Citations

Abstract Disclosure: M. Kakakhel: None. S. Cheema: None. R. Haider: None. F. Saleem: None. A. Ullah: None. M. Shah: None. A. Qadeer: None. M. Khawar: None. H. Jawad: None. H. Ahmad: None. F. Ahmed: None. S. N.l.: None. R. Alcaraz: None. Background: Diabetes mellitus (DM) is a global health burden, contributing significantly to mortality and comorbidities, including chronic kidney disease (CKD), which remains a leading cause of morbidity and mortality, especially in type 1 and type 2 diabetes. Methods: This retrospective cohort study examined mortality data from the United States from the CDC WONDER database, which gathers data from death certificates. Particularly, deaths from renal problems (ICD-10 codes E10.2-E14.2) linked to diabetes were the emphasis. Sex, age, race/ethnicity, and geographic location were among the important demographic factors that were used to stratify the data for the analysis. Temporal trends were examined using linear regression, while geographic disparities were assessed through chi-square tests and regression analysis. A significance level of p<0.05 was used for all tests. Results: According to the data, the death rate from diabetes increased significantly between 1999 and 2020. Between 1999 and 2020, the number of deaths from diabetes rose from 2,325 to 74,296. From 13.12 per 100,000 in 1999 to 276.92 per 100,000 in 2020, the age-adjusted death rate increased significantly, indicating a fivefold rise in mortality rates over the 21 years. After 2010, the pace of increase became especially steep, with notable yearly increases, especially in the years after 2013. In the Hispanic population alone, the number of deaths from diabetes more than quadrupled between 2013 and 2020, rising from 5,378 in 2013 to 10,395 in 2020. Male mortality rates were consistently greater than female death rates, indicating gender-based inequities as well. By 2020, there were 41,225 male deaths from diabetes (AAMR: 351.07). Notably, the death rate for men was almost 60% greater than that for women in 2020, with rates rising more sharply for men, particularly from 2010 onward. Significant differences in diabetes-related mortality were also observed by geography. Florida, Texas, and California had the largest mortality burdens; in California alone, 73,185 deaths, or 13.8% of all deaths, occurred. Lastly, more deaths were found to occur in medical facilities, especially inpatient settings, as well as at home or in hospice care, indicating a change in the place of death. Conclusions: The study shows a significant rise in diabetes-related mortality in the US over the past two decades, with males experiencing a higher mortality rate. Geographic disparities are also evident, with Southern states and large states like California, Texas, and Florida contributing disproportionately. The shift in death places suggests increased reliance on medical facilities, home care, and hospice care. The study calls for targeted public health interventions to address the growing burden of diabetes-related mortality, particularly in high-risk regions. Presentation: Sunday, July 13, 2025

Abstract Disclosure: M. Kakakhel: None. S. Cheema: None. R. Haider: None. F. Saleem: None. M.H. Alsaadi: None. H. Jawad: None. A.Z. Khan: None. A. Qadeer: None. M. Khawar: None. A. Ullah: None. M.F. Hemida: None. H. Ahmad: None. F. Ahmed: None. R. Alcaraz: None. Background: Globally, pediatric endocrine anomalies are a significant cause of early mortality. This study examines the discrepancies of paediatric endocrine abnormalities across various demographic and geographic groups in the United States, including gender, census region, race, age groups, and urbanization levels. Our objective was to identify disparities and trends to inform targeted interventions for the vulnerable demographics. Methods: Adhering to STROBES criteria, we extracted the mortality statistics due to endocrine disorders in the paediatric population from 1999-2020 from the Centre for Disease Control and Prevention-Wide Ranging Online Data for Epidemiologic research (CDC-WONDER). We calculated Age-Adjusted mortality rates (AAMRs) per 100,000 individuals. We calculated the Annual Percentage Change (APC) and Average Annual Percentage Change (AAPC) to assess trends stratified by age, gender, urbanization, race, census region, and states. Joinpoint regression model was used for analysis. Results: Between 1999 and 2020, 5161 deaths among children less than 1 year to 14 years of age in the United States were related to endocrine-related disorders. Most occurred due to thyroid disorders, diabetes mellitus, and other miscellaneous disorders related to glucose regulation and pancreatic internal secretion. The overall AAPC for endocrine-related mortality in the pediatric population increased from 1.4866 in 2008 to 2.4982 in 2020, with an AAPC of 0.1346 (95% CI: -3.1216 to -3.5003, p > 0.000001). The male population had higher AAPCs than the female population (women: -0.9938; men: 0.7223). There were notable racial differences: The Hispanic or Latino population had the highest AAPC (0.15), followed by the Black population (-0.195) and then the White population with the lowest AAPC (-0.166). According to the age group stratification, the age group with the highest AAPC was the 5-14 age group (2.03) and the lowest the less than 1 year age group, with AAPC (-2.007). There was a significant decrease in deaths in the less than 1 year age group from 1999 to 2020 (p value < 0.05) with APC (-2.0) (95% CI: -3.0191 to -0.9861) and for the 5-14 age group from 2010 to 2020 with APC (10.8861) (95% CI: 3.4843 to 18.8). Depending on the urban-rural categorization, metro areas had a higher AAPC (0.1508) compared to the non-metro areas with an AAPC (-0.4467). Conclusion: The overall decline in endocrine-related deaths in the pediatric population from 1999 to 2020 hails an improvement in efforts made to manage these disorders. However, the recent uptick in the Hispanic-Latino population emphasizes the necessity for targeted efforts to tackle these disparities and improve the health outcomes of affected communities as well as in children 1-4 years of age, depicting the increased vulnerability and need for added care in this age group. Presentation: Saturday, July 12, 2025

Abstract Disclosure: M. Kakakhel: None. F. Saleem: None. A. Ullah: None. K. Abdul Rehman: None. M.H. Alsaadi: None. H. Ahmad: None. A. Qadeer: None. R. Haider: None. M. Khawar: None. H. Jawad: None. J. Ikram: None. M. Filal: None. F. Ahmed: None. R.W. Kirchoff: None. Background: Thyroid disorders are common endocrine disorders including thyrotoxicosis, thyroiditis, hypothyroidism, and autoimmune thyroid diseases. Although it affects an estimated 200 million people globally, the epidemiology of this condition remains underexplored. Here we aimed to investigate mortality trends attributable to thyroid disorders in the US stratified by year, sex, race, and region. Methods: Using data from the CDC WONDER database, we calculated the crude and age-adjusted mortality rates (AAMRs) per 1,000,000 individuals (≥25 years) and used Joinpoint regression to determine annual and average annual percent change (AAPC) in the mortality trends. Results: Between 1999 and 2020, 561,089 deaths were recorded among patients with thyroid disorders in the United States, with an overall increasing trend in mortality (AAPC = 0.78). The age-adjusted mortality rate (AAMR) declined gradually from 1999 to 2018 (APC = -0.43, 95% CI: -0.65 to -0.22) before rising sharply from 2018 to 2020 (APC = 13.05, 95% CI: 6.39 to 20.12). Females had a higher AAMR (139.75) than males (81.99). Among racial and ethnic groups, Whites exhibited the highest AAMR (124.09), followed by American Indian or Alaskan Natives (121.36), Hispanics (89.13), Black or African Americans (88.44), and Asian or Pacific Islanders (59.36). Regional disparities showed the highest AAMR in the Midwest (130.45) and the lowest in the South (109.40). West Virginia (199.93) and North Dakota (198.82) reported the highest state-level AAMRs, while Nevada (60.18) and Florida (65.49) had the lowest. Non-metropolitan areas experienced higher mortality rates (141.04) than metropolitan areas (112.21), underscoring significant geographical and demographic variations in thyroid disorder-related mortality. Conclusion: Adults in the US have shown a rising trend in Thyroid-related deaths, with the highest mortality rates seen in females, Whites, and people living in the Midwest and Non-metropolitan areas. Our study also highlights the disparities across different demographics. Policy makers should focus on equitable access to healthcare to control this increasing mortality. Presentation: Monday, July 14, 2025

Abstract BACKGROUND IDH-mutant gliomas are diffusely infiltrating, slow-growing brain tumors that inevitably undergo malignant progression to grade IV astrocytoma - a highly aggressive and uniformly fatal disease with dismal prognosis. Despite extensive research, the molecular mechanisms underlying this transformation remain poorly understood. While alterations in canonical driver genes such as CDKN2A, PTEN, and PDGFRA have been implicated, the broader landscape of genetic drivers involved in IDH-mutant glioma progression is largely uncharacterized. In contrast to many other malignancies, IDH-mutant gliomas acquire relatively few somatic point mutations during progression - typically in the range of 30 to 40 - suggesting that alternative mechanisms may be driving tumor evolution. Notably, these tumors exhibit widespread chromosomal instability, resulting in a heterogeneous and highly prevalent pattern of copy number alterations (CNAs). This observation raises the possibility that gene dosage changes due to CNAs play a critical role in tumor progression. MATERIAL AND METHODS To investigate this hypothesis, we performed integrative analysis of IDH-mutant glioma cohorts from the Mayo Clinic and TCGA datasets (~1,500 patients), identifying twelve highly recurrent CNAs significantly associated with patient outcomes. We refer to these as malignant IDH-associated CNAs (mIDH CNAs). We hypothesized that these recurrent CNAs harbor driver genes that confer selective growth advantages, thereby contributing to their recurrence and prognostic relevance. To systematically assess the functional impact of genes within these regions, we developed a novel in vivo CRISPR-based screening platform, termed CRISPR-Knock Out and Activation Linked Assay (CRISPR-KOALA). This approach utilizes stereotaxic delivery of lentiviral gRNA libraries - targeting mouse orthologs of ~1,800 genes within recurrently deleted regions and ~1,200 genes within amplified regions - into the brains of a genetically engineered mouse model of IDH-mutant glioma (rs557mut;IDHmut;P53KO;AtrxKO;Cas9GFP: Yanchus et. al Science 2022). RESULTS Mice injected with mIDH CNA-targeting libraries developed tumors significantly more rapidly than control cohorts. Deconvolution of the CRISPR-KOALA screen identified several putative driver genes, including regulators of the NOTCH signaling pathway, lipid metabolism modulators such as THEM6, and the NMDA receptor modulator GRINA - implicating dysregulation of metabolism, immune response, and glutamatergic signaling in the malignant transformation of IDH-mutant gliomas. CONCLUSION Together, these findings provide critical insights into the genetic mechanisms driving the progression of IDH-mutant gliomas and identify new candidate genes and pathways for potential therapeutic targeting.

Importance Multiparametric magnetic resonance imaging (MRI), with or without prostate biopsy, has become the standard of care for diagnosing clinically significant prostate cancer. Resource capacity limits widespread adoption. Biparametric MRI, which omits the gadolinium contrast sequence, is a shorter and cheaper alternative offering time-saving capacity gains for health systems globally. Objective To assess whether biparametric MRI is noninferior to multiparametric MRI for diagnosis of clinically significant prostate cancer. Design, Setting, and Participants A prospective, multicenter, within-patient, noninferiority trial of biopsy-naive men from 22 centers (12 countries) with clinical suspicion of prostate cancer (elevated prostate-specific antigen [PSA] level and/or abnormal digital rectal examination findings) from April 2022 to September 2023, with the last follow-up conducted on December 3, 2024. Interventions Participants underwent multiparametric MRI, comprising T2-weighted, diffusion-weighted, and dynamic contrast–enhanced (DCE) sequences. Radiologists reported abbreviated biparametric MRI first (T2-weighted and diffusion-weighted), blinded to the DCE sequence. After unblinding, radiologists reported the full multiparametric MRI. Patients underwent a targeted biopsy with or without systematic biopsy if either biparametric MRI or multiparametric MRI was suggestive of clinically significant prostate cancer. Main outcomes and measures The primary outcome was the proportion of men with clinically significant prostate cancer. Secondary outcomes included the proportion of men with clinically insignificant cancer. The noninferiority margin was 5%. Results Of 555 men recruited, 490 were included for primary outcome analysis. Median age was 65 (IQR, 59-70) years and median PSA level was 5.6 (IQR, 4.4-8.0) ng/mL. The proportion of patients with abnormal digital rectal examination findings was 12.7%. Biparametric MRI was noninferior to multiparametric MRI, detecting clinically significant prostate cancer in 143 of 490 men (29.2%), compared with 145 of 490 men (29.6%) (difference, −0.4 [95% CI, −1.2 to 0.4] percentage points; P = .50). Biparametric MRI detected clinically insignificant cancer in 45 of 490 men (9.2%), compared with 47 of 490 men (9.6%) with the use of multiparametric MRI (difference, −0.4 [95% CI, −1.2 to 0.4] percentage points). Central quality control demonstrated that 99% of scans were of adequate diagnostic quality. Conclusion and relevance In men with suspected prostate cancer, provided image quality is adequate, an abbreviated biparametric MRI scan, with or without targeted biopsy, could become the new standard of care for prostate cancer diagnosis. With approximately 4 million prostate MRIs performed globally annually, adopting biparametric MRI could substantially increase scanner throughput and reduce costs worldwide. Trial registration ClinicalTrials.gov Identifier: NCT04571840

Background: Artificial Intelligence (AI)-driven Decision Support Systems (DSSs) promise improvements in diagnostic accuracy and clinical workflow efficiency, but their adoption is hindered by inadequate confidence calibration, limited transparency, and poor alignment with real-world decision processes, which limit clinician trust and lead to high override rates. Methods: We developed and validated a dynamic scoring framework to enhance trust in AI-generated diagnoses by integrating AI confidence scores, semantic similarity measures, and transparency weighting into the override decision process using 6689 cardiovascular cases from the MIMIC-III dataset. Override thresholds were calibrated and validated across varying transparency and confidence levels, with override rate as the primary acceptance measure. Results: The implementation of this framework reduced the override rate to 33.29%, with high-confidence predictions (90–99%) overridden at a rate of only 1.7%, and low-confidence predictions (70–79%) at a rate of 99.3%. Minimal transparency diagnoses had a 73.9% override rate compared to 49.3% for moderate transparency. Statistical analyses confirmed significant associations between confidence, transparency, and override rates (p < 0.001). Conclusions: These findings suggest that enhanced transparency and confidence calibration can substantially reduce override rates and promote clinician acceptance of AI diagnostics. Future work should focus on clinical validation to optimize patient safety, diagnostic accuracy, and efficiency.

Retrieval-Augmented Generation (RAG) offers a promising strategy to harness large language models (LLMs) for delivering up-to-date, accurate clinical guidance while reducing physicians’ cognitive burden, yet its effectiveness hinges on query clarity and structure. We propose an adaptive Self-Query Retrieval (SQR) framework that integrates three refinement modules—PICOT (Population, Intervention, Comparison, Outcome, Time), SPICE (Setting, Population, Intervention, Comparison, Evaluation), and Iterative Query Refinement (IQR)—to automatically restructure and iteratively enhance clinical questions until they meet predefined retrieval-quality thresholds. Implemented on Gemini-1.0 Pro, we benchmarked SQR using thirty postoperative rhinoplasty queries, evaluating responses for accuracy and relevance on a three-point Likert scale and for retrieval quality via precision, recall, and F1 score; statistical significance was assessed by one-way ANOVA with Tukey post-hoc testing. The full SQR pipeline achieved 87% accuracy (Likert 2.4 ± 0.7) and 100% relevance (Likert 3.0 ± 0.0), significantly outperforming a non-refined RAG baseline (50% accuracy, 80% relevance; p < 0.01 and p = 0.03). Precision, recall, and F1 rose from 0.17, 0.39 and 0.24 to 0.53, 1.00, and 0.70, respectively, while PICOT-only and SPICE-only variants yielded intermediate improvements. These findings demonstrate that automated structuring and iterative enhancement of queries via SQR substantially elevate LLM-based clinical decision support, and its model-agnostic architecture enables rapid adaptation across specialties, data sources, and LLM platforms.

Abstract Background/ Objectives: Agentic AI represents a promising evolution of AI technology applied to healthcare, with systems increasingly capable of operating autonomously to achieve defined clinical goals. However, the literature lacks conceptual clarity between “AI agents” and “agentic AI”, and few studies have rigorously explored their clinical applications. Therefore, this study aims to conduct a novel systematic review addressing this gap by examining agentic AI systems in healthcare settings, characterizing their applications, features, outcomes, and limitations, and clarifying the conceptual distinctions between AI agents and agentic AI systems using predefined and objective criteria. Methods A comprehensive search was conducted across PubMed, Embase, Cochrane, Scopus, and Google Scholar on April 6th, 2025. Studies were included if they involved AI systems in healthcare settings that demonstrated the following agentic features: autonomous operation, goal-directed behavior, and initiating action. Data on the clinical tasks achieved by the agents, key findings, features, and limitations were collected from the included studies. Screening and extraction followed PRISMA guidelines, with Risk of Bias assessed using ROBINS-I and Cochrane's Risk of Bias tools. Results Of 984 retrieved records, seven studies met the inclusion criteria, spanning domains such as emergency medicine, oncology, radiology, and rehabilitation. Multi-agent architecture was frequently used to decompose and coordinate complex workflows. Among the included studies, the AIs showed high accuracy in diagnosing cancer patients, conducting treatment plans, sending alerts, coaching messages, analysing image data, and adapting to challenging experimental scenarios. While demonstrating potential for improved efficiency, task accuracy, and patient engagement, significant limitations were noted: narrow task scope, lack of physical agency, limited clinical validation, and barriers to integration into real-world healthcare systems. Only one system had been deployed in a patient-facing trial setting. Conclusion The current literature suggests an emerging role and application of Agentic AI, holding promise with the potential to revolutionize diagnostics, triage, treatment planning, and patient management. However, real-world implementation and evaluations in the literature are limited. Future research must address critical validation, regulation, ethics, and clinical integration challenges to realize their full potential. Clear operational definitions and frameworks for evaluating agency are essential to support safe and effective deployment of these systems.

Background: Data accessibility remains a significant barrier in healthcare AI due to privacy constraints and logistical challenges. Synthetic data, which mimics real patient information while remaining both realistic and non-identifiable, offers a promising solution. Large Language Models (LLMs) create new opportunities to generate high-fidelity clinical conversations between patients and physicians. However, the value of this synthetic data depends on careful evaluation of its realism, accuracy, and practical relevance. Objective: To assess the performance of four leading LLMs: ChatGPT 4.5, ChatGPT 4o, Claude 3.7 Sonnet, and Gemini Pro 2.5 in generating synthetic transcripts of patient–physician interactions in plastic surgery scenarios. Methods: Each model generated transcripts for ten plastic surgery scenarios. Transcripts were independently evaluated by three clinically trained raters using a seven-criterion rubric: Medical Accuracy, Realism, Persona Consistency, Fidelity, Empathy, Relevancy, and Usability. Raters were blinded to the model identity to reduce bias. Each was rated on a 5-point Likert scale, yielding 840 total evaluations. Descriptive statistics were computed, and a two-way repeated measures ANOVA was used to test for differences across models and metrics. In addition, transcripts were analyzed using automated linguistic and content-based metrics. Results: All models achieved strong performance, with mean ratings exceeding 4.5 across all criteria. Gemini 2.5 Pro received mean scores (5.00 ± 0.00) in Medical Accuracy, Realism, Persona Consistency, Relevancy, and Usability. Claude 3.7 Sonnet matched the scores in Persona Consistency and Relevancy and led in Empathy (4.96 ± 0.18). ChatGPT 4.5 also achieved perfect scores in Relevancy, with high scores in Empathy (4.93 ± 0.25) and Usability (4.96 ± 0.18). ChatGPT 4o demonstrated consistently strong but slightly lower performance across most dimensions. ANOVA revealed no statistically significant differences across models (F(3, 6) = 0.85, p = 0.52). Automated analysis showed substantial variation in transcript length, style, and content richness: Gemini 2.5 Pro generated the longest and most emotionally expressive dialogues, while ChatGPT 4o produced the shortest and most concise outputs. Conclusions: Leading LLMs can generate medically accurate, emotionally appropriate synthetic dialogues suitable for educational and research use. Despite high performance, demographic homogeneity in generated patients highlights the need for improved diversity and bias mitigation in model outputs. These findings support the cautious, context-aware integration of LLM-generated dialogues into medical training, simulation, and research.

The increasing integration of large language models (LLMs) into healthcare presents significant opportunities, but also critical risks related to patient safety, accuracy, and ethical alignment. Despite these concerns, no standardized framework exists for systematically evaluating and stress testing LLM behavior in clinical decision-making. The PIEE cycle—Planning and Preparation, Information Gathering and Prompt Generation, Execution, and Evaluation—is a structured red-teaming framework developed specifically to address artificial intelligence (AI) safety risks in healthcare decision-making. PIEE enables clinicians and informatics teams to simulate adversarial prompts, including jailbreaking, social engineering, and distractor attacks, to stress-test language models in real-world clinical scenarios. Model performance is evaluated using specific metrics such as true positive and false positive rates for detecting harmful content, hallucination rates measured through adapted TruthfulQA scoring, safety and reliability assessments, bias detection via adapted BBQ benchmarks, and ethical evaluation using structured Likert-based scoring rubrics. The framework is illustrated using examples from plastic surgery, but is adaptable across specialties, and is intended for use by all medical providers, regardless of their backgrounds or familiarity with artificial intelligence. While the framework is currently conceptual and validation is ongoing, PIEE provides a practical foundation for assessing the clinical reliability and ethical robustness of LLMs in medicine.