Faculty Bibliography

BACKGROUND:Previous studies evaluated large language model (LLM) performance on the American Society of Nuclear Cardiology (ASNC) Board Preparation Exam. Without domain-specific context, the best model (GPT-4o) achieved 63.1%, below the estimated 65% passing threshold and the 78% mean score of human fellows-in-training (FITs). Providing textbook context improved GPT-4o to 73.8% on text-only questions, but still fell short of human trainees. Whether next-generation LLMs with retrieval-augmented generation (RAG) can exceed this gap is unknown. METHODS:Claude Opus 4.7 and GPT-5.5 were administered all 168 questions (141 text-only, 27 image-based) from the 2023 ASNC Board Preparation Exam across 5 iterations each, using RAG with a nuclear cardiology textbook, companion atlas, and ASNC clinical guidelines. Claude used local FAISS-based semantic retrieval; GPT-5.5 used Azure's cloud-hosted vector store. Performance was compared to prior LLM results and 13 human FITs. RESULTS:Across 5 iterations, Claude Opus 4.7 achieved a mean accuracy of 86.3% ± 1.4% (text 88.8%, image 73.3%). GPT-5.5 achieved 86.7% ± 2.2% (text 88.5%, image 77.0%) but refused a mean of 12.2 questions (7.3%) per iteration due to safety filters. Both models surpassed the human FIT mean (78.0%) and the estimated passing threshold. Compared to GPT-4o without context (63.1%), this represents a 23-percentage-point improvement in 18 months. CONCLUSION/CONCLUSIONS:Next-generation LLMs with RAG now surpass average human trainee performance on nuclear cardiology board preparation questions, suggesting significant potential as educational tools and knowledge-reference aids in cardiovascular imaging.

Sellar region neurocytoma (SELN) is a rare neoplasm whose relationship to other neurocytomas within the central nervous system (CNS) has remained unclear. Prior reports have variably classified SELN as a variant of extraventricular neurocytoma (EVN), while immunohistochemical and ultrastructural studies have suggested a hypothalamic origin. Here, we performed unsupervised clustering of DNA methylation data across a large pan-cancer reference set and identified SELN (n = 20) as distinct from other neurocytomas and regional mimics, as well as clustering with neuroendocrine tumors from other organ sites. SELN exhibited a CIMP-like phenotype, TTF1 negativity (0/8), and AVP (vasopressin) promoter hypomethylation, implicating a magnocellular hypothalamic cell of origin. In evaluable cases, a neuronal/neuroendocrine immunophenotype was observed (synaptophysin 8/8, chromogranin A 5/5) with absent pituitary transcription factor expression (PIT1 and TPIT negative 0/4). DNA sequencing (n = 5) and RNA-based fusion profiling (n = 4) did not detect recurrent mutations or gene fusions, respectively. Patients often presented with visual disturbances or headaches and spanned pediatric and older age groups (median 42 years, range 12.5-75), with no sex predilection. Despite locally aggressive imaging features in some cases (cavernous sinus invasion, carotid encasement, hydrocephalus), disease-free survival (n = 13) was comparable to central neurocytoma, with no disease-related deaths during the limited follow-up. Together, these findings support SELN as a molecularly distinct hypermethylated neuroendocrine-like epitype and clinicopathologic entity.

INTRODUCTION/BACKGROUND:The aims of this study were to evaluate the decision to extract or treat a tooth with initial root canal therapy (RCT) made by general dentists practicing in the United States; and evaluate the effect of clinical and non-clinical factors on decision-making. METHODS:Authors conducted a cross-sectional survey study. Participants read a clinical vignette presenting a questionably restorable tooth for consideration and provided their clinical recommendations. The primary treatment decision was analyzed as a binary variable: extraction or RCT. Other questions about clinical decision-making used the Likert-scale for capturing responses. Clinician related variables were evaluated: age, gender, years of experience, primary practice setting, and participation with payment systems. Kruskal-Wallis tests were used to compare the treatment decision with Likert-scale responses for clinical decision-making. A logistic regression model investigated the effect of clinician factors on the treatment decision. RESULTS:113 eligible dentists completed the survey. For the presented case, a maxillary first molar, 54% of general dentists recommended initial RCT and 46% recommended extraction. 37.2% of dentists responded that the patient's tooth was somewhat or very likely restorable. None of the evaluated clinician factors were statistically significant in predicting the treatment decision. Borderline significance (p=0.05) was noted for general dentists primarily practicing in academic settings; they had 3.57 times the odds of recommending tooth extraction than dentists primarily in private practice (p=0.05). CONCLUSIONS:This study found that general dentists' decision to either save a tooth with initial root canal therapy or extract was primarily associated with perceived tooth restorability.

PURPOSE/OBJECTIVE:Opioid overdose survivors are vulnerable to fatal overdose, particularly in the weeks following overdose, a period when health service use is often low. This study explores post-overdose health/medical services engagement among a diverse sample of survivors. METHODS:We conducted semi-structured interviews with 23 recent opioid overdose survivors across seven states. We asked about service utilization and drug use history, contextual factors related to the overdose, service use immediately after and in the months following, and the impact on wellbeing. RESULTS:Over half of participants described calls to 911 (n = 13), and one-third went to the hospital (n = 8). Six key themes guided by Levesque's patient-centered access to care framework emerged. Several themes explored why participants did not seek medical services following an overdose, including 1) the belief they could manage at home, 2) a lack of trust in responders, often based on prior negative experiences, 3) concern about unintended consequences such as legal issues, and 4) the financial, social, and opportunity costs associated with calling 911 or going to the hospital. Of those who did receive emergency services, some declined transport to the hospital for similar reasons, including potential repercussions from their treatment program if they were notified. A fifth theme focused on interactions with providers in the hospital, where individuals were rarely offered medications for opioid use disorder or connected to treatment. Interactions with providers were largely negative, though participants reported more positive experiences with social workers and peer support specialists. A few had early discharges, often due to feelings of vulnerability, unmanaged withdrawal symptoms, or poor interactions with providers. Last, the sixth theme focused on substance use treatment engagement. Treatment engagement was often delayed but was facilitated by factors like a social worker following up, longer hospital stays, or the participant seeking treatment on their own. CONCLUSIONS:Most of the sample did not engage with hospital or substance use treatment services after a nonfatal overdose. Negative prior experiences and fear of negative consequences were major barriers in this sample.

Lupus nephritis (LN), a severe manifestation of systemic lupus erythematosus (SLE), is a heterogeneous disease driven by diverse immune and tissue cell types. We obtained 538,194 single-cell and 142,881 single-nuclear profiles from kidney biopsies of 155 patients with LN and 30 preimplantation transplant biopsy controls, along with 327,326 single-cell blood profiles. We characterized key stromal and immune cell types and cell states; moreover, we distinguished cell states that were tissue specific from those that were also present in the blood. We observed that LN pathological features were associated with particular cell states. For example, after controlling for the effects of chronic tissue damage, we observed that expansion of glomerular and scar-associated macrophage populations correlated with increasing inflammatory disease activity. Scar-associated macrophages appear to drive LN fibrosis and, in active disease, infiltrate the glomeruli more than other myeloid cells. These observations support that therapeutic targeting of myeloid populations may offer a strategy to prevent renal inflammation and ongoing kidney damage in LN.

Primary human myeloid cells hold promise for immunotherapies, yet efficient, scalable technologies for engineering and screening in these cells remain limited. Here we present a virus-like particle (VLP)-based toolkit that delivers diverse CRISPR editing modalities to human monocytes, macrophages and dendritic cells with high efficiency while preserving viability and innate immune responsiveness. VLP-mediated delivery of ribonucleoproteins supports gene knockout, base editing and epigenetic silencing. Combined with adeno-associated virus-mediated donor delivery, this approach enables site-specific integration of large DNA sequences by homology-directed repair. We developed SLICeVLP, which pairs sgRNA delivery by VPX-lentivirus with Cas9 protein delivery by engineered VLPs, and used it for pooled loss-of-function and Perturb-seq screens in human macrophages. We uncovered regulators of tumor necrosis factor (TNF) and CD80 expression, converging on TNFAIP3 as a central regulator of inflammatory polarization. TNFAIP3 ablation drove a proinflammatory state resistant to suppressive repolarization and enhanced cytotoxicity in chimeric antigen receptor macrophages. This system enables unbiased functional genomics in primary human myeloid cells, with implications for myeloid cell therapy design.

Vasa previa is a rare but potentially catastrophic obstetric condition characterized by unprotected fetal blood vessels over or adjacent to the internal cervical os. Although advances in prenatal ultrasound have dramatically improved neonatal survival through planned cesarean delivery, many aspects of vasa previa remain poorly understood. Existing studies are limited by small sample sizes, single-center designs, heterogeneous diagnostic criteria, and inconsistent reporting of clinical outcomes, highlighting the need for large, systematically collected multicenter datasets. Here we describe, to our knowledge, the first United States multicenter registry of pregnancies complicated by vasa previa without concurrent placenta previa, the U.S. Vasa Previa Registry (US-VPR). This report outlines the rationale, design, and implementation of a large, multicenter, retrospective registry developed to characterize patient demographics, risk factors, placental pathology, natural history, antenatal management, maternal outcomes, neonatal outcomes, and healthcare utilization associated with vasa previa. The US-VPR includes contributions from 15 tertiary referral centers and represents the largest multicenter cohort of pregnancies complicated by vasa previa without concurrent placenta previa reported to date. By leveraging multicenter collaboration and standardized data collection, the registry provides a unique opportunity to characterize variation in clinical practice and outcomes across referral centers, strengthen the evidence base for this rare condition, and inform future prospective studies and evidence-based patient counseling.

BACKGROUND:Cannabis consumption modalities such as smoking, vaping, and consumption of edibles differ in pharmacokinetics and potency, and vary by demographic characteristics. Data linking real-world consumption modality patterns to use frequency and cannabis use disorder (CUD) remain limited. METHODS:Latent class analysis (LCA) was applied to identify modality-based classes using data from 17,732 adults reporting past 30-day cannabis use on the 2022-2023 National Survey on Drug Use and Health. Multinomial logistic regression delineated associations of class membership with demographic characteristics, cannabis use frequency, and the presence/severity of CUD. RESULTS:LCA yielded four classes based on each modality likelihood: primary smoking (48.2 %), smoking and vaping (20.5 %), mainly edibles (16.9 %), and polymodal use (14.4 %). Compared to adults aged ≥ 50, those aged 18-25 years were more likely to belong to the smoking and vaping (aRRR = 2.52, 95 % CI: 1.77-3.61) or polymodal classes (aRRR = 3.58, 95 % CI: 2.48-5.18). Females were more likely to be in the mainly edibles class (aRRR = 1.78, 95 % CI: 1.48-2.13). Daily cannabis use was more likely among polymodal (aRRR = 2.67, 95 % CI: 2.25-3.17) and less likely among mainly edibles (aRRR = 0.49, 95 % CI: 0.35-0.68) classes, relative to the primary smoking class. Compared to the primary smoking class, risk of having CUD was higher among people in the smoking and vaping class, and in the polymodal class, but lower in the mainly edibles class. CONCLUSIONS:Cannabis use modalities were associated with differences regarding demographic characteristics and severity of use. Future interventions could account for route-specific use to improve prevention and harm reduction efforts.