Citations

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.

Background: Intracardiac thrombosis (ICT) during liver transplantation (LT) is a rare but life-threatening complication with limited data regarding its risk factors, management, and outcomes. This study aimed to identify factors associated with ICT development, define predictors of intraoperative outcomes, and propose a novel classification and management algorithm for ICT. Methods: A multicenter, international retrospective case-control study was conducted on liver transplant recipients from seven centers between January 2005 and December 2023. ICT cases were identified through transesophageal echocardiography (TEE) or autopsy and matched 1:1 with controls. Results: Among 11 077 liver transplant recipients screened, 133 patients (1.2%) developed ICT. ICT occurred predominantly after reperfusion (58%). Compared to controls, ICT cases exhibited higher preoperative international normalized ratio ( P = 0.001) and fibrinogen levels ( P = 0.014). Post-reperfusion syndrome ( P < 0.001), disseminated intravascular coagulation ( P < 0.001), and primary graft failure ( P = 0.002) were also more common in the ICT group. Continuous veno-venous hemofiltration during surgery (odds ratio [OR] 8.20, 95% confidence interval [CI]: 2.09–38.1; P = 0.004) was associated with ICT development. ICT was associated with markedly higher rates of cardiac arrest ( P < 0.001) and intraoperative mortality ( P < 0.001), with thrombus location in the left heart chambers representing a significant predictor of mortality (OR 20.7, 95% CI: 6.25–81.3; P < 0.001). ICT grades 2 and 3 were associated with a markedly increased risk of intraoperative cardiac arrest ( P < 0.05). Importantly, delayed use of TEE was strongly associated with intraoperative death (OR 6.27, 95% CI: 1.45–27.1; P = 0.014). The 90-day mortality was 43% in ICT cases compared to 4.5% in controls, while 1-year mortality was 46% versus 10.5%, respectively. Conclusions: ICT during LT is a rare but severe complication associated with intraoperative mortality and poor short-term survival. Coagulation profile and intraoperative management play critical roles in ICT development. The early use of TEE enables timely diagnosis and intervention, improving patient outcomes. We propose a novel ICT classification offering a systematic framework for risk stratification and management to mitigate progression and mortality.

Abstract Objective: Electronic Health Record (EHR) data are critical for advancing translational research and AI technologies. The ENACT network offers access to structured EHR data across 57 CTSA hubs. However, substantial information is contained in clinical narratives, requiring natural language processing (NLP) for research. The ENACT NLP Working Group was formed to make NLP-derived clinical information accessible and queryable across the network. Methods: We established the ENACT NLP Working Group with 13 sites selected based on criteria including clinical notes access, IT infrastructure, NLP expertise, and institutional support. We divided sites into five focus groups targeting clinical tasks within disease contexts. Each focus group consisted of two development sites and two validation sites. We extended the ENACT ontology to standardize NLP-derived data and conducted multisite evaluations using the Open Health Natural Language Processing (OHNLP) Toolkit. Results: The working group achieved 100% site retention and deployed NLP infrastructure across all sites. We developed and validated NLP algorithms for rare disease phenotyping, social determinants of health, opioid use disorder, sleep phenotyping, and delirium phenotyping. Performance varied across sites (F1 scores 0.53–0.96), highlighting data heterogeneity impacts. We extended the ENACT common data model and ontology to incorporate NLP-derived data while maintaining Shared Health Research Informatics NEtwork (SHRINE) compatibility. Conclusion: This demonstrates feasibility of deploying NLP infrastructure across large, federated networks. The focus group approach proved more practical than general-purpose approaches. Key lessons include the challenge of data heterogeneity and importance of collaborative governance. This work also provides a foundation that other networks can build on to implement NLP capabilities for translational research.

Effective healthcare referral triage poses a significant challenge, especially extracting referral notes on time for scheduling. To offset this issue, this study aims to develop an optimized Random Forest classifier to recreate the referral notes content and identify patients requiring Spinal Cord Stimulation procedures. A performance index was introduced to improve accuracy, while a consistency index was designed to prevent overfitting. The optimized classifier, with the optimal selection at the number of estimates being 5 trees and tree depth being 4, achieved an accuracy of 0.74 with a lower standard deviation, demonstrating its reliability and effectiveness.

Abstract The advances in large language models (LLMs) have demonstrated sophisticated potential for automating complex tasks within the radiology workflow. From radiology report generation and report summarization to data collection for research trials, these models have proven to be powerful tools. However, optimal implementation of these models requires careful adaptation to the specialized medical domain. In addition, these models tend to generate information that is not truthful or factual, which can adversely affect patient care and clinical decisions. Strategies such as fine-tuning and prompt optimization have been shown to be impactful in eliminating these errors. Although these models undergo rapid updates and improvements, understanding the principles of prompt engineering and fine-tuning provides a foundation for evaluating and maintaining the performance of any LLM deployment. The current article aims to review the recent advancements in radiology using fine-tuning and prompt optimization to leverage LLMs’ capabilities. It delves into various techniques within each strategy, their advantages and limitations, and presents a framework to facilitate the practical integration of LLMs into radiology settings.

ABSTRACTBackgroundPercutaneous transluminal angioplasty and stenting (PTAS) in patients with carotid stenosis may have potential effects on brain parenchyma. However, current studies on parenchymal changes are scarce due to the need for advanced imaging modalities. Consequently, the alterations in brain parenchyma following PTAS remain an unsolved issue.PurposeTo investigate changes to the brain parenchyma using magnetic resonance elastography (MRE).Study TypeProspective.Population13 patients (6 women and 7 men; 39 MRI imaging sessions) with severe unilateral carotid stenosis patients indicated for PTAS were recruited between 2021 and 2024.Field Strength/SequenceNoncontrast MRI sequences including MRE (spin echo) were acquired using 3 T scanners.AssessmentAll patients underwent MRE before (preprocedural), within 24 h (early postprocedural) and 3 months after (delayed postprocedural) PTAS. Preprocedural and delayed postprocedural ultrasonographic peak systolic velocity (PSV) was recorded. MRE stiffness and damping ratio were evaluated via neural network inversion of the whole brain, in 14 gray matter (GM) and 12 white matter (WM) regions.Statistical TestsStiffness and damping ratio differences between each pair of MR sessions for each subject were identified by paired sample t tests. The correlations of stiffness and damping ratio with stenosis grade and ultrasonographic PSV dynamics were evaluated by Pearson correlation coefficients. The statistical significance was defined as p < 0.05.ResultsThe stiffness of lesion side insula, deep GM, and deep WM increased significantly from preprocedural to delayed postprocedural MRE. Increasing deep GM stiffness on the lesion side was positively correlated with the DSA stenosis grade significantly (r = 0.609). The lesion side insula stiffness increments were positively correlated with PSV decrements significantly (r = 0.664).Data ConclusionRegional brain stiffness increased 3 months after PTAS. Lesion side stiffness was positively correlated with stenosis grades in deep GM and PSV decrements in the insula.Evidence Level2.Technical EfficacyStage 2.

Abstract Single‐molecule immunoassay is a reliable technique for the detection and quantification of low‐abundance blood biomarkers, which are essential for early disease diagnosis and biomedical research. However, current single‐molecule methods predominantly rely on endpoint detection and necessitate signal amplification via labeling, which brings a variety of unwanted effects, like matrix effect and autofluorescence interference. This study introduces a real‐time mass imaging‐based label‐free single‐molecule immunoassay (LFSMiA). Featuring plasmonic scattering microscopy‐based mass imaging, a 2‐step sandwich assay format enables background reduction, minimization of matrix effect by dynamic tracking of single binding events, and fully leveraging real‐time data for improved measurement precision through a Bayesian Gaussian process model, the LFSMiA enables ultra‐sensitive and direct protein detection at the single‐molecule level in neat blood sample matrices. LFSMiA measurement is demonstrated for interleukin‐6 and prostate‐specific antigen in buffer, undiluted serum, and whole blood with sub‐femtomolar detection limits and eight logs of dynamic ranges. Moreover, comparable performance is achieved with an inexpensive miniaturized setup. To show its translational potential to clinical settings and point‐of‐care diagnostics, N‐terminal pro‐B‐type natriuretic peptide is examined in patient whole blood samples using the LFSMiA and results in a strong linear correlation (r > 0.99) with standard clinical lab results.

LBA3509 Background: Systemic inflammation has been implicated in colon cancer progression. We examined whether intake of a proinflammatory diet is associated with survival among patients (pts) with stage III colon cancer. Methods: This prospective cohort study was nested within a randomized phase III trial of adjuvant therapy in pts with stage III colon cancer who had undergone curative-intent resection. EDIP score – a validated tool to classify the inflammatory nature of diets – was derived as a cumulative exposure using data from food-frequency questionnaires completed within 6 weeks of randomization and 14-16 months (mo) after randomization. Cox proportional hazards regression was used to assess the associations of EDIP with disease-free survival (DFS) and overall survival (OS). Results: Of 1625 included pts, the mean (SD) age was 60.9 (10.5) years (yrs). Compared to pts in the lowest EDIP quintile, pts in the highest quintile (a more inflammatory diet) were younger (58.7 ± 10.8 vs 61.3 ± 9.5 yrs old) and more likely to be female (64.0% vs 48.9%) and have worse performance status (ECOG 1-2: 35.7% vs 19.4%). Pts in the highest quintile were less likely to be White (76.6% vs 92.0 %) and more likely to be Black (15.4% vs 3.7%). Baseline aspirin use, assigned chemotherapy (3 mo vs 6 mo), and assigned pharmacotherapy (celecoxib vs placebo) were not significantly different across EDIP quintiles. Compared with pts in the lowest EDIP quintile, pts in the highest quintile had significantly worse OS (multivariable hazard ratio [HR] 1.87, 95% confidence interval [CI] 1.26-2.77, P trend =0.01), but not DFS (HR 1.36, 95% CI 0.99-1.86, P trend =0.22). Diet and physical activity jointly influenced OS. Those with lower EDIP scores (quintiles 1-4, 80% of the study population) and higher physical activity (≥9 MET-h/wk) had the best OS (HR 0.37, 95% CI 0.25-0.53) compared with pts in the highest EDIP quintile (20% of the study population) and lower physical activity (<9 MET-h/wk) ( P interaction <0.001). The association between higher EDIP and OS was consistent when analyzed by celecoxib and placebo treatment arms ( P interaction 0.54). The relationship between EDIP and OS did not differ significantly according to aspirin use, with HR 1.60 (95% CI 0.71-3.60) among aspirin users and HR 2.01 (95% CI 1.27-3.16) among aspirin non-users ( P interaction 0.06). Conclusions: Our findings suggest that greater intake of a proinflammatory dietary pattern is associated with worse OS in pts with stage III colon cancer. Regular physical activity may attenuate the association, and further investigation of diet and physical activity intervention is warranted. Support: U10CA180821, U10CA180882, U24CA196171, U10CA180863, CCS 707213, UG1CA233234, U10CA180820, U10CA180868, U10CA180888; https://acknowledgments.alliancefound.org. Pfizer; ClinicalTrials.gov Identifier: NCT01150045.

Abstract Distinguishing high-risk intraductal papillary mucinous neoplasms (IPMNs), pancreatic cysts requiring surgery, from low-risk lesions remains a clinical challenge, often resulting in unnecessary procedures due to limited specificity of current methods. While radiomics and deep learning (DL) have been explored for pancreatic cancer, cyst-level malignancy risk stratification of IPMNs remains untapped. We conducted a multi-institutional study (seven centers, 359 T2W MRI images) to assess the feasibility of AI for predicting IPMN dysplasia grade using cyst-level image features. We developed and compared 2D and 3D radiomics-only, deep learning (DL)-only, and radiomics-DL fusion models, using expert radiologist scoring as a baseline reference. Model performance was evaluated using held-out test data. The radiomics-DL fusion model showed the highest discriminatory ability on the test set (AUC 0.692), outperforming the radiomics-only model (AUC 0.665). Expert accuracy varied widely (37.4%-66.7%). The fusion model integrating deep learning and radiomics features from routine T2W MRI (AUC: 0.692) demonstrates potential for objective, cyst-level risk stratification of IPMNs in a multi-center cohort, outperforming both radiomics-only models and expert radiologists. While performance requires improvement for standalone clinical use, this approach offers a scalable, non-invasive method to potentially improve diagnostic accuracy and reduce unnecessary surgical interventions.