Citations

Obstructive sleep apnea (OSA) is a common, heterogeneous disorder managed through a multidisciplinary pathway that connects recognition, physiologic testing, treatment selection, specialty referral, insurance coverage, and longitudinal reassessment. This narrative review provides a clinically oriented, system-level overview of the adult OSA care ecosystem in the United States for an international clinical audience, with particular emphasis on the evolving role of otolaryngology. We describe the organization and complementary functions of sleep physicians, otolaryngologists, dentists, and allied health professionals. We also outline the roles of polysomnography and home sleep apnea testing, along with major therapeutic pathways including positive airway pressure, oral appliance therapy, weight management, upper-airway surgery, and neurostimulation. Otolaryngology provides an important interface between physiologic diagnosis and airway-specific treatment by integrating anatomic evaluation, sleep-state airway phenotyping, intervention selection, and post-treatment reassessment. Recent advances in neurostimulation, pharmacologic weight management, wearable and home-based monitoring, airflow analysis, and artificial intelligence are expanding opportunities for accessible and phenotype-informed care. Important challenges include underdiagnosis, treatment attrition, fragmented transitions, variable coverage, and reliance on the apnea–hypopnea index. Although health-care structures differ internationally, multidisciplinary coordination, individualized treatment selection, and longitudinal reassessment provide broadly applicable principles for developing patient-centered OSA programs.

Functional constipation (FC) and irritable bowel syndrome with constipation (IBS-C) are prevalent disorders of gut–brain interaction with limited effective treatments. The vagus nerve exerts considerable influence on the gastrointestinal (GI) tract, indicating that noninvasive electrical neuromodulation targeting this nerve could provide a promising therapeutic avenue for both FC and IBS-C. This review explores several emerging noninvasive neuromodulation techniques involving the vagus nerve, such as transcutaneous auricular vagal nerve stimulation, percutaneous electrical nerve field stimulation, and transcutaneous electrical acustimulation. The clinical applications of these methods in managing FC and IBS-C are evaluated. Existing evidence supports the efficacy and safety of noninvasive neuromodulation in relieving symptoms of these conditions. Future investigations should aim to refine these techniques, advance device technology, and further clarify the mechanisms of action.

Ex vivo liver perfusion (EVLP) sustains human or large animal livers outside the body under near-physiological conditions, enabling functional monitoring for lactate clearance, bile production, and oxygen consumption and allowing targeted therapeutic interventions. Originally developed to optimize donor grafts for transplantation, EVLP has evolved into a powerful translational research platform bridging preclinical discovery and early clinical translation. This review discusses EVLP as a platform for gene therapy, immunotherapy, pharmacology, and personalized medicine, with particular emphasis on gene- and immune-based interventions as mechanistically mature exemplars. We consolidate advances in pharmacological testing and toxicity modeling, viral and non-viral gene delivery, genome engineering, and immunomodulation using perfused livers. We further describe emerging applications, including autologous EVLP pathways for organ-confined therapy, ex vivo liver surgery, and bioengineering strategies such as biliary organoid repair, RNA interference, and mitochondrial transfer. We highlight how these applications align with a paradigm shift in biomedical research, including the NIH’s recent initiative to prioritize human-based experimental models over animal-only studies. By leveraging transplant-declined or surgically resected organs that would otherwise be unused, ex vivo perfusion bridges the gap between pre-clinical testing and clinical practice, enabling real-time evaluation of interventions in functional human tissue. We discuss both the scientific opportunities afforded by EVLP and the technical, biosafety, and ethical challenges that must be addressed to enable responsible clinical translation.

Background: Colorectal cancer (CRC) remains one of the common causes of cancer-related morbidity and mortality in the United States. Emerging increases in early-onset CRC and persistent disparities remain major public health concerns. Objective: To characterize long-term CRC incidence and mortality trends in the United States, with particular emphasis on identifying demographic groups experiencing rising or plateauing disease burden. Methods: We conducted a population-based ecological study using CRC incidence data from the United States Cancer Statistics database (1999–2022) and mortality data from CDC WONDER (1999–2023). Temporal trends were evaluated using Joinpoint regression to estimate annual percent change (APC) and average annual percent change (AAPC), stratified by age, sex, race/ethnicity, and urbanization level. Results: Overall CRC incidence and mortality declined during the study period (both AAPC approximately −2.08%). However, incidence increased significantly among adults younger than 55 years, particularly those aged 20–24 years. Mortality also increased among younger adults aged 25–44 years and remained elevated among individuals aged 45–54 years. Non-Hispanic Black individuals consistently experienced the highest incidence and mortality rates. Mortality declines plateaued in several older age groups after 2018–2020. Conclusions: Despite overall improvements in CRC burden, increasing incidence and mortality among younger adults and persistent disparities highlight the need for risk-stratified prevention, earlier detection strategies, and tailored therapeutic approaches.

Abstract Objectives To comprehensively evaluate the validity of International Classification of Diseases, Tenth Revision, Clinical Modification (ICD-10-CM) codes for both prevalent diagnoses and less common diseases, and to assess the performance of a large language model (LLM)-based system in validating these codes. Materials and Methods This retrospective study analyzed hospital admissions from the Medical Information Mart for Intensive Care IV (MIMIC-IV) database. We developed a validated LLM-based system using GPT-4o, refined through iterative prompt engineering, to assess ICD-10-CM code validity. We measured the positive predictive value (PPV) of ICD-10-CM codes, PPV of principal and secondary diagnoses, and the performance of an LLM-based system in code validation. Results Among 865 079 assigned codes, the PPV was 84.6% (95% CI, 84.5%-84.6%). Principal diagnoses had a PPV of 93.9% (95% CI, 93.7%-94.1%), while secondary diagnoses had a PPV of 83.8% (95% CI, 83.7%-83.9%). The LLM system demonstrated high performance in validating ICD codes, achieving 93.6% accuracy, 95.4% sensitivity, and 85.2% specificity. Among correctly assigned secondary diagnoses, the majority (67.9%) represented historical or baseline conditions, while 32.1% reflected active conditions that deviated from baseline status; 22.3% of these emerged after hospital admission. PPV decreases with later diagnosis positions, with the largest decline occurring between principal and secondary diagnoses. Discussion and Conclusion In this large-scale evaluation, ICD-10-CM codes exhibited generally high accuracy, though variability existed by position and condition type. A validated LLM system performed comparably to physician review and offers a scalable means to improve coding accuracy. These findings support the potential for integrating LLM-based auditing into routine workflows to strengthen the quality of administrative and research data.

ABSTRACT Objective To comprehensively evaluate the validity of ICD-10-CM codes for both prevalent diagnoses and less common diseases, and to assess the performance of a large language model (LLM)-based system in validating these codes. Materials and Methods This retrospective study analyzed hospital admissions from the Medical Information Mart for Intensive Care (MIMIC-IV) database. We developed a validated LLM-based system using GPT-4o, refined through iterative prompt engineering, to assess ICD-10-CM code validity. We measured the PPV of ICD-10-CM codes, PPV of principal and secondary diagnoses, and the performance of an LLM-based system in code validation. Results Among 865,079 assigned codes, the PPV was 84.6% (95% CI, 84.5%-84.6%). Principal diagnoses had a PPV of 93.9% (95% CI, 93.7%-94.1%), while secondary diagnoses had a PPV of 83.8% (95% CI, 83.7%-83.9%). The LLM system demonstrated high performance in validating ICD codes, achieving 93.6% accuracy, 95.4% sensitivity and 85.2% specificity. Among correctly assigned secondary diagnoses, the majority (67.9%) represented historical or baseline conditions, while 32.1% reflected active conditions that deviated from baseline status; 22.3% of these emerged after hospital admission. PPV decreases with later diagnosis positions, with the largest decline occurring between principal and secondary diagnoses. Discussion and Conclusion In this large-scale evaluation, ICD-10-CM codes exhibited generally high accuracy, though variability existed by position and condition type. A validated LLM system performed comparably to physician review and offers a scalable means to improve coding accuracy. These findings support the potential for integrating LLM-based auditing into routine workflows to strengthen the quality of administrative and research data.

Objectives: Diabetes mellitus and chronic pancreatitis are established risk factors for pancreatic ductal adenocarcinoma (PDAC). The co-occurrence of exocrine and endocrine pancreatic dysfunction may increase the risk of progression to PDAC and could play a role in early detection. We sought to explore this hypothesis by comparing PDAC risk among patients with diabetes mellitus (DM) and low fecal elastase (FE) in a real-world cohort of patients. Materials and Methods: A retrospective cohort study was conducted using the TriNetX research network including adults with continuous follow-up from 2016 to 2023. The cohort was stratified into 4 groups based on the presence or absence of DM, and normal FE (>200 µg/g) or low FE (<200 µg/g). The 4 groups included: (1) low FE and DM, (2) low FE without DM, (3) normal FE and DM, (4) normal FE without DM. Propensity score matching was performed to balance covariates including age, sex, race, comorbidities to allow for a 6-way comparison of PDAC risk between groups. Time-to-event analyses were conducted with Kaplan-Meier curves and Cox proportional hazards models to estimate hazard ratios (HR) and 95% CIs for the primary outcome of PDAC incidence. Results: Of the study cohort, 29,207 had fecal elastase levels. Low FE (<200 µg/g) was observed in 33% (n = 9585) of these patients. After 6-way comparisons, we observed a higher PDAC risk among individuals with both DM and low FE compared with other groups. The highest PDAC risk was observed in patients with both a low FE (<200 µg/g) and DM compared with those with a normal FE and without a history of DM (HR: 3.68; 95% CI, 2.07–6.53; P < 0.001). Sensitivity analysis using a stricter fecal elastase cutoff of <100 µg/g showed a similar finding, with the highest hazard ratio for PDAC among patients with both DM and very low FE levels (HR: 3.07, 95% CI, 1.37–6.91; P = 0.004). Conclusions: The coexistence of low fecal elastase and DM amplifies the risk of pancreatic ductal adenocarcinoma in our cohort of patients. These results suggest that fecal elastase values may be used to enrich the diabetic population for future PDAC risk stratification, though more studies are needed to validate these findings.