Citations

Background: Tumor mutational burden (TMB) is an FDA-approved biomarker for immune checkpoint inhibitor (ICI) therapy. However, its predictive value varies among tumor types and molecular contexts. We investigated whether a very high TMB identifies a biologically distinct subset and whether a higher cutoff provides additional clinical insights beyond the conventional high TMB threshold. Methods: We analyzed 133 patients with advanced solid tumors and TMB ≥ 10 mutations/Mb (mut/Mb) who underwent tumor genomic profiling using a 523-gene DNA/RNA next-generation sequencing panel. Tumors were stratified into prespecified TMB categories: 10–20 mut/Mb (TMB-H) and >20 mut/Mb (TMB-VH). The clinical characteristics, ICI outcomes (in the treated subset), and pathway-level genomic features were compared between groups. Results: TMB-VH was observed in 42/133 (31.6%) patients and spanned more than 20 tumor types. MSI was markedly more prevalent in TMB-VH than in TMB-H tumors (38.1% vs. 2.2%; Fisher’s exact p = 8.9 × 10−8). Pathway-level comparisons did not identify statistically significant differences after false discovery rate correction (all q > 0.05), and the observed patterns were descriptive in nature. In the ICI-treated subset with complete follow-up, objective response did not differ according to the TMB group. Overall survival (OS) was also similar between groups, whether measured from metastatic diagnosis (log-rank p = 0.937) or from ICI initiation (log-rank p = 0.814), although OS was numerically longer in the TMB-VH group in both analyses without reaching statistical significance. Conclusions: In this cohort study, TMB-VH was strongly associated with MSI but not independently associated with improved ICI outcomes. Larger multicenter cohorts are needed to validate pathway-oriented patterns and clarify the clinical utility of extreme TMB thresholds across various histologies. Integrating the functional context (e.g., MSI status, gene-level context, and pathway-level features) with TMB magnitude may enable more robust, tumor-aware biomarker models for immunotherapy selection.

OBJECTIVES: To evaluate and compare the ability of three popular open-source artificial intelligence platforms to diagnose common trauma-related fractures using radiologic imaging. METHODS: Design: Retrospective diagnostic performance comparison study. Setting: Publicly accessible online radiologic imaging databases. Patient Selection Criteria: Five common orthopedic trauma fractures were assessed: ankle, tibial plateau, intertrochanteric, femoral neck, and humerus. Radiographs and computed tomography (CT) images were collected. Images were randomly selected from confirmed diagnoses on Radiopaedia.org. Outcome Measures and Comparisons: ChatGPT 5, Grok 3, and Claude 4.5 Sonnet were queried with each image. Diagnostic accuracy, sensitivity, specificity, positive and negative predictive values, and performance by modality (X-ray vs. CT) were assessed. The reference standard was the expert-verified diagnosis provided by Radiopaedia.org, limited to cases labeled with a “diagnosis certain” tag. RESULTS: Each model was provided with 30 radiographs and 20 CT images whenever possible. ChatGPT 5, Grok 3, and Claude 4.5 Sonnet accurately diagnosed diseased images in 26.8%, 18.8%, and 22.4% of cases, respectively. By fracture type, ChatGPT 5 demonstrated the highest correct classification rates for ankle (10%), femoral neck (38%), humerus (40%), and tibial plateau (44%) fractures. Grok 3 demonstrated the highest correct classification rate for intertrochanteric fractures (6%). Overall sensitivities were 0.267, 0.187, and 0.223 for ChatGPT 5, Grok 3, and Claude 4.5 Sonnet, respectively. ChatGPT 5 and Grok 3 outperformed Claude 4.5 Sonnet (both P < 0.001). No modality-based performance differences were observed for any model. CONCLUSIONS: Among the publicly available large language models evaluated for radiologic interpretation of orthopedic trauma imaging, ChatGPT 5 demonstrated the highest overall diagnostic accuracy, followed by Claude 4.5 Sonnet and Grok 3. Despite relative variation between the models, overall diagnostic accuracy for fracture detection was low across all platforms (<27%). In their baseline forms, these publicly accessible large language models are not recommended for radiologic imaging interpretation. LEVEL OF EVIDENCE: Diagnostic Level III. See Instructions for Authors for a complete description of levels of evidence.

How is postural orthostatic tachycardia syndrome (POTS) diagnosed? What about adolescents who seem to have POTS but do not meet the diagnostic criteria? How can we treat POTS and related conditions? How can we best respond to common questions of frustrated patients and parents and guardians? This article provides evidence- and expert-based answers to questions that frequently arise when caring for patients with POTS and related conditions.

Abstract Background: Clostridioides difficile infection (CDI) requiring colectomy carries substantial mortality risk, with optimal timing of surgery remaining poorly defined. We examined temporal trends in colectomy among inpatients with CDI, identified predictors of surgical intervention and postoperative mortality, and evaluated the association between surgical timing and patient outcomes. Methods: A retrospective cohort study was conducted using the National Inpatient Sample database from 2018 to 2022. We compared patients undergoing colectomy with those managed medically. To minimize confounding by hospital-onset cases, the analysis of surgical timing and mortality was restricted to patients undergoing colectomy within 8 days of admission. Predictors were identified using survey-weighted logistic regression and LASSO regression models. Results: Among 240,564 CDI hospitalizations (representing 1,207,995 weighted nationally), 717 patients underwent colectomy (3,585 weighted). CDI prevalence declined from 0.99% (2018) to 0.76% (2022), while colectomy rates increased from 0.28% to 0.34%. Peritonitis (OR 5.42; 95% CI, 4.46–6.59), coagulopathy (OR 4.96; 95% CI, 3.76–6.55), and sepsis/septic shock (OR 3.89; 95% CI, 3.39–4.47) were the strongest predictors of colectomy. Among patients undergoing colectomy within 8 days (2,830 weighted), in-hospital mortality was 26.5% overall, increasing from 21.0% (2018) to 30.7% (2022). Sepsis/septic shock (OR 8.20; 95% CI 2.92–23.07) and coagulopathy (OR 7.27; 95% CI 3.31–15.97) predicted mortality. Each additional day from admission to colectomy was associated with a 16% (OR 1.16; 95% CI 1.04–1.28) increased mortality risk. Conclusions: In this nationally representative cohort, surgical timing was an independent and modifiable determinant of survival in patients with CDI requiring colectomy. Our findings underscore the importance of early surgical consultation for CDI patients with peritonitis, sepsis, and coagulopathy.

Background: Perianal fistulizing Crohn’s disease (CD) is associated with significant morbidity and remains difficult to treat. Although advanced medical therapies are widely used, much of the available evidence derives from heterogeneous fistula populations or luminal CD trials, with limited perianal-specific synthesis and inconsistent outcome definitions. We conducted a systematic review focusing exclusively on perianal-specific clinical, radiologic, and composite outcomes in adults with perianal fistula (PAF) CD. Methods: We performed a systematic review in accordance with PRISMA 2020. Electronic databases were searched from inception through November 2025. We included randomized controlled trials and cohort studies enrolling adults with CD reporting outcomes specific to PAF. Interventions included biologics and small-molecule therapies, compared with placebo or other therapies. Due to substantial heterogeneity in outcome definitions and study designs, a meta-analysis was not performed. Risk of bias was assessed using Risk of Bias 2 (RoB 2) for randomized trials and the Newcastle–Ottawa Scale for observational studies. Results: Seven studies including >1200 participants with PAF-CD met inclusion criteria. Follow-up ranged from 24 weeks to 5 years. Across studies, outcome definitions and assessment modalities varied. Upadacitinib demonstrated significantly higher clinical fistula closure compared with placebo across multiple dose regimens at 52 weeks. In observational comparisons, ustekinumab and vedolizumab were associated with higher clinical closure rates than anti-TNF therapies. However, infliximab demonstrated higher closure rates than adalimumab as a first-line treatment. The definition for radiologic remission was less consistent across studies and often did not parallel clinical outcomes. Composite clinical–radiologic remission and response were reported in a limited number of studies, with filgotinib showing higher composite outcomes in comparison to placebo in a phase 2 trial. Surgical interventions, relapse outcomes, biomarkers [C-reactive protein (CRP)/fecal calprotectin (FCP)], and patient-reported outcomes were variably reported and not consistently significant across comparisons. Conclusions: Evidence for advanced therapies in PAF CD remains limited by heterogeneity in endpoint definitions, subjectivity in clinical observation, inconsistent radiologic reporting, and reliance on subgroup or observational comparisons. While anti-TNF therapy remains the most established option in guidelines, emerging data suggest significant benefits with ustekinumab, vedolizumab, and JAK inhibitors in selected patients. There is a need for PAF-specific, adequately powered randomized trials using standardized composite clinical and radiologic endpoints.

Introduction Translating findings from cell and animal models to the clinic is difficult because these systems insufficiently capture human disease biology. Computational approaches, notably pathways modeling, have overcome interspecies differences by mapping shared signaling pathways between species. We considered whether this approach could be generalized to bridge animal and virtual drug screening data to effects measured in patients. We emphasized translational psychiatry datasets because of the unmet need for effective therapies. Methods We conducted three parallel analyses using published drug screening data in zebrafish, structure-function computational screening data, and clinically-reported efficacy data for experimental and approved antipsychotic drugs and predicted drug pathways using the PathFX algorithm Results Despite choosing screens developed for unique aspects of psychiatry and without careful curation of shared drugs, we found pathways associated with drugs that performed “well” across these distinct platforms—specifically, among drugs measured as favorable (expected to improve disease) or unfavorable (expected to aggravate disease) across screens, these drugs connected to distinct pathway proteins; however, these drugs are connected to similar protein families and shared Gene Ontology functional terms. By comparing screens, we discovered that favorable drugs may influence G-proteins, solute carrier family proteins, adrenoceptors, and steroid hydroxylase activity and that unfavorable drugs affect serotonin receptors and phosphodiesterase activity. Discussion This suggests that predictive platforms could emphasize functional information as features that could overcome differences in distinct screening platforms to eventually improve translational approaches.

Abstract Background Automated phlebotomy has the potential to improve patient outcomes and address phlebotomist workforce challenges. The Autonomous Blood Drawing Optimization and Performance Testing trial (ADOPT, ClinicalTrials.gov NCT05878483) assesses performance, safety, and patient experience of an autonomous robotic phlebotomy device (ARPD) that utilizes multimodal imaging. Methods Patients were recruited from outpatient phlebotomy departments in the Netherlands. In cohort 1, analytical equivalence was assessed by comparing within-subject test results from ARPD-collected and manually collected samples using Passing–Bablok regression analysis. In cohort 2, the ARPD was assessed in routine use. The primary performance endpoint was first-stick success rate; the primary safety endpoint was the rate of adverse events. Exploratory endpoints assessed pain and patient preference. Results One hundred fifty-three patients were enrolled in cohort 1. No statistically significant differences were observed in activated partial thromboplastin time, prothrombin time, lactate dehydrogenase, aspartate transaminase, or platelet count, indicating analytical equivalence. A total of 1633 patients were enrolled in cohort 2. The overall first-stick success rate, in patients where the ARPD could identify a suitable vein, was 94.5% (95% CI, 93.3–95.5) and remained high in patients with difficult venous access (92.7%), obesity (97.4%), and age ≥65 years (93.4%). Adverse events, all mild, occurred at a rate of 0.6%. Ninety percent of patients reported far less (19%), less (32%), or similar pain (39%) compared with manual phlebotomy. Eighty-two percent indicated they would strongly prefer (16%) or prefer (31%) use of the ARPD in the future or had no preference (35%). Conclusions Automated phlebotomy using the ARPD shows a favorable performance and safety profile and may improve patient experience. ClinicalTrials.gov Registration Number: NCT05878483