Citations

Abstract Background: Recent approvals and clinical trials have established covalent BTK inhibitors (BTKis), including ibrutinib, acalabrutinib, and zanubrutinib, as effective agents for the treatment of mantle cell lymphoma. Multiple BTKi-based regimens are recommended by experts and clinical guidelines in both the first-line and later-line settings. Additional approvals of the noncovalent BTKi pirtobrutinib and multiple CAR T-cell therapies have also advanced the management of patients who have progressed on covalent BTKi therapy. In light of these developments in treatment strategies for mantle cell lymphoma (MCL), this study compares trends in treatment choice among experts and healthcare professionals (HCPs) in an online treatment decision support tool. Methods: An online decision support tool, developed with 5 US-based experts in MCL, was launched in April 2023 and updated in April 2025. The tool provided treatment recommendations for various clinical scenarios involving newly diagnosed (ND) or relapsed/refractory (R/R) MCL, incorporating factors such as age, fitness, previous therapies, and response duration. HCPs entered specific patient characteristics along with their intended treatment choice. The tool then displayed expert recommendations for that scenario, and HCPs were asked whether these recommendations influenced their treatment decision. HCP-entered data and expert recommendations were analyzed to identify common trends and divergence in current BTKi therapy use. Results: The 2023 online tool was used by 119 HCPs who entered 148 patient scenarios, and the 2025 online tool was used by 69 HCPs who entered 96 patient scenarios. In 2023, for patients with ND MCL who had later-stage disease or were not eligible for radiation therapy (n = 75), 64% of expert-recommended treatments included BTKis, with the ECHO and TRIANGLE regimens representing 12% and 17% of recommendations, respectively. In 2025, 81% of expert recommendations were BTKi-based regimens, with ≥3 experts recommending a BTKi regimen in 66 of 67 unique ND case scenarios. The ECHO and TRIANGLE regimen each represented 28% to 29% of expert recommendations. In addition, 51% of recommendations for patients with 17p deletions or TP53 mutations consisted of the BOVen regimen. The percentage of HCPs who indicated that they were considering a BTKi-based regimen in the setting of ND MCL was 17% (13 of 75 ND cases) in 2023 and 66% (44 of 67 ND cases) in 2025. The use of both the ECHO and TRIANGLE regimens by HCPs increased from 5% and 4%, respectively, in 2023 to 19% for each in 2025. In 2025, 6% of HCPs considered the BOVen regimen for patients with TP53 mutation (n = 17). In 2023 and 2025, approximately 50% of the HCP-entered patient cases for previously treated MCL had not received a BTKi (n = 48 and n = 18, respectively). For these scenarios, 94% and 98% of expert recommendations consisted of covalent BTKis in 2023 and 2025, respectively. HCPs selected covalent BTKis for 68% of cases in 2023 and 89% of cases in 2025. For patients with R/R MCL previously treated with a BTKi, CAR T-cell therapy represented 36% of expert recommendations in 2023 (23 scenarios) and 60% in 2025 (18 scenarios). By contrast, 4% and 11% of HCPs considered CAR T-cell therapy for these patients in 2023 and 2025, respectively. For HCPs reporting on the tool's clinical impact (n = 96), 35% indicated expert recommendations changed their intended therapy, 41% stated their treatment plan matched the recommendations, 10% cited barriers to implementing recommendations, and 14% remained undecided on optimal treatment. Conclusions:This analysis highlights the ongoing adoption of expert-recommended treatments into real-world clinical practice. Our data suggest that expert and HCP use of BTKi therapy in the ND MCL setting has increased in recent years. In R/R MCL, CAR T-cell therapy appears to be considered infrequently by HCPs. Barriers to implementation of expert recommendations warrant further investigation. Enhanced decision support tools and targeted educational initiatives could help increase the use of expert treatment recommendations and potentially improve patient outcomes.

Abstract Introduction: The prognosis of patients (pts) with histological transformation (HT) of Waldenström macroglobulinemia (WM) remains dismal with standard chemoimmunotherapy. CD19-targeted chimeric antigen receptor T-cell therapies (CAR-T) can lead to durable responses in relapsed/refractory (R/R) large B-cell lymphoma (LBCL) and are approved in the second or third line settings. Pts with HT-WM were largely excluded from CAR-T trials. We previously reported high response rates in a series of 23 pts. The aim of the present study was to evaluate a larger cohort of pts with R/R HT-WM treated with CAR-T with a longer follow-up. Methods: We performed a multicenter retrospective study of pts with biopsy-proven transformed WM/lymphoplasmacytic lymphoma treated with CAR-T in the French DESCAR-T registry (NCT04328298), 4 centers from ECWM and 4 US centers. This study includes updated data from previously published cases (Durot et al, Blood, 2024). Responses were assessed according to the Lugano 2014 criteria. Cytokine release syndrome (CRS) and immune effector cell-associated neurotoxicity syndrome (ICANS) were graded according to the ASTCT 2019 consensus criteria. Hematological toxicity and infections were graded according to the NCI CTCAE (version 5.0). Progression-free survival (PFS) and overall survival (OS) were calculated from the date of CAR-T infusion and estimated using the Kaplan-Meier method. Results: 50 pts were included (39 from DESCAR-T centers, 7 from US centers and 4 from ECWM centers). Median age was 60 (range: 32-72) and 68 (range: 42-82) years at WM diagnosis and CAR-T infusion, respectively. MYD88L265P and CXCR4 mutations were present in 79% (23/29) and 15% (2/13) of patients with available data, respectively. The median time from WM to HT diagnosis was 5.3 years (range: 0-32). Median prior lines of therapy was 1 (range: 0-9) for WM and 2 for HT (range: 0-4). Twenty (40%) pts previously received a BTK inhibitor for WM or HT, 12 (24%) underwent prior autologous SCT and 1 pt allogeneic SCT. 40% of pts presented primary refractory disease and 44% were refractory to last treatment before CAR-T. 40 pts (80%) received bridging therapy, of whom 17 (44%) responded (7 complete responses (CR)). 35 (70%), 10 (20%), and 5 (10%) pts received axicabtagene ciloleucel (axi-cel), tisagenlecleucel (tisa-cel), and lisocabtagene maraleucel (liso-cel), respectively. The best overall response rate (ORR) was 94%, with 44 (88%) pts reaching CR. With a median follow-up of 25.7 months (range: 1-69) from CAR-T infusion, the median PFS was 1.2 years (95% CI: 0.5-4.5) and median OS was 4.2 years (95% CI: 1.2-not reached). Median duration of response was 1.5 years (95% CI: 0.4-4.4), with a median of 4 years (95% CI: 1.5-NR) for pts still in CR at 6 months after CAR-T. 22 pts experienced relapse/progression after CAR-T, 19 related to LBCL and 3 to WM. Median survival after relapse was 7.9 months (95% CI: 3.8-25.3). At LBCL relapse after CAR-T, salvage therapy consisted of CD20xCD3 bispecific antibodies for 9 pts, lenalidomide +/- anti-CD20 or anti-CD19 antibody for 6 pts, and BTKi for 3 pts. Among 19 pts who died, 14 (74%) were due to progressive disease and 5 (26%) from infections. 37 (76%) pts experienced CRS, including 2 (4%) grade 3 events, and 22 (45%) had ICANS, with 2 (4%) ≥ grade 3 events. 17 (35%) pts presented infections, including 12 ≥ grade 3. 21 (43%) pts developed ≥ grade 3 prolonged cytopenias (ie, not resolved within 30 days post infusion). CNS involvement due to LBCL was present in 11 pts (22%) at the time of CAR-T. 6 of them had ICANS, with 1 grade 4. The best ORR was 91% (best CR 91%). Only 1 patient did not obtain a response, with stable disease at 1 month and progression at 3 months. The median PFS was 1.2 year (95% CI: 0.2-NR) and median OS was NR (95% CI: 0.5-NR). On univariate analysis, only elevated LDH was correlated with worse PFS (HR 2.5, p=0.04) and OS (HR 3.3, p=0.04). Other variables, including number of prior lines of therapy, MYD88 mutation status, primary refractory disease, refractory to last treatment, CNS involvement before CAR-T, response after bridging therapy and type of CAR-T received, were not associated with inferior PFS or OS. Conclusion: This cohort represents the largest series of pts with transformed WM treated with anti-CD19 CAR-T. We confirm the high efficacy previously reported with an acceptable safety profile, including for pts with CNS involvement due to LBCL at CAR-T infusion.

Abstract Introduction: Reporting patient outcomes by CAR-T treatment centers is part of the accreditation requirement under the FACT (Foundation for Accreditation of Cell Therapy) Immune Effector Cell (IEC) Program standard. Currently, this is a time-consuming and manual process, making it increasingly difficult to sustain with the growing CAR-T applications and patient volume. The objective of this study is to evaluate how a Large Language Model (LLM) can be used to automate the abstraction and summarization of medical record for CAR-T toxicities. Methods: Clinical data from patients treated with CAR-T between January 2018 and June 2025, including hematology clinical notes, lab results, flowsheets, vital signs, and medication administration record, were used as the source of patient information. The American Society of Transplant and Cell Therapy (ASTCT) CAR-T toxicity grading guideline for cytokine release syndrome (CRS) and immune effector cell associated neurologic syndrome (ICANS) were used to define these toxicities. With the large context window of the LLM (Gemini 2.5 pro), we created a prompt by embedding both the patient data and the ASTCT guidelines for CRS and ICANS, along with an outline of specific information we aimed to abstract from the data. Additionally, the LLM was instructed to provide contextual justification for the data it extracted from patient information. The LLM-generated output included event onset date, resolution date, maximum grade, date of maximum grade, and medications given for management. The CRS and ICANS outcome reports generated by the LLM were then compared to the IEC compliance program data. Any discrepancies identified were reviewed manually. Results: Among patients (pt) treated with CAR-T in the specified timeframe, 47 were selected as training cohort. This cohort included pt with toxicities events, no events, and a mix of events of varying grades. For the test cohort, 66 pts who received CAR-T from March to June 2025 were included. Demographics for the training and test cohorts are reflective of evolving clinical practice (training, test cohorts: median age 69, 65; male 55%, 74%; lymphoma/leukemia 66%, 48%; multiple myeloma 34%, 52%; CD19 CAR-T 46%, 50%). Compared to data from the IEC compliance program, the LLM-generated summaries for CRS events demonstrated an accuracy of 90%, precision 95%, sensitivity 100%, specificity 80%, F1 score 97%, and Matthew's correlation coefficient 87%. For ICANS event, the LLM achieved an accuracy of 91%, precision 92%, sensitivity 92%, specificity 90%, F1 score 92%, Matthew's correlation coefficient 83%. Maximum grade, for both CRS and ICANS, were identified with >90% accuracy. However, dates for CRS and ICANS were less accurate, with the lowest accuracy noted in identifying the event resolution date, 45% for CRS and 67% for ICANS. Medical records were reviewed for all discrepancies between the LLM and the IEC Compliance Program data. The most common discrepancy was a mismatch in event resolution date, determined to be due to differing criteria used by the compliance team and the LLM prompt. Of note, a few discrepancies were attributed to missed information by the IEC Compliance Team, which was correctly identified and categorised by the LLM. An updated LLM prompt (LLMv2) was developed and applied to the test cohort. In version 2, the prompt was revised to include clearer instructions on how to summarize CRS and ICANS events, as well as new instructions on additional data elements to be summarized. The performance of LLMv2 on this cohort for CRS event identification was 100% across all categories (accuracy, precision, sensitivity, specificity, F1 score, and Matthew's correlation coefficient). For ICANS the LLMv2 achieved accuracy of 96%, precision 69%, sensitivity 100%, specificity 93%, F1 score 82%, Matthew's correlation coefficient 80%. Accuracy in identifying event resolution date improved significantly from 64% to 99%. Conclusions: Our study demonstrates that LLMs can significantly reduce the time required to abstract and summarize CAR-T toxicities, while maintaining high accuracy. One limitation of our study is that it is based on data from a single institution. However, plans are already underway to evaluate the use of LLMs in a multi-institutional setting.

Abstract Background: RUNX1 mutations are established high-risk features in acute myeloid leukemia (AML), associated with poor response to induction therapy and reduced overall survival (OS). Current prognostic models, including the ELN 2022 classification, incorporate RUNX1 mutation status but are based largely on cohorts of European ancestry. This raises concerns about their generalizability to underrepresented populations, particularly African American patients, who continue to experience worse outcomes in AML across treatment settings. Methods: We analyzed publicly available datasets, including TCGA-LAML, Beat AML, and a recent multi-omic study of 100 African American AML patients (Stiff et al., Nat Genet 2024). We evaluated RUNX1 mutation frequency, OS outcomes, and co-mutation patterns, stratified by race. To contextualize survival differences by ancestry, we performed a survival analysis of adult AML patients using the SEER 17 registries (2010–2020), accessed through SEER*Stat. Kaplan-Meier estimates and median OS were compared by race using disease-specific survival. Results: RUNX1 mutations were present in approximately 10–15% of AML cases and consistently associated with inferior OS (median 16–24 months) in both de novo and AML-MRC subsets. Co-mutations in DNMT3A and TET2 further worsened the prognosis. In the African American cohort described by Stiff et al., the frequency of RUNX1 mutations was similar to that in White patients; however, African American patients showed enrichment for IDH2, RAS, and WT1 mutations and lower incidence of NPM1 mutations. Despite comparable mutation rates, African American patients demonstrated significantly worse OS, even under standardized treatment. SEER 17 survival analysis of 5,932 AML cases (683 African American) confirmed shorter median OS in African American patients compared to non-African American patients (10.4 vs. 15.7 months, p<0.01).Conclusions: RUNX1 mutations confer adverse prognosis in AML, but current risk models may not adequately reflect risk in African American patients due to differing co-mutation patterns and underrepresentation in genomic datasets. These findings highlight the need for ancestry-aware risk stratification and inclusive genomic profiling to reduce disparities in AML outcomes. References: Lew-Derivry et al. Leukemia. 2023 Jalili et al. PMCID: PMC5980915 Stiff et al. Nat Genet. 2024 Bhatnagar et al. PMCID: PMC7933110 Tang et al. Blood. 2009 Mendler et al. JCO. 2012 He et al. Scand J Clin Lab Invest. 2020 Döhner et al. Blood. 2022. PMID: 35926292

Abstract Background: Monoclonal B-cell lymphocytosis (MBL), a precursor to chronic lymphocytic leukemia (CLL), is an understudied yet common premalignant condition present in ~8 million adults in the United States. Based on flow cytometry, MBL can be classified into two immunophenotypes: CLL-type (CD5, CD23, and CD20 co-expression) and non-CLL-type (CD5-negative). While the majority of the data to date pertains to CLL-type MBL, the available evidence indicates non-CLL-type MBL is also associated with an increased risk of hematologic malignancies. Clonal hematopoiesis of indeterminate potential (CHIP) is another precursor condition associated with hematologic malignancies. The relationship between non-CLL-type MBL and CHIP and their combined effect remains unknown. Herein, we report findings on these two precursor conditions in the largest screening cohort to date. Methods: Using flow cytometry, we screened for MBL in the Mayo Clinic Biobank participants who were 40 years or older with no prior history of hematologic cancers. Using whole-exome sequencing of whole-blood DNA, CHIP was screened on 291 genes related to myeloid (56 genes) or lymphoid malignancies (235 genes). Incident hematologic malignancies were detected during follow-up using International Classification of Diseases (ICD) codes in the medical record. All diagnoses were manually reviewed for confirmation. Multivariable logistic regression was used to evaluate the association between non-CLL-type MBL and CHIP. Time to incident hematologic malignancy was calculated from date of MBL screening to the earliest date of incident malignancy, withdrawal, death, last medical visit, or study end. Overall survival (OS) was calculated from date of MBL screening to date of death or last known alive. Multivariable Cox regression, adjusted for age, sex and residency, was used to investigate associations between non-CLL-type MBL and CHIP with clinical outcomes; with results reported using hazard ratios (HR) and 95% confidence intervals (CI). Results: In total, 8,802 individuals with MBL screening and CHIP data were analyzed. Non-CLL-type MBL was detected in 249 (2.8%) individuals; 121 (49%) were male, and median age was 78 years (range, 47-95). We identified 757 (8.6%) individuals with CHIP in at least one of the 291 genes tested. The most frequently mutated CHIP genes were DNMT3A, TET2, ASXL1, PPM1D, and TP53. A total of 47 (0.5%) individuals had both non-CLL-type MBL and CHIP; the four genes with the highest number of CHIP variants in these individuals were DNMT3A (n=16), TET2 (n=10), ASXL1 (n=3), and ATM (n=3). An association between non-CLL-type MBL and any CHIP (OR=1.6, 95%CI:1.2-2.3, P=0.006) was observed. When analyzing CHIP in genes related to myeloid malignancies, the association with non-CLL-type MBL was also significant (OR=1.6, 95%CI:1.1-2.3, P=0.012); however, the association was not statistically significant for genes related to lymphoid malignancies (OR=1.6, 95%CI:0.8-3.3, P=0.165). With a median follow-up of 5.4 years, 104 individuals developed incident hematologic malignancies (66 lymphoid and 38 myeloid malignancies). Non-CLL-type MBL (HR=3.5, 95%CI:1.9-6.5, P<0.001) and CHIP (HR=3.2, 95%CI:2.1-5.1, P<0.001) were each associated with increased risk of incident hematologic malignancies. Co-existence of both precursor conditions was associated with a substantially higher risk of incident hematologic malignancies (HR=10.3, 95%CI:4.0-26.2, P<0.001) compared to individuals with neither precursor condition. Among individuals with non-CLL-like MBL, any CHIP was a borderline significant predictor for incident hematologic malignancy (HR=3.6, 95%CI:1.0-12.9, P=0.049). Compared to individuals with no precursor conditions, OS was not significantly different in those with non-CLL-type MBL-only (HR=1.1, 95%CI:0.8-1.5, P=0.40), those with CHIP-only (HR=1.2, 95%CI:0.99-1.4, P=0.06), or those with both CHIP and non-CLL-type MBL (HR=1.3, 95%CI:0.8-2.1, P=0.39). Conclusions: In the largest screening cohort to date evaluating the relationship of CHIP and non-CLL-type MBL, we found that non-CLL-type MBL and CHIP are associated with each other. Individuals with non-CLL-type MBL were more likely to have CHIP variants in DNMT3A, TET2, ASXL1, and ATM. Among individuals with non-CLL-type MBL, the presence of any CHIP was associated with increased risk of developing hematologic malignancy.

Abstract Background: Despite recent therapeutic advances in AS cHL, PET-adaptive chemotherapy-based regimens remain a first-line treatment option, especially for patients (pts) without access to novel agents. Most PET-adaptive regiments start with either 2 cycles of escBEACOPP (escBEACOPP2) or ABVD (ABVD2) prior to interim PET assessment (iPET), followed by treatment de-escalation or intensification by iPET result. There have been no direct comparisons of escBEACOPP2 vs ABVD2 PET-adaptive regimens. Additionally, the AS-Hodgkin Lymphoma International Prognostication Index (A-HIPI) score (Rodday JCO 2023) has not been fully explored in escBEACOPP2 regimens. Therefore, we utilized multiple analytic approaches to compare outcomes of PET-adaptive escBEACOPP2 and ABVD2 regimens, while adjusting for and assessing the effect of the baseline A-HIPI score. Methods: Through the global HoLISTIC Consortium (www.hodgkinconsortium.com), we obtained individual patient data from 4 AS clinical trials of newly diagnosed cHL pts treated with PET-adaptive escBEACOPP2 (AHL2011) or ABVD2 (SWOG0816, RATHL, HD0607) regimens. Pts were restricted to age 18 to 65 years (y), stage IIB-IV, and PET-adapted treatment arms. Positive iPET was based on Deauville score >3. Treatment effects of escBEACOPP2 vs ABVD2 (reference) on 5y progression-free survival (PFS) and overall survival (OS) were first evaluated using Cox models. Importantly, baseline disease risk was adjusted for using the 5y PFS or OS A-HIPI score (including stage, sex, age, bulk, lymphocyte count, albumin, white blood cell count), where higher scores indicate higher predicted risk (scale 1-100). The A-HIPI was modeled per 1 standard deviation (SD) increase. Our multistate model (MSM) comprised 4 health states: diagnosis, sustained remission at 1yr, treatment failure, and death from any cause. Pts all started in the diagnosis state and transitioned to other states without return to prior states. We assessed the impact of treatment on transitions (except binary sustained remission at 1y) that involved ≥5 pts per treatment group, adjusting for A-HIPI for all transitions. All effects are reported as adjusted hazard ratios (aHRs) with 95% confidence intervals (CI). Results: 2,381 AS cHL pts were included (n=372 for escBEACOPP2, n=2009 for ABVD2). For escBEACOPP2 and ABVD2 groups (median follow-up 60 and 56 months), respectively, the mean baseline A-HIPI scores (i.e., predicted 5y PFS event rate) were 25.1 (SD=6.3) and 22.9 (SD=6.5), the rates of positive iPET were 11.3% and 17.0%, 5y PFS was 87.0% and 79.5%, and 5y OS was 96.3% and 94.6%. In Cox models, escBEACOPP2 was associated with significantly improved PFS compared to ABVD2 (aHR=0.54, 95% CI=0.39-0.73); the aHR for OS was 0.67 (95% CI=0.37-1.19). Notably, higher risk A-HIPI score was associated with significantly worse PFS (aHR=1.42, 95% CI=1.31-1.54) and OS (aHR=1.60, 95% CI=1.47-1.76), independent of treatment. In the MSM, the effect of escBEACOPP2 compared to ABVD2 regimens was modeled in 3 transitions with adjustment for A-HIPI score: those treated with escBEACOPP2 had lower likelihood of treatment failure with (aHR=0.38, 95% CI=0.22-0.66) or without (aHR=0.61, 95% CI=0.40-0.92) sustained remission at 1y; the aHR for treatment failure to death was 0.55 (95% CI=0.25-1.22). Higher risk baseline A-HIPI score was associated with higher likelihood of treatment failure and death without sustained remission at 1y (aHR=1.43, 95% CI=1.28-1.61 and aHR=2.50, 95% CI=1.80-3.49, respectively). Furthermore, higher risk baseline A-HIPI score was associated with higher likelihood of treatment failure after achieving sustained remission at 1y (aHR=1.29, 95% CI=1.12-1.48) as well as death after treatment failure (aHR=1.54, 95% CI=1.27-1.86). Conclusions: Comparing newly diagnosed adult AS cHL PET-adaptive regimens with adjustment for baseline A-HIPI score, we found pts treated with escBEACOPP2 on AHL2011 had improved PFS versus pooled pt data from 3 ABVD2 trials. Using MSM, we further demonstrated that escBEACOPP2 was associated with a lower likelihood of treatment failure with or without achieving sustained remission at 1yr. In addition, baseline A-HIPI score was associated with PFS and OS, which was independent of treatment. Finally, higher baseline A-HIPI scores maintained prognostic impact throughout the disease course, including increased risk of treatment failure for pts in remission at 1yr and death after treatment failure.

Abstract Introduction: Ciltacabtagene autoleucel (cilta-cel) is a highly effective BCMA-redirected CAR T-cell therapy for RRMM. While the pivotal CARTITUDE-1 trial established a target dose of 0.75 x 10^6 CAR+ viable T-cells/kg, the approved cell dose is 0.5-1.0 x 10^6 and the optimal dosing strategy to balance efficacy and toxicity remains unclear. This question is of interest given that a dose-response association has been demonstrated for idecabtagene vicleucel (a related BCMA-directed CAR T product). It has also been hypothesized that lower doses of cilta-cel may be associated with reduced toxicity. This study aims to evaluate the association between cilta-cel dose and key efficacy and toxicity outcomes using a large, real-world dataset from the US Multiple Myeloma Immunotherapy Consortium. Methods: We conducted a multicenter, retrospective analysis of 751 RRMM patients treated with commercial cilta-cel across 15 US academic centers. The administered cilta-cel dose (million cells/kg) was categorized into four groups: low (<0.5), standard-low (0.5 to <0.6), standard (0.6 to <0.8), and standard-high (0.8 to 1.0). We compared incidence rates of cytokine release syndrome (CRS), immune effector cell-associated neurotoxicity syndrome (ICANS), delayed neurotoxicity (DNT), and immune effector cell-associated hemophagocytic lymphohistiocytosis-like syndrome (IEC-HS) across these dose categories using logistic regression. Best response rates and progression-free survival (PFS), estimated using the Kaplan-Meier method, were also compared, with differences assessed using the log-rank test for the latter. Results presented are derived from multivariable models adjusting for extramedullary disease, ECOG status, penta-refractory disease and receipt of bridging therapy. Results: Among 751 infused patients, the dose distribution was: <0.5 (n=33), 0.5 to <0.6 (n=131), 0.6 to <0.8 (n=550), and 0.8 to 1.0 (n=37). The four dose groups were well-balanced in terms of baseline patient and disease characteristics, including age, high-risk cytogenetics, and extramedullary disease, with the exception of t(14;16), which was more common in the standard-low and standard dose groups (p=0.006). There were no statistically significant differences in the rates of any grade CRS (p=0.11), any grade ICANS (p=0.84), grade ≥3 ICANS (p=0.36), duration of CRS (p=0.18) or ICANS (p=0.64), or IEC-HS (p=0.16) across the four dose categories. Time to onset of CRS was shorter for the standard-high dose compared to the low dose (6 days vs. 8 days; p=0.08). Importantly, there was no significant difference in the rate of DNT (3.0%, 9.2%, 8.8%, 5.4%; p=0.24). Similarly, there was no significant difference in overall response rates (97%, 96%, 96%, 88%; p=0.50) or complete response rates (75%, 65%, 75%, 66%; p=0.10) by dose. The median follow-up was 12.3 months (IQR: 6.9, 20.5). For patients who received a standard-low dose (0.5, 0.6), the median PFS was 17.5 (IQR: 13.5, NR) months. For the standard dose range of (0.6, 0.8), the median PFS was 28.7 (95% CI: 19.8, NR) months, while the median PFS for the standard-high dose (0.8, 1.0) was not reached. For the low dose range of <0.5, the median follow up was only 8.7 months, which did not allow reliable estimation of the median PFS. While the overall difference in PFS among the groups did not reach statistical significance by the log rank test, there was a clear trend toward improved PFS with higher cell doses in the multivariable model, with outcomes clearly ordered in the same direction as the cell dose categories. A similar trend was also evident in the rates of 12-month PFS which showed an improvement with increasing dose (62% ,67%, 73%, 80%). Conclusions: In this large, real-world cohort of RRMM patients treated with cilta-cel, a lower cell dose was not associated with reduced incidence or severity of CRS, ICANS, DNT, or IEC-HS. Conversely, a dose-response correlation with PFS was observed, with higher doses trending toward longer median and 12-month PFS. It remains unknown if a lower cell dose in the commercial setting reflects poor T cell quality, yet these findings suggest that dose reduction of cilta-cel may not be an effective strategy to mitigate toxicities and might be associated with a trend toward lower efficacy. Therefore, the standard target dose should be utilized to maximize patient outcomes and further prospective dose exploration is warranted.

Abstract Introduction: The GPRC5D-directed bispecific antibody (BsAb) Talquetamab (Talq) has shown promising results for patients (pts) with relapsed/refractory multiple myeloma (RRMM) in the MonumenTAL-1 clinical trial (Chari et al. NEJM 2022). However, there are limited data on the use of Talq in the real-world setting. Methods: All pts with RRMM who started Talq between 8/2023-3/2025 across 15 US academic centers were included. High-risk cytogenetic abnormalities (HRCA) were defined as del(17p), t(4;14), t(14;16) and/or gain/amplification 1q21. Time-to-event analyses were estimated using the Kaplan-Meier method. Toxicities including CRS and ICANS were graded using CTCAE v5.0 or ASTCT criteria. Results: A total of 484 pts were included, with a median age of 65 years (range 35-89), and 83 pts (17%) were aged ≥75 years. 20% of pts were black. 21% had baseline CrCl<40, 65% had HRCA and 29% had >1 HRCA (i.e Ultra high-risk MM). The median number of prior lines of therapy was 6 (IQR 4, 8). 50% were penta-refractory, 69% received at least one prior BCMA-directed therapy (46% CART; 32% BsAb). Prior to initiation of Talq, 11 pts (2%) had CNS MM, 29 (6%) had PCL and 126 (26%) had true extramedullary disease (EMD). 70% would have been ineligible for the MonumenTAL-1 trial. For most pts (n=405; 84%) the dosing strategy from Cycle 2 was 0.8 mg/kg every 2 weeks. and the 0.4 mg/Kg weekly dosing was used in 28 patients (6%). In 6% of pts, an additional drug was added to Talq, mostly pomalidomide (n=20) or daratumumab (n=10). 101 pts (21%) received Talq as bridging for CART. Median duration of Talq treatment was 102 days. 62% developed CRS (Grade ≥3, 2%; including one Grade 5 event). 15% developed ICANS (Grades 3-4, 4%). Steroids were used in 213 patients (44%), tocilizumab in 145 (30%) and anakinra in 9 (2%). 22 patients (5%) required ICU stay during the first cycle, for a median of 6 days (range 1-31). A total of 200 pts (41%) had infections (median 1 infection each, range 1-6), including 99 (20%) with bacterial infections, 113 (23%) viral infections and 21 (4%) fungal infections. IVIG was used in 261 pts (54%). 79% of patients (n=384) developed Talq-related oral toxicity, 55% (n=269) skin toxicity and 199 (41%) nail toxicity. 157 patients died after a median follow up of 11.5 months (mos), and the leading causes of death were MM progression (n=120) and infection (n=10). 8 pts (1.7%) developed invasive second primary malignancies. The ORR was 73%, including 46% ≥VGPR. The median PFS (mPFS) for the entire cohort was 7.8 mos (95% CI 6.9-9.4) and the 12 months PFS rate was 39%. The median OS had not been reached and the 12 months OS rate was 63%. For pts who would have been eligible for MonumenTAL-1, the mPFS was 12.2 mos (95% CI 9.7 – not reached (NR); for trial-ineligible pts, the mPFS was 6.9 mos (5.8 – 8.0; p=0.003). For BCMA-naive pts, the mPFS was 15.7 mos (11.8-NR); for BCMA-exposed pts, the mPFS was 6.8 mos (6.0-7.8; p=0.0002). For patients who received Talq as bridging prior to CART, the median PFS was NR (14.1 – NR); for those who did not receive Talq as bridging the mPFS was 6.5 mos (5.3-7.6; p<0.0001). In multivariable analyses, worse OS and PFS were associated with trial ineligibility [(HR 1.38, 95% CI 1.03-1.86; p=0.03) and (2.33, 1.31-4.14; p=0.004)], BMI<30 [(1.42,1.06-1.89; p=0.02) and (1.50, 1.01-2.22; p=0.045)] and true EMD at baseline [(1.95, 1.49-2.56; p<0.001) and (1.65, 1.16-2.36; p=0.005). On the other hand, use of Talq as bridging to CART was associated with better PFS (0.27, 0.17-0.43; p<0.001) and OS (0.13, 0.05-2.56; p<0.001). Ultra high-risk MM (1.40, 1.06-1.84; p=0.02) and baseline albumin <3 (1.68, 1.18-2.37; p=0.004) were associated with worse PFS; ECOG ≥2 (1.68, 1.18-2.40; p=0.004), and baseline Plt<50K (2.14, 1.48-3.10; p<0.0001) was associated with worse OS. Conclusions: In this large real-world analysis of 484 Talq recipients, the ORR was comparable to MonumenTAL-1 but the median PFS was shorter. This likely reflects differences in patient selection, with 70% of patients in our cohort being trial-ineligible. Patients who received Talq as pre-CAR-T bridging had superior survival outcomes. Rates of toxicities were similar to MonumenTAL-1, including >70% of patients developing dysgeusia.

Abstract Introduction: Elranatamab, a BCMA-CD3 bispecific antibody, has shown efficacy in relapsed multiple myeloma (MM), but real-world data on its safety and effectiveness are limited. Methods: We conducted a retrospective, multi-center study of patients (pts) treated with commercial elranatamab from 8/2023–3/2025 at 9 U.S. academic centers. Logistic and Cox regression were used for binary and time-to-event outcomes, respectively. Multivariable analysis (MVA) adjusted for prior lines of therapy (LOTs), baseline Hgb (per g/dL), and LDH (per 100 U/L) as continuous variables; and for BCMA exposure, extramedullary disease (EMD; extraosseous soft tissue lesions), and ECOG ≥2 as binary variables. Results: Of 130 pts, median age was 71 yrs (range 39–95); 57% were female, 82% white, 13% black, and 35% had ECOG ≥2 (11% ≥3). Median LOTs was 6; 91% were triple-class refractory, 49% penta-refractory, 49% BCMA-exposed (median 15.5 mo [IQR 11.8–20.5] since last exposure), and 49% had prior anti-BCMA CAR-T. Most recent BCMA exposures included: CAR-T (84%), TCE (9.4%), and ADC (6.3%). Only 22% met MagnetisMM-3 cohort A eligibility (which excluded BCMA-exposed pts). High-risk cytogenetics (del(17p), t(4;14), t(14;16), or amp(1q)) were present in 44% (25%, 13%, 6.3%, and 7.9%, respectively). EMD was seen in 22%. Median treatment duration was 4.0 mo (95% CI 2.93–6.3) overall and 7.9 mo (95% CI 5.83–NA) with ≥VGPR. Discontinuation due to toxicity or non-relapse mortality occurred in 12% but was significantly lower among those receiving IVIg (5% vs. 17%, p=0.031); IVIg was given in 47%. The overall response rate (ORR) was 65%, including ≥VGPR in 46% and ≥CR in 36%, consistent with MagnetisMM-3 (ORR 61%, ≥CR 35%; Lesokhin et al. Nat Med. 2023). After median follow-up of 7.5 mo, median PFS was 4.3 mo (95% CI 3.4–11.3), OS was 14.6 mo (95% CI 8.6–NA), and DOR was 12.2 mo (95% CI 6.1–NA). In contrast, MagnetisMM-3 (Prince et al., ASH 2024) reported median PFS and OS of 17.2 and 24.6 mo, respectively, with DOR not reached despite 28.4 mo of follow-up. In univariate analysis, ECOG ≥2 was associated with lower ORR (45% vs. 71%; OR 0.34, 95% CI 0.16–0.73, p=0.006), shorter OS (4.3 mo vs. NA; HR 2.41, 95% CI 1.33–4.35, p=0.004), and shorter PFS (1.7 vs. 6.5 mo; HR 1.77, 95% CI 1.09–2.87, p=0.020). Prior BCMA-directed therapy was linked to reduced odds of ≥VGPR (OR 0.38, 95% CI 0.19–0.77, p=0.008) and ≥CR (OR 0.43, 95% CI 0.20–0.89, p=0.025). Higher baseline LDH was linked to shorter OS (HR 1.37, 95% CI 1.23–1.53, p<0.001) and PFS (HR 1.31, 95% CI 1.18–1.44, p<0.001). Higher baseline Hgb was associated with higher ORR (OR 1.46, 95% CI 1.20–1.81, p<0.001) and ≥CR (OR 1.54, 95% CI 1.26–1.92, p<0.001). Among BCMA-exposed pts, recent exposure (<1 [n=15] vs. ≥1 yr [n=32]) was associated with shorter OS (4.1 mo vs. NA; HR 3.66, 95% CI 1.26–10.6, p=0.017). In MVA, higher baseline LDH remained independently associated with shorter OS (aHR 1.36, 95% CI 1.19–1.54, p<0.001) and PFS (aHR 1.27, 95% CI 1.13–1.43, p<0.001); higher Hgb with greater odds of ≥CR (aOR 1.44, 95% CI 1.14–1.85, p=0.003) and longer PFS (aHR 0.84, 95% CI 0.72–0.99, p=0.034); and BCMA exposure with lower odds of ≥CR (aOR 0.32, 95% CI 0.10–0.91, p=0.037). CRS occurred in 40% (grade ≥2 in 12.3%, grade 3 in 2.3%), and ICANS in 17% (grade ≥2 in 7.7%). These rates contrast with MagnetisMM-3 (no grade ≥3 CRS, 3.4% ICANS). ICANS risk was lower in BCMA-exposed pts (9.4% vs. 24%; OR 0.32, 95% CI 0.11–0.85, p=0.029), and decreased with more LOTs (OR 0.73, 95% CI 0.55–0.92, p=0.016) and higher Hgb (OR 0.71, 95% CI 0.52–0.94, p=0.023). Tocilizumab and steroids were used in 36% (treatment 28%, prophylaxis 9.2%) and 23%, respectively. Steroid use was less common in BCMA-exposed pts (11% vs. 34%; OR 0.23, 95% CI 0.08–0.60, p=0.004) and declined with more LOTs (OR 0.68, 95% CI 0.52–0.85, p=0.002). Infections occurred in 40% of the cohort, more common in pts with CrCl <40 mL/min (59% vs. 36%; HR 2.58, 95% CI 1.02–6.80, p=0.049). IVIg was linked to improved infection-free survival (HR 0.59, 95% CI 0.37–0.93, p=0.023). Conclusions: Elranatamab demonstrated meaningful activity in real-world relapsed MM, though PFS and OS were shorter than in MagnetisMM-3, likely reflecting an older, frailer population with higher disease burden. ICANS was more frequent than in MagnetisMM-3. Baseline anemia and elevated LDH, markers of poor marrow reserve and higher disease burden, were strongly linked to worse outcomes.

Abstract PAX5 is genetically altered in over 30% of B-ALL cases, with point mutations representing the second most common type of alteration. These point mutations are highly enriched in the DNA binding (paired) domain and have been recognized as founder events in PAX5 P80R and PAX5alt B-ALL subtypes, which account for over 12% of B-ALL. However, the molecular features and underlying mechanisms of these mutations in B-ALL remain largely undefined, limiting the development of targeted therapies and resulting in poor outcomes, especially in adult patients. We systematically investigated the functional roles and oncogenic mechanisms of PAX5 paired domain mutations. Integrated CRISPR screening and genomic profiling revealed that the paired domain is critical for PAX5 function. Disruptions in this region significantly reduced cell viability in B-ALL cell lines, and analysis of 2,955 B-ALL samples showed that recurrent PAX5 missense mutations are highly enriched in this domain, further underscoring its functional importance. Functional analysis of knock-in mice with PAX5 mutations revealed that R38H and R140L severely impair DNA binding and block B-cell differentiation, while P34L and P80R retain binding ability but alter binding specificity and gene regulation. Mice carrying homozygous (P34L, R38H, P80R, R140L) or heterozygous (R38H, P80R) mutations developed spontaneous B-ALL within one year. To model co-occurring mutations observed in patients, Pax5 R38H/P34L and R38H/R140L mice were generated. These compound mutations exhibited greater leukemogenic potential than single mutants. Transcriptomic analysis of leukemia samples showed consistent up-regulation of metabolic pathways and Myc target genes. Notably, biallelic PAX5 alterations and additional genetic lesions in signaling pathways were detected in all PAX5R38H/+ and PAX5P80R/+ leukemias, faithfully recapitulating the genetic alteration features of patient B-ALL driven by PAX5 mutations. Among Pax5-mutant mouse strains, Pax5P80R/+ mice develop B-ALL with short latency and complete penetrance, making them an ideal model to study the mutational evolution and leukemogenesis of B-ALL. To characterize the stepwise leukemogenesis process, we performed serial bone marrow aspirations in the same Pax5P80R/+ mice. In all 17 mice analyzed, we identified early preleukemic clones with only Pax5 WT deletions, followed by the emergence of leukemic clones harboring Jak1/3 mutations. Single-cell RNA/BCR-seq revealed that Pax5 WT allele deletions relieve metabolic constraints, thereby promoting clonal expansion of preleukemic B cells blocked at the pre-B stage. These changes coincide with downregulation of metabolic checkpoint genes and upregulation of glycolytic enzymes, promoting genetic instability and facilitating malignant transformation. Clinically, PAX5 P80R patients have better outcomes than PAX5alt cases. Using our mouse models, we observed that leukemic cells driven by the PAX5 P80R mutation exhibit greater sensitivity to dexamethasone compared to the ones carrying other PAX5 mutations in PAX5alt subtype, potentially due to their heightened dependence on glucose uptake and glycolysis, suggesting that metabolic rewiring driven by PAX5 mutants may influence therapeutic response. Interestingly, we found that an aberrant isoform of MEGF10 (aMEGF10) is highly and specifically expressed in PAX5 P80R B-ALL. This N-terminal truncated isoform encodes a protein that retains the intracellular immunoreceptor tyrosine-based activation motifs (ITAMs). Using CUT&Tag, we showed that PAX5 P80R directly binds to the aMEGF10 promoter. By engineering human CD34⁺ cells, we established a humanized PAX5 P80R leukemia model that recapitulates the gene expression profile of patient samples, including robust activation of aMEGF10. Functionally, aMEGF10 binds to and activates SYK, thereby promoting leukemic cell proliferation. Importantly, this oncogenic effect depends on intact ITAM motifs and SYK signaling, revealing a novel targetable vulnerability in PAX5 P80R leukemias. Overall, the study establishes that PAX5 DNA binding domain mutations contribute to leukemogenesis through stepwise loss of PAX5 WT alleles and acquisition of signaling pathway mutations, involving impaired differentiation, metabolic deregulation, and novel oncogene activation. These findings not only deepen the understanding of PAX5-driven leukemias but also highlight aMEGF10–SYK signaling as a promising therapeutic target.