Citations

Abstract Background In aging and dementia research, the relationships between modifiable risk factors and biomarkers often vary across cohorts. We aim to evaluate the consistency and generalizability of the association between the area deprivation index (ADI) ‐ an indicator of socioeconomic/resource disadvantage within a specific area ‐ and cognition and Alzheimer's disease (AD) biomarkers across three distinct research cohorts, two from the U.S. (U.S. POINTER Imaging, ADNI) and one from Spain (UBRAIN). Method We included a total of 3091 participants without dementia, 1879 from U.S. POINTER Imaging, 282 from ADNI and 930 from UBRAIN (Table 1). The ADI was calculated using validated indices from the US and Spain, with neighborhood deprivation ranked as Low, Moderate, and High in the US, and Low and High in Spain. Outcomes of interest included (1) the preclinical Alzheimer's cognitive composite (PACC‐5), (2) MRI‐derived gray matter volumes in AD‐vulnerable regions and (3) PET or CSF‐derived amyloid status (+/‐). We ran logistic regression models with ADI level as predictor adjusted by sex, age, and APOE4 and considered education and race or ethnicity (only US cohorts) in our analyses. Result Participants in more deprived areas had lower education across cohorts and were more likely to be female (U.S. POINTER) and African American (U.S. POINTER and ADNI, Table 1). Independent of age, sex and APOE4 status, higher ADI was associated with lower PACC in U.S. POINTER (Moderate: β=‐0.165, p <0.001; High β=‐0.412, p <0.001) and UBRAIN (ADI: High β=‐0.096, p = 0.022) but not in ADNI (Figure 1). After controlling for education, this association remained significant in U.S. POINTER (Moderate: β=‐0.166, p = 0.002; High: β=‐0.248, p = 0.007) but not in UBRAIN (β=‐0.021, p = 0.66). In U.S. POINTER, there was a trend towards an association of higher ADI with lower GM volume in AD‐vulnerable regions (β=‐0.126, p = 0.089) that was not observed in ADNI or UBRAIN (Figure 2). No associations between ADI and amyloid status were found across cohorts. Conclusion We observed differing associations in neighborhood deprivation and cognition across US cohorts, as well as between the US and Spain cohorts. Country‐specific factors and sample representativeness must be considered when investigating the pathophysiologic mechanisms underlying social determinants and disparities in AD.

Abstract Background Ten years after the original publication, the Centiloid framework is now broadly used to harmonize amyloid‐PET quantification, facilitate data sharing and comparison across cohorts, and even assist visual interpretation in clinical settings. We evaluated the global implementation of Centiloids by comparing their distribution and corresponding positivity thresholds across cohorts. Methods We gathered data from publicly available cohorts and reached out to investigators across the world to collect cross‐sectional Centiloids, demographic and clinical information, and visual read data. Gaussian mixture models (GMM, k=2) were fitted to Centiloid values for each cohort and cutoffs were calculated as mean + 2SD of the lower Gaussian. When visual reads were available, we determined Centiloid cutoffs that maximized correspondence with visual reads (Cohen’s kappa). Data was combined across cohorts using random effects meta‐analyses. Results As of January 2025, we included 37 cohorts ( n = 41,678 participants) with heterogeneous pipelines, radiotracers, and clinical and demographic characteristics (Table‐1). The low Gaussian peaks ranged from ‐9 to 10CL; the meta‐analysis identified a common peak at 1CL. The second peaks were more heterogeneous (range=38‐102CL; meta‐analysis outcome=64CL). Across cohorts, the proportion of cognitively unimpaired versus impaired participants impacted the position of both peaks, with better separation in cohorts enriched in impaired individuals (Figure‐1C). The meta‐analysis indicated a GMM‐based cutoff of 19CL (95%CI: 16‐21CL, Figure‐1B); subgroup analyses showed no evidence of significant effect between single versus multicenter settings (17 versus 20CL, p = 0.30), MRI‐based or PET‐only processing (18 versus 19CL, p = 0.88), and no evidence of difference across radiotracers (Flutemetamol: 16CL; PIB: 17CL, Flutafuranol: 18CL, Florbetaben: 19CL, Florbetapir: 20CL, p = 0.78). In a subset of 29,496 participants with visual reads available, binary visual reads corresponded well to Centiloids (common kappa=0.86, Figure‐2A). The visual read‐based cutoff of 24CL (95%CI: 21‐27CL, Figure‐2B) maximized correspondence between visual read and quantification and was slightly higher than the GMM‐based cutoff. All meta‐analysis models showed high non‐random heterogeneity (I 2 >80%) across studies, suggesting non‐random differences in peaks and cutoffs. Conclusions Meta‐analysis‐based cutoffs align well with thresholds from the existing literature. High heterogeneity among studies underscores the need to investigate contributing factors, raising concerns about applying common cutoffs.

Abstract Combination therapy incorporating programmed cell death protein 1 (PD-1) blockade results in unprecedented response rates in both frontline and relapsed/refractory (R/R) classical Hodgkin lymphoma (cHL). Previous retrospective studies have suggested benefit for PD-1 blockade before autologous stem cell transplant (ASCT) but included few patients receiving PD-1 blockade with cytotoxic chemotherapy. To explore the impact of anti–PD-1 based salvage on outcomes for patients with R/R cHL, we retrospectively reviewed 1280 patients with R/R cHL who underwent ASCT from 2010 to 2022 at 6 transplant centers, none of whom received PD-1 blockade as part of frontline therapy. Overall, 25% received a PD-1 inhibitor at any point before ASCT (10% in conjunction with chemotherapy), 28% received salvage brentuximab vedotin (BV) without PD-1 blockade, and the rest received salvage chemotherapy alone. Patients who received PD-1 inhibitors at any point before ASCT had a significantly higher 2-year progression-free survival than those who received BV without PD-1 inhibitors or patients receiving chemotherapy alone (88.2%, 70.2%, and 67.4%, respectively; P&lt; .0001). When restricted to patients in complete response before ASCT, the benefit of PD-1 blockade remained significant. PD-1 blockade before ASCT is independently associated with superior post-ASCT outcomes and patients proceeding to ASCT should be treated with PD-1–based salvage.

An ST-segment elevation myocardial infarction is a life-threatening condition within the spectrum of acute coronary syndrome, with well-recognized implications for short-term and long-term health outcomes. Retrospective review of our institutional electronic health records identified patterns of inaccurate acute coronary syndrome diagnostic assignment, including the presence of erroneous ST-segment elevation myocardial infarction diagnoses in discharge records. Such erroneous diagnoses in the durable health record may increase risk to patients through inadequate risk assessment and inappropriate follow-up care. Misapplication of a shortcut in our electronic health record system for rapidly ordering emergent cases was identified as a potential cause of the quality gap. To address this issue, a nurse-led quality improvement project was developed by using the define, measure, analyze, improve, and control framework. The project intervention decreased the percentage of patients with erroneous diagnoses from 24% to 7% and marginally improved nursing staff satisfaction. Our project successfully addressed an institutional quality concern. Our goal of reducing erroneous discharge diagnoses with minimal disruption to our existing practice was achieved, positively affecting our patients. Including both nurse leaders and frontline health care workers ensured the success of our project.

The treatment landscape for soft tissue sarcomas (STS) is evolving constantly. These NCCN Guidelines Insights focus on recent updates to the NCCN Guidelines for STS specific to surgical management and systemic therapy recommendations for various subtypes.

Abstract Background Women with a pathogenic variant in BRCA1 or BRCA2 are at high risk of developing ovarian cancer, and it is often recommended that they undergo bilateral salpingo-oophorectomy at an early age, resulting in surgical menopause. Menopausal hormone replacement therapy (HRT) is an effective way to mitigate the adverse outcomes of early menopause; however, the safety of menopausal HRT on breast cancer risk in this population has not been established. Methods We conducted a prospective matched analysis of HRT use following menopause and breast cancer risk in BRCA carriers. Women who initiated HRT were matched one-to-one with women who had not initiated menopausal HRT by gene, year of birth, and age at menopause, resulting in 676 matched pairs. Menopausal HRT use collected by questionnaire included formulation and mode of administration. Results After a mean of 5.6 years, there were 87 (12.9%) incident breast cancer cases in the 676 exposed women and 128 (18.9%) cases in the 676 unexposed women (P = .002). Compared with unexposed matched control individuals, women who used estrogen alone experienced a statistically significantly decreased risk of breast cancer (hazard ratio = 0.37, 95% CI = 0.24 to 0.57). No protective or adverse effect was associated with the use of estrogen plus progestogen (hazard ratio = 0.94, 95% CI = 0.54 to 1.63). Conclusions Our findings suggest no substantial increase in the risk of breast cancer in BRCA carriers with the use of HRT and that estrogen alone might be protective.

Abstract INTRODUCTION New methods estimate amyloid positivity onset age (EAOA) from amyloid positron emission tomography (PET). We explore the genetics of EAOA to identify molecular factors underlying the earliest Alzheimer's disease (AD) changes. METHODS Harmonized amyloid PET data from 4216 participants were used in genome‐wide survival, tissue‐specific gene expression, and genetic covariance analyses of EAOA. RESULTS Variants in apolipoprotein E ( APOE ), ABCA7 , and RASGEF1C associated with earlier EAOA. APOE ε4/ε4 and ε3/ε4 converted 6.3 and 5 years earlier than ε3/ε3, respectively. ε2 was protective against earlier EAOA. rs4147929, an expression quantitative trait locus for ABCA7 , associated with a 4 year earlier EAOA. This variant was associated with lower brain expression of ABCA7 , which was associated with increased amyloid pathology at autopsy. Multiple immune‐related diseases shared genetic covariance with EAOA. DISCUSSION APOE , ABCA7 , and RASGEF1C associated with earlier EAOA, with supporting evidence from tissue‐specific expression analyses, offering insights into intervenable targets at early stages of AD. Highlights Novel methods estimate how long ago a patient converted to amyloid positivity. Estimating this amyloid clock allows us to determine the onset of the earliest Alzheimer's disease changes. We evaluated what genes influence when someone converts to amyloid positivity. Apolipoprotein E ( APOE ) , ABCA7 , and RASGEF1C associated with earlier age of amyloid positivity. Genetic results were supported by tissue‐specific expression analyses.

Abstract Background: For large language models (LLMs) to reach their potential as information technology tools that make medication use safer, clinically relevant benchmarks capable of automated grading and designed specifically to measure the performance of LLMs for medication tasks are required. The purpose of this study was to design a suite of benchmarking tests reflective of Comprehensive Medication Management (CMM; the standard of care for medication optimization) and quantify the baseline performance of the latest LLMs. Methods: We established six benchmarks representing critical stages of the CMM process: drug formulation matching, drug order (sig) generation, drug route matching, drug-drug interaction identification, renal dose identification, and drug-indication matching. For each benchmark, we curated a clinician-annotated dataset comprising 250 standardized input-output pairs including both inpatient and outpatient medications. We evaluated the clinical knowledge retrieval capabilities of three LLMs: GPT-4o-mini, MedGemma-27B, and LLaMA3-70B. We employed a zero-shot prompting strategy, excluding in-context examples, to assess the models’ internal clinical knowledge rather than their few-shot learning potential. To check reliability, each model was run three times using a temperature of 0.7 (a mid-range value of an LLM setting controlling text generation randomness). Performance was assessed using task-specific evaluation metrics including precision (positive predictive value), recall (sensitivity), F1-score, accuracy, and correctness consistency across trials. Results: Across six benchmarks, LLaMA3-70B demonstrated the highest performance in four tasks: drug-formulation matching (F1, 54.0% [95 CI: 50.1-58]), drug-order generation (accuracy, 88.0%), drug-route identification (F1, 74.3% [95 CI: 71-78]), and drug-indication identification (accuracy, 97.6% [95 CI: 95.6-99.2]). In the drug–drug interaction task, GPT-4o-mini achieved the highest overall accuracy (70.4% [95 CI: 64.8-75.7]). For renal dose–adjustment identification, GPT-4o-mini demonstrated the highest F1 score (83.3% [95 CI: 77.6-88]). Correctness-consistency scores ranged from 8.0% to 97.6% across benchmarks, with no model exhibiting uniformly superior consistency. Conclusions: Model performance varied substantially across medication-related tasks. LLaMA3-70B demonstrated promising baseline performance in tasks involving formulation, ordering, route, and indication. GPT-4o-mini showed potential advantages in drug–drug interaction detection and renal dose adjustment. These findings underscore the need for task-specific evaluation when deploying models for medication-focused clinical decision support.

Abstract Background The purpose of this study was to develop a series of benchmarks specifically for large language model (LLM) performance in the identification of drug-drug interactions (DDIs). Methods We evaluated six LLMs using a clinician-designed dataset of 750 DDI scenarios. Tasks included: (1) a pointwise two-drug classification task (2) a pairwise three-drug discrimination task and (3) a listwise 4–6 drug selection task. Zero-shot prompting was applied and performance was assessed using precision, recall, F1 score, accuracy, and self-consistency. Results Model performance varied by task structure and interaction severity. GPT-5-Chat demonstrated the highest recall and F1 score in the pointwise task (80.7% and 80.2%), whereas DrugGPT achieved superior accuracy in the pairwise (93.7%) and listwise task (92.6%). All models had improved performance with more severe compared to less severe DDIs (accuracy 72.0-99.2% vs. 58.3-86.7%). All models tended to have higher self-consistency in the pairwise and listwise experiments compared to the pointwise experiment (68.8-93.6% vs. 31.0-74.2%). No model exhibited uniformly high reliability across all tasks. Conclusions Current LLMs show inconsistent capabilities in identifying DDIs. Performance degrades as reasoning space expands with limited stability. These findings emphasize the need for multi-format evaluation frameworks and reliability-aware assessment when considering LLMs for medication-safety applications.

ABSTRACT Background Previous studies on the relationship between prediabetes, HbA1c and cognitive trajectories show mixed results. Therefore, we investigated the association of prediabetes, diabetes and HbA1c with change in global and domain‐specific cognitive scores in cognitively unimpaired (CU) adults, as well as the potential effect modification by sex or age. Methods This longitudinal study included 4236 CU persons aged ≥ 50 years from the population‐based Mayo Clinic Study of Aging. We ran linear mixed‐effect models with baseline prediabetes, diabetes and HbA1c predicting longitudinal global and domain‐specific (i.e., memory, language, attention/executive function, and visuospatial skills) cognitive z‐scores and raw scores. Models were adjusted for age, sex, education, medical comorbidity, repeated cognitive testing, and ApoEɛ4. We additionally ran models with sex and age interactions. Results Compared to normoglycemic individuals, individuals with prediabetes and diabetes showed poorer performance on cognitive tests, i.e., TMT‐B, category fluency, and block design subtest over time. Additionally, the presence of prediabetes, diabetes, longer diabetes duration and higher HbA1c were associated with faster global and domain‐specific cognitive decline over a median follow‐up of 6.4 years (range 1–19). Three‐way interactions showed that the effect of diabetes on decline in global cognition, attention and visuospatial domains was more pronounced in women than in men. Age did not modify the effect of diabetes or higher HbA1c on cognition. Conclusions The results suggest that prediabetes and diabetes significantly shorten the timeframe before a potentially clinically noticeable change in global and domain‐specific cognition is reached. Thus, lifestyle modification to reverse prediabetes and prevent diabetes could potentially reduce the rate of cognitive decline in aging populations. The findings differed by sex (i.e., the effect of diabetes on decline in global cognition, attention and visuospatial domains was more pronounced in women than men).