Citations

Abstract Functional magnetic resonance imaging (fMRI) data are inherently complex, characterized by high dimensionality, intricate inter-regional dependencies, and substantial individual variability across experimental paradigms. Traditional linear mixed models (LMMs) provide a principled framework that models population-level fixed effects while estimating variance components arising from subject-level random effects; however, they often fail to adequately capture nonlinear relationships inherent in neuroimaging data. To address these limitations, we introduce the nonlinear mixed model (NMM) approach, an innovative extension of the LMM framework that integrates neural networks to flexibly model complex fixed-effect relationships while preserving the random-effects structure to account for individual differences. NMM advances fMRI analysis by: (1) identifying robust functional connectivity (FC) patterns consistently observed across multiple paradigms; (2) leveraging SHapley Additive exPlanations (SHAP) analysis to provide post-hoc interpretability of the nonlinear fixed effects, quantifying how age, sex, and paradigm contribute to predicted FC and how these effects are distributed across large-scale brain networks; and (3) using subject-specific random effects as neural fingerprints that not only show systematic variability across attention and default mode systems but also predict standardized cognitive scores, demonstrating biological relevance. Applied to the Philadelphia Neurodevelopmental Cohort (PNC) across emotion, n-back, and resting-state paradigms, NMM achieved superior model fit relative to classical LMMs, as evidenced by lower mean squared error (MSE) in predicting FC. This framework offers a statistically rigorous and practically explainable approach for modeling large-scale FC from modest covariates while explicitly separating population-level effects from stable individual variability in functional brain organization.

Abstract Inflammatory bowel diseases (IBD), principally Crohn’s disease (CD) and ulcerative colitis (UC), are common chronic disorders involving inflammation and often progressive tissue damage. Genome-wide association studies have mapped many risk signals, but the causal variants, effector genes and relevant cellular contexts remain difficult to resolve, limiting mechanistic interpretation and therapeutic translation. Here we performed a multi-ancestry GWAS meta-analysis of 125,992 individuals with IBD and more than 1.2 million controls, identifying 619 independent association signals (374 novel) at 420 IBD regions that account for 77–80% of SNP-based heritability. Fine-mapping resolved 81 high-confidence variants, 41 not previously reported. Although most signals were shared between CD and UC, 39% showed IBD subtype specificity, with UC signals showing stronger enrichment in functional annotations from intestinal epithelial, secretory and enteroendocrine cells, and CD showing stronger genetic correlations with circulating inflammatory biomarkers, including C-reactive protein and glycoprotein acetylation. Latent causal modelling supported a causal effect of decreased high-density lipoprotein on CD risk. By integrating bulk and single-cell eQTL and pQTL resources using colocalisation and Mendelian randomisation, together with coding-variant evidence from exome sequencing, we prioritised 664 candidate effector genes across 341 signals, including 390 newly implicated IBD genes, revealing new biological mechanisms and candidate therapeutic targets supported by human genetics.

PURPOSE We sought to characterize outcome associations of poly ADP-ribose polymerase inhibitors (PARPi) in patients with ovarian cancer (OVCA) whose cancers contain homozygous loss (deletions) of homologous recombination repair (HRR) pathway genes, an alteration class not uniformly reported on clinical testing and not susceptible to acquired resistance via secondary reversion mutations. METHODS This study used US-based deidentified Flatiron Health-Foundation Medicine OVCA Clinico-Genomic Database. Patients with advanced OVCA and genomic profiling by Foundation Medicine were assembled into cohorts evaluating PARPi monotherapy use, and separately, PARPi maintenance therapy use, with outcomes associated with HRR gene alterations: BRCA1 / 2 , PALB2 , RAD51B/C/D , BARD1 , and BRIP1 . RESULTS 1.5% (82/5,404) had HRR gene homozygous loss, with BRCA1 and BRCA2 loss being most common. Among all OVCAs with BRCA1/2 alterations, 7.5% (63/836) were homozygous losses. Most BRCA1 (34/38, 89%) and BRCA2 (18/25, 72%) homozygous losses were intragenic, with breakpoints not spanning entire gene. Maintenance cohort: 22 patients had HRR gene homozygous loss, with strongly improved progression-free survival (PFS) for PARPi versus no PARPi use (HR, 0.06, 95% CI, 0.01 to 0.37). Monotherapy cohort: patients without HRR alterations (n = 198), HRR mutations or rearrangements (n = 93), and HRR homozygous loss (n = 8) had median PFS of 4.9, 10.6, and 25.4 months, and median overall survival of 18.5, 22.5, and 51.8 months, respectively. CONCLUSION We observe durable PARPi effectiveness for patients with OVCA and HRR gene homozygous loss in routine practice. We demonstrate the importance of detecting and reporting homozygous loss in HRR genes, an aspect not standard to all sequencing assays.

Our knowledge regarding baroreflex regulation of subpopulations of differently sized sympathetic action potentials (APs) remains incomplete. This study tested the hypothesis that baroreflex sigmoid morphology varies across AP subpopulations. In 12 healthy individuals (6 females; 32 ± 5 years), we measured diastolic blood pressure (DBP; brachial artery catheter) and muscle sympathetic APs (peroneal microneurography; continuous wavelet transform) during a modified Oxford protocol (100 μg nitroprusside bolus, 150 μg phenylephrine bolus). Baroreflex sigmoids (four-parameter logistic functions) modelled the relationship between DBP and sympathetic AP discharge probability for 10 normalized AP clusters. For each AP cluster, baroreflex sigmoid morphology was quantified by max gain, threshold, centering, and saturation pressures, and operating and response ranges. One-way analyses of variance and Dunnett post-hoc tests were performed. Data are mean ± SD. At the centering pressure of the baroreflex sigmoid, max gain was greater for medium APs (e.g., Cluster 3: -5.6 ± 2.9%/mmHg) compared to small (e.g., Cluster 1: -1.4 ± 1.1%/mmHg; P < 0.01) and large APs (e.g., Cluster 9: -0.6 ± 0.4%/mmHg; P < 0.01). Discharge probability at the centering pressure was also greater for medium APs (e.g., Cluster 3: 61 ± 22%) compared to small (e.g., Cluster 1: 12 ± 10%; P < 0.01) and large APs (e.g., Cluster 9: 3 ± 3%; P < 0.01). There were no differences in the saturation pressure (i.e., DBP where any further reduction does not elicit an increase in AP discharge probability) between AP clusters (main effect P = 0.22). The threshold pressure (i.e., DBP where any further reductions elicit an increase in AP discharge probability) was greater for medium (e.g., Cluster 3: 78 ± 8mmHg) compared to large APs (e.g., Cluster 9: 74 ± 11mmHg; P = 0.03). The operating range (i.e., Threshold – Saturation DBP) was greater for medium (e.g., Cluster 3: 14 ± 3mmHg) compared to large APs (e.g., Cluster 9: 8 ± 5mmHg; P < 0.01). There were no differences in the threshold pressure or operating range of medium compared to small APs (both P ≥ 0.82). The response range (i.e., Saturation – Threshold discharge probability) was greater for medium (e.g., Cluster 3: 26 ± 13%) compared to small (e.g., Cluster 1: 5 ± 4%; P < 0.01) and large APs (e.g., Cluster 9: 1 ± 1%; P < 0.01). These data confirm prior work using linear modeling and spontaneous approaches to quantify sympathetic AP baroreflex gain and extend to demonstrate variation in the morphology of the baroreflex sigmoid. During large changes in blood pressures, we identify differences in broader features of baroreflex control (e.g., threshold pressure, operating range, and response range) between AP subpopulations that support strong baroreflex regulation over medium-sized APs. This abstract was presented at the American Physiology Summit 2026 and is only available in HTML format. There is no downloadable file or PDF version. The Physiology editorial board was not involved in the peer review process.

PURPOSE For patients with breast cancer undergoing breast conservation, escalating the dose (boost) of radiation to the lumpectomy cavity after whole-breast irradiation (WBI) reduces ipsilateral breast recurrence (IBR) but extends treatment duration. This phase III trial investigated whether boost delivery during WBI versus after WBI provides noninferior IBR and preserves cosmetic appearance. METHODS NRG/RTOG 1005 randomly assigned patients at higher risk for IBR after lumpectomy and axillary surgery to either a sequential boost of 12 Gy in six fractions(F) or 14 Gy in 7F after WBI of 50 Gy in 25F or 42.7 Gy in 16F (sequential arm) or a concurrent boost of 8 Gy in 15F of 0.53 Gy per day with WBI of 40 Gy in 15F (concurrent arm) using 3-dimensional conformal radiation therapy (RT) or intensity-modulated RT. Based on 1.59% 5-year IBR for the sequential arm, defining the noninferiority margin as a hazard ratio upper limit on the 90% CI of 2.12, 2,312 patients provide 80% power for noninferiority of IBR as first recurrence for the concurrent arm. Secondary end points included disease-free survival and overall survival, adverse events (AEs), and cosmetic outcomes. RESULTS Between May 24, 2011, and June 20, 2014, 2,354 patients were randomly assigned, with 2,255 eligible for analysis (sequential arm, n = 1,118; concurrent arm, n = 1,137). With median follow-up of 7.3 years, there were 56 IBR events; 5- and 7-year IBR were 2.1% and 2.2% on the sequential arm and 1.9% and 2.6% on the concurrent arm, respectively. The noninferiority criterion was met: HR (90% CI): 1.31 (0.84 to 2.04), P = .037. No differences were observed in AEs, cosmetic outcomes, or survival between arms. CONCLUSION Concurrent boost during WBI results in noninferior IBR compared with sequential boost without worsening toxicity or cosmetic outcomes and reduces overall treatment time.

Abstract Background The Global Bipolar Cohort (GBC) was established to identify existing bipolar disorder (BD) cohorts worldwide and foster collaborations focused on descriptive and analytic outcomes relevant to BD. A distributed analytic framework has been implemented to engage multiple sites without the need for central data pooling. This report describes the GBC endeavor and global functional impairment patterns. Cross-cohort comparisons of functional correlates are limited by heterogeneous measures and data-sharing constraints. Large, culturally diverse comparisons are needed to distinguish broadly reproducible correlates from cohort-specific effects. Participating sites completed a 28-item descriptive survey covering diagnostic methods, cognition, genetics, treatment, functioning, and follow-up strategies. We implemented a harmonized local logistic regression model of dichotomized functional outcome and shared summary statistics only. Results We identified 69 cohorts across five continents. Thirty-seven cohorts contributed functional outcome analyses from 17,130 participants. Outcome measures included clinician-rated disability scales and social indicators such as employment and marital status. The proportion classified with poor functioning ranged from 16% to 77% (mean 50%). In 32 of 37 cohorts, the overall regression model significantly explained variance in functioning. Current depressive symptoms were the most robust and reproducible correlate of poor functional outcome: they were assessed in 29 cohorts, significant in 22 (75.8%), ranked among the top three correlates in 22, and were the top-ranked correlates in 19. Associations between depressive burden and poor functioning were observed across clinician-rated disability scales and work or social indicators, and across geographically diverse cohorts. Comorbid substance use disorder and medication-related variables were associated with poorer functioning in subsets of cohorts, whereas sex, ancestry, bipolar subtype, psychosis history, and premorbid IQ showed weak or inconsistent associations. Cognitive measures, available in a minority of regression models, showed modest and non-uniform effects. Conclusions Across heterogeneous international cohorts, current depressive symptom burden emerged as the most consistent correlate of poor functioning in bipolar disorder. These findings replicate earlier multisite work at a larger scale, show that protocol-based distributed analyses can identify reproducible clinical signals without sharing individual-level data, and support prioritizing detection and treatment of depressive symptoms when aiming to improve real-world functioning. Future work should expand longitudinal harmonization and representation of under-studied populations.

Abstract Concurrent thyrotropin (TSH) secreting and growth hormone (GH) secreting pituitary adenomas are rare entities. Although tumors of the PIT-1 lineage may coexpress TSH and GH, concurrent biologically active secretion resulting in central hyperthyroidism and acromegaly is uncommon. We report the case of a 42-year-old woman presenting with weight loss, palpitations, amenorrhea, and progressive acromegalic features. Biochemical evaluation revealed central hyperthyroidism with elevated free T4 and inappropriately normal TSH, markedly elevated insulin-like growth factor 1 (IGF-1) and GH, hyperprolactinemia, and elevated α-subunit. Pituitary magnetic resonance imaging demonstrated a 2.7-cm macroadenoma with cavernous sinus extension. She underwent transsphenoidal resection without complications. Pathology confirmed a mature plurihormonal PIT-1 lineage adenoma immunopositive for TSH, GH, and prolactin with a Ki-67 of 3%. Immediately postoperatively, thyrotoxicosis and hyperprolactinemia resolved, IGF-1 levels declined, acromegalic features improved, and menses resumed. Over the ensuing months, thyroid and prolactin tests remained within normal range, but the IGF-1 level remained elevated. This case illustrates the rare presentation of a PIT-1 lineage adenoma secreting biologically active TSH, GH, and prolactin. Comprehensive hormonal evaluation and long-term biochemical and radiologic surveillance are essential, given the risk of incomplete remission despite apparent surgical cure.

Background: Objective performance assessment is essential in surgical skill training, yet current methods are labor-intensive and focus on observing the trainee rather than the end-product of the procedure. Machine learning (ML) methods offer reproducible feedback but have mainly relied on kinematic or video data, often reducing assessment to binary or ternary classification. Our objective was to compare ML regression models predicting expert-assigned scores of vascular anastomoses from computational fluid dynamics (CFD) features of the final product. Additionally, we aimed to assess biomechanical plausibility of predictions. Methods: A total of 146 participants performed 419 end-to-side anastomoses on case-specific three-dimensional (3D) printed simulators. Anastomoses were digitized via 3D scanning, ranked by experts, and characterized using CFD-derived hemodynamic features. These served as input for linear models (Ridge, Partial Least Squares), support vector machines, and tree-based ensembles (Random Forest, Extremely Randomized Trees, and Extreme Gradient Boosting [XGBoost]), evaluated using 10-fold nested cross-validation with genetic hyperparameter optimization. Results: Inter-rater reliability of expert indicated strong agreement (intraclass correlation coefficient ICC3k = 0.846). XGBoost achieved the lowest mean root mean squared error of 0.758 (95% bootstrap CI: 0.722–0.799) and a coefficient of determination (R2) of 0.673 (0.617–0.725), with the most stable performance across folds. Shapley additive explanations (SHAP) identified the wall shear stress gradient, transverse wall shear stress, and maximum pressure as the most influential features—variables associated with intimal hyperplasia and atherosclerotic remodeling. Conclusions: Tree-based ensemble methods, particularly XGBoost, effectively modeled biomechanical properties against expert scores. Combining CFD and ML can provide reproducible, mechanistically relevant feedback in vascular surgical skill training.

Abstract Cutaneous IgM/IgG immune complex vasculitis is a newly recognized provisional subtype of single-organ vasculitis that is clinically indistinguishable from IgA vasculitis. This retrospective case series evaluated the presence, clinicopathological characteristics and outcomes of patients with IgM and/or IgG dominance or codominance on direct immunofluorescence testing (2010–23). This review represents only the second study to date to characterize this population. Early analysis of 14 patients revealed that cutaneous IgM/IgG immune complex vasculitis was mainly skin limited, and treatment with topical and oral corticosteroids typically resulted in complete resolution without recurrence. Additionally, this study demonstrated that nearly two-thirds (64%) of patients with cutaneous small-vessel vasculitis with IgM and/or IgG dominance or codominance on direct immunofluorescence were ultimately found to have an underlying autoimmune connective tissue disease, cryoglobulinaemia, antineutrophil cytoplasmic antibody-associated vasculitis, infection or a drug-related cause. Therefore, it is imperative that clinicians perform a thorough investigation of this unique population to evaluate for potential underlying diseases and/or triggers.

Abstract Aims Heart failure is prevalent; however, there is no cost-effective screening option. Against formal echocardiography, we assessed the diagnostic performance of focused cardiac ultrasound (FoCUS), performed by novice users guided and interpreted by artificial intelligence (AI) for the assessment of left ventricular ejection fraction (LVEF). Methods and results We prospectively enrolled 496 adults referred for diagnostic echocardiography. Novice operators (without clinical or imaging experience) underwent a 4-h imaging workshop, then performed FoCUS using a point-of-care device (Philips Lumify) utilizing real-time AI-image guidance (UltraSight). Images were assessed by AI-image interpretation software (Mayo Clinic), and two expert blinded echocardiologists. Median age was 67, 39% female, and body mass index range 17–56 kg/m2. Forty-one subjects (8.3%) exhibited moderate or greater LV dilatation and 28 (5.6%) had LVEF &lt;40%. The median scan time was 4 min (IQR 3–5). Adequate views were achieved in 95.0% and 97.4% of subjects for AI interpretation and expert analysis, respectively. A two-step diagnostic screening process for low EF in which only abnormal AI reads or uninterpretable scans were reviewed by experts (15.1%) had a sensitivity 96.2%, specificity 95.4%, PPV 54.3%, and NPV 99.8%. A subsequent prospective validation study of 344 subjects demonstrated similar results (sensitivity 100%, specificity 99.4%). Conclusion In this early feasibility investigation, AI-guidance technology embedded on a handheld device, enabled novice users without clinical or imaging experience to acquire sufficient quality images to accurately assess LVEF with minimal training. The technology evaluated in this study may hold promise for AI-guidance and interpretation to facilitate low-cost screening for cardiac dysfunction.