Citations

Abstract Background Accurate diagnosis and risk stratification in myeloproliferative neoplasms (MPNs) rely on expert interpretation of complex clinical records. Outside major academic centers, specialist expertise is scarce, and time-consuming chart reviews extend clinic visits and delay treatment decisions, underscoring the need for automated, point-of-care phenotyping for MPN patients. Methods Patients with the ICD codes of PV, ET and MF were identified from the electronic health record. Cases were defined as patients meeting the 2017 WHO diagnostic criteria for PV, ET or MF or those with a definitive MPN diagnosis determined by a hematologist. Controls were the patients who met neither the diagnostic criteria nor had a hematologist-assigned diagnosis. Two clinicians extracted demographic, clinical (symptoms, thrombotic history), laboratory (complete blood count, serum chemistry), genetic (mutations, karyotype), bone marrow morphology, diagnostic mentions and prognostic score information from unstructured clinical notes and pathology reports. Prognostic scores included the conventional risk model for PV, the International Prognostic Score of Thrombosis (IPSET) for ET and the Dynamic International Prognostic Scoring System (DIPSS), DIPSS-plus or Mutation-enhanced International Prognostic Scoring System (MIPSS-70) for MF. Discrepancies in data extraction were resolved by a consensus. A random sample of 510 patients (PV: 85, ET: 85, MF: 85, controls: 255) was split into prompt development (n=60) and test (n=450) sets for iterative workflow engineering and evaluation. The workflow included: (i) a rule-based algorithm for automated retrieval of demographic details, the first bone marrow biopsy report, the first clinical note after biopsy and the latest laboratory results within 90 days preceding biopsy (ii) a HIPAA-compliant LLM (GPT-4 turbo) for data extraction from the clinical notes and pathology reports (iii) a regular expressions-based algorithm for post-processing extracted data into structured tables, (iv) a clinician-informed source prioritization algorithm to select data from preferred sources, (v) a modified WHO-informed diagnostic characterization tool incorporating hematologist-assigned diagnoses to distinguish cases from controls and; (vi) an NCCN guidelines-based algorithm for prognostic tool selection and risk-stratification based on available data. Structured prompts were used to extract demographic, clinical, laboratory, genetic and pathological data in a zero-shot deterministic setting. The performance of automated data extraction, diagnostic characterization and prognostication was compared against the manually curated data. Accuracy and F1 scores were computed for data extraction and prognostication tasks while sensitivity and specificity were calculated to assess diagnostic performance. Results A total of 450 (PV: 150, ET: 150, MF: 150) pathology reports and 50 (PV: 52, ET: 55; MF: 65) clinical notes were identified for 450 test set patients. The test set data extraction accuracy was 95% (reports: 97%, notes: 92%) for PV, 93% (reports: 92%, notes: 94%) for ET and 98% (reports: 99%, notes 97%) for MF. The test set F1 score was 0.95 (reports: 0.96, notes: 0.93) for PV, 0.95 (reports: 0.95, notes: 0.96) for ET and 0.98 (reports: 0.99, notes: 0.97) for MF. PV patients were diagnosed with 100% sensitivity and 98% specificity whereas ET and MF patients were diagnosed with 100% sensitivity and specificity. The accuracy of risk stratification was 100% for PV and ET and 94% for MF. The F1 score for risk stratification was 1.00 for PV and ET and 0.94 for MF. Conclusions The first-in-class automated framework successfully identified pertinent clinical information, performed accurate data extraction from unstructured records and enabled clinician-guided source prioritization for accurate diagnosis and risk stratification of MPN patients. Prospective validation of the framework is required for widespread clinical implementation.

Abstract In this study, we investigated tumor intrinsic mechanisms of resistance to cevostamab, a bispecific T-cell engager targeting Fc receptor-like protein 5 (FCRL5) and CD3 that has shown promising clinical activity and durable responses in advanced MM. Longitudinal bone marrow aspirates from nine MM patients treated with cevostamab were analyzed. All patients had previously been treated with both immunomodulatory drugs (IMiDs) and proteasome inhibitors (PI) and eight patients had received daratumumab prior to cevostamab, which was, on average, the 6th line of treatment (mean treatment duration 220±273 days). Five of nine patients responded with ≥PR. Whole genome and transcriptome sequencing was performed on sorted CD138+ MM cells at baseline, and whole exome and transcriptome sequencing was performed on five and three samples, respectively, following cevostamab discontinuation. Baseline FCRL5 expression levels did not correlate with clinical response. However, in one non-responder patient, FCRL5 expression increased four-fold after 28 days of treatment. Given the location of the FCRL5 gene on chromosome 1q—commonly amplified in advanced MM—we evaluated the presence of Gain(1q) in our study cohort. All patients whose best response was ≤PR (n=6) harbored a Gain(1q), whereas it was absent in those who achieved VGPR and sCR (n=3). Crucially, one of the patients that achieved sCR was initially treated with 60 mg of cevostamab for one year and then escalated to a higher dose cohort (90 mg) in the second year. At progression, a clonal FCRL5E115A missense mutation (allelic frequency = 56%) and a subclonal FCRL5G455Efs*19(allelic frequency = 5%) frameshift deletion mutation were detected, both absent at baseline. The missense mutation was located on the extracellular domain of FCRL5 targeted by cevostamab. To validate the functional impact of FCRL5E115Aand FCRL5G455Efs*19 on the sensitivity to cevostamab in vitro, we generated a FCRL5 knockout (KO) model utilizing CRISPR-Cas9 technology in OPM2 MM cells. In vitro cevostamab cytotoxicity assay with PBMCs and pan T cells from healthy donors at various effector to target (E:T) ratios confirmed a complete loss of cevostamab sensitivity in our FCRL5KO models. cDNA encoding patient-derived FCRL5 alterations (FCRL5E115Aand FCRL5G455Efs*19) as well as FCRL5wtwere cloned into mammalian expression vector and reintroduced into FCRL5KO cells. Cevostamab sensitivity was reinstated by FCRL5wt expression in KO cells. In contrast, FCRL5E115A conferred complete resistance to cevostamab (p<0,001) and FCRL5G455Efs*19 depicted a significant decline in cevostamab sensitivity relative to FCRL5wt expressing models. Additional experiment utilizing direct stochastic optical reconstruction microscopy (dSTORM) are ongoing to monitor the membrane localization of FCRL5 mutants. By integrating genomic, transcriptomic and functional data with clinical outcomes, this work is the first to report on acquired FCRL5 mutations causing resistance to cevostamab. We hypothesize that the prolonged treatment duration may have contributed to the acquisition of the target mutations and note a subsequent durable sCR response to BCMA-targeted immunotherapy in the same patient, suggesting immune-related factors such as T cell exhaustion, were not drivers of cevostamab resistance. Due to the limited sample size, these findings require validation in larger cohorts. Nonetheless, our findings begin to elucidate tumor intrinsic mechanisms associated with cevostamab resistance.

Abstract Introduction: Achieving rapid and deep hematological responses is essential for improving organ function and reducing early mortality in AL amyloidosis. While daratumumab has transformed the management of newly diagnosed disease, many daratumumab-treated patients will not clear amyloidogenic serum free light chains or will relapse. Chimeric antigen receptor (CAR) T-cell therapy is a highly active modality in multiple myeloma, capable of inducing deep and durable remissions. Whether this strategy can be safely and effectively adopted in AL amyloidosis remains an important unanswered question. Methods: This multicenter retrospective study involved 8 US academic medical centers within the US Multiple Myeloma Immunotherapy Consortium. We included patients with systemic AL amyloidosis who received BCMA-targeted CAR T therapy to treat AL amyloidosis with concurrent myeloma to investigate the safety, and efficacy of this approach. Results: Between September 2021 and March 2025, we identified 33 patients with AL amyloidosis treated with CAR T (n=9 ide-cel, n=24 cilta-cel). The median age was 69-years (IQR: 60, 72); 48% were female. Most patients were kappa light-chain isotype (61%, n=20), 4 patients (13%) had the t(11;14) translocation. Although baseline plasma cell burden was unknown in 2 cases, all patients had a concurrent diagnosis of myeloma, and 88% of patients had ≥20% bone marrow plasma cells (BPMC) at diagnosis. The median number of prior lines was 6 (range: 3-13), all patients had previous anti-CD38 exposure, and 27 patients (82%) had received previous autologous stem cell transplant. Regarding organ involvement, 11 patients (33%) had cardiac, 9 (27%) renal, 8 (24%) gastrointestinal, 3 (9%) peripheral nerve and 1 (3%) autonomic nerve involvement. Five patients had AL amyloid deposition without known organ involvement. Prior to CAR T, cardiac staging was unavailable or unknown in 19 cases. Of the remaining 14 cases, 6 cases were European Modification of Mayo 2004 stage 3A with no stage 3B cases. Prior to lymphodepletion the median BMPC was 6% (IQR: 1, 20). The overall rate of cytokine release syndrome (CRS) was 64% (Grade 1, G1: n=15, Grade 2, G2: n=4, Grade 3, G3: n=2), and immune cell associated neurological syndrome (ICANS) was 15% (G1: n=3, G3: n=2). The rate of CRS was 56% for ide-cel (G1: 33%, G2: 22%) and 67% for cilta-cel (G1: 50%, G2: 8%, G3: 8%). The median time to CRS onset was 1 day for ide-cel and 6 days for cilta-cel. The median duration of CRS was 1 day for ide-cel and 2 days for cilta-cel. The rate of ICANS was 11% for ide-cel (G1: n=1), and 16% for cilta-cel (G1: n=2, G3: n=2). There were no cases of non-ICANS neurotoxicity. Ten patients (30%) experienced infections (≥Grade 3 = 9%). The overall response rate (ORR) by AL amyloidosis response criteria (Palladini et al. 2012) was 91%, including n=5 very good partial responses (VGPRs) and n=24 complete responses (CRs). The ORR for ide-cel was 78% (n=6 CRs, n=1 VGPR), and for cilta-cel was 96% (n=18 CRs, n=4 VGPRs). Among responders, the median time to VGPR or better was 30 days (IQR: 29, 32). One cilta-cel recipient died of infection on day 17 prior to disease reassessment. At a median follow up of 17.8 months, the 12 and 18-month overall survival rates were 88% (95%CI: 76%-100%) and 83% (95%CI: 68%-100%), respectively. There were 5 deaths; 3 were related to underlying disease, 1 related to infection, and 1 unrelated. Seven patients were eligible for cardiac response assessment (NTproBNP>650 OR BNP>150) at the time of CAR T, and of those, three obtained a cardiac response (2 Cardiac PR, 1 Cardiac VGPR). Three patients were eligible for renal response assessment (>1g proteinuria/day) at the time of CAR T; of those, two obtained a renal PR. Conclusion: We demonstrate that commercial BCMA-targeted CAR T-cell therapy induces deep and rapid hematological responses in a cohort of heavily pretreated patients with AL amyloidosis. Although all patients exhibited overlapping features with concurrent multiple myeloma (≥10% plasma cells at diagnosis), CAR T-cell therapy was feasible to administer, with modest rates of CRS and only isolated cases of severe CRS or ICANS. Nearly all patients achieved a deep hematological response, which appeared to translate into meaningful organ improvement among those who were evaluable.

Abstract Background: Graft-versus-host disease (GVHD) remains the leading cause of non-relapse morbidity and mortality in allogeneic hematopoietic-stem-cell transplantation (allo-HSCT) recipients. Clinical scoring systems (e.g., EBMT, HCT-CI, DRI, etc.) can stratify risk but show modest predictive performance (C-statistics 0.49–0.67) when it comes to GVHD. To address this gap, we developed GVHD-Intel 1.0, a real-time, scalable, and interpretable machine learning (ML) framework from an international study to predict both acute (aGVHD) and chronic GVHD (cGVHD), providing individual and bulk predictions using common input variables and HLA alleles. Methods: This retrospective global study, approved by respective institutional review boards, included electronic health records from three tertiary care centers in Abu Dhabi, UAE, and King Hussein Cancer Center (KHCC), Jordan, from 2009 to 2023. The model was developed in phases. The UAE399 cohort (399 HSCT recipients) served as the baseline dataset. To enhance model generalizability, 55 adult allo-HSCT recipients from KHCC were added to the training set. Two distinct test cohorts were used: UAE46 (46 allogeneic HSCT recipients from UAE) and KHCC150 (150 adult allogeneic HSCT recipients from KHCC), collected later and independently of training data. Variables included donor type, sex mismatch, disease status, time from diagnosis to transplant, and agents used in conditioning and GVHD prophylaxis (e.g., TBI, ATG, alemtuzumab use etc.). HLA alleles (A, B, C, DQ, DR) were symbolically encoded and transformed via one-hot encoding. GVHD events were classified as aGVHD (<100 days) or cGVHD (≥100 days). Missing values were excluded, and class imbalance was handled via class weighting. GVHD-Intel 1.0, co-developed with MBZ University of Artificial Intelligence, features modular ensemble architecture, drug-level feature engineering, novel HLA encoding for low-resolution compatibility scoring, and built-in interpretability using SHAP and LIME. The framework is cloud-deployable for both real-time and batch predictions. Results: The GVHD-Intel 1.0 was externally tested on two independent, temporally and institutionally distinct cohorts (UAE46 and KHCC150), demonstrating strong and consistent predictive performance for both acute and chronic GVHD. For cGVHD prediction, the model achieved an AUC (area under the curve or C-statistic for binary outcomes) of 0.832, F1 score of 0.896, accuracy of 82.7%, PPV of 0.930, sensitivity of 0.864, Brier score of 0.178, and log loss of 0.543. For aGVHD, performance metrics included an AUC of 0.802, F1 score of 0.860, accuracy of 77.6%, PPV of 0.930, sensitivity of 0.799, Brier score of 0.196, and log loss of 0.581. Model calibration remained strong across both tasks, with Brier scores <0.20. Conclusions: GVHD-Intel 1.0 is a robust, interpretable machine learning framework developed from real-world, multicenter data. By integrating widely available clinical variables and HLA allele data, it outperforms conventional clinical risk scores in predicting both acute and chronic GVHD in allo-HSCT recipients. Slightly reduced performance in aGVHD prediction is likely due to the lack of relevant variables. Next steps include prospective validation across multiple transplant centers globally and continued refinement through expanded data integration. GVHD-Intel 1.0 holds strong potential as a clinically valuable, scalable decision-support tool that can be enacted widely and easily by practicing transplant physicians worldwide in all types of settings.

Abstract Background: Chronic Lymphocytic Leukemia (CLL) is the most prevalent leukemia among adults in Western populations, disproportionately affecting the elderly, with a median age at diagnosis of approximately 70 years. Over the past two decades, significant advancements in therapeutic options, including Bruton tyrosine kinase (BTK) inhibitors, BCL-2 antagonists, and monoclonal antibodies, have improved clinical outcomes. However, despite these advances, limited data exist on long-term mortality trends and disparities in CLL outcomes among older adults in the United States. Understanding such patterns is critical to guiding geriatric oncology care, identifying gaps in healthcare delivery, and informing public health strategies. Methods: We conducted a population-based analysis using the Centers for Disease Control and Prevention Wide-Ranging Online Data for Epidemiologic Research (CDC WONDER) Underlying Cause-of-Death database. We identified all deaths among individuals aged ≥65 years from 1999 to 2020 in which CLL (ICD-10 code C91.1) was listed as the underlying cause. Mortality data were stratified by sex, age group (65–74, 75–84, ≥85), U.S. Census region, urban-rural classification, state, and place of death. We calculated crude mortality rates (CMRs) and age-adjusted mortality rates (AAMRs) per 100,000 persons, standardized to the U.S. 2000 population. Joinpoint regression analysis was used to assess temporal trends in AAMRs and compute annual percent changes (APCs) and average annual percent changes (AAPCs), with statistical significance set at p<0.05. Results: Between 1999 and 2020, a total of 85,371 CLL-related deaths occurred among U.S. adults aged ≥65 years. The overall AAMR declined from 11.18 per 100,000 in 1999 to 7.75 in 2020, reflecting a significant downward trend (AAPC: -1.86%; 95% CI: -2.18 to -1.54; p<0.0001). Men consistently exhibited higher mortality rates (AAMR: 13.48) than women (AAMR: 6.45), though the mortality decline was more pronounced among women (AAPC: -2.10% vs. -1.65%, respectively; both p<0.0001). Regional disparities were evident, with the Midwest reporting the highest AAMR (10.63), and the West the lowest (8.71). Urbanization analysis revealed higher mortality in non-metropolitan areas (AAMR: 10.33) than metropolitan ones (AAMR: 9.08), though both experienced significant declines over time. Age-stratified analysis showed the steepest declines in the 65–74 age group (AAPC: -3.91%; p<0.0001), followed by those aged 75–84 years (AAPC: -1.96%) and ≥85 years (AAPC: -0.25%). States with the highest mortality included Minnesota, Iowa, and South Dakota, while Hawaii, Nevada, and New Mexico had the lowest. Most deaths occurred in medical facilities (43.8%), followed by home (26.7%), nursing homes/long-term care (19.5%), and hospice (5.6%). Conclusion: CLL-related mortality among the U.S. geriatric population declined significantly between 1999 and 2020, likely reflecting the impact of therapeutic innovations and improved disease management. However, substantial disparities persist across sex, age, region, urbanization, and state-level subgroups. These findings underscore the importance of equitable access to modern CLL treatments and the need for targeted interventions to reduce mortality in high-risk, underserved older populations. Efforts to close these gaps will be essential in optimizing outcomes in the era of personalized cancer care for aging Americans.

Abstract Background Adult xanthogranuloma (AXG) is a rare clonal histiocytic neoplasm (HN) with a poorly defined mutational landscape. Clinically and immunophenotypically, AXG can resemble other HNs such as Erdheim-Chester Disease (ECD) which frequently harbor MAPK pathway alterations and carry BRAF V600E mutations in close to 50% of cases. However, due to the rarity of AXG, the prevalence of the BRAF V600E mutation and the broader mutational landscape is not well known. In this study, we characterize the mutational profile of AXG. Methods We conducted a retrospective analysis of patients diagnosed with AXG at our institution between January 1, 2016, and July 15, 2025. The study was approved by the institutional review board. To ensure inclusion of true AXG cases, only patients with a confirmed histopathologic diagnosis and without multisystemic involvement or characteristic imaging findings of ECD, such as symmetric long bone osteosclerosis, were included. BRAF alterations were identified by immunohistochemistry (IHC) and/or next-generation sequencing (NGS). We also performed an extensive literature review of mutational studies in AXG and included cases that underwent comprehensive NGS. Results Seventeen patients with AXG were identified. The median age at diagnosis was 40 years (range: 18–75 years). Out of the 17 patients, 11 (65%) reported the face as the primary site of disease, and 11 (65%) had solitary lesions. Of the 17 patients, 13 (76%) underwent BRAF V600E IHC staining, and none tested positive. Twelve (71%) patients had tissue samples analyzed via NGS, and none exhibited the BRAF V600E mutation. Eight of the 12 (67%) patients with NGS testing harbored other alterations involving the MAPK pathway. These included BRAF fusions (n=3), KRAS mutations (n=2), and individual cases with TRAF7 mutation, ALK1 fusion and FGFR1 fusion. One patient was found to have a mutation in WRN, which is not previously reported in AXG. Three patients had no identifiable pathogenic alterations on NGS. Ten patients were monitored without treatment, 2 underwent surgical resections and 3 received chemotherapy (1 had a complete response [CR], 2 achieved a partial response). One patient with FGFR1 fusion was treated with pemigatinib and achieved a CR for over 3 years. One patient with TRAF mutation obtained a CR after treatment with rituximab. We also identified 51 cases of AXG with adequate mutational testing reported in the literature, which reported findings similar to ours. In a retrospective study from France involving 29 patients, MAPK alterations were observed in 55% of cases: however, only 10% harbored BRAF V600E mutations (1). Another analysis of 270 patients with various HN subtypes included 13 individuals with AXG, none had BRAF mutations. MAP2K1 mutations were seen in 31% of patients (2). Taken together, across a combined cohort of 68 AXG patients, the most common alterations were: KRAS mutations (16%), fusions (14%), MAP2K1 mutations (7%) and BRAF V600E (6%). Among fusions, the most involved gene was BRAF (50%), and GAB2 was the most common fusion partner (n=2). In total, 59% of patients had MAPK alterations. These findings collectively underscore the diverse mutational landscape of AXG and highlight the rarity of BRAF V600E mutations relative to closely related HNs. Conclusion We present a series of AXG cases with mutational profiling, from our institution and the published literature. We demonstrate that MAPK alterations are frequently observed in AXG and are driven by heterogeneous mutations. In contrast to other HNs where BRAF V600E mutations are common, we identified no such mutations in our cohort and found only rare cases in the literature. Our institutional data, along with findings from the existing literature, suggest that AXG may have a distinct molecular profile compared to other HNs. A deeper understanding of its mutational landscape may offer insights into disease biology and identify therapeutic targets. ReferencesAnnabi E, Mahévas T, Chasset et al. Clinical, Histopathologic, and Molecular Features and Treatment of Multiple Xanthogranuloma in Adults. JAMA Dermatol. 2025;161(7):765-768. doi:10.1001/jamadermatol.2025.0886.Durham BH, Lopez Rodrigo E, Picarsic J, et al. Activating mutations in CSF1R and additional receptor tyrosine kinases in histiocytic neoplasms. Nat Med. 2019;25(12):1839-1842. doi:10.1038/s41591-019-0653-6

Abstract Background:Erdheim-Chester Disease (ECD) is a rare, incurable histiocytic neoplasm. Since its recognition in 1930, our understanding of ECD's pathophysiology has evolved. The identification of its clonal nature after 2015 established ECD as a cancer, prompting increased disease awareness and treatment advances. Targeted therapies show promising responses, but changes in ECD mortality over time have not been extensively documented. Methods:A multi-institutional retrospective cohort study analyzed adults with ECD consecutively seen at Mayo Clinic MN, AZ, and FL and the University of Alabama at Birmingham between 2007 and 2024. Clinical and prognostic data were abstracted. Overall survival was estimated using the Kaplan-Meier method and compared to expected survival using U.S. Census data, based on age at diagnosis (by calendar year), date of diagnosis, and sex. Standardized mortality ratios (SMRs) were calculated by dividing the number of observed deaths in the study population by the expected number of deaths using U.S. population data. Risk of death was compared by date of diagnosis (pre- and post-2015) using Cox proportional hazards analysis. Statistical analysis was performed with R software. Results:A total of 159 ECD patients were included (median age 58 years [range 19-82], 65% male). Forty-nine patients were diagnosed between 1990-2014 and 110 between 2015-2024. Median follow-up for the entire cohort was 4.0 years (95% CI: 3.2-4.4). Median follow-up for those diagnosed between 1990-2014 was 10.4 years (CI: 3.2-16.8), compared to 3.5 years (CI: 3.0-4.2) for those diagnosed between 2015-2024. The 5-year overall survival rate was 80% (95% CI: 72-88). Patients with ECD had an increase in mortality over expected, with 28 deaths compared to an expected 11.2, an SMR of 2.5 (95% CI: 1.7-3.6, p<0.001). Of 17 patients with an identified cause of death, death was attributable to ECD in 11 (65%), secondary malignancy in 1 (7%), and unrelated causes in 5 (29%). Cause of death was unknown in 11 patients. Among deaths, 11 (39%) were treated with targeted therapy, 8 (29%) with chemotherapy, and 1 (4%) received both. One patient (4%) received steroids, 1 received immunomodulator and targeted therapy (4%), and 6 (21%) received no disease-directed therapy. Of the patients whose death was related to ECD, 10 (91%) had multisystem disease. For 49 patients diagnosed pre-2015, 14 deaths occurred during the follow-up, significantly higher than the expected 5.2 deaths based on U.S. population data (SMR of 2.7 [95% CI: 1.5-4.4, p=0.001]). Among those diagnosed between 2015 and 2024, 14 deaths occurred, compared to expected 6.1 (SMR of 2.3 [95% CI: 1.3-3.8, p=0.006]). In an age- and sex-adjusted Cox proportional hazards model, diagnosis after 2015 vs. diagnosis prior to 2015 yielded a hazard ratio of 0.59 (95% CI: 0.27-1.31, p=0.19). Conclusion:The mortality rate of ECD patients remains higher than the general population, despite advancements in disease recognition and development of targeted therapies. While our data suggest a possible trend toward improved survival among patients diagnosed after 2015, we did not find statistically significant evidence of decreased mortality compared to those diagnosed prior to 2015. Larger cohorts with longer follow-up periods are needed to draw definitive conclusions regarding mortality.

Abstract Introduction: CD20xCD3 bispecific antibodies (BsAb) have been approved for use in relapsed/refractory (RR) large B-Cell lymphomas (LBCLs) with manageable toxicities. The main adverse events of clinical interest are cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS). The risk factors for CRS and ICANS are not well understood. Knowledge of variables that contribute to increased risk may help clinicians identify and tailor management for high-risk patients (pts) and increase the comfort level for providers using these therapies in community practices. Methods: We performed a multicenter retrospective study at 21 US centers evaluating pts with RR LBCLs receiving commercially available epcoritamab (epco) or glofitamab (glofit) between 2023 and 2025. We obtained baseline characteristics, safety and efficacy outcomes. Univariate analysis was performed to identify associations between CRS, ICANS, and clinical variables of interest. Results: As of May 31, 2025, a total of 312 pts with RR LBCLs were treated with BsAb, Epco n=193 (62%), Glofit n= 119 (38%); median age 68; 63% male; 70% ECOG PS 0-1. Most common histologic type was diffuse large B-cell lymphoma, not otherwise specified (DLBCL, NOS) (n=239, 84%) followed by high-grade B-cell lymphoma (HGBL) (n=46, 16%). 179 pts (57%) were previously treated with chimeric antigen receptor T-cell (CAR-T) therapy, with 47% receiving CAR-T <6 months (mo) prior to BsAb. Among pts with prior CAR-T, CRS with CAR-T therapy of any grade occurred in 63%, with 26% ≥ grade 2. ICANS of any grade (G) occurred in 25% of pts (9.4% ≥ G3). Overall and complete response (ORR and CR) rates to BsAb were 51% and 29%, respectively. With a median follow up of 5 mo, median progression-free and overall survival (PFS and OS) were 2.3 and 7.3 mo, respectively. With BsAb treatment, CRS of any grade occurred in 117 (39%) patients: 24% G1, 8.7% G2, 4% G3, 1% G4 and 1% G5. The max grade CRS occurred most frequently after the third dose of epco (35%) and the first dose of glofit (23%). CRS was managed with additional steroids in 24% of pts and tocilizumab in 22%. 13% of pts experienced ICANS: 5.8% G1, 3.4% G2, 2% G3, 1% G4 and 0.3% G5. Factors that increased the risk for any grade CRS included: bulky disease ≥ 7.5 cm at time of BsAb treatment (OR 2.04, 95% CI 1.17 – 3.55, p = 0.012), elevated LDH (OR 2.56, 95% CI 1.45– 4.76, p = 0.002) and treatment with epco compared to glofit (OR 1.69, 95% CI 1.04- 2.77, p = 0.036). As G ≥2 CRS generally requires inpatient monitoring and management, we next compared the risk factors between pts with no CRS or G1 CRS and G ≥2 CRS. Factors associated with increased risk for high grade CRS included: bone marrow involvement by lymphoma (OR 3.47, 95% CI 0.97 – 11.3, p = 0.043), platelet count ≤ 75 (OR 2.46, 95% CI 1.21 – 4.87, p = 0.011), ≥ 2 extra nodal sites (OR 2.04, 95% CI 1.07 – 3.97, p = 0.032) and ≥G2 CRS with prior CAR-T (OR 3.35, 95% CI 1.13 – 9.97, p = 0.027). Factors that increased the risk for ICANS included: bulky disease (OR 2.67, 95% CI 1.28 – 5.49, p = 0.008), elevated Cr ≥ 1.5x ULN (OR 3.79, 95% CI 1.25 – 10.4, p = 0.012), elevated AST/ALT ≥ 3x ULN (OR 4.65, 95% CI 0.92 – 19.9, p = 0.042), low albumin (OR 2.85, 95% CI 1.49 - 5.5, p = 0.002), and co-occurrence of CRS with BsAb (OR 9.1, 95% CI 3.85- 25, p < 0.001). No other clinical variables including older age, histologic subtype, ECOG PS, B symptoms, IPI score, cell of origin or histologic subtype, baseline labs such as ALC, CRP, fibrinogen or D-dimer were significantly associated with CRS or ICANS risk. There was no difference in ORR or CR rates in patients who experienced CRS or ICANS. Conclusions: Our analysis of pts treated with commercial BsAb for LBCL shows an increased risk for CRS and ICANs based on clinical variables that are routinely obtained at baseline including tumor bulk, sites of disease, laboratory parameters, and history of prior CRS with CAR-T therapy. Differences in baseline populations receiving epco and glofit treatments in this real-world analysis have been previously reported (Brooks et al, Blood 2025) and the comparative risk for CRS between these treatments remains unclear. Future plans including multivariable analysis may allow for development of a predictive model to identify pts at high risk of CRS and ICANS with BsAb treatment, potentially leading to prospective clinical trials evaluating prophylactic interventions for high-risk pts.

Abstract Introduction: Epcoritamab (epco) and glofitamab (glofit) are bispecific antibodies (BsAbs) approved for the treatment of patients (pts) with multiply relapsed or refractory (r/r) large B-cell lymphoma (LBCL). Though response rates are relatively high, most patients (pts) eventually experience progression of disease (POD). We sought to investigate the treatment patterns and clinical outcomes following POD post-BsAb for r/r LBCL Pts, as this has not previously been described. Methods: We performed a multicenter retrospective study including 21 US centers evaluating pts with r/r LBCL receiving commercially available epco or glofit between 2023 and 2025. Baseline characteristics and outcomes were extracted from medical records. Efficacy and survival outcomes were assessed by site investigators and included overall response rate (ORR), complete and partial response (CR, PR) rates, progression-free survival (PFS) and overall survival (OS), analyzed using the Kaplan Meier method. Results: As of May 15, 2025, a total of 312 pts with r/r LBCLs were treated with BsAb, of whom a total of 168 (53.8%) pts experienced POD [104 of 193 (53.8%) treated with prior epco and 64 of 119 (53.8%) treated with prior glofit]. Pts with POD had a median age at BsAb start of 67 years (interquartile range [IQR] 58-76), most were male (68.4%), and the most common histology was diffuse large B-cell lymphoma not otherwise specified (n=122 [71.8%]). Sixty-seven of 168 pts (39.9%) with POD after BsAb had experienced primary refractory disease defined as failure to achieve PR or CR to frontline therapy. A total of 102 (60.7%) pts with POD to BsAb had received chimeric antigen receptor (CAR) T-cell therapy prior to BsAbs, with 33.3% of these pts being refractory to CAR T and 30.4% progressing within 6 months of CAR T-cell infusion. The median time to progression following BsAb start was 48 days. Following POD, 70 pts received no further anti-lymphoma therapy. Among 98 pts receiving subsequent therapy, 85 initially received systemic therapy and 13 initially received local therapy with radiation (concurrently with systemic therapy in 8 pts). The most commonly-used next line of systemic therapies included: loncastuximab tesirine in 19 pts (ORR 16.7%, all PR); commercial CAR T in 11 pts (ORR 50% [CR 36.4%, PR 18.2%); tafasitamab with lenalidomide in 7 (no responses); and other chemotherapy in 46 pts (ORR 27.6% [CR 24.1%]). The median follow-up time from POD was 9.2 months among survivors (95% confidence interval [CI] 7.8-13.3). Among those who initiated subsequent-line therapy following BsAbs, the median progression-free survival (PFS2) was 1.9 months (95% CI 1.5-2.4) and median overall survival (OS) from initiation of subsequent-line therapy was 3.9 months (95% CI 2.4-5.5). The estimated 1-year OS following initiation of next therapy was 12%. Conclusions: We report the outcomes of largest cohort of pts with POD post-BsAbs in r/r LBCL. Progression events occurred early after initiation of single agent BsAbs, and almost half of pts who progressed did not receive subsequent therapy. For those who received standard subsequent-line therapy after POD, response rates and survival outcomes were poor. Pts with POD after BsAB should be strongly considered for clinical trials of novel therapeutics or combination therapies.

Abstract Background: Sickle cell crisis (SCC) remains a leading cause of emergency hospitalization and acute morbidity in patients with sickle cell disease (SCD). With improved survival in SCD, long-term complications such as cardiovascular disease are emerging as critical determinants of morbidity and mortality. While right-sided heart failure has been adequately explored in SCD, the specific contribution of the other cardiovascular conditions to in-hospital mortality during sickle cell crisis remains unclear. This study evaluates the independent impact of cardiovascular comorbidities on in-hospital mortality among adults hospitalized for SCC using a large, nationally representative database. Methods: This is a retrospective analysis involving adult hospitalizations (≥ 18 years old) for SCC in the United States (U.S) using 2016-2020 National Inpatient Sample (NIS). The study was exempt from institutional review board approval as the NIS database contains deidentified patient information. ICD-10 codes used for SCC were: D570, D5721, D5741, D5743, D5745, D5781. Regarding cardiovascular co-morbidities, the study considered age, race (African American, Hispanic and Caucasian), dyslipidemia, hypertension, diabetes mellitus (both type 1 and type 2), systolic and diastolic heart failure, rheumatic valvular heart disease, atrial fibrillation/flutter, pulmonary hypertension, chronic coronary artery disease and ischemic stroke. Univariable Cox regression was performed to identify independent predictors associated with in-hospital mortality among patients admitted for SCC. All factors from the univariable analysis with p<0.20 were included in a multivariable Cox regression model. All the outcomes were adjusted for age, sex, and Charlson Comorbidity Index. P-value of ≤0.05 was used to determine statistical significance. Results: The current study analyzed 424,540 SCD patients admitted for SCC (mean age 32.4 years, 55.4% females), from 2016 to 2020. Each additional year of age beyond 18 was associated with a linearly increasing in-hospital mortality among study patients hospitalized for SCC. This observation was confirmed in both univariate (HR: 1.06, 95% CI: 1.05–1.07, p<0.001) and multivariate (HR: 1.05, 95% CI: 1.04–1.06, p<0.001) analyses. On univariable analysis of cardiovascular co-morbidities in patients admitted for SCC, a higher in-hospital mortality was conferred by: hypertension (HR: 2.40; 1.95 – 2.94, p <0.01) dyslipidemia (HR: 2.38; 1.54 – 3.65, p <0.01), diabetes mellitus (HR: 2.15; 1.51 – 3.04, p <0.01, systolic (HR: 5.20; 3.70 – 7.31, p <0.01) and diastolic heart failure (HR: 2.11; 1.44 – 3.08, p <0.01), atrial fibrillation/flutter (HR: 5.43; 4.04 – 7.32, p <0.01), pulmonary hypertension (HR: 2.34; 1.75 – 3.13, p <0.01) and ischemic stroke (HR: 5.75; 3.30 – 10.03, <0.01); a lower in-hospital mortality was associated with the African American race (HR: 0.43; 0.20 – 0.91, p = 0.027). In multivariable analysis of cardiovascular comorbidities that demonstrated nominal significance (p < 0.20) in univariable analysis, only the following remained significantly associated with in-hospital mortality: systolic heart failure (HR: 2.65; 1.77 – 3.98, p <0.01), atrial fibrillation/flutter (HR: 2.01; 1.41 – 2.8), p <0.01) and ischemic stroke (HR: 3.25; 1.85 – 5.71, p <0.01). Although not statistically significant in univariable analysis (p = 0.065), chronic coronary artery disease (CAD) emerged as a significant predictor of reduced in-hospital mortality in the multivariable analysis (HR: 0.52; 0.29–0.91, p = 0.021). Conclusions: Among adults hospitalized for sickle cell crises, select cardiovascular comorbidities, including systolic heart failure, atrial fibrillation/flutter, and ischemic stroke, were independently associated with significantly increased in-hospital mortality. Interestingly, CAD was linked to reduced mortality in adjusted models. This may be attributed to closer medical surveillance and management in patients with established CAD, along with the potential protective mechanism(s) in SCD that limit atherosclerotic progression — warranting further investigation