Identification of an ANCA-associated vasculitis cohort using deep learning and electronic health records.

Background: ANCA-associated vasculitis (AAV) is a rare but serious disease. Traditional case-identification methods using claims data can be time-intensive and may miss important subgroups. We hypothesized that a deep learning model analyzing electronic health records (EHR) can more accurately identify AAV cases.

Detecting early cognitive deficits in preclinical Alzheimer's disease using a remote digital multi-day learning paradigm.

Remote, digital cognitive testing on an individual's own device provides the opportunity to deploy previously understudied but promising cognitive paradigms in preclinical Alzheimer's disease (AD). The Boston Remote Assessment for NeuroCognitive Health (BRANCH) captures a personalized learning curve for the same information presented over seven consecutive days. Here, we examined BRANCH multi-day learning curves (MDLCs) in 167 cognitively unimpaired older adults (age = 74.

POCUS-first in acute diverticulitis: Quantifying cost savings, length-of-stay reduction, and radiation risk mitigation in the ED.

Background: Recent studies have validated the efficacy of point-of-care ultrasound (POCUS) as an alternative diagnostic imaging approach to computed tomography (CT) for patients with suspected acute diverticulitis. This study aimed to quantify the national impact of this approach in cost savings, ED length-of-stay (LOS), and radiation risk mitigation using a POCUS-first approach for acute diverticulitis in the emergency department (ED).

Methods: Using published data, we constructed a Monte Carlo simulation model to compare two POCUS-first strategies (nonselective and selective approaches) for evaluating patients with suspected acute diverticulitis in the ED.

Intraoperative FiO and risk of impaired postoperative oxygenation in lung resection: A propensity score-weighted analysis.

Study Objective: To assess whether, in a lung resection cohort with a low probability of confounding by indication, higher FiO is associated with an increased risk of impaired postoperative oxygenation - a clinical manifestation of lung injury/dysfunction.

Design: Pre-specified registry-based retrospective cohort study.

Setting: Two large academic hospitals in the United States.

The history of Women in Radiology (WIR) programs at two academic institutions: How we did it and how we merged best practices.

Rationale: The percentage of women in radiology has remained at 25 % for decades. Our institutions identified the need to recruit, support and retain women in our departments in order to change this status quo.

Methods: Our institutions created two different frameworks for Women in Radiology (WIR) programs in order to meet the needs of each distinct department.

Measurement of circulating viral antigens post-SARS-CoV-2 infection in a multicohort study.

Objectives: To determine the proportion of individuals with detectable antigen in plasma or serum after SARS-CoV-2 infection and the association of antigen detection with postacute sequelae of COVID-19 (PASC) symptoms.

Methods: Plasma and serum samples were collected from adults participating in four independent studies at different time points, ranging from several days up to 14 months post-SARS-CoV-2 infection. The primary outcome measure was to quantify SARS-CoV-2 antigens, including the S1 subunit of spike, full-length spike, and nucleocapsid, in participant samples.

Spatially resolved analysis of pancreatic cancer identifies therapy-associated remodeling of the tumor microenvironment.

In combination with cell-intrinsic properties, interactions in the tumor microenvironment modulate therapeutic response. We leveraged single-cell spatial transcriptomics to dissect the remodeling of multicellular neighborhoods and cell-cell interactions in human pancreatic cancer associated with neoadjuvant chemotherapy and radiotherapy. We developed spatially constrained optimal transport interaction analysis (SCOTIA), an optimal transport model with a cost function that includes both spatial distance and ligand-receptor gene expression.

Risk of Neonatal Hypoglycemia in Infants of Mothers With Gestational Glucose Intolerance.

Objective: To examine the relationship between gestational glucose intolerance (GGI) and neonatal hypoglycemia.

Research Design And Methods: This was a secondary analysis of 8,262 mother-infant dyads, with delivery at two hospitals between 2014 and 2023. We categorized maternal glycemic status as normal glucose tolerance (NGT), GGI, or gestational diabetes mellitus (GDM).

What are the barriers to health professionals' training on gender-affirming care from patients' and clinicians' perspectives?

Objective: The aim of the study is to identify the barriers to gender-affirming health care education for providers from the perspectives of patients and providers.

Methods: A qualitative study based on grounded theory was conducted. Participants included transgender and gender diverse (TGD) patients seeking care, as well as resident physicians and attending physicians involved in care of patients seeking gender-affirming care.

Empagliflozin and Risk of Incident Gout: Analysis from the EMPagliflozin Comparative Effectiveness and SafEty (EMPRISE) Cohort Study.

Background: Hyperuricemia is frequently observed in patients with type 2 diabetes (T2D) and is associated with increased risk of gout and cardiovascular disease (CVD). Empagliflozin lowers serum urate levels by enhancing its urinary excretion.

Objective: To compare initiators of empagliflozin vs dipeptidyl peptidase-4 inhibitor (DPP4i) and initiators of empagliflozin vs glucagon-like peptide-1 receptor agonist (GLP-1RA) with respect to the risk of incident gout events.

Joint genotypic and phenotypic outcome modeling improves base editing variant effect quantification.

CRISPR base editing screens enable analysis of disease-associated variants at scale; however, variable efficiency and precision confounds the assessment of variant-induced phenotypes. Here, we provide an integrated experimental and computational pipeline that improves estimation of variant effects in base editing screens. We use a reporter construct to measure guide RNA (gRNA) editing outcomes alongside their phenotypic consequences and introduce base editor screen analysis with activity normalization (BEAN), a Bayesian network that uses per-guide editing outcomes provided by the reporter and target site chromatin accessibility to estimate variant impacts.

Facilitating the transition to residency: A resident-as-coach pilot program.

Background: The transition from medical school to residency is a critical developmental phase; coaching may help students prepare for this role transition.

Aims: We explored whether near-peer coaching could improve a specific workplace skill prior to residency.

Methods: A resident-as-coach program was piloted for the medicine sub-internship, an advanced acting internship rotation.

Determinants of long-term physical and mental health outcomes after intensive care admission for trauma survivors.

Introduction: Collectively, studies from medical and surgical intensive care units (ICU) suggest that long-term outcomes are poor for patients who have spent significant time in an ICU. We sought to identify determinants of post-intensive care physical and mental health outcomes 6-12 months after injury.

Methods: Adult trauma patients [ISS ≥9] admitted to one of three Level-1 trauma centers were interviewed 6-12 months post-injury to evaluate patient-reported outcomes.

Disease modifying therapy in the treatment of tumefactive multiple sclerosis: A retrospective cohort study.

Tumefactive multiple sclerosis (TMS) is characterized by large demyelinating brain lesions. This was a retrospective cohort study of 67 patients with TMS between January 2015-2023, examining different disease modifying therapy impact on expanded disability scale score change at follow-up. Median age was 36 with a female predominance.

Expert-centered Evaluation of Deep Learning Algorithms for Brain Tumor Segmentation.

Purpose To present results from a literature survey on practices in deep learning segmentation algorithm evaluation and perform a study on expert quality perception of brain tumor segmentation. Materials and Methods A total of 180 articles reporting on brain tumor segmentation algorithms were surveyed for the reported quality evaluation. Additionally, ratings of segmentation quality on a four-point scale were collected from medical professionals for 60 brain tumor segmentation cases.

Monte Carlo based dosimetry of extraoral photobiomodulation for prevention of oral mucositis.

Photobiomodulation therapy (PBMT) is recommended for prevention and treatment of oral mucositis, a painful condition that occurs in cancer patients. Intraoral PBMT is limited to treating distal oral mucosa and oropharynx. Extraoral PBMT may provide a more efficient intervention.

Viral and host mediators of non-suppressible HIV-1 viremia.

Non-suppressible HIV-1 viremia (NSV) is defined as persistent low-level viremia on antiretroviral therapy (ART) without evidence of ART non-adherence or significant drug resistance. Unraveling the mechanisms behind NSV would broaden our understanding of HIV-1 persistence. Here we analyzed plasma virus sequences in eight ART-treated individuals with NSV (88% male) and show that they are composed of large clones without evidence of viral evolution over time in those with longitudinal samples.

Deep learned representations of the resting 12-lead electrocardiogram to predict at peak exercise.

Aims: To leverage deep learning on the resting 12-lead electrocardiogram (ECG) to estimate peak oxygen consumption (V˙O2peak) without cardiopulmonary exercise testing (CPET).

Methods And Results: V ˙ O 2 peak estimation models were developed in 1891 individuals undergoing CPET at Massachusetts General Hospital (age 45 ± 19 years, 38% female) and validated in a separate test set (MGH Test, n = 448) and external sample (BWH Test, n = 1076). Three penalized linear models were compared: (i) age, sex, and body mass index ('Basic'), (ii) Basic plus standard ECG measurements ('Basic + ECG Parameters'), and (iii) basic plus 320 deep learning-derived ECG variables instead of ECG measurements ('Deep ECG-V˙O2').

A deep learning approach for transgender and gender diverse patient identification in electronic health records.

Background: Although accurate identification of gender identity in the electronic health record (EHR) is crucial for providing equitable health care, particularly for transgender and gender diverse (TGD) populations, it remains a challenging task due to incomplete gender information in structured EHR fields.

Objective: Using TGD identification as a case study, this research uses NLP and deep learning to build an accurate patient gender identity predictive model, aiming to tackle the challenges of identifying relevant patient-level information from EHR data and reducing annotation work.

Methods: This study included adult patients in a large healthcare system in Boston, MA, between 4/1/2017 to 4/1/2022.