Citations

Importance Limited granular data exist on breast cancer in transgender, nonbinary, and/or gender-diverse (TGD) individuals. Objective To describe patient demographic and tumor clinicopathological variables, clinically relevant risk factors, method of breast cancer detection, treatment characteristics, and locoregional outcomes and to estimate a 5-year breast cancer–specific survival (BCSS). Design, Setting, and Participants This cohort study obtained data on patients treated from 1990 to 2023 from 22 US academic medical centers using institutional databases, electronic medical records, and tumor registries. Included participants were TGD individuals 18 years or older with stage 0 to IV breast cancer. A comparative breast cancer cohort was selected from the 2016 to 2021 Surveillance, Epidemiology, and End Results Program (SEER) breast cancer dataset. Data were analyzed from January to October 2025. Exposure Breast cancer. Main Outcomes and Measures Estimated 5-year BCSS and locoregional recurrence rates. Wilcoxon rank sum and χ 2 tests were used to compare continuous and categorical variables. Kaplan-Meier analysis was used to estimate 5-year BCSS. Results A total of 112 TGD persons (median [IQR] age at diagnosis, 42.5 [36.5-51.0] years; 104 individuals [92.9%] were assigned female at birth) with breast cancer and 1 with bilateral disease were included in the analysis. Forty-three patients (38.4%) used gender-affirming hormone therapy before diagnosis. Among the entire cohort, 58 patients (51.8%) detected their own breast cancer, 68 (60.7%) had a familial breast cancer history, and 16 (14.3%) had a pathogenic germline variant. Tumors were predominantly hormone receptor positive (96 [85.7%]) and early stage (29 [25.7%] ductal carcinoma in situ [DCIS]; 51 [45.1%] stage I). The TGD cohort, compared with patients with breast cancer in the SEER dataset (N = 480 915), was younger (median [IQR] age, 42.5 [36.5-51.0] years vs 62.0 [52.0-72.0] years), had a higher proportion of persons assigned male at birth (8 [7.1%] vs 3522 [0.7%]), more often had progesterone receptor–positive disease (89 [79.5%] vs 335 491 [69.8%]), and presented with earlier-stage disease (eg, DCIS: 29 [25.9%] vs 79 604 [16.6%]). With a median (IQR) follow-up time for the TGD cohort of 38 (24-74) months, 5-year BCSS was favorable at 96.2% (95% CI, 85.3%-99.0%). Conclusions and Relevance This cohort study of TGD patients with breast cancer found that this population was young and often detected their own tumors via self-examination, suggesting opportunities to improve early detection through screening for individuals at elevated risk. Although short-term BCSS appeared favorable, prospective research is essential to define evidence-based guidelines and elucidate long-term outcomes.

Abstract Background Large language models (LLMs) are increasingly used in medication-related tasks despite limited evidence supporting their accuracy and safety. LLMs are vulnerable to adversarial attacks and may confabulate in response to fabricated information in the prompt. Given the substantial risk of harm from medication errors, the purpose of this study was to establish a benchmark task to evaluate LLM performance in distinguishing fabricated from real medications. Methods Two datasets (brand and generic), each consisting of 250 medication lists were developed, each including 4-6 medications and one fabricated medication (specifically a Pokémon character), with complete dosing information provided. Six LLMs were evaluated (GPT-5-Chat, GPT-4o-mini, DrugGPT, Gemma-3-27B-IT, Llama-3.3-70B-Instruct, and Qwen3-32B). For each medication list, models were queried for dosing information and disease indication across three experimental conditions: standard decoding, standard decoding with mitigation prompting, and deterministic decoding (temperature = 0). Each experiment was performed in triplicate. Outputs were deemed confabulations if the LLM failed to recognize and reject the fabricated medication. The primary outcome was the rate of confabulations. Exact paired-permutation tests were used to compare confabulation rates across LLMs and prompting approaches. Results Confabulation rates under standard conditions ranged from 2.7% to 99.6%. DrugGPT demonstrated the lowest baseline confabulation rates (2.7-6.4%), whereas other models demonstrated substantially higher rates (47.2-99.6%). Mitigation prompting significantly reduced confabulation rates compared to standard and deterministic decoding across all tasks and datasets (p < 0.001): GPT-5-Chat achieved confabulation rates of 0% with mitigation prompting from baseline rates of 66.4-88.5%. Deterministic decoding did not substantially impact confabulation rates. Conclusions LLMs performing medication-related tasks are susceptible to confabulation in response to invalid or fabricated inputs, raising important safety concerns for clinical use. While mitigation prompting reduced confabulation rates, this study underscores the importance of robust safeguards when deploying LLMs in medication-related decision-making.

PURPOSE The combination of the Bruton tyrosine kinase inhibitor ibrutinib with bendamustine-rituximab for first-line treatment of mantle cell lymphoma (MCL) prolonged progression-free survival (PFS), but without improvement in overall survival (OS), likely because of toxicity. Acalabrutinib was shown to be efficacious and less toxic than ibrutinib in a head-to-head trial in chronic lymphocytic leukemia and therefore might lead to better outcomes in MCL. METHODS Patients 65 years and older with previously untreated MCL received acalabrutinib (100 mg twice daily) or placebo (until disease progression or unacceptable toxicity), plus six cycles of bendamustine (90 mg/m 2 once daily; days 1 and 2) and rituximab (375 mg/m 2 as a single dose; day 1) followed by rituximab maintenance in responding patients for 2 years. Crossover to acalabrutinib at disease progression was permitted. The primary end point was PFS per the independent review committee; overall response rate and OS were secondary end points. RESULTS In total, 598 patients were randomly assigned, with 299 in each arm. At a median follow-up of 49.8 months using the reverse Kaplan-Meier method, the median PFS was 66.4 months in the acalabrutinib arm and 49.6 months in the placebo arm (hazard ratio [HR], 0.73 [95% CI, 0.57 to 0.94]; P = .0160). Benefit was seen across all subgroups, including those with high-risk features. Overall response/complete response rates were 91.0%/66.6% and 88.0%/53.5% in the acalabrutinib and placebo arms, respectively. OS was not significantly different (HR, 0.86 [95% CI, 0.65 to 1.13]; P = .27). Grade 3 or greater adverse events were reported in 88.9% and 88.2% in the acalabrutinib and placebo arms, respectively. CONCLUSION The combination of acalabrutinib with bendamustine-rituximab significantly improved PFS. Clinical benefit of acalabrutinib with bendamustine-rituximab was achieved with manageable toxicity.

Background: Patients with skin lesions suspicious for skin cancer or atypical melanocytic nevi of uncertain malignant potential often present to dermatologists, who may have variable dermoscopy triage clinical experience. Objective: To evaluate the clinical utility of a digital dermoscopy image-based artificial intelligence algorithm (DDI-AI device) on the diagnosis and management of skin cancers by dermatologists. Methods: Thirty-six United States board-certified dermatologists evaluated 50 clinical images and 50 digital dermoscopy images of the same skin lesions (25 malignant and 25 benign), first without and then with knowledge of the DDI-AI device output. Participants indicated whether they thought the lesion was likely benign (unremarkable) or malignant (suspicious). Results: The management sensitivity of dermatologists using the DDI-AI device was 91.1%, compared to 84.3% with DDI, and 70.0% with clinical images. The management specificity was 71.0%, compared to 68.4% and 64.9%, respectively. The diagnostic sensitivity of dermatologists using the DDI-AI device was 86.1%, compared to 78.8% with DDI, and 63.4% with clinical images. Diagnostic specificity using the DDI-AI device increased to 80.7%, compared to 75.9% and 73.6%, respectively. Conclusion: The use of the DDI-AI device may quickly, safely, and effectively improve dermoscopy performance, skin cancer diagnosis, and management when used by dermatologists, independent of training and experience.

Abstract Variation in de-adoption of ineffective or unsafe treatments is not well-understood. We examined de-adoption of erythropoiesis-stimulating agents (ESA) in anemia treatment among patients with chronic kidney diseases (CKD) following new clinical evidence of harm and ineffectiveness (the TREAT trial) and the FDA's revision of its safety warning. We used an interrupted time series approach to estimate changes in use of epoetin alfa (EPO) and darbepoetin alfa (DPO) in the commercial and Medicare Advantage (MA) and Medicare fee-for-service (FFS) populations. We also examined how changes in both trends and levels of use were associated with physicians’ characteristics. Study cohort included patients with CKD stages 3 to 5 during 2007-2015. Use of DPO and EPO declined over the study period. There were no consistent changes in DPO trend across insurance groups, but the level of DPO use decreased right after the FDA revision in all groups. The decline in EPO use trend was faster after the TREAT trial for all groups. Nephrologists were largely more responsive to evidence than primary care physicians. Differences by physician's gender, and age were not consistent across insurance populations and types of ESA.

Abstract Defining patient-to-patient similarity is essential for the development of precision medicine in clinical care and research. Conceptually, the identification of similar patient cohorts appears straightforward; however, universally accepted definitions remain elusive. Simultaneously, an explosion of vendors and published algorithms have emerged and all provide varied levels of functionality in identifying patient similarity categories. To provide clarity and a common framework for patient similarity, a workshop at the American Medical Informatics Association 2019 Annual Meeting was convened. This workshop included invited discussants from academics, the biotechnology industry, the FDA, and private practice oncology groups. Drawing from a broad range of backgrounds, workshop participants were able to coalesce around 4 major patient similarity classes: (1) feature, (2) outcome, (3) exposure, and (4) mixed-class. This perspective expands into these 4 subtypes more critically and offers the medical informatics community a means of communicating their work on this important topic.

AbstractObjectivePerforate the expanded polytetrafluoroethylene membrane of the GORE® CARDIOFORM Septal Occluder (GSO) and GORE® CARDIOFORM ASD Occluder (GCA) after implantation.BackgroundPercutaneous transseptal access to the left atrium is necessary for many structural and electrophysiological procedures. The potential need to access the left atrium may influence decision‐making for patent foramen ovale or atrial septal defect closure.MethodsSixteen canines underwent implantation of equal number GSO or GCA devices. A transseptal crossing procedure was performed through the device 85 (±1) days postoccluder implantation. The crossing procedure was performed utilizing commercially available equipment: radiofrequency/SureFlex sheath and standard needle/Mullin's sheath. Progressive dilation of the perforation was performed to allow passage of a 12 French Mullin's sheath into the left atrium.ResultsLeft atrial access was achieved in all cases. Postmortem analysis demonstrated passage through both occluder discs in all radiofrequency/SureFlex sheath cases (4 GSO, 4 GCA) and half of the standard needle/Mullin's sheath cases (3 GSO, 1 GCA). The remaining standard needle/Mullin's sheath cases demonstrated perforation through the right atrial disc but passage around the septal aspect of the left atrial disc, thus not perforating the left atrial disc. No acute embolic complications from the procedure were observed.ConclusionsLeft atrial access may be achieved through the GSO or GCA devices after implantation and endothelialization. The combination of a radiofrequency needle and steerable sheath provides benefit over a standard needle and Mullin's sheath in accomplishing passage through both occluder discs.

Heart disease is the leading cause of mortality in the U.S. with approximately 610,000 people dying every year. Effective therapies for many cardiac diseases are lacking, largely due to an incomplete understanding of their genetic basis and underlying molecular mechanisms. Zebrafish (Danio rerio) are an excellent model system for studying heart disease as they enable a forward genetic approach to tackle this unmet medical need. In recent years, our team has been employing electrocardiogram (ECG) as an efficient tool to study the zebrafish heart along with conventional approaches, such as immunohistochemistry, DNA and protein analyses. We have overcome various challenges in the small size and aquatic environment of zebrafish in order to obtain ECG signals with favorable signal-to-noise ratio (SNR), and high spatial and temporal resolution. In this paper, we highlight our recent efforts in zebrafish ECG acquisition with a cost-effective simplified microelectrode array (MEA) membrane providing multi-channel recording, a novel multi-chamber apparatus for simultaneous screening, and a LabVIEW program to facilitate recording and processing. We also demonstrate the use of machine learning-based programs to recognize specific ECG patterns, yielding promising results with our current limited amount of zebrafish data. Our solutions hold promise to carry out numerous studies of heart diseases, drug screening, stem cell-based therapy validation, and regenerative medicine.