Yi, Cecilia Arana

Introduction: Telemedicine has been widely utilized in the healthcare marketplace since the beginning of the COVID-19 pandemic. Over the last 3 years, virtual care has been widely employed in the care of patients with polycythemia vera (PV), essential thrombocythemia (ET) and myelofibrosis (MF) at Mayo Clinic. These rare disorders have few experts; therefore, telemedicine becomes an optimal approach for patients receive expert care even when they are not close to the medical center. 1 Additionally, telemedicine reduces financial burden for the patients. We aim to estimate potential cost savings to patients utilizing telemedicine. Methods: We identified patients with MPN seen at Mayo Clinic, Arizona who have had at least 2 visits (1 or more virtual visits) between March 2020 and June 2022. We extracted retrospective data from electronic medical records (EMR) including sociodemographic data, disease related data and number of visits - both virtual and in person - of the patients identified. The address recorded from the EMR was assumed to be the patient's location at the time of virtual visit. Driving distances were determined using zip codes in Google Maps and the patients were stratified into quartiles based on round trip travel distances. Patients in quartiles 1 through 3 were assumed to have driven while patients in quartile 4 were assumed to have traveled by plane. Patients in quartile 1 were assumed to have 1 travel day including the follow-up visit and no accommodation, patients in quartile 2 were assumed to have 2 travel days including the follow up visit and 1 night of accommodation whereas patients in quartiles 3 and 4 were assumed to have 2 travel days, 1 day for the follow up visit and 2 nights of accommodation. Mileage rate costs were based on US Internal Revenue Service (IRS) standard mileage rates for medical or moving purposes for 2021. 2 Owing to extreme variability in flight costs, flying costs were assumed to be equal to the driving costs. Per diem costs for accommodation and meals and incidentals were based on the US General Services Administration (GSA) 2021 rates for Phoenix 3 with the cost being 75% of the full per diem rate on traveling days via GSA formulaic approach. Results: One hundred and eight patients having 282 virtual visits were identified. Patients ranged from 30 - 87 years with the mean and median age of 65 years and 67 years respectively. Majority of the patients were female (62%) and white (98%) with gender distribution varying by distance. Descriptive statistics of the estimated out-of-pocket costs for patients by distance quartiles are summarized in Table 1a. Mean estimated cost savings ranged from $46.8 to $877.3 with the savings of $43.8 - $1372.2 at extremes. The total cost savings by all patients was $90,306.7. The breakdown of costs by category for each quartile is shown in Table 1b which shows that the majority of costs saved by patients in quartile 1 were due to avoiding costs spent on meals and incidentals (~90% of mean costs) whereas costs saved due to meals and incidentals for patients in quartile 4 were only ~16% of the mean costs. Conclusions: Our research highlighted telemedicine as a cost-effective approach for patients with MPN who received virtual care from March 2020 - June 2022. Study limitations included not weighing costs for childcare or accompanying individuals and using driving costs instead of flying costs, however, these considerations potentially underestimate the actual cost savings to the patient. Additionally, since the median age of this population was 65, we didn't include loss of wages to the analysis, which may further underestimate the costs. Despite concerns over billing, reimbursement, and state licensure, virtual care is being established as a critical part of medicine moving forward. 1. Doolittle, G.C. and A.O. Spaulding, Providing Access to Oncology Care for Rural Patients via Telemedicine. Journal of Oncology Practice, 2006. 2(5): p. 228-230. 2. Internal Revenue Service. Standard mileage rates for 2021. Available at: https://www.irs.gov/tax-professionals/standard-mileage-rates 3. US General Services Administration. Per diem rates. Available at: https://www.gsa.gov/travel/plan-book/per-diem-rates/per-diem-rates-results?action=perdiems_report&fiscal_year=2021&state=AZ&city=&zip=85054

Background: Venetoclax (Ven) in combination with hypomethylating agent (HMA) is the preferred treatment for elderly/unfit patients with newly-diagnosed acute myeloid leukemia (AML). Current prognostication for AML is based on European LeukemiaNet (ELN) genetic risk stratification which has limited applicability in patients treated with Ven-HMA ( Blood, 2022). Accordingly, our primary objective was to determine predictors of treatment response and survival outcomes in treatment-naïve AML patients receiving Ven-HMA. Methods: Our study population was recruited from Mayo Clinic (MN, FL, AZ), after institutional review board approval and based on documentation of newly-diagnosed AML treated with Ven-HMA outside of clinical trials between November 2018 and April 2023. Cytogenetic and molecular studies were performed by conventional karyotype, and next-generation sequencing, respectively. Response was assessed according to the 2022 ELN criteria Results: Patient characteristics A total of 301 newly-diagnosed AML patients (median age 73 years, 66% male, 62% de novo) received a median of 3 cycles (range 1-48) of azacitidine 75 mg/m2 days 1-7 ( n=100) or decitabine 20 mg/m2 days 1-5 ( n=201) with Ven. ELN 2022 cytogenetic risk included favorable (1%, n=4), intermediate (61%, n=184) or adverse (38%, n=113). Mutations involved TP53 in 77/301 (25%), ASXL1 in 54/279 (19%), RUNX1 in 54/296 (18%), NPM1 in 44/299 (15%), DNMT3A in 40/296 (13%), K/NRAS in 41/296 (14%), IDH1 in 20/301 (6%), IDH2 in 30/301 (10%), FLT3-ITD in 25/301 (8%) and DDX41 in 9/247 (4%) of informative cases. Predictors of response Complete remission, with (CR) or without (CRi), count recovery, was documented in 182 (60%) patients, including CR in 36% and measurable residual disease (MRD) negative by flow cytometry in 77% of 105 informative cases that achieved CR/CRi. Median time to CR/CRi was 1.4 months. In univariate analysis, CR/CRi was more likely to occur in the presence of NPM1 mutation, (86% vs. 56%; p < .01), IDH2 mutation, (80% vs. 58%; p = .02), DNMT3A mutation, (78% vs. 57%; p = .01), or DDX41 mutation, (100% vs. 58%; p <.01), and in the absence of adverse karyotype, (71% vs. 42%; p < .01), TP53 mutation, (67% vs. 40%; p <.01), FLT3-ITD mutation, (63% vs. 36%; p = .01), or RUNX1 mutation, (64% vs. 44%; p = .01); in multivariable analysis, adverse karyotype, ( p < .01), FLT3-ITD ( p < .01), RUNX1 ( p < .01), NPM1 ( p = .03), IDH2 ( p = .04), and DDX41, ( p = .01) mutations remained significant. Predictors of survival After a median follow-up of 8.5 months (range 0.5-56), 13.2 months for living patients, 174 (58%) deaths, 73 relapses (40% of those achieving CR/CRi), and 41 (14%; including 35 in CR/CRi) allogeneic hematopoietic stem cell tranplants (AHSCT) were documented. In univariate analysis of pre-treatment variables, thrombocytopenia < 100 x 10 9/l, (HR 1.5, 95% CI 1.03-2.2; p = .03), adverse karyotype, (HR 2.8, 95% CI 2.1-3.8; p < .01) TP53 mutation, (HR 2.5, 95% CI 1.8-3.5; p < .01), or absence of IDH2 mutation, (HR 3.5, 95% CI 1.7-7.6; p < .01) predicted inferior survival. Failure to achieve CR/CRi (HR 4.5, 95% CI 3.3-6.1; p < .01) was also associated with inferior survival and in multivariable analysis, failure to achieve CR/CRi, (HR 3.4, 95% CI 2.5-4.8; p< .01), presence of adverse karyotype, (HR 1.6, 95% CI 1.1-2.6; p = .02), TP53 mutation, (HR 1.6, 95% CI 1.002-2.4; p = .04), and absence of IDH2 mutation (HR 2.2, 95% CI 1.004-4.7; p = .04) were identified as risk factors for survival. Subsequent HR-weighted scoring resulted in three-tiered risk stratification: low (0-1 point; n= 130), intermediate (2-3 points; n= 105), and high (4-5 points; n= 66), with respective median survival (3-year rate) of 28.9 (48%), 9.6 (6%), and 3.1 (0%) months ( p < .01) (Figure). AHSCT had an independent favorable impact on survival (HR 0.3, 95% CI 0.1-0.6; p < .01), most apparent in low ( p =0.04), and intermediate ( p <.01), as opposed to high ( p = .06) risk. Conclusions: In the current single institutional series of Ven-HMA treated newly-diagnosed AML, response to Ven-HMA was the foremost predictor of survival. Additional risk factors for survival included adverse karyotype, presence of TP53 and absence of IDH2 mutations. These observations allowed for a three-tiered genetics-enhanced survival model with CR/CRi as a backbone, and also confirmed survival advantage from AHSCT, regardless of risk category.

Background: Outcomes of adults with relapsed/refractory acute myeloid leukemia (RR-AML) have remained poor. CD38 is a transmembrane glycoprotein that is highly expressed in AML and is potentially targetable with CD38 blocking agents [Naik J et al. Haematologica 2019, Tembhare et al. J Immunother Cancer 2020]. XmAb18968 is a novel CD38-CD3 bi-specific T-cell engager with Fc domain modified to minimize Fcγ receptor binding and non-selective T-cell activation resulting in reduced cytokine release without compromising target cell killing. We designed a phase I clinical trial to assess the safety and tolerability of XmAb18968 in patients with RR-AML. Methods: We conducted a multicenter, open label, phase I dose escalation study of XmAb18968 (NCT05038644). Patients aged ≥18 years with relapsed/refractory AML, ECOG PS 0-2 and adequate organ function were eligible. Major exclusion criteria are hematopoietic cell transplantation within 6 months of enrollment, active acute graft-versus-host disease, and acute promyelocytic leukemia. Dose escalation was conducted using 3+3 design and 4 separate dose levels were studied [0.8mg, 1mg, 1.3mg, 1.5mg]. XmAb18968 was administered intravenously on days 1, 8, 15, 22 on a 28-day cycle. During cycle 1 alone, day 1 dose was fractionated and administered over 2 days to minimize the risk of significant cytokine release syndrome (CRS). Treatment was continued until progression or unacceptable toxicity. The primary objective was to determine the safety and dose limiting toxicity (DLT) and secondary objective was to determine the preliminary efficacy. Measurable residual disease (MRD) status was assessed using multicolor flow cytometry with 0.01% sensitivity. Correlative studies included pharmacokinetics, characterizing changes in serum cytokines and phenotypic expression of activated T-cells and leukemic cells, correlation of the phenotypic expression with changes in transcriptome at the single-cell level, proteomics evaluation of cytokine secretion at the single-cell level and correlation of response with N-glycan profiling, quantitative site-occupancy, and direct glycopeptide analysis. Results: Thirteen patients with RR-AML were enrolled, 12 of whom were evaluable (1 patient came off study before DLT period). Median age was 63 years (45-77), 77% with adverse risk disease, 38% with secondary AML, and received median two (range 2-7) prior lines of therapy. Twelve of 13 patients (92%) were previously treated with venetoclax based regimen and 31% had prior allogeneic HCT. Adverse risk mutations included TP53 (31%), RUNX1 (23%), and ASXL1 (15%). Grade 3 or higher adverse events included anemia (8%), neutropenia (15%), and thrombocytopenia (23%) (Table 1). CRS grades 1-2 were noted in 62% of patients, no grade ≥3 CRS or neurotoxicity (any grade) was seen. There was no dose limiting toxicity (DLT). Of the 12 evaluable total patients, clinically meaningful response was noted in two patients who were MRD positive by flow cytometry prior to study entry and achieved MRD negative CR (Table 2). These two patients who achieved MRD negative CR subsequently proceeded to allogeneic HCT. Median time to response was 1 cycle. Median OS of the cohort was 6.2 months (1.8-NR). Conclusions: XmAb18968 is safe and well tolerated in patients with RR-AML. Dose escalation yielded MRD negative CR in RR-AML patients who were MRD positive at study entry. Details on correlative studies examining mechanisms of therapeutic efficacy and resistance will be reported in the main meeting.

BACKGROUND: Recently proposed World Health Organization (WHO, 5 th edition) and International Consensus Classification (ICC) recognized myeloid neoplasms (MN) harboring TP53-mutated ( TP53mut) as a separate entity. However, there are critical differences between the two classifications regarding allelic-status, variance allele frequency (VAF) cut-off and blast categories, that may lead to under- or overestimation of the risk. METHODS: We analyzed MN harboring TP53mut (VAF ≥2%, n=536) managed at Mayo Clinic (USA) and South Australia Health Network (Australia). ICC (Arber et al, Blood 2022) and WHO (Bluebook, Beta v2) definitions were used to stratify MN harboring TP53mut. RESULTS: Of 536 cases, 379 (71%) had myelodysplastic syndromes (MDS) (<20% blasts) and 157 (29%) had acute myeloid leukemia (AML). Using WHO, 160 (42%) of 379 were classified as MDS-bi TP53 as they harbored 2 TP53mut (n=115; 30%) or 1 TP53mut plus confirmed loss of heterozygosity (LOH) or copy neutral LOH (cnLOH) of TP53 locus (n=45; 12%). Sixty-two (16%) TP53mut MDS with VAF >49% were classified as presumptive MDS-bi TP53. Of the 157 cases with 1 TP53mut VAF ≤49%, 17 confirmed to have an intact wild-type TP53 copy and were classified as monoallelic. While 140 (89%) cases did not have copy number variation (CNV) analysis. We next evaluated the number of TP53mut cases classified differently by WHO and ICC. As WHO does not recognize TP53mut AML as a separate entity, 157 (29%) TP53mut AML were not classifiable. 46% (174/379) of the TP53mut MDS including 48 (30%), 12 (19%) and 79 (65%) of bi TP53, presumptive bi TP53 and monoallelic TP53mut MDS according to WHO were classified differently by ICC ( Figure 1A). Discordance between the two classifications were predominantly driven by: (i) The difference in VAF cut-off. Sixty patients (11%) with TP53mut VAF <10% were excluded from the ICC classification of TP53mut MN as it mandates TP53mut VAF cut-off ≥10%. 11 (18%) and 41 (68%) of these had bi TP53, monoallelic TP53mut MDS by WHO criteria while 8 (13%) had AML. Importantly, within the VAF <10% cohort, median overall survival (OS) of 2 TP53mut or 1 TP53mut with 17p13.1 (hereafter 17p) deletion or complex karyotype (CK, n=30) was significantly poor compared to 1 TP53mut VAF <10% (11.6 vs 32.4 months; P =0.002, Figure 1B), suggesting that ICC may underestimate the poor prognosis of multi-hit or multi-hit equivalent TP53mut MN with VAF <10%. (ii) The difference in bone marrow blast cut-off. The underlying assumption by WHO is that prognosis of bi TP53mut MDS is poor irrespective of blast percentage category (0-9% vs 10-19%). In contrast, ICC considers that TP53mut MDS/AML (blasts 10-19%) is associated with poor prognosis irrespective of TP53 allelic status. Median OS of TP53mut MDS with 10-19% blasts was significantly poor compared to blast 0-9% (6.4 vs 11.6 months, P <0.0001). Second, median OS of bi TP53mut was significantly poor than monoallelic TP53mut in MDS- blast 0-9% (10.1 vs 15.9 months; P =0.0001) but not in blasts 10-19% (6.8 vs 6.2 months; P =0.19). Hence, exclusion of monoallelic TP53mut MDS with blast 10-19% as TP53mut MN from the WHO classification may underestimate the poor prognosis. (iii) The difference in the definition of biallelic and multi-hit TP53 loss. 35 cases with 1 TP53mut VAF ≤49%, had 17p deletion but did not have CNV analysis to confirm the loss of TP53 locus and could not be classified further using WHO. ICC classified 18 (50%) of these cases as multi-hit. Median OS of 1 TP53mut VAF ≤49%, and 17p deletion on karyotype was significantly poor than 1 TP53mut (VAF ≤49%without 17p deletion on karyotype (6.4 vs 14.6 months, P =0.001). (iv) The difference in the definition of multi-hit equivalent. 32% (33/104) WHO defined monoallelic cases without available CNV data, were classified as multi-hit-equivalent by ICC as they harbored CK. However, WHO does not consider CK as presumptive evidence of bi TP53. Median OS of 1 TP53mut VAF ≤49% with CK was significantly poor compared to 1 TP53mut without CK (8.2 vs 26.9 months, P =0.0008). CONCLUSION: Our study provides evidence that poor prognosis of ‘multi-hit’ TP53 with VAF <10% is underestimated by ICC. On the other hand, poor prognosis of monoallelic TP53 MDS with 10-19% blasts as well as 1 TP53mut VAF ≤49% in the presence of CK is underestimated by WHO. These data suggest that both classifications can be further refined.