Citations

Background: Multiple Myeloma (MM) is predominantly a disease of older adults. However, older patients vary widely in physiological status and treatment tolerance. The simplified frailty index (Facon 2020) incorporates age, ECOG performance status, and comorbidities (Charleston Comorbidity Index [CCI]) to classify patients as frail or non-frail. While frailty is associated with increased toxicity with conventional systemic therapies, its impact on outcomes of older patients treated with bispecific antibodies (BsAbs) has not been established. Thus, we conducted a retrospective study to evaluate the efficacy and toxicity of BsAbs in older patients based on frailty. Methods: This is a single institution study. We included patients with MM ≥65 years who initiated treatment with a BsAb at Mayo Clinic in Rochester, MN from August, 2018 to January, 2024. Patients who received a BsAb on trial or as standard of care were included. For patients who received 2 different BsAbs at different timepoints, we included the first treatment only. We grouped patients into frail and non-frail based on the simplified frailty index (score ≥2 = frail). We evaluated the overall response (ORR) and ≥very good partial response (VGPR) rates, progression-free (PFS) and overall (OS) survival, early mortality (within 90 days) including early non-relapse mortality (NRM), and toxicities including cytokine release syndrome (CRS), immune effector cell associated neurotoxicity syndrome (ICANS) and healthcare utilization. P values were insignificant (>0.05) unless otherwise stated. PFS and OS were calculated from the day of BsAb start. Results: A total of 99 patients age 65-89 years were included. The products used were teclistamab (N=43), talquetamab (N=2), and investigational agents (N=54); 67% were ≥70 and 29% were ≥75 years; 39% were female, 95% were non-Hispanic White and 4% were Black/African American. Seventy patients (71%) were considered frail and 29 (29%) were non-frail. At the time of BsAb start, a higher proportion of frail patients had platelets <75 x 10^9/L (30% vs 8%, P=0.04). Otherwise, there was no significant difference between frail and non-frail patients in age (median: 72 vs 70), hemoglobin <8 g/dL (19% vs 12%), absolute neutrophil count <1 x 10^9/L (6% vs 8%), presence of high-risk FISH (70% vs 60%), ISS stage 3 (32% vs 41%), high LDH (58% vs 61%), presence of extramedullary disease (39% vs 25%), serum creatinine ≥2 mg/dL (15% vs 8%), bone marrow plasma cell% (20% in both), triple- (83% vs 90%) and penta-class refractoriness (63% vs 48%), number of prior lines of therapy (median: 6 in both), prior exposure to BCMA-directed therapy (24% vs 21%, P=0.80) and prior autologous stem cell transplant (73% vs 86%). Step-up dosing was used in 81% vs 48% of patients (P=0.001), and treatment was initiated in the outpatient setting in 46% vs 28% (P=0.12) in frail and non-frail patients, respectively. The median follow-up was 1.9 (95%CI: 1.0-2.3) years. There was no significant difference between frail and non-frail patients in ORR (52% vs 62%), ≥VGPR (46% vs 48%), 1-year PFS (36% vs 48%, P=0.26), or 1-year OS (57% vs 61%, P=0.30), respectively. Age and CCI alone were not associated with PFS or OS, but ECOG >1 was associated with decreased PFS (hazard ratio: 3.1, P=0.001) and OS (hazard ratio: 3.1, P<0.001). There was no significant difference between frail vs non-frail patients in grade 3/4 thrombocytopenia (39% vs 17%, P=0.06), grade 3/4 neutropenia (27% vs 35%), total days of hospitalization (median: 8 vs 5, P=0.35), number of unplanned admissions in 6 months (median: 1 in both), or total hospital days for unplanned admissions (1 vs 2 days). There was no significant difference between frail and non-frail patients in early mortality (23% vs 10%, P=0.17), including early NRM (4% in both), CRS (59% vs 66%), grade 3-4 CRS (3% vs 7%), or CRS recurrence (21% vs 10%). ICANS occurred in 5 (9%) frail patients and 1 non-frail patient (grade 2); among frail patients, grades were: 1 in 3 patients, 3 in 1 patient, and 4 in 1 patient. Similarly, there was no significant difference in CRS/ICANS between the 2 groups when patients with and without step-up dosing were analyzed separately. Conclusion: Frailty, defined using the simplified frailty index, was not significantly associated with inferior efficacy or increased toxicity with BsAb therapy among older patients with MM. Poor performance status predicted inferior PFS and OS, while age and comorbidity scores did not.

Background: Immunosuppressive therapy (IST) is considered the standard front-line treatment for acquired severe aplastic anemia (SAA) for pts >40 years. For pts aged <40 years who have a suitable human leukocyte antigen (HLA)-matched sibling donor (MSD), hematopoietic stem cell transplantation (HSCT) is recommended. This recommendation is based on the toxicity associated with alternative donors; haploidentical (HAPLO) and unrelated donor (URD) HSCT for patients who lack a MSD. However, the hematopoietic response after IST is ~ 70-80%, even after augmentation with eltrombopag in the front-line setting, and failure-free survival (FFS, defined as patients alive and in remission without clonal disease more than 10 years after IST) remains < 40%. HSCT is often employed as a second line salvage therapy for those who do not respond to IST or have disease recurrence. However, outcomes after HSCT are often compromised because of IST-related complications such as clonal evolution, development of donor specific antibodies, infectious adverse events and organ dysfunction. In contrast, a successful allogeneic HSCT leads to relatively prompt hematopoietic and immune reconstitution and mitigates the risk of secondary clonal disease. Innovations in recent decades, such as improved graft versus host disease (GVHD) prophylaxis and supportive care, have steadily reduced the morbidity and mortality of adults undergoing HSCT from unrelated or mismatched donors and expanded the donor pool such that almost all pts with SAA have a potential donor available. The use of post-transplant cyclophosphamide (PTCy) is now a well-established approach to facilitate engraftment and mitigate the risk of GVHD. It was successfully used in relapsed/refractory SAA pts using HAPLO donors in the multi-center BMT CTN #1502 trial (NCT02918292). Similar PTCy based approaches have been effective in the upfront setting with HAPLO and URD in small single center studies. Opened in July 2024, the CureAA trial (NCT06517641), seeks to increase donor options for SAA pts who have not yet received IST by using PTCy based GVHD prophylaxis for HSCT using HAPLO or URD. Study Design and Methods: This BMT CTN study (#2207) is a prospective multi-center phase II trial to assess efficacy of HSCT using HAPLO or URD bone marrow with a reduced intensity, thymoglobulin (ATG)-containing preparative regimen and PTCy GVHD prophylaxis in previously untreated SAA pts. This trial has two parallel cohorts for each donor type, HAPLO and URD, with an accrual goal of 30 participants in each cohort. All will receive ATG (0.5mg/kg on day -9 and 2mg/kg on days -8 and -7), cyclophosphamide (14.5mg/kg on days -6 and -5), fludarabine (30mg/m2 on days -6 to -2), 400 cGy total body irradiation (day -1 in a single fraction), a marrow graft (goal >2.5 x 108 nucleated cells per kg of recipient) on day 0 and PTCy (50mg/kg on days +3 and +4). Additionally, participants receive tacrolimus (until day +180) and mycophenolate mofetil days +5-35. The primary objective of the trial is to estimate the GVHD-free failure-free survival (GFFS) rate at 1 year after initiation of conditioning. GFFS is a composite endpoint of survival without grade III-IV acute GVHD, chronic GVHD requiring immunosuppression, or primary or secondary graft failure requiring second definitive therapy within 1 year after initiation of conditioning. Secondary objectives will evaluate patient survival, engraftment, transfusion dependence, cumulative incidence of GVHD and hematologic response. Health related quality of life, healthcare utilization and financial toxicity data will also be collected. These endpoints will allow clinical comparison with current results achieved with upfront IST in other studies. We hypothesize that this HSCT approach will improve survival without significant transplant-associated morbidity by using a low toxicity regimen that reduces rates of GVHD and provides rapid hematologic recovery without graft failure, thereby making HAPLO and URD HSCT a viable option for initial upfront treatment of SAA. The ability to use mismatched donors will also ensure that a HSCT option is more widely available to all racial/ethnic groups. This collaborative study was funded by grants U10HL069294, U24HL138660 to the BMT CTN from the NHLBI/ NCI, along with funding from Sanofi. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH.

Background: A deeper understanding of the cytogenetic features has enabled robust risk stratification for patients with newly diagnosed multiple myeloma (NDMM). However, heterogeneity persists in the currently available prognostic tools and there is further room for personalization. We assess the utility of PC-PRO, a test to measure the proportion of monotypic plasma cells in the synthetic phase (S-phase) of cell cycle, as a prognostic marker for NDMM. Methods: Patients with active MM diagnosed between January 2013 and January 2023 with a flow cytometry-based S-phase assessment at diagnosis were included. Flow cytometric immunophenotyping was performed using the following antibodies: CD19, CD38, CD45, CD138, cytoplasmic kappa and lambda immunoglobulin, and DAPI (4',6-diamidino-2-phenylindole). The percentage of plasma cells in S-phase was determined by measuring the proportion of cells with DNA content of the monotypic plasma cells between the G0/G1 and G2/M peaks. Plasma cell clonality was detected through demonstrating CD38 and CD138 positivity along with immunoglobulin light chain restriction, abnormality of CD19/ CD45 expression and/or ploidy difference by DAPI staining. High-risk cytogenetic abnormalities (HRCA) by FISH were defined as follows: presence of deletion 17p (del17p), t(4;14), t(14;16), t(14;20), 1q gain/amplification, 1p deletion (when available). Additionally, the second iteration of the revised ISS (R2-ISS), incorporating both laboratory and cytogenetic features as well an adaptation of the recently proposed International Myeloma Working Group (IMWG) Classification for high-risk disease, was utilized for risk stratification. All time-to-event analyses were performed using the Kaplan Meier method and compared using the log-rank test. Univariable and multivariable Cox proportional hazards analyses were performed to assess the independent prognostic impact of S-phase. Results: A total of 823 patients were included in the study. The median follow-up was 6 years [95% Confidence interval (CI): 5.5-6.3 years] and estimated median overall survival (OS) was 8.3 years. The median S-phase value for the cohort was 0.8 % (interquartile range: 0.5-1.4%), with 135 (16%) patients harboring an S-phase of ≥2% at diagnosis. Applying the R2-ISS prognostic model to our cohort, 11% (79/706) patients were stratified to the R2-ISS high-risk cohort; 93 patients (out of 788, 12%) had 2 or more HRCA at diagnosis. Patients with a S-phase ≥2% at diagnosis had a median OS of 3.9 years (95% CI: 2.9-5.7) vs. 9.2 years [95% CI: 7.9-not reached (NR)] for patients with a S-phase <2% [Hazard ratio (HR): 2.2 (95% CI: 1.7-2.8), p<0.0001]. The median OS was inferior for cohorts with an increasing S-phase value: S-phase of 0.1-1%: HR: 1 (standard), S-phase 1.1-3%: HR 1.7 (95% CI: 1.3-2.1), S-phase >3%: HR 3.0 (95% CI: 2.2-4.1)]. On a multivariable analysis including S-phase ≥2%, age at diagnosis of MM (>65 years) and R2-ISS high-risk status, S-phase ≥2% was an independent predictor of inferior OS [hazard ratio (HR) 1.9 (95% CI: 1.4-2.6), p<0.0001]. Age >65 years [HR 1.9 (95% CI: 1.5-2.5), p<0.0001] and R2-ISS high-risk status [HR 1.8 (1.3-2.6), p=0.001] were also independent predictors of OS. Within the sub cohort of patients with 2 or more of the listed high-risk cytogenetic features (n=93, 12%), a S-phase ≥2% (n=24) was associated with a worse median OS of 1.6 years (95% CI: 0.7-NR) vs. 6.2 years (95 % CI:3.3-NR) for S-phase <2% (p=0.002). Conclusion: Flow cytometry-based determination of the proportion of monotypic plasma cells in S-phase is an easy-to-study and powerful prognostic tool for NDMM. A S-phase of ≥2% at diagnosis is a predictor of OS independent of age and R2-ISS high-risk stage. An S-phase of ≥2% was able to identify a distinct group with markedly inferior outcomes even within the sub cohort of patients with 2 or more high-risk cytogenetic features.

Chronic lymphocytic leukemia (CLL), one of the most common types of leukemia among adults in the US, is characterized by the accumulation of CD5+CD19+ monoclonal B cells. Current clinical treatment approaches for progressive CLL patients focus on novel agents targeting Bruton Tyrosine Kinase (BTK) and Bcl-2. Although clinical trials combining novel agent have shown significant efficacy with high overall response and, in some cases, achievement of high rates of undetectable measurable residual disease (MRD), CLL remains incurable with a majority of patients eventually experiencing relapse of their disease. Therefore, it remains of high interest to expand therapeutic options. Transcriptional chemical inducers of proximity (TCIPs) are small molecules designed and developed by Dr. Gerald R. Crabtree at Stanford University. TCIPs are a new class of anticancer drugs that recruit endogenous cancer drivers or a downstream transcription factor to the promoters of cell death genes and activate their expression, resulting in apoptosis of cancer cells. Current TCIPs are produced by covalently linking small molecules that bind B-cell lymphoma 6 (BCL6) to those that bind to transcriptional activators, such as Bromodomain-containing protein 4 (BRD4) or cyclin-dependent kinases 9 (CDK9). The Stanford team has demonstrated TCIPs generate remarkable killing in diffuse large B cell lymphoma (DLBCL) cell lines, including chemotherapy-resistant, TP53-mutant lines, with a half maximal effective concentration (EC50) of 1-10nM at 72hours (h) and exhibited cell-specific and tissue-specific effects (Gourisankar et al, 2023). We have initiated TCIPs cytotoxicity assessment in primary leukemic cells from untreated CLL patients. Testing two types of TCIPs with a flow cytometry-based apoptosis assay, we found time and dose-dependent drug-associated killing in primary CLL B cells collected from untreated CLL patients over a 24-72h interval. The IC50 for BRD4-BCL6 TCIP JWZ-7-133 at 72h was 560nM while CDK9-BCL6 TCIP BAK-04-232 at 72h was 10nM. After 24, 48, and 72h exposure of CLL B cells to TCIP JWZ-7-133, the IC50 was 1050, 820 and 560nM, respectively. We also found that TCIP JWZ-7-133 exhibited a similar ability to inhibit leukemic cell viability with an IC50 of 1070nM using a cell viability assay, compared to a flow cytometry-based cell death assay (24h IC50 1050nM). CDK9-BCL6 TCIPs are designed and produced by covalently linking BI-3812 (binding to BCL6) and SNS-032 (binding to CDK9) with a rigid linker. BI-3812 is a potent and efficacious BCL6 inhibitor, inhibiting the BTB domain of BCL6, and SNS-032 is a potent and selective inhibitor of CDK9. When CLL B cells were treated with only BI-3812 or SNS-032 or TCIP BAK-04-232 as single agents for 72h, we found that TCIP BAK-04-232 had superior and significant killing effects compared to the single agents, with an IC50 of 10nM. Specifically this killing activity was more than 10-fold compared to SNS-032 alone (IC50 over 100nM), and more than 1,000-fold compared to BI-3812 alone (IC50 over 10,000nM). There remains an urgent to identify new categories of potent drug, inhibitor, or small molecules to expand potential clinical treatment options for patients with CLL. The results presented here indicate that TCIPs targeting BCL6 and BRD4/CDK9 have a strong capacity to induce high levels of cell death in CLL B cells at nanomolar concentrations, and this remarkable killing effect may be due to its unique gain-of-function drug mechanism. TCIPs provide a completely new drug direction and strategy for the design of anticancer drugs: to rewire cancer drivers to activate apoptosis in cancer cells. This study is the first of a kind demonstration of TCIPs in robustly targeting CLL B cells. Further studies to fully validate the preclinical activity of TCIPs and investigate the molecular mechanism(s) for TCIP mediated killing of leukemic B cells will help to define the role of TCIPs as a novel therapeutic approach in CLL.

Background and Rationale: Systemic mastocytosis (SM) is a clonal hematologic neoplasm characterized by abnormal accumulation of mast cells which can lead to symptoms related to mast cell degranulation and/or, in the case of advanced systemic mastocytosis (AdvSM), organ dysfunction due to mast cell infiltration. AdvSM is a category of SM comprised of aggressive systemic mastocytosis (ASM), mast cell leukemia (MCL), and SM with an associated hematologic neoplasm (SM-AHN). Among these, SM-AHN is the most common subtype, accounting for approximately 70% of AdvSM. A KIT D816V mutation found in up to 95% of SM cases is the primary oncogenic driver, making it a key therapeutic target. Avapritinib, a highly potent and selective inhibitor of D816V mutant KIT, has demonstrated high rates of rapid, deep, and durable responses in all AdvSM subtypes, including SM-AHN (Gotlib et al. Nature Medicine, 2021; DeAngelo et al., Nature Medicine, 2021). However, despite showing prolonged OS and low risk of progression of the SM component, some patients (20%) exhibit progression of the AHN component, including transformation to AML. In some patients, combination therapy with avapritinib and AHN-directed therapy may therefore be necessary for optimal control of SM-AHN. Hypomethylating agents (HMA) form the backbone of therapy for advanced myelodysplastic syndrome (MDS) and myelodysplastic/myeloproliferative neoplasm (MDS/MPN); hematologic neoplasms most likely to co-occur with SM in patients (pts) with SM-AHN. To date, the combination of HMA and KIT inhibitors has not been clinically evaluated. In this study, we aim to assess the safety and tolerability of the combination of avapritinib and decitabine in patients with SM-AHN. Study design and Methods: The primary objective of the study is to establish the recommended dose (RP2D) of avapritinib in combination with decitabine based on dose-limiting toxicity, overall safety, and clinical activity data. A dose evaluation steering committee (DESC) comprised of participating investigators will review the totality of safety and efficacy data to determine RP2D. Secondary endpoints include SM overall response rate by the end of 6 cycles by the m-IWG consensus criteria and proposed ECNM-AIM criteria (Gotlib et al., J Allergy Clin Immunol Pract, 2022) and incidence/frequency of adverse events. Exploratory endpoints include patient reported outcomes, PFS, OS, and correlative biomarkers linked to mast cell disease or AHN. Pts with a diagnosis of SM-AHN by the WHO 2022 criteria will be eligible for inclusion. Pts must have any of the following myeloid AHN per WHO 2022 criteria: MDS, CMML, MDS/MPN with neutrophilia, MDS/MPN with SF3B1 mutation and thrombocytosis, or MDS/MPN-NOS. Patients with MDS must have intermediate, high- or very-high risk by IPSS-R. Patients with very-low or low-risk MDS by IPSS-R must be intolerant/refractory to ESA or have serum EPO > 200 U/L. A baseline platelet count of ≥ 75 x 109/L and ≥ 25 x 109/L will be required for inclusion in the dose-finding and the dose-expansion phases, respectively. Key exclusion criteria include a diagnosis of AML, history of intracranial hemorrhage, prior treatment with avapritinib or decitabine with documented progression in SM or AHN component, respectively, or treatment with alternative KIT inhibitor or azacitidine within 4 weeks of study treatment initiation. In the dose finding part of the study, a Bayesian adaptive design will be employed to assess different doses of avapritinib (50 to 200 mg PO daily) in combination with fixed doses of decitabine (oral decitabine/cedazuridine: 35/100 mg/day or intravenous decitabine: 20 mg/m2/day on days 1-5) in a 28-day treatment cycle. Close attention will be paid to treatment-emergent thrombocytopenia with protocol-driven treatment holds and dose modification strategies employed when platelet count drops below 50 x 109/L. After RP2D is determined, an additional 10 patients will be enrolled in the dose expansion phase. Patients with platelet count ≥ 25 x 109/L and < 75 x 109/L, will receive lead-in dosing with decitabine or decitabine/cedazuridine with the ability to add avapritinib if sufficient platelet thresholds are achieved. Planned correlative studies include serial single-cell multi-omics utilizing scDNA-seq and immunophenotyping to assess impact of combination treatment on a molecularly complex disease state. This study is currently open to accrual (NCT06327685).

Background: JAK2 mutations are invariably associated with polycythemia vera (PV); more than 50% of affected patients also harbor non-JAK2 mutations, some considered prognostically relevant (BJH 2020;189:291). We queried a large single institutional database of patients with PV, in order to examine the incidence and distribution of non-JAK2 mutations, across different disease stages, in order to assess their impact on disease phenotype and prognosis, in the context of other risk factors. Methods: Mayo Clinic databases (1998-2023) were searched to identify next-generation sequencing (NGS)-annotated patients with PV, with diagnosis retrospectively fitted to meet the 2022 International Consensus Classification (ICC) criteria (Blood 2022; 140:1200). NGS was performed either through research-based targeted capture assays or a clinically ordered 42-gene panel. Conventional statistical methods were used for group comparisons and survival analyses. Results: i) Mutational frequencies in chronic vs. transformed disease phases 319 patients with PV were considered; NGS was performed at time of i) diagnosis or before fibrotic or leukemic transformation (N=270; 84.6%), ii) fibrotic progression (N=37; 11.6%), or iii) leukemic transformation (N=12; 3.8%). Mutational frequencies in newly-diagnosed/chronic phase PV were 19% for TET2, 10% for ASXL1, 4% for SRSF2, ⁓3% each for SF3B1, DNMT3A, and SH2B3, and 1-2% each for TP53 and IDH2. Significant differences in mutation distribution were apparent for TP53/SRSF2/IDH1/U2AF1 in blast (50%/25%/17%/8%) vs. fibrotic (8%/0%/0%/5%) vs. chronic (2%/4%/2%/0.4%) phase (p<0.01, <0.01, =0.03, and =0.03, respectively). ii) Phenotypic correlates of specific mutations among patients with NGS information obtained during chronic phase (N=270) Median age was 61.4 years (51% males) and incidence rates included 22% for leukocytosis ( ≥15 x 10(9)/L) and 25%/14% for arterial/venous thrombosis history. Numerically prevalent or prognostically relevant mutations were selected for comparison of clinical or laboratory characteristics: TET2, ASXL1, SRSF2, TP53, IDH2. Significant associations were apparent for ASXL1 (younger age distribution), SRSF2 (older age distribution, leukocytosis), and TP53 (leukocytosis) mutations. Median age was 66.4, 58.7, 76.8, 63.9, and 68.6 years for TET2, ASXL1, SRSF2, TP53, IDH2 mutations, respectively (p<0.01). The corresponding median leukocyte counts were 12.1, 11.4, 14.4, 16.7, and 11.5 x 10(9)/L (p<0.01). iii) Prognostic impact of mutations detected at diagnosis or chronic phase (N=270) At median follow-up of 9.5 years (range 0.1-42.4), 115 (43%) deaths, 12 (4%) leukemic transformations, and 53 (20%) fibrotic progressions were documented. In univariate analysis of mutations, overall survival (OS) was adversely affected by SRSF2 and IDH2 mutations, leukemia-free survival (LFS) by ASXL1, SRSF2, IDH2, RUNX1, and CEBPA mutations, and myelofibrosis-free survival (MFS) by SRSF2 mutation. Age-adjusted (p<0.01) multivariable analysis, including mutations and other risk variables, identified SRSF2 (HR 4.2, 95% CI 1.9-9.5), ASXL1 (HR 2.0, 1.1-3.7), IDH2 (HR 5.3, 1.8-15.3) mutations and leukocytosis ≥15 x 10(9)/L (HR 2.0, 1.3-3.1) as independent predictors of inferior OS. SRSF2 and ASXL1 mutations were also independently associated with inferior LFS (p=0.02 and 0.06, respectively) and SRSF2 (p<0.01) and karyotype (p<0.01) with inferior MFS (p<0.01). Median survival for 235 patients without any adverse mutation (ASXL1, SRSF2, or IDH2) was 17.8 years vs. 8.8 years for 35 patients with any one of the three adverse mutations (p=0.01; HR 1.8, 95% CI 1.1-2.9). The number of non-JAK2 mutations, in general, was also associated with significantly different survival outcomes: median OS was 20, 15.1, and 9.3 years for patients with non-JAK2 mutations numbering zero, one, or two or more, respectively (p<0.01) Conclusions: SRSF2 mutations in PV cluster with older and ASXL1 with younger age distribution. SRSF2 and TP53 mutational frequencies were higher in blast-phase disease. SRSF2, ASXL1, and IDH2 mutations were significantly associated with inferior OS, independent of each other and other risk factors. SRSF2 mutations were also detrimental to LFS and MFS and ASXL1 to LFS. Such information might allow further refinement of current risk models for PV, which in future studies should also account for variant allele frequency.

Background Pacritinib (PAC) is a relatively selective JAK2 vs. JAK1 inhibitor (JAKi) and is one of four FDA-approved JAKi for use in myelofibrosis (MF). PAC is indicated for MF patients with thrombocytopenia (Platelets <50 x 109/L). In a previously published controlled study of MF patients with platelet count <100 x 109/L (PERSIST-2), including JAKi-naïve and ruxolitinib-experienced, PAC therapy at the recommended dose of 200 mg BID was associated with a ≥35% spleen volume reduction rate of 22%; response rate was lower in ruxolitinib-experienced patients for both 200 mg BID (9.3%) and 100 mg BID (1.8%) dosing (JAMA Oncol 2018;4:652); a more recent post-hoc analysis suggested an additional benefit in anemia with 25-37% of transfusion-dependent patients achieving a ≥12-week transfusion-free period (Blood Adv 2023;7:5835). Herein, we describe our experience with the drug outside of a clinical trial setting. Methods Study patients were retrospectively recruited from Mayo Clinic (USA) databases, after approval from IRB. Diagnostic criteria were according to the International Consensus Classification (Blood 2022;140:1200). Study inclusion criteria included documentation of PAC therapy outside of a clinical trial setting. Spleen response was adjudicated by the 2013 International Working Group-European LeukemiaNet (IWG-ELN) response criteria (Blood 2013;122:1395). Anemia response was rated by the revised 2024 IWG-ELN criteria (manuscript submitted; transfusion-free period x 12 weeks and/or 12-week average increase in hemoglobin ≥1.5 g/dL). Conventional statistical methods were employed. Results A total of 60 MF patients (median age 70 years; males 62%; primary MF 63%) received PAC upfront (N=19; 32%) or after failing treatment with another JAKi (N=41; 68%), including ruxolitinib (N=39; 65%): JAK2 60%, CALR 27%, MPL 10%. Reasons for discontinuation of previous JAKi treatment included drug intolerance (41%), relapse after initial response (33%), and suboptimal response (26%). Median interval between diagnosis and treatment with PAC was 4 years (range 0-22). At PAC treatment initiation, DIPSS-plus risk distribution was high in 48%, intermediate-2 47%, and intermediate-1 (5%). Transfusion-dependent anemia (TDA; ≥3 units during 12 weeks prior to enrollment) was present in 40% of patients, platelet count <50 x 109/L 25%, leukocyte count >11 x 109/L 40%, spleen size >15 cm 76%, constitutional symptoms 55%, complex/monosomal karyotype 27%, and mutations in SRSF2 11% or ASXL1 33%. PAC initial dose was 100 (N=41) or 200 (N=19) mg BID. Dose was increased in 44% of patients started at 100 mg BID and decreased in 55% of those started at 200 mg BID. Among 47 patients evaluable for anemia response, only one of 23 with non-TDA (2% overall and 4.3% at 200 mg BID dosing) and none of 24 with TDA responded. In evaluable patients, spleen response rate was 13% (95% CI 5.9-24.7) overall and higher in JAKi-naïve (29%; 95% CI 11.7-54.7) vs. JAKi-experienced (6%; 95% CI 1.6-19.1; p=0.04). Spleen response rate was higher in the absence of leukocytosis (21% vs. 0%; p=0.01) or complex/monosomal karyotype (17% vs. 0%; p=0.05) but similar between 100 and 200 mg BID initial dosing (p=0.9). At a median follow-up of 14 months (range 1-27), PAC was discontinued in 80% of the patients: intolerance 58%, suboptimal response 54%, transition to allogeneic stem cell transplants (ASCT) 10%. Major side effects were more frequent with 200 vs. 100 mg BID dose and included diarrhea 43%, nausea/vomitting 17%, dyspnea 25%, peripheral edema 28%, allergic reaction 13%, grade 3/4 neutropenia 22%, and grade 4 thrombocytopenia 34%. Up to the time of this writing, 17 deaths (28%), 11 (18%) ASCT, and 3 (5%) leukemic transformations were documented. In multivariable analysis, post-PAC survival was adversely affected by TDA (HR 11.9; p<0.01), platelet count <50 x 109/L (HR 7.4; p<0.01), and SRSF2/ASXL1 mutations (HR 10.7; p<0.01), all independent of ASCT, which was independently associated with longer survival (HR 9.0; p<0.01). Conclusions: The observations from the current real-world study, regarding the moderate activity of PAC for reducing spleen size, are consistent with those of the aforementioned controlled clinical trial (PERSIST-2). However, despite recent suggestions of the drug's mechanistically unique erythropoietic activity, the anemia response rate seen in the current study was underwhelming.

Background and Significance When patients (pts) with follicular lymphoma (FL) present with a high tumor burden, they require systemic treatment with rituximab (R) in combination with chemotherapy (chemo). Given the known cumulative toxicities of chemotherapy and the older age of pts at the time of diagnosis, novel chemo-free treatment approaches are desired by pts as well as physicians. R-lenalidomide (R-Len), a chemo-free alternative to R-based chemoimmunotherapy, has recently been listed as a chemo-free option in the National Comprehensive Cancer Network® (NCCN) guidelines. Golcadomide (GOLCA) is a potential first-in-class oral cereblon E3 ligase modulator (CELMoD™) agent purposefully designed for the treatment of lymphoma. GOLCA co-opts cereblon to induce targeted degradation of the transcription factors Ikaros/Aiolos, inducing deep, rapid, and extensive tumor killing, independent of cell of origin, and enhanced immunostimulatory activity. Compared with immunomodulatory (IMiD) agents like lenalidomide, GOLCA more efficiently drives the cereblon complex to the closed active state resulting in more potent degradation (Watson ER et al, Science 2022), which is expected to translate into improved clinical benefit. In addition, the phase 1/2 study CC-99282-NHL-001 showed that GOLCA alone or in combination with R (GOLCA-R) was well tolerated and showed promising activity in pts with heavily pre-treated relapsed/refractory (R/R) FL, including in those with prior Len-based therapies and/or T cell-redirecting therapies (Chavez et al, EHA 2024). Based on this, GOLSEEK-2 (NCT06425302) is a randomized phase 2 trial assessing the efficacy and safety of GOLCA-R as a chemo-free option in pts with newly diagnosed advanced-stage FL. Study Design and Methods After a screening period of ≤ 5 weeks, approximately 90 pts with newly diagnosed advanced-stage FL will be randomized 1:1:1 between GOLCA plus rituximab at a dose of 0.2 mg or 0.4 mg once daily from day 1 to 14 every 28 days for 12 cycles or investigator's choice R-chemo (R-CHOP or R-bendamustine). R maintenance could be considered (as per pt and physician decision) for pts who achieve a partial response or better after completion of initial therapy for up to one year in GOLCA-R arms and up to one and a half years in the R-chemo arm. Pts with histologically confirmed diagnosis of grade 1, 2, 3a FL, or classic FL according to WHO classification would be eligible for the study. Other key inclusion criteria include no prior systemic treatment for FL, stage II-IV disease, the requirement for treatment as per investigator's judgment, and measurable disease as defined by Lugano 2014 criteria. Key exclusion criteria include other lymphoma subtypes, clinical evidence of transformed lymphoma, and documented or suspected central nervous system involvement. The primary endpoint is complete metabolic response (CMR) during the GOLCA plus rituximab combination treatment period, assessed by the investigator and based on Lugano 2014 Response Criteria. Secondary endpoints include progression-free survival, overall survival, overall response rate, duration of response, complete response rate at 30 months, CMR rate at 6 months and 12 months, and safety outcomes.

Background Survival risk models in essential thrombocythemia (ET) include the International Prognostic Scoring System (IPSET; Blood 2012;120:1197) and the more recently described “triple-A” prognostic score that considers Age and absolute neutrophil (ANC) and lymphocyte (ALC) counts, as risk variables (AJH 2023;98:1829). Monocytosis, defined by absolute monocyte count (AMC) of ≥0.8 x 109/L, was recently reported to be associated with inferior survival in polycythemia vera (ASH 2024 abstract). The current study examined the prognostic contribution of monocytosis in ET, in the context of other clinical and genetic risk factors. Methods Study patients were recruited from Mayo Clinic (USA) databases, based on availability of information on ANC, ALC, and AMC (N=649). Diagnostic criteria were according to the International Consensus Classification (Blood 2022;140:1200). Mutations were screened by next-generation sequencing (NGS). Conventional statistical methods were employed (JMP Pro 17.0.0, SAS Institute, Cary, NC, USA) and optimal cut-points for age, ANC, ALC, and AMC were determined by Receiver Operating Characteristic (ROC) analysis. Akaike Information Criterion (AIC) and ROC plots were used to estimate predictive accuracy of risk models. Results A total of 649 patients (median age 60 years; females 65%) were informative for AMC, ANC, and ALC, with median (range) values of 0.6 x 109/L (0.2-2.5), 5.7 x 109/L (1.5-26.5), and 1.9 (0.4-5.7), respectively. At presentation, IPSET risk distribution was low in 34%, intermediate 47%, and high 19%; triple-A risk distribution was low 21%, intermediate-1 50%, intermediate-2 13%, and high 16%; driver mutation distribution was 63.8% JAK2, 24.0% CALR, 3.7% MPL, and 8.5% triple-negative; median for platelets were 728 x 109/L, hemoglobin 13.9 g/dL, and leukocytes 8.5 x 109/L ; leukocytosis (leukocytes ≥11 x 109/L) was documented in 18.8% of patients, palpable splenomegaly 7%, arterial thrombosis (AT) history 14.5%, venous thrombosis (VT) history 9.2%, and major hemorrhage 7.8%. At median follow-up of 7 years (range 0.02-31.4), 146 (22.5%) deaths, 52 (8%) fibrotic progressions, 22 (3.4%) leukemic transformations, 71 (11%) AT events, 36 (5.6%) VT events, and 64 (10%) major bleeds were recorded. In age-adjusted univariate analysis, higher AMC was significantly associated with inferior survival, with a ROC determined cutoff value of ≥0.8 x 109/L (N=154; 24%); significance was sustained (HR 2.3; p<0.01) during multivariable analysis (MVA) that also identified age >70 years (HR 16.6; p<0.01), age 50-70 years (HR 4.3; p<0.01), ANC ≥8 x 109/L (HR 1.9; p<0.01), ALC <1.7 x 109/L (HR 1.8; p<0.01), and AT history (HR 1.6; p=0.01). A similar MVA conducted on an external cohort from the University of Florence (N=514) confirmed the significant contribution from all of these risk variables, except AT (p=0.8). An HR-based 4-variable prognostic model based on Age >70 years (4 points) or 50-70 years (2 points), AMC ≥0.8 x 109/L (1 point), ANC ≥8 x 109/L (1 point), and ALC <1.7 x 109/L (1 point), henceforth referred to as the AAA+A model, was subsequently developed using the Mayo Clinic cohort and validated by the Florence cohort: Low (0-1 points; N= 180; median not reached), intermediate-1 (2-3 points; N=266; median 21.6 years), intermediate-2 (4-5 points; N=151; median 12.2 years), and high (6-7 points; N=52; median 8.7 years); HR values were 2.1 for high vs. intermediate-2, 5.4 for intermediate-2 vs. intermediate-1, and 3.8 for intermediate-1 vs. low (p<0.001in all instances). The 20-year survival predictive performance of AAA+A (AIC 131; AUC 0.88) was superior to that of AAA (AIC 143; AUC 0.85) and IPSET (AIC 149; AUC 0.84). In MVA that included AAA+A risk variables, AMC ≥0.8 x 109/L (p=0.04) and ALC <1.7 x 109/L (p=0.06) appeared to affect leukemia-free survival and ANC ≥8 x 109/L (p=0.04) myelofibrosis-free survival. In univariate analysis, AMC ≥0.8 x 109/L was associated with AT (p=0.03) but not VT (p=0.14) history and showed borderline significance for venous (p=0.09) and arterial (p=0.07) thrombosis-free survivals; however, significance was lost during MVA. Conclusions: Monocytosis is an adverse prognostic factor in ET and its inclusion as a risk variable enhances the predictive performance of the AAA risk model and is equally simple to apply worldwide. Analyses that include mutations will be presented at meeting.

Introduction Prognostication in myelofibrosis (MF) currently relies on manual assessment of clinical, pathological, and laboratory variables, which is inefficient and error-prone. The potential of large language models (LLMs) and rule-based programming to automate data extraction and prognostication in these patients remains unexplored and merits investigation. Methods Patients with MF were identified from a retrospective chart review. Data regarding clinical (age, constitutional symptoms, transfusion requirements), laboratory (hemoglobin, circulating white blood cell, platelet, and blast counts), and pathological (grade of bone marrow fibrosis, chromosomal analyses, cytogenetics) features was extracted from the latest free-text clinical notes and pathology reports using the generative pretrained transformer-4 (GPT-4) model in a zero-shot setting with the prompting parameters set to generate deterministic responses. The patients were randomly sampled into development (15%) and test (85%) sets, with the documents of patients in the development set used for iterative prompt engineering. The finalized prompts were used for data extraction in the test set. Rule-based programming was used for post-processing of the extracted data and ascertaining prognosis via Dynamic International Prognostic Scoring System (DIPSS), DIPSS-plus (DIPSS+), and Mutation-enhanced International Prognostic Scoring System-70 (MIPSS-70). Automatically extracted data and prognostic categories were compared with the manually generated human annotations. Mean accuracy, precision, and recall with 95% confidence intervals (CI) were computed. Detailed error and sensitivity analyses were conducted to assess performance. Results The study included 100 patients (development: 15; test: 85) with 100 clinical notes (development: 15; test: 85) and 90 pathology reports (development: 13; test: 77). The mean age of these patients was 70 (SD: 10.6); 55 (55%) patients were men, and 93 (93%) patients were White. In the development set, the mean accuracy, precision, and recall for data extraction from clinical notes were 94% (95% CI: 90-98%), 92% (84-99%), and 99% (97-100%). For data extraction from pathology reports, the mean accuracy, precision, and recall were 96% (93-100%), 98% (92-100%), and 83% (63-100%). For prognostication, the mean accuracy, precision, and recall were 87% (79-94%), 80% (73-87%), and 100% using DIPSS (accuracy: 80%; precision: 73%; recall: 100%), DIPSS+ (accuracy: 93%; precision: 83%; recall: 100%) and MIPSS-70 (accuracy: 87%, precision: 83%; recall: 100%). In the test set, the mean accuracy, precision, and recall for data extraction from clinical notes were 96% (95% CI: 93-98%), 91% (87-96%), and 98% (97-100%). For data extraction from pathology reports, the mean accuracy, precision, and recall were 95% (92-98%), 93% (87-99%), and 90% (80-100%). For prognostication the mean accuracy, precision, and recall were 95% (94-96%), 89% (88-90%) and 99% (96-100%) using DIPSS (accuracy: 94%; precision: 89%; recall: 100%), DIPSS+ (accuracy: 95%; precision: 88%; recall: 95%) and MIPSS-70 (accuracy: 94%, precision: 90%; recall: 100%). Error analyses identified incorrect extractions by GPT-4 were related to the circulating blasts and constitutional symptoms variables from the clinical notes. Sensitivity analyses excluding the associated variables improved overall accuracy from 94% to 97% and 96% to 97% in the development and test sets, respectively. Similarly, incorrect extractions from the pathology reports were related to the CALR gene mutation variables. Sensitivity analyses excluding the associated variables improved the overall accuracy from 96% to 98% and 95% to 97% in the development and test sets, respectively. Conclusion LLMs and rule-based programming can accurately extract data from the electronic health records and prognosticate MF patients via DIPSS, DIPSS+, and MIPSS-70. The automated approach could significantly streamline clinical and research workflows, providing an efficient and less resource-intensive alternative to manual prognostication.