subtypes of motor neurons with specialized anatomical and functional properties are required for facilitating movement. Exploring the relationships among developmental, anatomical and functional properties of spinal motor neuron subtypes unveiled fundamental insights into the mechanisms governing locomotion. It is not yet clear to what extent these organizing principles shape motor control outside the spinal cord. Extraocular motor neuron subtypes are similarly anatomically and functionally diverse. Defining the links between development, anatomy, and the role in behavior for individual extraocular motor neurons and how heterogeneity shapes population function is crucial for uncovering the fundamental principles underlying motor system organization. To date, the complexity and developmental inaccessibility of most vertebrate models precludes linking anatomy, birthdate, and function of extraocular motor neurons, thereby limiting our understanding of their individual and collective contributions to behavior. Using the small and accessible model vertebrate – the larval zebrafish – where functional extraocular motor neuron subtypes are genetically accessible therefore provides a unique opportunity to test the hypothesis that general principles organize vertebrate motor systems. Our lab’s previous work links motor neuron pool identity to birthdate and evaluates function by engaging the vestibulo- ocular reflex. My preliminary data indicates that, when mature, extraocular motor neuron responses vary considerably within individual motor pools. The goal of this proposal is to reveal the organization within heterogenous populations of extraocular motor neurons, and to define the behavioral impacts of such variability. Aim 1 will reveal the generality of principles that link developmental, anatomical, and functional properties of individual motor neurons. Aim 2 will define if/how heterogeneity within a population contributes to behavior, illuminating how subtypes of motor neurons work together to control movement. As my long-term goal is to investigate the neural basis of movement in health and disease, this work is foundational training on my path to an independent position. Beyond training, my data stand to speak to the generality of principles of spinal motor circuit organization. Completion of these experiments will take a critical step towards understanding motor system development and function in health and disease.

NIH Research Projects · FY 2025 · 2025-09

PROJECT SUMMARY Toll-like receptors (TLRs) are an evolutionarily ancient class of cell-surface molecules with well characterized roles in morphogen signaling and innate immunity. In these contexts, TLRs use a canonical signaling mechanism that involves recruitment of adapter proteins to the cytoplasmic domain of TLRs and subsequent activation of a kinase cascade that regulates transcriptional responses. Recently, we and others have found that TLRs are required for normal development and function of sensory circuits. Unlike morphogen signaling and innate immunity, sensory circuit development does not require canonical TLR signaling mechanisms. The non-canonical mechanisms by which TLRs regulate the development of sensory circuits are unknown. I propose studies of the sole TLR encoded by the C. elegans genome, TOL-1, to reveal how non- canonical TLR signaling regulates sensory system development and function and to identify molecular mechanisms of non-canonical TLR signaling. My preliminary data show that TOL-1 is highly enriched in in the developing neuropil of embryonic nervous system. Using a genetically engineered conditional allele of tol-1, I found that depletion of TOL-1/TLR from sensory neurons causes marked defects in chemotaxis behavior. My preliminary data further show that TOL-1 does not require its intracellular domain to function in sensory neurons, indicating that its function is independent of canonical TLR signaling mechanisms. l will use C. elegans genetics, genomics, and high resolution / super-resolution microscopy to (1) determine how TLRs regulate the establishment and maintenance of sensory neuron architecture and function, (2) test the hypothesis that TLRs that lack the ability to directly interact with intracellular partners are nevertheless able to regulate gene expression in sensory neurons, and (3) test the hypothesis that TLR function in sensory neurons requires interactions with specific extracellular partners. The proposed studies will advance understanding of how a highly conserved family of cell-surface receptors functions in sensory circuits.

NIH Research Projects · FY 2025 · 2025-09

SUMMARY Externalizing problems, characterized by disruptive conduct, aggression, hyperactivity, attentional problems, and the inability to adequately adapt to new situations, are leading contributors to health burden amongst children and adolescents globally. A recent mental health surveillance among children in the United States indicates one in 11 children have an attention deficit hyperactivity disorder (ADHD) diagnosis. Despite the extraordinarily improved outcomes associated with early diagnosis and targeted intervention, externalizing disorders such as ADHD and oppositional defiance disorder (ODD) are usually only diagnosed after symptoms have clearly emerged and are causing significant impairments in daily functioning. This critical need for early- life prediction of externalizing disorders can now be addressed, thanks to explosive advances in artificial intelligence (AI) and widespread accessibility of remote technology. There is untapped opportunity to bring simple, scalable and accessible solutions to families to improve detection and intervention of early emerging externalizing disorders. Our proposal is responsive to the Individually Measured Phenotypes to Advance Computational Translation in Mental Health Initiative, with the key objective of using computer vision and machine learning based approaches to facilitate early identification of externalizing disorders. To achieve this, we propose synergistic aims: Aim 1: Identify maternal-infant health and social determinants of health predictors of externalizing disorders (ADHD, ODD, and conduct disorder (CD)) by employing a suite of machine learning models to leverage an expansive nation-wide, representative cohort of >9M children with birthing parent linkage (Epic Cosmos). Aim 2: Isolate a parsimonious set of infant and toddler behaviors, derived via computer vision microcoding, by leveraging the expansive Healthy Brain and Child Development (HBCD) study to integrate video, behavioral, cognitive, and survey data from 3,100 infants 9–15 months of age to predict antecedents of psychopathology at 2 years of age; further, evaluate if specific patterns of infant affect, arousal, and attention, as measured by microcoding, will significantly improve the prediction of externalizing problems at age 2 compared to models using caregiver reports and EHR data. Aim 3: Develop and validate a scalable, remote infant phenoscreening to characterize individual prediction models of externalizing behaviors. We will combine and extend findings from SA1 and SA2 to develop a fully remote, low- cost, brief parent-administered smartphone/web-based screening assessment of infant behavior in a new, nationally representative cohort of 1,500 toddlers at 9–15 months of age. The innovation of this project is three- fold: (i) identification of maternal and infant health predictors of externalizing disorders using large-scale, multi- level data, (ii) addressing structural inequities in healthcare access and time to diagnosis through evaluation of a highly scalable portal for multi-trait assessment at infant age 12 months, and (iii) use of AI to automatically quantify infant behaviors predictive of future externalizing outcomes. The translational significance of this work will be advancing low cost, accessible, scalable and accurate population screening in infancy; development of more precise interventional targets; and alleviating socioeconomic disparities in diagnosis and intervention.

NIH Research Projects · FY 2025 · 2025-09

Project Summary/Abstract Astrocyte gap junctions are necessary for memory formation, synaptic plasticity, coordination of neuronal signaling, and closing the visual and motor critical periods; we also know signaling through astrocyte gap junctions can drive changes between behavioral states. Studies on networks of gap-junction-connected astrocytes have been limited to in vitro methods, slice electrophysiology, and pan-astrocyte knockouts. However, these methods cannot deeply examine the complexity and function of distinct, 3D, in vivo astrocyte networks, and largely limit mechanistic work to calcium dynamics. However, studies in other model systems have found that gap junctions flux many molecules known to facilitate neuroplasticity – especially connexin 43 (cx43), which has the largest pore limit and is the predominant astrocyte connexin. To address this gap in neuroscience, I designed and built a vector that marks functional astrocyte networks by biotinylating molecules fluxed by an infected astrocyte's gap junctions. Using a tissue-clearing method I optimized to preserve astrocyte morphology, I was able to map functional astrocyte network communication across whole, cleared brains. I found that there are many independent astrocyte networks across the brain, each connecting specific brain regions rather than diffusing indiscriminately. Through mass spectrometry, I also identified over 200 biotinylated metabolites, antioxidants, and peptides that traverse astrocyte networks. I will now build on these findings to test a central hypothesis: astrocyte networks circulate pools of antioxidants that buffer oxidatively stressed local environments and facilitate neuroplasticity. I will study neuroplasticity in adult mouse barrel cortex induced by whisker trim, a robust model of neuronal remodeling. Using my vector, I found that astrocyte networks usually connect several brain regions. However, barrel cortex astrocyte networks are almost exclusively limited to barrel cortex itself. This makes barrel cortex an optimal system to establish the fundamental organizational and mechanistic properties of astrocyte networks. Here, I first use state-of-the-art spatial transcriptomics and mass spectrometry imaging to establish the fundamental organizational properties of astrocyte networks. I then use chemogenetics in neurons to challenge broad astrocyte networks to functionally adapt to extremes in local environments and determine mechanisms by which astrocyte networks support the varied needs of the multiple brain regions they traverse. Finally, the first experiments in my independent laboratory will manipulate astrocyte network wide antioxidant pools to determine if these pools can rescue neuronal remodeling during periods of oxidative stress, then eliminate astrocyte networks in the brain to determine if antioxidant support is the sole way astrocyte networks support remodeling neurons. Together, these experiments provide a foundational understanding of astrocytic networks while exposing mechanisms necessary for neuronal plasticity to occur.

NIH Research Projects · FY 2025 · 2025-09

Expansion of NCI-sponsored clinical research in hematological malignancies to Brooklyn and Queens The Perlmutter Cancer Center (PCC) at NYU Langone serves a heterogeneous population across Manhattan, Brooklyn, Queens, Staten Island, and Long Island, a catchment area of nine million. This region faces high cancer disparities, with significant unmet needs in Brooklyn and Queens. One of the key pillars of the PCC mission is ensuring broader access to high-quality care and research throughout the New York area. Efforts include a robust clinical trials portfolio, community outreach, and research aimed at understanding cancer risk factors, biology, and advancing education within our communities. The PCC Clinical Trials Office (CCTO) oversees 200+ trials across the National Clinical Trials Network (Alliance, NRG, COG), Experimental Therapeutics Clinical Trials Network, and other sponsors, supporting over 550 annual enrollees. Dr. Peter Martin focuses on caring for patients with lymphoma, with clinical practices in Manhattan and Brooklyn. An experienced clinical and translational researcher, Dr. Martin is a productive member of the Alliance Lymphoma Committee and NCI Lymphoma Steering Committee. He has led and contributed to multiple NCTN trials, and his investigator-initiated research has provided the framework for larger trials within the Alliance. Further advancing PCC’s mission, the Dr. Martin collaborates with organizations to promote patient education around hematological malignancies and clinical research. He is actively engaged with MCC's Office of Community Outreach and Engagement and Research Informatics team with the objective of increasing accessibility to clinical research. Moreover, he will continue collaborations with population scientists at WCM and through the NCI-funded Lymphoma Epidemiology of Outcomes cohort study to better understand barriers to research participation and to develop strategies to overcome them.

NIH Research Projects · FY 2025 · 2025-09

Project Summary / Abstract Maternal illness during the postpartum period introduces significant challenges to caregiving, with implications for both maternal and offspring survival. Postpartum sickness is linked to an increased risk of maternal anxiety, depression, impaired bonding, and breastfeeding deficits. However, the mechanisms by which sickness alters maternal neurobiology remain poorly understood. The paraventricular nucleus (PVN) of the hypothalamus is a critical regulator of maternal caregiving behaviors, housing neurons that synthesize and respond to neuropeptides such as oxytocin, vasopressin, and corticotropin-releasing hormone (CRH). This proposed research investigates how sickness disrupts PVN neuropeptidergic circuits to influence maternal caregiving. Using an acute inflammatory challenge induced by injection of the endotoxin lipopolysaccharide (LPS), I will examine caregiving behaviors in primiparous and multiparous mice. Preliminary behavioral analyses suggest that first-time mother mice exhibit reduced caregiving during sickness, increasing the risk of pup mortality. In contrast, experienced mother mice demonstrate behavioral adaptations that maintain caregiving under similar conditions. Preliminary in vivo photometry imaging of CRH activity suggests that PVN CRH neurons are highly active during sickness, as well as changes to PVN oxytocin neuron activity during nursing bouts. This proposal will test the hypothesis that neuroplasticity in PVN neuropeptide circuits during the postpartum period modulates caregiving behaviors under inflammatory stress. The proposed research has two aims: (1) to determine the effects of sickness on PVN neuropeptide circuit dynamics during caregiving and (2) to investigate how neuroplasticity in the PVN neuropeptidergic system promotes caregiving in experienced mothers despite illness. These findings will enhance our understanding of the neurobiological mechanisms underlying maternal caregiving and their resilience to physiological challenges, advancing knowledge relevant to maternal mental health and offspring survival.

NIH Research Projects · FY 2025 · 2025-09

Project Summary: Uncontrolled type 2 diabetes (T2D) poses a significant health challenge causing rise in morbidity and mortality in the US, especially impacting minoritized populations, who endure disproportionately higher rates of health complications. Behavioral health intervention leveraging low-cost, ubiquitous digital technology (e.g., mobile phones and text messages) can capture and address the complex psychosocial (e.g., stress) and behavioral factors (e.g., adherence to self-care) that significantly impact glycemic control. Text- messaging interventions have shown promise in improving disease management among patients with T2D; however, they are hampered by low engagement and high attrition, especially among minoritized populations, limiting their efficacy. The goal of this Pathway to Independence Award (K99/R00) is to accelerate the candidate’s transition to an independent investigator on digital health behavioral interventions (DHBI) with specific expertise in optimizing user engagement for improved health outcomes. In the K99 phase of this award, the candidate will leverage data from the recently completed i-Matter and newly renewed i-Matter2 trials (in formative phase), two AHRQ-funded text messaging DHBI for T2D patients to: 1) enhance his machine learning (ML) experience with additional training for developing phenotypes of individual’s engagement behavior, 2) obtain expertise in mixed- method designs and user-centered approaches for robust evaluation of the engagement phenotypes, and 3) obtain training in causal inference approaches for evaluating when and how engagement with DHBI becomes effective for intended T2D health outcomes. The candidate will recruit individuals who completed i-Matter for a survey (n=114) and purposefully sample 12 of them from the i-Matter2 formative phase for qualitative interviews. Using a mixed-method approach, the survey and interviews will gather insights into participants’ engagement with the DHBI, complementing their objective system use data, and will be used to evaluate their engagement phenotypes. In the R00 phase, the candidate will use acquired skills and training to develop a text-messaging intervention tailored to individual’s engagement phenotypes and test the efficacy of the tailored intervention against a non-tailored group (control) receiving standard text messages in a pilot randomized two-arm clinical trial (n=96). Following baseline assessments of demographics, health literacy, digital literacy, and adherence to T2D self-care behaviors, participants in both arms will receive text messages with questions related to T2D self- care for 9 months. Both groups will also participate in 3-, 6-, and 9-mo study visits when adherence to self-care behavior will be measured with a validated scale via REDCap surveys. The primary aim of this trial is to evaluate the efficacy of the tailored intervention on HbA1c reduction (primary outcome), and improved adherence to self- care behaviors (secondary outcome). Both groups will also be invited to participate in a post-study engagement survey which will be used to explore potential mechanisms that influence engagement’s effect on the clinical and behavioral outcomes (exploratory aim).

NIH Research Projects · FY 2025 · 2025-09

Project Summary During development, fixed genetic programs generate the neuronal diversity required for the complex circuits that generate and modulate behavior. Other developmental mechanisms work in tandem with these fixed programs and regulate gene expression in response to neuronal activity and other extrinsic cues. How these different developmental processes are controlled and coordinated during neurodevelopment remains a major question. The nematode C. elegans is a powerful model for the study of neurodevelopment, as most C. elegans neurons quickly assemble into circuits during embryogenesis. By using C. elegans we can (1) observe neurodevelopment in real time, (2) deploy powerful genetic and genomic tools, and (3) observe the consequences of genetic manipulations on nervous system structure and function. This project will focus on the development of a pair of C. elegans chemosensory neurons, the BAGs, which control behavioral responses to aversive cues in the environment. Prior studies revealed an unexpected role for the p38 MAP kinase PMK-3 during a critical period in embryonic development when BAG neurons establish expression of genes required for their function. Neural activity during this critical period also influences BAG neuron development. To determine how PMK-3 regulates gene expression during neurodevelopment, we performed a genetic screen for modifiers of the gene expression and functional defects caused by mutation of PMK-3. Through fine mapping a suppressor mutation, I identified the highly conserved E3 ubiquitin ligase HECD-1 as a factor that acts either downstream of or in parallel to p38 MAPK. The proposed research plan will determine the mechanism by which HECD-1 regulates gene expression during nervous system development. Because mutation of the human homolog of HECD-1 has recently been linked to severe neurodevelopmental disorders, these studies will advance understanding of molecular mechanisms that go awry in neurodevelopment to cause disease.

NIH Research Projects · FY 2025 · 2025-09

Project Summary/Abstract Delirium, characterized by fluctuating disturbances of cognition, is common in older adults undergoing hip fracture surgery. Anesthesia may exacerbate delirium severity, especially in patients with dementia who are at heightened risk for delirium onset. Monitored anesthesia care with soft-tissue infiltration with local anesthesia (MAC-STILA) is a novel approach that involves anesthesiologist titration of intravenous sedation to a level that preserves spontaneous breathing followed by surgeon administration of a local anesthetic injected directly into the surgical site. MAC-STILA may provide significant benefit to hip fracture patients by avoiding adverse effects of traditional anesthesia approaches but has not been widely tested. We propose a multi-center clinical trial to demonstrate feasibility of randomizing older hip fracture patients to MAC-STILA versus the current standard of care involving general anesthesia and to determine efficacy of MAC-STILA for reducing incidence and severity of post-operative delirium. The trial will be conducted at 6 hospitals and coordinated by the Major Extremity Trauma Research Consortium. Using the Delphi approach, we will first achieve consensus on standards for MAC-STILA among a panel of orthopedic surgeons and anesthesiologists at 6 participating hospitals to facilitate reproducibility and evaluation. We will then enroll and randomize 140 patients, stratified by baseline cognitive impairment. We will examine aspects of trial feasibility including recruitment rate, rate of crossover to a different anesthesia approach, factors associated with treatment crossover, and data completeness. We will examine the effect of MAC-STILA compared with general anesthesia on delirium incidence and severity using validated cognitive assessments through post-operative day 5 or time of discharge. We will also explore the treatment effect on delirium among patients with baseline cognitive impairment and by sex. Secondary outcomes include intraoperative physiologic parameters (e.g. blood pressure, heart rate), post operative complications, postoperative pain and narcotic use, cognitive function at 30- and 90-days post discharge, and 30 and 90-day mobility, disability, and mortality. The goal of this study is to demonstrate feasibility and establish the preliminary data needed to support a larger, rigorously designed multi-site effectiveness trial.

NIH Research Projects · FY 2025 · 2025-09

Summary: This grant aims to understand the plasticity of melanocyte stem cells (McSCs) and mature melanocytes (Mcs) in adult hair follicles (HF). Our recent study revealed that all McSCs in early anagen undergo partial differentiation and then make the decision to revert to a stem cell (SC) phenotype or go on to full differentiation. This newly described mechanism challenges the previous belief that once the differentiation program is initiated in SCs, it is not reversed in normal homeostasis. Many questions still remain unexplored in this new model, including: What are the steps driving McSC partial differentiation and then either reversion to a SC state or full melanocyte maturation? When does McSC plasticity end? i.e. can fully mature Mcs also undergo dedifferentiation? How does this newly observed plasticity translate to human melanocytes? Addressing these questions will provide an essential platform to advance this new SC paradigm. Further, understanding McSCs in particular is relevant because melanocytes (Mcs) are critical for hair and skin pigmentation, and when dysregulated, they can result in gray hair, skin depigmentation and melanoma. In Aim 1, we will use innovative techniques and tools for live imaging to fully delineate the fate decision process of McSCs. We will track the movement of individual McSCs from the onset of hair growth until their fate is determined using a live imaging system. The regulatory mechanisms governing the ultimate fate of McSCs remain largely unexplored. This is crucial, particularly in light of the understanding that the ultimate fate of SCs involves dedifferentiation from a partially differentiated state within the McSC system. Our live imaging will help us understand how McSCs interact with their environment during early hair growth and when they undergo final fate decisions. In Aim 2, we have found that all McSCs undergo partial differentiation in the onset of the hair follicle growth phase, but they do not go on to full differentiation despite the fact that essential differentiation pathways have been initiated. In examining McSC scRNAseq datasets for regulons, Foxd3, a TF known to repress melanocyte embryonic development, came up in the 1st position. We will use Foxd3 KO and over-expression models to see how its perturbation affects McSC differentiation. We predict that loss-of-function will induce full differentiation in early anagen McSCs, and gain-of-function will promote a more undifferentiated state of McSCs. In Aim 3, we will ask at what stage McSCs are no longer capable of complete dedifferentiation to a SC state. Our preliminary results suggest that mature Mcs retain the potential to undergo full reversion to a SC state in ex vivo assays. We will examine this potential in great detail, and both mouse and human Mcs will be investigated for this potential.

NIH Research Projects · FY 2025 · 2025-08

SUMMARY/ABSTRACT Research involving recently deceased humans physiologically maintained following declaration of death by neurologic criteria (“DNC research”) is the most promising method for advancing xenotransplantation, and has potential to become the optimal model for studying countless interventions that carry risks we cannot justify imposing on living human research participants. DNC research has also been used to test novel pharmaceuticals and medical devices, and is going to be used to test gene therapies. During recent DNC studies including one that lasted 8 weeks, our institution gained novel insights impossible to obtain from any other translational model about function, human immune response, and survival of xenogeneic organs. DNC research involves testing experimental interventions in “brain dead” humans. Although the person is deceased, the body and physiological functions are maintained through medical interventions delivered by bedside healthcare staff (such as nurses) with training and experience caring for living patients, raising a number of potential concerns related to the rights and well-being of family members who authorize DNC research (“DNC family”) and healthcare staff. There is known confusion about the definition of brain death, and healthcare staff and DNC families have concerns about zoonotic disease, questions about prioritization of clinical care versus research, and what constitutes respectful treatment of the body in DNC research. These issues could negatively impact attitudes toward DNC research and promote distrust of research, thereby limiting donor availability. No guidelines exist for engagement with DNC research stakeholders, and there is no consensus approach for how to obtain DNC research authorization from donor families. Our aims are to (1) understand experiences, attitudes, and perceptions of DNC family members and healthcare staff involved in DNC research at the bedside, (2) conduct a focused ethnography of healthcare staff participating in DNC research, and (3) to design new protocols to guide engagement with DNC family members and healthcare providers during the conduct of DNC research. DNC research could serve as the most effective translational bridge to an unlimited supply of organs to treat end stage organ failure and could also become the most effective model for research on countless other life- saving interventions. However, the rights and wellbeing of families and healthcare staff must be prioritized during DNC research. Protocols developed in this study will inform transplant programs, OPOs, and all other entities preparing to perform DNC research in the US on how to engage ethically with donor families and healthcare staff in this novel translational research.

NIH Research Projects · FY 2026 · 2025-08

SUMMARY Hexavalent chromium, Cr(VI), is widely used in industrial processes such as stainless steel production, electroplating, textile manufacturing, wood preservation, and leather tanning. Occupational or environmental exposures to Cr(VI) have been associated with a variety of adverse respiratory, cardiovascular, gastrointestinal, hematological, and hepatic effects. However, very little is known about the impact of Cr(VI) on skeletal muscles, the largest organ in the human body comprising about 40% of body mass. The 3-month toxicity study by National Toxicology Program (NTP) reported a dose-related increase of serum creatine kinase (CK) activities in both male and female rats exposed to Cr(VI) via drinking water compared to the control group, indicating muscle injury. Consistently, a significant increase of serum CK activity has been reported in chrome plating workers compared to the control subjects. Despite a clear sign of muscle injury has been reported in both human and animal study, the impact of Cr(VI) exposure on muscle injury and regeneration is completely unknown. Our recent study reveals that acute exposure of mouse C2C12 myoblast cells to Cr(VI) inhibits myogenic differentiation in a dose- dependent manner. Additionally, Cr(VI) inhibition of myogenesis is further confirmed in freshly isolated mouse primary myoblasts ex vivo. Moreover, the expression of key myogenic regulatory factors (MRFs) was significantly altered in Cr(VI)-treated cells. Taking together, our results demonstrate an inhibitory effect of Cr(VI) in myogenic differentiation in vitro and ex vivo. Given that in vitro myogenic differentiation largely recapitulates in vivo myogenesis during early embryo development and adult muscle regeneration, we hypothesize that Cr(VI) adversely affects skeletal muscle development and regeneration in vivo. In this proposal, we aim to address whether Cr(VI) lead to defective myogenesis and impaired muscle regeneration with the following specific aims. First, we will examine the effects of acute or chronic Cr(VI) exposure on adult muscle injury and regeneration in vivo. Second, we will explore the impact of Cr(VI) on skeletal muscle development by assessing primary and secondary myogenesis in mouse embryos prenatally exposed to Cr(VI). Lastly, to elucidate the mechanism underlying Cr(VI) inhibition of myogenic differentiation, we will analyze the role of MyoD and Myf5 in Cr(VI)- induced defective myogenic differentiation, and explore whether special AT-rich sequence-binding protein 2 (Satb2), a known Cr(VI) target gene in lung carcinogenesis, acts as an important target or a functional mediator of Cr(VI) in myogenic differentiation. The success of the proposed research will generate valuable evidence to support the adverse impact of Cr(VI) on skeletal muscle development and regeneration, and promote more epidemiological research on Cr(VI) exposure as a significant risk factor for skeletal muscle health.

NIH Research Projects · FY 2025 · 2025-08

SUMMARY In the US, obesity affects 42% of the population and is associated with over $260 billion in direct medical costs annually. Highly efficacious anti-obesity medications, such as incretin mimetics medications (IMMs: e.g. GLP-1, semaglutide), hold great promise for the treatment of the epidemic of weight-related chronic diseases. However, patients on IMMs often have poor adherence and do not achieve medication maintenance dosing, limiting the impact of these highly efficacious medications. These medications are associated with a high clinical management burden and healthcare systems are currently unequipped to handle the extension of their use to additional patient populations. Overburden can hinder health care provider (HCP) ability to adhere to prescribing guidelines and limits patient access to care, including side effect management. Innovative solutions are needed to increase IMM adherence while reducing the burden of medication management placed on HCPs. Generative artificial intelligence (GenAI) has the potential to address the clinical and administrative demands associated with the management of patients on IMMs. This increases HCP capacity and addresses both clinical and administrative demands. GenAI could reduce a bottleneck that could impede a patient’s likelihood of achieving maintenance dose. Through its inherent flexibility to incorporate and synthesize multiple data sources, GenAI has the potential to address multiple aspects of medication management, including streamlining patient- clinician communication, supplying personalized patient advice for management of common side effects, providing clinical decision support (CDS) on optimal dose titration, and giving prescribing guidance based on non-clinical factors such as insurance coverage and medication availability. Under this proposal, we will test the effect of implementing AIManage, an electronic health record (EHR)- integrated CDS tool enhanced with a GenAI-driven side-effect management chatbot, on IMM adherence and achievement of maintenance dosing, as well as reduction of HCP burden. We will conduct this study in two phases: 1) a formative phase to refine and user-test AIManage in real-world settings, and 2) a randomized clinical trial (RCT). The clinical trial phase will use a hybrid type 1 RCT to evaluate the effectiveness of AIManage vs. usual care (UC) on maintenance dose achievement and medication adherence at 12-months among 810 patients on IMMs. Using the extended RE-AIM framework, we will also apply an equity lens to measure the reach, adoption, and implementation (i.e., fidelity, cost) of AIManage. This project aims to: (Aim 1) Refine and implement AIManage for management of patients on IMMs in outpatient primary care and bariatric medicine practices; (Aim 2) Estimate AIManage impact on medication persistence and maintenance dose achievement; and (Aim 3) Evaluate the implementation outcomes of reach, adoption, and fidelity including the cost and cost-effectiveness of AIManage vs. UC from a health care system perspective to provide insights into barriers and facilitators for an equitable full-scale implementation.

NIH Research Projects · FY 2025 · 2025-08

Maternal mortality rates from drug overdoses have increased, especially among pregnant and postpartum women aged 35 to 44. Despite the essential need for effective screening, education, and referrals, there is limited understanding of how well substance use screening works during pregnancy. Additionally, pregnant women frequently encounter barriers to accessing prenatal care due to variations in perinatal drug toxicology testing practices across patient population. This K01 application focuses on examining variation in perinatal drug toxicology screening policies and practices using the Socio-cultural Framework for the Study of Health Service Disparities “SCF-HSD" (Alegría, 2010). The study will progress through three phases: (1) use multilevel statistical models with NYS Medicaid claims data to identify predictors of perinatal toxicology testing and characterize hospital variations across over 600,000 mother-infant dyads in 225 hospitals from 2021 to 2024; (2) conduct 1:1 interviews with hospital administrators and staff to collect and analyze toxicology testing policies and practices, capturing multiple perspectives on testing attitudes and adherence; (3) integrate quantitative and qualitative findings from Aims 1 and 2 using a mixed methods design, incorporating patient perspectives via focus group sessions to develop hospital policy recommendations. Dr. Choi's career development plan involves receiving interdisciplinary mentorship to advance her skills in data-driven mixed methods research. In the era of big data, there is a growing need for mixed methods research training that equips individuals with the skills to harness big datasets and employ qualitative approaches informed by stakeholders (e.g., OB/GYN, social workers, hospital administrators, and patients) to understand complex computations. This includes training in 1) evaluating policy and system-level contributors to perinatal health research; 2) advanced multi-level modeling techniques; 3) qualitative research; and 4) mixed methods research. The plan aims to build Dr. Choi's expertise in perinatal health, with a focus on examining patient-centered approaches for exploring perinatal health care quality and access. Her career goals include enhancing the understanding of substance use screening effectiveness, improving healthcare practices, and developing targeted interventions to improve perinatal and infant health. The career development award will support Dr. Choi in a collaborative and resource-rich academic environment at New York University Grossman School of Medicine. She will actively participate in professional development activities to disseminate research findings and foster collaborations. She will have access to Medicaid data, experienced mentors, and advisors, which will support impactful research and the implementation of findings to improve health outcomes for all populations, with a particular focus on perinatal substance use.

NIH Research Projects · FY 2025 · 2025-08

Project Summary Supplemental breast screening ultrasound is recommended for women with dense breasts, accounting for nearly half of U.S. women over age 40. While breast ultrasound improves cancer detection rates, its cost-effectiveness remains low for women at average risk of breast cancer, with a cost-effectiveness ratio of $728,000 per quality- adjusted life year (QALY) gained. This limited efficiency is driven by high false-positive rates in breast ultrasound interpretation, leading to unnecessary diagnostic workups and biopsies, and by one-size-fits-all guidelines that recommend supplemental ultrasound for all women with dense breasts. This proposal aims to improve the diagnostic accuracy and cost-effectiveness of breast cancer screening by building an AI system to determine the necessity of supplemental screening ultrasound. Our core hypothesis is that in a substantial portion of the screening population, it is possible to forgo currently guideline-recommended supplemental ultrasounds without impacting diagnostic accuracy. To verify this hypothesis, we propose the following aims. First, we will build a multi-modal deep neural network, Multi-Modal Diagnoser (MMD), that integrates current and prior mammographic and ultrasound imaging data to improve detection of breast cancer. To train MMD, we will assemble a substantial dataset of over 2.12 million exams (420,000 patients) from our institution, as well as two external datasets from other institutions for independent evaluation. MMD will serve as the foundation for our second model, the Ultrasound Benefit Predictor (UBP). UBP will utilize screening mammography images to determine the necessity of supplemental ultrasound for women with dense breasts through an interpretability mechanism that highlights specific mammographic regions where ultrasound could add diagnostic value. Finally, we will critically assess these two elements of our AI system through a retrospective study, examining their impact on radiologists' performance and cost-effectiveness through the lens of patients’ QALY. This project seeks to pioneer the integration of AI in breast cancer screening, leading to more personalized patient care and optimized healthcare resource allocation.

NIH Research Projects · FY 2025 · 2025-08

Abstract The skin protects animals and humans from harm. To this end, the skin needs to detect and repair wounds. Slowed or failed wound detection and repair causes different skin disorders and signaling pathways that detect and repair wounds are potential targets for clinically important therapeutics. The goal of this project is to identify the pathways and mechanisms that the skin uses to detect and repair wounds. For this, we combine genetics and imaging in zebrafish embryos with computational modeling. Identical to the human embryonic skin, the zebrafish embryonic skin consists of two cell layers – an outer and a basal layer. Due to its transparency, the zebrafish embryonic skin can be imaged at high resolution, which permits quantitative measurements and data-based computational modeling. We have found a G-protein coupled receptor (GPCR) signaling pathway that is activated in the skin cells at the margin of wounds, and we have built tools to image the force-generating and force-transmitting machinery in skin cells. Using these tools, we will address three questions. First, we will decipher how the skin cells use the GPCR signaling to detect wounds. Second, we will ask how the two layers of the skin cooperate to move and close wounds. Third, we will use the quantitative imaging data and measurements of physical properties of the skin to generate a physical model of a two-layered skin that we will test and refine using physical and genetically perturbations. Since epithelial integrity is essential for many organs including the lining of the gut, blood vessels and lungs, the proposed studies will provide necessary context to better understand epithelial barrier function in general and should inform us about strategies of how to correct epithelial disorders.