Citations

Abstract Background Delirium is often underdiagnosed in clinical practice and is not routinely coded for billing. While manual chart review can identify delirium, it is labor-intensive and impractical for large-scale studies. Natural language processing (NLP) can analyze unstructured text in electronic health records (EHRs) to extract meaningful clinical information. Methods To support national integration of NLP for EHR-based delirium identification across different institutions, we launched the Delirium Interest Group within the national Evolve to Next-Gen Accrual to Clinical Trials (ENACT) NLP Working Group. This paper outlines our initial efforts to standardize, evaluate, and translate an NLP-based delirium detection model into the i2b2/ENACT platform. Results Multisite contextual inquiry identified several key challenges, including variations in local screening practices (eg, tools used, documentation frequency, and quality control), the need for harmonized definitions in the context of EHRs, and the complexity of modeling temporal logic. Multisite NLP evaluation revealed variable performance degradation driven by differences in delirium screening practices, clinical documentation patterns and semantics, and note syntactic structures. Conclusion Our work represents an important first step toward enabling scalable and standardized NLP-based delirium detection across institutions. By engaging diverse institutions through the ENACT NLP Working Group, we identified shared challenges and site-specific variations that impact model implementation and performance. Our collaborative approach enabled the development of a more robust framework for delirium identification across heterogeneous EHR systems. Future efforts will build on this foundation to enhance the validity, usability, and translational impact of delirium detection.

Abstract Diffuse large B-cell lymphoma (DLBCL) poses a challenge in hematology given its varied symptoms, and the complex interplay between disease and treatment effects on health-related quality of life (HRQoL). The phase 3 POLARIX study demonstrated superior progression-free survival and a similar safety profile with polatuzumab vedotin plus rituximab, cyclophosphamide, doxorubicin, and prednisone (Pola-R-CHP) vs R-CHOP (rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone) in patients with previously untreated DLBCL. Here, we evaluate HRQoL through patient-reported outcome (PRO) instruments to fully characterize the patient experience in the POLARIX study. Changes from baseline in HRQoL, lymphoma symptoms, and gastrointestinal (GI) symptoms were assessed, as well as incidence and severity of common symptoms by PROs vs clinician-reported adverse events (AEs). Baseline characteristics of PRO-evaluable patients (N = 874) were consistent. Comparison between PROs and clinician-reported AEs revealed a notable discordance; patients generally reported a higher incidence of symptoms than clinicians, emphasizing the need for patient-centric tools to accurately capture the patient experience. Both treatments exhibited rapid and sustained improvements in HRQoL and lymphoma symptoms, with the most substantial improvements seen in global health status/QoL, lymphoma symptoms, fatigue, role, emotional, and social functioning. GI symptoms (diarrhea, constipation, nausea, and vomiting) were generally similar between treatment arms and returned to baseline levels after treatment completion. These HRQoL data underscore the complementarity of PROs, as an adjunct to clinician-reported AEs, in evaluating the efficacy and tolerability of new treatments, including Pola-R-CHP, which may represent a new benchmark for patient-reported HRQoL in previously untreated DLBCL. This trial was registered at www.clinicaltrials.gov as NCT03274492.

Abstract Objective Increased frequency of neuroimaging has led to enhanced identification of small nonfunctioning pituitary adenomas (NFPAs) leading, in many cases, to extensive follow-up. However, the value of ongoing monitoring of these incidental lesions remains unclear. The study aims to determine the need for surgical intervention and assess the risks of developing new endocrinopathies during follow-up in patients with conservatively treated micro-NFPAs. Design A systematic review and individual participant data (IPD) meta-analysis. Methods We conducted a bibliographical search of PubMed and EMBASE to identify relevant studies. Authors of eligible studies were invited to share IPD. Cohort studies including patients with conservatively treated micro-NFPAs with at least 1 follow-up magnetic resonance imaging were considered eligible. Fourteen studies met inclusion criteria. Six authors provided IPD (N = 647). Data were reanalyzed for verification. In cases of discrepancies the original authors were contacted for authentication. Results Risk estimates were reported as number of events per 100 person-years (PYs). Estimates were pooled using the 2-step approach. Overall probability of surgery was 0.2/100 PYs (95% CI: 0.0-0.4; I2 = 28%). Probability of surgery due to visual impairment was 0.1/100 PYs (95% CI: 0.0-0.2; I2 = 0%). Both were independent of baseline tumor size (≥6 or <6 mm), sex, or age (P values >.40). Risk of developing a new endocrinopathy was 1.0/100 PYs (95% CI: 0.4-1.6; I2 = 0%). Data for classical meta-analysis were available for 7 studies (N = 1089) and supported the IPD results. Conclusions These data suggest that routine follow-up of micro-NFPAs can be reduced significantly and that available guidelines should be revisited.

Background Endovascular thrombectomy (EVT) has transformed acute ischemic stroke (AIS) care, with onset-to-puncture (OTP) time widely recognized as a critical determinant of outcome. However, emerging evidence suggests that in-hospital procedure time (PT)—from arterial puncture to final recanalization—may have an equally or more significant impact. This study examines the relative contribution of PT versus OTP to functional outcomes in patients with AIS undergoing EVT. Methods A retrospective analysis was conducted of 6644 patients with AIS treated at 44 international stroke centers from the Stroke Thrombectomy and Aneurysm Registry (STAR; 2016–2023). Multivariable regression, time-equivalence analysis, and marginal effects modeling were used to assess associations between PT, OTP, and 90-day modified Rankin Scale (mRS) outcomes. Centers were stratified by procedural efficiency and compared using propensity score matching (PSM). Mediation analysis evaluated whether PT accounted for inter-center differences. Results PT and OTP were independently associated with functional outcomes; however, PT had a significantly stronger effect (adjusted OR for mRS 0–2: PT=0.56 vs OTP=0.96 per hour). Each 5 min increase in PT was equivalent to 78–100 min of additional OTP in outcome impact. Centers with faster average PT had higher rates of functional independence (number needed to treat (NNT)=10), fewer complications, and lower symptomatic intracranial hemorrhage rates. PT significantly mediated the relationship between center tier and outcomes (Sobel’s P<0.001). Conclusion While minimizing OTP remains important, PT exerts a greater influence on outcomes after EVT. Procedural efficiency should be emphasized in stroke systems of care and included in center performance metrics to improve patient outcomes.

BACKGROUND AND OBJECTIVES: Flow-diverting stents (FDSs) have improved treatment options for intracranial aneurysms. Although FDS have proven successful in treating aneurysms that would previously have required complex microsurgical or endovascular repairs, treatment failures remain a concern. In this multicenter study, we aim to analyze the predictors of residual aneurysmal filling at early angiographic follow-up. METHODS: A retrospective analysis of a prospectively maintained neuroendovascular database at 14 high-volume US centers was conducted for all patients who underwent FDS placement for intracranial aneurysms between 2011 and 2019. Aneurysms were graded on the O'Kelly-Marotta scale (OKM). RESULTS: A total of 2277 patients underwent FDS embolization for intracranial aneurysms during the study period. The patient's mean age was 55 (SD ± 13.5) years, and 82% were female. The median maximal aneurysm diameter was 6.7 mm (IQR 4.10-11 mm). Complete aneurysm obliteration (OKM grade 4) occurred in 1109 patients (68%), and 146 (9%) were OKM grade 3, 188 (12%) were OKM grade 2, and 174 (11%) were OKM grade 1. In the adjusted multivariate analysis, individuals in the 65-74 years and 75+ years age groups had significantly increased risk of residual aneurysm filling at 6-month follow-up compared with the 18-44 years age group (odds ratio [OR] 1.92 (95% CI: 1.33-2.78) and 2.92 (95% CI: 1.79-4.77), respectively). Aneurysms arising from the superior cerebellar artery, the basilar artery, and the carotid terminus were more likely to have a residual filling (OR 12.6, 95% CI: 1.39–14.7; OR 4.27, 95% CI: 1.39–14.7; OR 3.16, 95% CI: 1.24–8.06, respectively). CONCLUSION: As the utilization of FDS for intracranial aneurysms becomes more widespread, the efficacy of aneurysm obliteration needs to be assessed. Our results indicate that several patient-dependent factors, such as age, aneurysm location, and size, predict the likelihood of residual aneurysm filling 6 months post-treatment.

Abstract Measurable residual disease (MRD) assessment by flow cytometry (FC) plays an essential role in prognosis and therapy escalation of B-cell acute lymphoblastic leukemia (B-ALL). However, the high degree of expertise and manual analysis time required limits the availability of this assay. To overcome this limitation, we developed a data-enhancing artificial intelligence (AI) pipeline that accelerates and simplifies MRD analysis. Unaltered FC files from 171 B-ALL MRD-positive and 89 MRD-negative cases were processed through an AI pipeline trained with 31 expert-gated negative controls. Cluster-informed downsampling reduced FC files from 1.2 million to 155 884 cells per case, on average, (87% cellularity reduction), whereas preserving small MRD populations (median, 100% retention for MRD of &lt;1%) and allowing for true percentage MRD estimates using a correction factor. A deep neural network cell classifier automatically identified normal hematopoietic subsets (macro-averaged F1 score of 0.86); and an AI measure of anomaly discriminated B-ALL from benign mononuclear (area under the curve [AUC] of 0.98) or B-lymphoid cells (AUC of 0.94). Manual analysis of AI-enhanced files was completed in only 1.01 minutes per case, on average (standard deviation of ±0.57); with 100% positive agreement with conventional analysis (for MRD of ≥0.01%), 100% negative agreement, and excellent quantitative correlation (R2 = 0.92). Our cloud-based AI-enhancement solution accelerates B-ALL MRD identification without compromising test performance and has the potential of facilitating B-ALL MRD analysis by more clinical laboratories.

Abstract Background Lung adenocarcinoma in never-smokers is a major public health burden, especially among East Asian women. Polygenic risk scores (PRSs) are promising for risk stratification but are primarily developed in European-ancestry populations. We aimed to develop and validate single- and multi-ancestry PRSs for East Asian never-smokers to improve lung adenocarcinoma risk prediction. Methods PRSs were developed using genome-wide association study summary statistics from East Asian (8002 cases; 20 782 controls) and European (2058 cases; 5575 controls) populations. Single-ancestry models included a PRS with 25 single nucleotide variants (PRS-25), PRS clumping and thresholding (CT), and LDpred2; multi-ancestry models included LDpred2 plus European PRS-128, PRS-CSx, and clumping and thresholding, super learning, empirical Bayes (CT-SLEB). Performance was evaluated in independent East Asian data from the Female Lung Cancer Consortium and externally validated in the Nanjing Lung Cancer Cohort. We assessed predictive accuracy via area under the curve operating procedure (AUC), with 10-year and (ages 30-80 years) absolute risks estimates. Results The best multi-ancestry PRS, using East Asian and European data via CT-SLEB, outperformed the best East Asian-only (LDpred2; AUC = 0.629, 95% confidence interval [CI] = 0.618 to 0.641), achieving an AUC of 0.640 (95% CI = 0.629 to 0.653) and odds ratio of 1.71 (95% CI = 1.61 to 1.82) per standard deviation increase. Nanjing Lung Cancer Cohort validation confirmed robust performance (AUC = 0.649, 95% CI = 0.623 to 0.676). The top 20% PRS group had a 3.92-fold higher lung adenocarcinoma risk than the bottom 20%. Further, the top 5% PRS group reached a 6.69% lifetime absolute risk. Notably, this group reached the average population 10-year lung adenocarcinoma risk at age 50 years (0.42%) by age 41 years, 9 years earlier. Conclusions Multi-ancestry PRS approaches enhance lung adenocarcinoma risk stratification in East Asian never-smokers, with consistent external validation, suggesting future clinical utility.

Abstract Error analysis is a critical step in evaluating and improving clinical concept extraction models, the most common clinical natural language processing (NLP) task. Unlike corpus annotation, which follows standardized protocols for creating gold-standard datasets, error analysis requires nuanced judgment grounded in both clinical expertise and NLP knowledge. This is especially important given the heterogeneity of clinical text, where variations in documentation style, note structure, and terminology can substantially influence model behavior. Despite its importance, there is currently no standardized, user-level framework to support systematic error analysis in clinical concept extraction task. In this study, we developed and validated MedError, a machine-assisted, human-in-the-loop framework designed to standardize and enhance error analysis for clinical concept extraction tasks. We collected and manually curated a corpus of 1,187 unique errors from a total of 4,237 notes across three different distinct hospitals. The error categories were defined using our previously validated error taxonomy and included 480 false negatives and 707 false positives across 25 error types and 48 clinical concept categories. We evaluated the performance of three proprietary and three open-source large language models (LLMs) in automatically classifying these errors into 26 and 15 predefined categories. We further developed a machine-assisted framework, MedError, which integrates best practices in error analysis, LLM-assisted classification and reasoning, and a user-friendly interface to enable more efficient, reproducible, and context-aware error analysis. The framework supports both single-site and federated multisite error analysis, facilitating the effective translation of clinical NLP systems into real-world settings.

ABSTRACT Background Ventricular tachycardia (VT) is a major contributor to sudden cardiac death among patients with ischemic heart disease (IHD). While catheter ablation reduces the arrhythmic burden, the impact on long‐term outcomes remains uncertain. We assess the efficacy of catheter ablation in patients with IHD. Methods We performed a meta‐analysis by systematically searching electronic databases from inception to April 2025. Pairwise, reconstructed time‐to‐event data meta‐analyses and Trial Sequential analyses were conducted. Results A total of seven studies (1192 participants) were included. Among patients with IHD with ICD, VT ablation was associated with a significant reduction in ICD shocks (RR = 0.50, 95% CI = [0.34–0.74]), a trend toward reduced VT storm (RR = 0.64, 95% CI = [0.40–1.04], p = 0.07), and lower rates of cardiovascular hospitalization (RR = 0.73, 95% CI = [0.53–1.01], p = 0.06) compared to ICD alone. However, no significant differences were observed when compared to patients receiving ICD plus antiarrhythmic drugs (AADs). VT ablation showed no significant impact on mortality or VT/VF recurrence compared to either ICD alone or ICD with AADs. Reconstructed Kaplan–Meier analysis for all‐cause mortality showed no significant difference between VT ablation and AAD + ICD. Trial Sequential Analysis provided conclusive evidence for VT/VF recurrence, while further data is required for other outcomes. Conclusion Among patients with IHD with ICD, VT ablation reduces ICD shocks, VT storm, and cardiovascular hospitalizations compared to ICD‐only therapy, but offers no significant advantage over ICD combined with AADs. VT ablation does not impact overall mortality or the recurrence of VT/VF. Trial sequential analysis confirmed conclusive evidence for VT/VF recurrence, while additional data is needed for other outcomes.

Abstract Background Patients with end-stage heart failure and chronic kidney disease requiring dual-organ transplantation (DOT) face significant challenges in utilizing durable mechanical circulatory support due to the risks associated with renal replacement therapies (RRTs) and multi-organ failure. Given the limited options available for long-term support in this patient population, there remains a critical need for alternative strategies to optimize end-organ function and bridge patients safely to transplant. With prolonged waitlist times for DOT, we present our experience with the Impella 5.5 as temporary mechanical circulatory support, demonstrating its potential to provide hemodynamic stability and support as a bridge to transplantation (BTT) in this complex cohort. Methods A single-centre retrospective review was completed of all patients listed for single-organ transplantation or DOT between December 2019 and November 2022 at Mayo Clinic in Florida, supported by the Impella 5.5 intended as BTT. The focus of this analysis was patients requiring RRT or listed for heart/kidney transplantation. Data were extracted from the electronic health record at baseline and during their transplant episode of care after institutional review board approval as exempt status for retrospective data collection. Results A total of 41 patients were supported with Impella 5.5, intended as BTT. All patients underwent successful transplantation. We focus on the 10 patients with Impella support who underwent DOT. In the DOT group, the median age at transplantation was 63 years (59–66), with nine males and one female. The baseline median ejection fraction was 19% (15–22), with 50% Caucasian and 50% African American and an even split between ischaemic and non-ischaemic aetiology. Median body mass index was 30 kg/m2 (26–31), and 60% were in blood group O. The median time on the waitlist for DOT patients was 53 days (29–75). Perioperative management of DOT Impella patients demonstrated baseline haemodynamics of RA 11 mmHg (7–16), mean PA 36 mmHg (32–47), PCWP 29 mmHg (21–35), mixed venous saturation (SVO2%) 51 (46–61) and Fick CI 2.03 L/min/m2 (1.66–2.5). Post-Impella placement haemodynamics demonstrated significant improvements in RA pressure to 5 mmHg (4–8), P = 0.02, SVO2 to 70% (65–72), P = 0.002, and Fick CI to 5.5 (5.2–8), P = 0.03. The average duration of support was 44 days (range 10–94). The median glomerular filtration rate at baseline was 27 mmol/L (16–29). Twenty-four hour urine protein averaged 168 mg/24 h (range 87–328), with the 24 h creatinine clearance of 29 mg/24 h (range 24–35). Eight of the 10 patients required continuous or intermittent RRT before DOT. The median total duration of RRT (including Impella support) was 36 days (9–72). DOT recipients had a 1 year survival of 90%, with an average follow-up of 432 days. Conclusions Our findings demonstrate that prolonged use of the Impella 5.5 provides safe and effective haemodynamic support for patients with end-stage heart and kidney failure awaiting dual-organ transplantation. With a 1 year survival rate of 90%, our data suggest that Impella 5.5 can be a viable alternative to traditional support strategies, particularly in patients who are otherwise limited by RRTs. As dual-organ transplantation becomes more prevalent, the Impella 5.5 offers a promising bridge to transplant, improving both short-term and long-term outcomes in this complex patient population.