Influence of somatic mutations and pretransplant strategies in patients allografted for myelodysplastic syndrome or secondary acute myeloid leukemia

REFERENCE 1. Le CH. The prevalence of anemia and moderate-severe anemia in the US population (NHANES 2003-2012). PLoS One. 2016;11:e0166635. 2. Auerbach M, Henry D, Derman RJ, Achebe MM, Thomsen LL, Glaspy J. A prospective, multi-center, randomized comparison of iron isomaltoside 1000 versus iron sucrose in patients with iron deficiency anemia; the FERWON-IDA trial. Am J Hematol. 2019;94:1007-1014. 3. Bhandari S, Kalra PA, Berkowitz M, Belo D, Thomsen LL, Wolf M. Safety and efficacy of iron isomaltoside 1000/ferric derisomaltose versus iron sucrose in patients with chronic kidney disease: the FERWON-NEPHRO randomized, open-label, comparative trial. Nephrol Dial Transplant. 2020;gfaa011. https://doi.org/10.1093/ ndt/gfaa011. Epub ahead of print. 4. Wolf M, Rubin J, Achebe M, et al. Effects of iron isomaltoside vs ferric carboxymaltose on hypophosphatemia in iron-deficiency anemia: Two randomized clinical trials. JAMA. 2020;323:432-443. 5. Pollock RF, Biggar P. Indirect methods of comparison of the safety of ferric derisomaltose, iron sucrose and ferric carboxymaltose in the treatment of iron deficiency anemia. Expert Rev Hematol. 2020;13:187-195. 6. Go AS, Chertow GM, Fan D, McCulloch CE, Hsu CY. Chronic kidney disease and the risks of death, cardiovascular events, and hospitalization. N Engl J Med. 2004;351:1296-12305.

interaction. In accordance with recently reported findings 3,4 we show that the presence of poor risk mutations facilitates refinement of the prognostic information of CK with a dismal prognosis in patients carrying both CK and poor-risk mutations ( Figure S19).
Focusing in a next step on other disease-related, patientassociated and transplant-related factors (Tables S9,S10; Figures S20-S23) we found that pretransplant strategy was the most prominent factor significantly influencing OS. Patients, who were transplanted F I G U R E 1 Effects of "poor risk mutations" and pretransplant strategy on posttransplant outcome. A, Illustrates posttransplant relapsefree and overall survival depending on "poor risk mutation" status and pretransplant strategy as well as the interplay between both parameters. B, Illustrates effects of "poor risk mutation" status and pretransplant strategy on outcome after HMA-based salvage therapy for posttransplant relapse in terms of complete remission rate and overall survival in 41 relapsed patients. "Poor-risk mutation" status and pretransplant strategy are depicted in indicated colors and line pattern respectively. Hazard ratios are relative to those patients without any poor risk mutation and relative to those receiving upfront transplantation, respectively. Allo-HSCT, allogeneic hematopoietic stem cell transplantation; CR, complete remission; HMA, hypomethylating agents; HR, hazard ratio; mut, mutation without prior cytoreduction, had longer survival compared to patients who received pretransplant cytoreduction ( Figure 1A, Figure S15). In addition, patients undergoing upfront transplant had a trend towards a lower relapse rate ( Figure 1A). Except for age (OS and RFS) and CK (RFS) no other factor impacted OS, RFS, relapse incidence or NRM (Tables S9-S10; Figures S20-S23).
In multivariate analyses the presence of poor-risk mutations as well as pretransplant strategy confirmed their prognostic role with poor-risk mutation carriers and pretreated patients having inferior survival and higher risk of relapse. Additionally, BCOR and EZH2 mutations were associated with relapse and NRAS and SF3B1 mutations with NRM (Tables S11-S13).
Based on this, we addressed the hypothesis that there might be an interaction between these two disease-related and procedurerelated factors and re-analyzed our cohort by combining the information about mutation status of these four genes and pretransplant strategy ( Figure 1A). Indeed, the outcome of patients with poor-risk mutations, who had received pretransplant therapy, was poor with 5-year OS and RFS of 23% and 5.7% and significantly inferior compared to patients with poor-risk mutations in the upfront group (5-year OS: 45%, 5-year RFS: 27%). In contrast, among patients without poor-risk mutations the upfront group had a favorable 5-year OS of 83% compared to 62% in pretreated patients, while 5-year RFS was comparable (5-year RFS 61% vs 54%). These data were confirmed by a separate analysis of patients with MDS ( Figure S24).
Relapse, mainly driven by poor-risk mutations and pretransplant strategy (Table S12), was the major cause of treatment failure ( Figure S2) with 62% of deaths being attributable to relapse. Therefore, we finally asked whether these two variables also influenced response to and survival following salvage therapy with HMA (Aza n = 40, decitabine n = 1) and donor lymphocytes infusions (DLI, Figure S25; Table S14). The negative prognostic impact of poor-risk mutations was indeed abrogated after salvage therapy, as indicated by a comparable CR rate (43% vs 53%) and survival (2-year OS 43% vs 53%) in mutation carriers and patients without poor-risk mutations. In contrast, the CR rate was significantly higher (73% vs 21%) and survival (2-year OS 69% vs 27%) significantly longer in the upfront group compared to pretreated patients ( Figure 1B

SUPPORTING INFORMATION
Additional supporting information may be found online in the Supporting Information section at the end of this article. There has been increasing interest in real-world data in AML which reflect treatments and outcomes in unselected patient populations. Recent real-world registry data from Europe have reported that a significant proportion of older patients do not receive any anti-leukemic therapy at the time of diagnosis, other than best supportive case (BSC). [1][2][3] However, there has been a paucity of reported analysis on why such patients do not receive treatment. In