Citations

Abstract Introduction Chimeric antigen receptor (CAR) T-cell therapy and bispecific antibodies have emerged as transformative options for patients with relapsed or refractory multiple myeloma (MM). Both modalities commonly target B-cell maturation antigen (BCMA) and have demonstrated high overall response rates in pretreated populations ranging from 65-83%. However, the optimal sequencing and comparative effectiveness of these two novel therapies remain unclear. Understanding their treatment outcomes is critical to inform treatment decisions in this rapidly evolving therapeutic landscape. Method A retrospective cohort study was conducted using TriNetX, a global federated health research network that provides access to electronic medical records from approximately 132 million patients, primarily in the United States. Adult patients (≥18 years) diagnosed with relapsed or refractory MM between January 1, 2022, and January 1, 2024, were identified using ICD-10-CM codes. Patients were categorized into two cohorts: those who received CAR T-cell therapy including idecabtagene vicleucel and ciltacabtagene autoleucel and those who received bispecific antibodies including teclistamab and elranatamab. Propensity score matching (1:1) using nearest neighbor matching with a 0.1 pooled standard deviation caliper was performed for demographics, comorbidities, prior stem cell transplantation, medications, and performance status. Study outcomes included 1-year rates of all-cause mortality, all-cause hospitalization, cytokine release syndrome (CRS), immune effector cell-associated neurotoxicity syndrome (ICANS), anemia, neutropenia, thrombocytopenia, sepsis, and pneumonia. Statistical analyses were conducted on the TriNetX platform, with significance set at P<0.05 (two-sided). TriNetX calculates hazard ratios (HRs) and cumulative incidences (CIs) using R survival package version 3.2-3 with the proportional hazard assumption tested using Schoenfeld residuals. This study does not require Institutional Review Board review or informed consent due to the use of de-identified data. Results Our study identified 2,023 patients, including 911 patients who received CAR T-cell therapy and 1,112 who received bispecific antibiotics. After propensity score matching (PSM), each cohort included 649 patients. The CAR T-cell therapy group had a mean age of 66.8 ± 8.4 years, with 41.4% female, 68.0% White, 20.6% African American, and 2.3% Asian. The bispecific antibody group had a mean age of 66.3 ± 9.3 years, with 41.1% female, 67.3% White, 21.1% African American, and 2.2% Asian. CAR T-cell therapy group experienced a 64% lower risk of all-cause mortality (HR 0.36; 95% CI, 0.26–0.48) and a 44% lower risk of all-cause ICU admission (HR 0.56; 95% CI, 0.42–0.75) compared with bispecific antibodies group. However, CAR T-cell therapy group was associated with higher risk of CRS (HR: 1.91; 95% CI: 1.57–2.31), ICANS (HR: 1.50; 95% CI: 1.12–2.01), neutropenia (HR: 2.32; 95% CI: 1.96–2.76), and thrombocytopenia (HR: 1.80; 95% CI: 1.43–2.28) compared to bispecific antibodies group. There was no significant difference observed in anemia (HR: 0.96; 95% CI: 0.82–1.13) between these two treatment groups. Additionally, CAR T-cell therapy group was associated with 46% lower risk of sepsis (HR: 0.54; 95% CI: 0.39–0.75) and 68% lower risk of pneumonia (HR: 0.32; 95% CI: 0.17–0.61) compared to bispecific antibodies group. ConclusionIn this real-world analysis, CAR T-cell therapy demonstrated a substantial lower mortality and reduced ICU utilization compared with bispecific antibodies in patients with relapsed or refractory MM, despite its higher risks of CRS, ICANS, and hematologic toxicities. An important limitation of our study is the lack of data on the number of prior lines of therapy, which likely differed between groups, as bispecific antibodies are typically used in the fourth line or beyond, while CAR T-cell therapies are approved to be used as the second line. This discrepancy may have influenced both mortality and infection rates. Despite the limitation, our findings underscore the need for prospective clinical trials directly comparing CAR T-cell therapy and bispecific antibodies to validate these observations and define the optimal sequencing of these modalities. Furthermore, research into predictive biomarkers and personalized selection algorithms will be crucial to optimize treatment decisions in this rapidly evolving therapeutic landscape.

Abstract Introduction Mature T-cell lymphomas, including aggressive nodal PTCL, advanced stage CTCL, and extranodal cytotoxic lymphomas, are associated with poor outcomes. While karyotype (KT) abnormalities are frequently detected in these diseases, limited data exist on their prognostic significance. We evaluated the impact of KT abnormalities, including complex KT and specific chromosomal lesions, along with clinical and demographic factors, on overall survival (OS) in a single-center cohort. Methods We retrospectively reviewed 189 patients (pts) diagnosed with T-cell lymphoma who had successful metaphase KT at diagnosis from peripheral blood (PB), bone marrow (BM), or tumor tissue. T-cell lymphomas were classified according to WHO 2022 into nodal, extranodal/cytotoxic, and cutaneous subtypes. CTCLs were stratified into indolent (Early-stage CTCL, IA–IB) and aggressive (Sézary syndrome, transformed mycosis fungoides). KTs classified as normal, abnormal, and complex (≥3 abnormalities), and recurrent abnormalities were recorded. Baseline clinical and lab data were collected. OS was estimated using the Kaplan-Meier (KM) method with 95% confidence intervals (CI). Univariate Cox proportional hazards (CPH) model identified factors associated with OS, and significant variables were entered into multivariate CPH models. Results The median age was 66 years (IQR 55-74); 65% were male and 35% were female. Overall, 58% of pts were alive at analysis. Nodal PTCL was the most common subtype (113 pts, 60%), followed by early CTCL (30, 16%), aggressive CTCL (24, 13%), and extranodal lymphomas (22, 12%). KT was obtained from 166 BM (86%), 26 PB (13%), and 2 tissue (1%) specimens. Among abnormal KTs, 63% were from BM, 35% from PB, and 2% from tissue. Specific abnormalities included gain of chr1 (5%), gain of chr3 (5%), gain of chr7 (5%), gain of chr8 (5%), monosomy 5 or del(5q) (4%), monosomy 9 or del(9q13q22) (4%), Monosomy 10 (4%), loss of chrY (5%), and presence of marker chromosomes (5%).Of the 189 pts, 48 (25%) had an abnormal KT, while 141 (75%) had a normal KT. Complex KT was observed in 31 (16%) pts. Abnormal non-complex and complex KTs were seen respectively in: aggressive CTCL (8%, 42%), nodal PTCL (4%, 13%), extranodal (9%, 14%), and early CTCL (27%, 10%). KM survival analysis showed a median OS for the entire cohort of 95 months (95% CI: 65-245). Pts with a normal KT had a median OS of 143 months (95% CI: 81-inf) compared with 34 months (95% CI: 18-83) for those with an abnormal KT. Pts without a complex KT had a median OS of 144 months (95% CI: 85-inf), whereas those with a complex KT had 14 months (95% CI: 4-25). For specific abnormalities, pts with gain of chr7 had OS of 3 months, gain of chr8 had 19 months, monosomy 5 or del(5q) had 4 months, loss of chrY had 30 months, gain of chr1 had 4.5 months, gain of chr3 had 7.5 months, monosomy 9 or del(9q13q22) had 2 months, monosomy 10 had 17 months, and marker chromosome presence had 5.5 months. Univariate CPH identified complex KT (HR = 3.9, p < 0.001), gain of chr1 (HR = 5.7, p < 0.001), gain of chr3 (HR = 4.6, p < 0.001), monosomy 5 or del(5q) (HR = 4.6, p < 0.001), gain of chr7 (HR = 4.2, p < 0.001), gain of chr8 (HR = 3.3, p < 0.001), monosomy 9 or del(9q13q22) (HR = 4.9, p < 0.001), monosomy 10 (HR = 2.8, p = 0.008), and marker chromosomes (HR = 4.6, p < 0.001) as adverse factors. Higher platelet count was protective (HR = 0.9, p < 0.001), while greater number of abnormalities increased hazard (HR = 1.1, p < 0.001). In multivariate CPH model, complex KT (HR = 3.1, 95% CI: 1.2-8.2, p = 0.019), monosomy 9 or del(9q13q22) (HR = 4.4, 95% CI: 1.2-15.4, p = 0.022), gain of chr1 (HR = 3.5, 95% CI: 1.1-11.8, p = 0.042), and low platelet count (HR = 0.98, 95% CI: 0.97-0.99, p = 0.006) remained independent predictors of poor OS. Conclusion: We demonstrate that cytogenetic abnormalities detected in PB or BM, most notably complex KTs, confer a markedly poor prognosis in aggressive T-cell lymphomas. The integration of KT data with clinical parameters, such as platelet count, offers a powerful approach to refine risk stratification, identify biologically high-risk subgroups, and potentially guide personalized therapeutic strategies in this challenging disease spectrum. To our knowledge, this is one of the largest single-center studies assessing the prognostic significance of KT in mature T-cell lymphomas, providing novel insights into the role of specific chromosomal lesions in patient outcomes.

Importance The ideal duration of androgen deprivation therapy (ADT) for treating localized prostate cancer is unknown due to variable adherence and treatment durations tested in clinical trials. Objective To determine the ideal duration of ADT for patients with prostate cancer treated with radiotherapy. Data Sources This individual patient data meta-analysis of 13 randomized phase 3 clinical trials evaluated the use of radiotherapy alone or with ADT. It included patients with a median follow-up of 11.3 (IQR, 9.5-14.5) years and ADT duration of 0 to 36 months. Most patients (7392 [72%]) included had National Comprehensive Cancer Network high-risk or very high-risk disease. Study Selection For this meta-analysis, a systematic literature search from 1980 to 2020 was performed in trial registries (Cochrane Central Register of Controlled Trials and ClinicalTrials.gov), MEDLINE (1966-2020), Embase (1982-2020), Web of Science, and Scopus to identify trials. Data Extraction and Synthesis Intention-to-treat and as-treated analyses were performed. The number needed to treat to prevent 1 distant metastasis at 10 years was calculated based on prognostic risk group. The analyses were conducted from January 5 to August 15, 2023. Main Outcomes and Measures The primary end point for this study was overall survival, defined as time to death or last follow-up from randomization. Secondary end points included biochemical recurrence, distant metastasis (DM), prostate cancer–specific mortality, and other-cause mortality. Results The median (IQR) age among the 10 266 male patients was 70 (65-74) years. Longer durations of ADT were associated with nonlinear improvement in relative benefits of DM, prostate cancer–specific mortality, and overall survival, with reduced estimated benefits beyond 9 to 12 months of ADT based on the end point. There was a near-linear increase in other-cause mortality associated with long-term ADT use (hazard ratio, 1.28; 95% CI, 1.09-1.50; P = .002 for 28 vs 0 months of ADT). The optimal ADT duration based on 10-year DM was 0, 6, 12 months, and undefined for patients with 1 intermediate-risk factor, 2 or more intermediate-risk factors, and National Comprehensive Cancer Network high-risk and very high-risk disease, respectively. Conclusions and Relevance The results of this meta-analysis suggest that, for men with localized prostate cancer treated with definitive radiotherapy and ADT, there are relative and absolute benefits from increasing durations of ADT that help provide individualized risk estimates.

Abstract Cerebrospinal fluid (CSF) cell-free DNA (cfDNA) is under evaluation as a liquid biopsy tool across multiple neuro-oncologic indications, including diagnosing lesions of uncertain etiology, delineating progression from pseudoprogression, and confirming suspected leptomeningeal disease (LMD). We evaluated the real-world performance of CSF cfDNA sequencing for identification of variant alleles (VA) using targeted panel sequencing (PredicineCARE) and/or copy number alterations (CNA) using low-pass whole-genome sequencing (LPWGS; PredicineSCORE) in a diverse cohort across three aforementioned indications. From 82 adult patients, 93 samples (92.4% from lumbar puncture and 7.6% from ventriculoperitoneal shunt taps) were analyzed. For CNS progression, gliomas were most frequently evaluated (60%); for LMD, metastatic disease was most common (86.7%). Median CSF volume utilized for analysis was 4.60 mL (range: 0.40-7.50 mL). Median cfDNA yield was 0.29 ng/mL (range: 0.00–1037.74 ng/mL). For lesions of uncertain etiology (n=40 patients, 1 sample/patient), VAs were identified in 10 samples (25%); an additional 10 samples had sufficient cfDNA for LPWGS (25%). 50% of samples in this indication had insufficient cfDNA for testing. For suspected progression (n=26 patients, 35 samples), VAs were identified in 6 samples (17.1%); an additional 14 samples underwent LPWGS (40%); 15 had insufficient cfDNA for further testing (42.9%). For suspected LMD (n=16 patients, 18 samples), VAs were quantifiable in 4 samples (22.2%); LPWGS could be performed in an additional 9 (50%); 5 had insufficient cfDNA (27.8%). Across indications, identification of known tumor-associated alterations in CSF, including those confirmed previously by tissue, served as a non-surgical diagnostic adjunct. In some patients, targetable mutations could be identified via CSF. Conversely, absence of such alterations and/or low cfDNA yields may lower suspicion for active neuro-oncologic disease, when evaluated alongside clinical history. Despite modest cell-free DNA yield, CSF cfDNA evaluation provides valuable information to aid clinical decision-making across neuro-oncologic indications.

Abstract Background The associations between different types of diabetes, characterized by distinct pathophysiology and genetic architecture, and pancreatic ductal adenocarcinoma (PDAC) risk are not understood. Methods We investigated associations of genetic susceptibility to type 2 diabetes (T2D), 8 T2D mechanistic clusters, type 1 diabetes (T1D), and maturity-onset diabetes of the young (MODY) with PDAC risk. We used genome-wide association study (GWAS) summary-level statistics for T2D (242 283 cases, 1 569 734 controls), T1D (18 942 cases, 501 638 controls), and PDAC (10 244 cases and 360 535 controls) in individuals of European ancestry. Results Two-sample Mendelian randomization (MR) using the Robust Adjusted Profile Score (MR-RAPS) method indicated that genetically predicted T2D was associated with PDAC risk (OR = 1.10; 95% CI = 1.05 to 1.15), particularly the T2D obesity (OR = 1.28; 95% CI = 1.15 to 1.42) and lipodystrophy (OR = 1.25; 95% CI = 1.03 to 1.51) clusters. No association was observed for T1D with PDAC risk (OR = 1.01; 95% CI = 0.99 to 1.02). Pathway/gene-set analysis using the summary-based Adaptive Rank Truncated Product (sARTP) method revealed a significant association between the MODY gene-sets and PDAC risk (P = 1.5 × 10−8), which remained after excluding 20 known PDAC GWAS loci (P = 7.6 × 10−4). HNF1A, FOXA3, and HNF4A were the top contributing genes after excluding the previously identified GWAS loci regions. Conclusions Our results from this genetic association study support that T2D, particularly the obesity and lipodystrophy mechanistic clusters, and MODY genomic susceptibility regions play a role in the etiology of PDAC.

Background The eccentric stenosis morphology in patients with intracranial atherosclerotic stenosis (ICAS) has a critical impact on the efficacy of endovascular treatment. This study aimed to establish a rabbit model of eccentric carotid atherosclerotic stenosis that mimics the pathophysiological and morphological characteristics of eccentric plaques in humans. Methods Rabbits received localized cryoinjury of the right carotid artery after 2 weeks of a high fat, high cholesterol diet, followed by an additional 8 weeks on the same diet. On the experimental endpoint, comprehensive analyses of the model’s pathological changes were conducted using serum lipid analysis, DSA, optical coherence tomography (OCT), histopathological analysis, and immunohistochemistry. Results The high fat, high cholesterol diet significantly elevated serum lipid levels in rabbits. Hematoxylin and eosin staining results at 24 hours postoperatively revealed focal intimal damage induced by 5 s of cryoinjury. After localized cryoinjury and high fat, high cholesterol diet feeding, DSA results demonstrated significant carotid stenosis, with a stenosis degree of 47±5% and a length of approximately 9 mm. OCT images revealed eccentric fibrous plaque formation at the stenotic site. Morphological analysis of the plaque samples demonstrated a plaque eccentricity index of 0.66±0.05 and a plaque burden of 60.68±7.41%. Histopathological analysis showed that atherosclerotic lesions developed in all rabbits of the model group. Among them, three rabbits had intermediate lesions, one rabbit had atheroma, and eight rabbits had fibroatheroma. Conclusions This study successfully established a rabbit model of eccentric carotid atherosclerotic stenosis, providing a novel experimental platform for transformation research of neurointerventional devices.

Abstract Background Delirium is often underdiagnosed in clinical practice and is not routinely coded for billing. While manual chart review can identify delirium, it is labor-intensive and impractical for large-scale studies. Natural language processing (NLP) can analyze unstructured text in electronic health records (EHRs) to extract meaningful clinical information. Methods To support national integration of NLP for EHR-based delirium identification across different institutions, we launched the Delirium Interest Group within the national Evolve to Next-Gen Accrual to Clinical Trials (ENACT) NLP Working Group. This paper outlines our initial efforts to standardize, evaluate, and translate an NLP-based delirium detection model into the i2b2/ENACT platform. Results Multisite contextual inquiry identified several key challenges, including variations in local screening practices (eg, tools used, documentation frequency, and quality control), the need for harmonized definitions in the context of EHRs, and the complexity of modeling temporal logic. Multisite NLP evaluation revealed variable performance degradation driven by differences in delirium screening practices, clinical documentation patterns and semantics, and note syntactic structures. Conclusion Our work represents an important first step toward enabling scalable and standardized NLP-based delirium detection across institutions. By engaging diverse institutions through the ENACT NLP Working Group, we identified shared challenges and site-specific variations that impact model implementation and performance. Our collaborative approach enabled the development of a more robust framework for delirium identification across heterogeneous EHR systems. Future efforts will build on this foundation to enhance the validity, usability, and translational impact of delirium detection.

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.

Abstract Current discussion surrounding the clinical capabilities of generative language models (GLMs) predominantly center around multiple-choice question-answer (MCQA) benchmarks derived from clinical licensing examinations. While accepted for human examinees, characteristics unique to GLMs bring into question the validity of such benchmarks. Here, we validate four benchmarks using eight GLMs, ablating for parameter size and reasoning capabilities, validating via prompt permutation three key assumptions that underpin the generalizability of MCQA-based assessments: that knowledge is applied, not memorized, that semantic consistency will lead to consistent answers, and that situations with no answers can be recognized. While large models are more resilient to our perturbations compared to small models, we globally invalidate these assumptions, with implications for reasoning models. Additionally, despite retaining the knowledge, small models are prone to answer memorization. All models exhibit significant failure in null-answer scenarios. We then suggest several adaptations for more robust benchmark designs more reflective of real-world conditions.

Abstract Background Breast cancer genome-wide association studies (GWAS) have identified more than 200 susceptibility loci, but most studies are dominated by European and East Asian populations. Methods We analyzed breast cancer GWAS summary statistics from African (AFR), East Asian (EAS), European (EUR), and Hispanic/Latina (H/L) samples (159,297 cases and 212,102 controls). We estimated logit-scale SNP-based heritability, polygenicity, and cross-ancestry genetic correlation, partitioned heritability across functional annotations, and integrated GWAS results with the Tabula Sapiens single-cell atlas using scDRS+. Results The logit-scale heritability of breast cancer ranged from ℎ 2 =0.47 (SE = 0.07) in EAS to AFR ℎ 2 =0.61 (SE = 0.10), with no significant differences across ancestries (p=0.63). The estimated number of susceptibility markers in a sparse normal-mixture effects model also varied from 4,446 (SE = 3,100) in EAS to 8,308 (SE = 2,751) in AFR, but differences were not significant across ancestries (p=0.55). Cross-sample genetic correlations varied, with the strongest correlation between EUR and EAS (𝜌 = 0.79, SE = 0.08) and weakest between AFR and H/L (𝜌 = 0.26, SE = 0.24). Regulatory annotations were enriched for breast cancer heritability across samples. Integration with single-cell expression profiles implicated ancestry-shared associations with innate immune, secretory epithelial, and stromal cell types. Conclusion These results indicate substantial cross-ancestry sharing of breast cancer polygenic architecture, highlight a consistent contribution of regulatory variation, and identify convergent cellular contexts that motivate functional follow-up and inform expectations for the transferability and attainable performance of common-variant risk prediction across populations.