Search

In advanced systemic mastocytosis (advSM), disease evolution is often triggered by KITmutations (D816V in >80% of cases) and by additional mutations (eg, in SRSF2, ASXL1, and/or RUNX1[S/A/Rposin >60% of cases]). In a recently reported phase 2 study, midostaurin, a multikinase/KIT inhibitor, demonstrated an overall response rate (ORR) of 60% in advSM but biomarkers predictive of response are lacking. We evaluated the impact of molecular markers at baseline and during follow-up in 38 midostaurin-treated advSM patients. The median overall survival (OS) was 30 months (95% confidence interval, 6-54) from start of midostaurin. ORR and OS were significantly different between S/A/Rneg(n = 12) and S/A/Rpos(n = 23) patients (ORR: 75% vs 39%, P= .04; OS: P= .01, HR 4.5 [1.3-16.2]). Depending on the relative reduction of the KITD816V expressed allele burden (EAB) at month 6, patients were classified as KITresponders (≥25%, n = 17) or KITnonresponders (<25%, n = 11). In univariate analyses at month 6, reduction of KITD816V EAB ≥25%, tryptase ≥50%, and alkaline phosphatase ≥50% were significantly associated with improved OS. In multivariate analysis, only KITD816V EAB reduction ≥25% remained an independent on-treatment marker for improved OS (P= .004, HR 6.8 [1.8-25.3]). Serial next-generation sequencing analysis of 28 genes in 16 patients revealed acquisition of additional mutations or increasing variant allele frequency in K/NRAS, RUNX1, IDH2, or NPM1associated with progression in 7 patients. In midostaurin-treated advSM patients, the complexity and dynamics of mutational profiles significantly affect response, progression, and prognosis.

Systemic mastocytosis with an associated hematologic neoplasm (SM-AHN) is the most common subtype of advanced SM (advSM), diagnosed in up to 80% of patients. The AHN is most frequently diagnosed as a myeloid neoplasm, e.g., SM-MDS/MPNu or SM-CMML. Acquired mutations in KIT(usually KITD816, KITD816mut) are detectable in >90% of patients. The basis for the SM-AHN phenotype is usually the multi-lineage involvement, e.g. monocytes, eosinophils and other non-mast cell lineages, of KITmutations. Core binding factor (CBF) positive AML (CBFposAML) represents a distinct subtype and is identified in 5-8% of all AMLs. KITmutations, most frequently KITD816mut, are detectable in up to 45% of CBFposAML patients and are associated with an adverse prognosis. There is, however, only little information on KITD816mut/CBFnegAML. We therefore evaluated a) clinical and molecular characteristics, b) response to treatment and, c) survival and prognostic factors in 40 KITD816mut/CBFnegpatients with histologically proven SM and associated AML (SM-AML), collected at 4 centers of the European Competence Network on Mastocytosis (ECNM). Molecular analyses (n=32) revealed at least one additional somatic mutation (median, n=3) apart from KITD816, most frequently SRSF2(n=12, 38%), RUNX1(n=11, 34%), TET2(n=11, 34%), ASXL1(n=10, 31%), or NPM1(n=7, 22%). At least one mutation in SRSF2, ASXL1or, RUNX1(S/A/Rpos) was identified in 21/32 (66%) patients. At diagnosis of SM-AML 21/40 (52%) patients had an aberrant karyotype. Secondary AML evolved in 29/40 (73%) patients from SM ± associated myeloid neoplasm and longitudinal molecular analyses revealed acquisition of new somatic mutations (TP53, n=2; NPM1, n=1; RUNX1, n=1, ASXL1, n=1; BCOR, n=1; IDH1/2, n=1) and/or karyotype evolution in 15/16 (94%) patients at the time of SM-AML. Thirty-one of 40 (78%) patients were treated with intensive chemotherapy (ICT) with a complete response (CR) rate of 40%. Allogeneic stem cell transplantation (SCT) was performed in 12/40 (30%) patients with durable CR in 6/12 (50%) patients. S/A/Rposand/or the presence of a poor-risk karyotype were adverse predictive markers for response to treatment. To further investigate whether KITD816mut/CBFnegAMLdefines a distinct AML subtype associated with SM, two independent AML databases (AMLdatabases) were retrospectively screened and 69 KITD816mut/CBFnegAML patients identified. The comparison between KITD816mut/CBFnegSM-AML from ECNM (n=40) centers with KITD816mut/CBFnegAMLdatabases(n=69) revealed remarkable similarities: a) a high KITD816 variant allele frequency (VAF) (median 34% vs. 29%), b) with the exception of SRSF2(38 vs. 18%), a highly similar mutation landscape, rather comparable to that of advSM (Jawhar et al., Blood 2017) than to that of de novoAML, c) in contrast to de novoAML, a low frequency of FLT3mutations (3 vs. 7%), and d) a high frequency of an aberrant karyotype (52 vs. 42%). The median overall survival (OS) of 40 KITD816mut/CBFnegSM-AML and 17 evaluable KITD816mut/CBFnegAMLdatabaseswas 5.4 (95% confidence interval, CI [1.7-9.1]) and 26.4 (95% CI [0-61.0]) (P=0.015) months, respectively (Figure 1). However, if only the patients with ICT ± allogeneic SCT were compared, median OS between the two groupswas not different (16.7 vs. 26.4 months, P=0.4). In multivariate analyses, S/A/Rposand a poor-risk karyotype remained the only independent poor-risk factors with regard to OS. These results were independent of treatment modalities. We conclude that KITD816mut/CBFnegAML is a new poor-risk subtype associated with SM (SM-AML). The remarkable clinical, genetic and prognostic similarities between SM-AML and AMLdatabasessuggest that a significant proportion of the AMLdatabasespatients may in fact have SM-AML. We therefore strongly recommend to determine serum tryptase and KITD816 mutation status in all AML patients, and to perform bone marrow histology in KITD816mutpatients. These simple diagnostic measures would allow reclassification to SM-AML and inclusion of KIT inhibitors in established treatment modalities of AML.

In systemic mastocytosis (SM), clinical phenotype, response to treatment and prognosis are influenced by the KITD816V mutation (80-90% of patients positive), additional mutations in SRSF2, ASXL1or RUNX1(S/A/Rpos) and chromosomal aberrations. In core-binding factor acute myeloid leukemias (CBF-AML), KITmutations are identified in about 30% of cases, conferring an adverse prognostic impact on survival. However, patients with AML are not routinely checked for the contemporaneous presence or absence of SM. We sought to evaluate a) clinical and molecular characteristics, b) responses to treatment and survival, and c) prognostic factors in 44 patients with SM and associated AML (SM-AML). The median age was 64 years (range 28-83), 64% were male. In the majority of patients (30/44, 68%), AML evolved a median of 24 months (range 2-153) after diagnosis of SM without (6/30, 20%) or with (24/30, 80%) an associated hematologic neoplasm (AHN) such as myelodysplastic syndrome (SM-MDS), myeloproliferative neoplasm (SM-MPN) or MDS/MPN (SM-MDS/MPN). Based on histopathology, SM was contemporaneously diagnosed with AML in 14/44 (32%) patients. At diagnosis of SM-AML, the median bone marrow infiltration by mast cells and myeloid blasts was 15% (range 5-65; ≥20% in 45% of patients) and 39% (range 20-95), respectively, and the median serum tryptase level was 92 μg/l (range 2-885; ≥150 in 39% of patients). In the majority of patients, material was available for molecular analyses and revealed the following markers at diagnosis of SM-AML: a) KITD816V in 36/41 (88%) patients, b) additional mutations by targeted next-generation sequencing (NGS, 28 myeloid gene panel) in 32/33 (97%) patients, the most frequent mutations were: RUNX1(n=12, 36%), TET2(n=11, 33%), SRSF2(n=11, 33%), ASXL1(n=9, 27%), IDH1/ 2(n=6, 18%), NPM1(n=6, 18%), BCOR(n=5, 15%), DNMT3A(n=4, 12%), and, c) an aberrant karyotype in 25/35 (71%) patients (in analogy to related myeloid neoplasms, e.g. MDS or AML, patients were classified according to their karyotype into three groups: favorable-, intermediate- and poor-risk; only 1 patient with a CBF translocation). Patients with a known history of SM were frequently S/A/Rpos(19/22, 86%) and acquisition of additional mutations (NPM1, n=2; IDH1, n=2; BCOR/ JAK2/ TP53/ PHF6/ RUNX1, n=1) or karyotype evolution at time of SM-AML was observed in 8/15 (53%) or 9/13 (69%) patients, respectively. NGS on DNA derived from CD34+ myeloid blasts of 6 KITD816V positive patients revealed the presence of KITD816V positive blasts in only 1/6 (17%) patients while additional mutations were found in blast cells of all 6 patients. Eight patients received supportive care due to age and/or comorbidity. In 36 patients, disparate combinations of AML-type chemotherapy ± hypomethylating agents (HMA) ± cladribine (n=16) were administered. Cytarabine-based consolidation treatment was administered in 4/36 (11%) patients in complete remission (CR). Allogeneic stem cell transplantation (SCT) was performed in 15/36 (42%) patients, only 6/15 (40%) patients were in CR prior to allogeneic SCT. Overall, 5/10 (50%) patients in CR of AML showed persistence of SM. From diagnosis of SM-AML, the median observation was 10 months (range 0-200); 32/44 (73%) patients died during observation. The median overall survival (OS) was 11 months (95% confidence interval [CI], 4-17) with median OS of 2 (range 0-5), 4 (range 0-12) and 74 (range 0-149) months for patient cohorts who were treated with supportive care, intensive chemotherapy/HMA and intensive chemotherapy followed by allogeneic SCT, respectively (P=0.0002). On multivariate analysis, the only independent prognostic marker for OS was favorable-risk karyotype (n=14) vs. intermediate-/poor-risk karyotype (n=21; median OS 21 vs. 7 months; HR 3.2 [1.3-7.8], P=0.007). We conclude that a) SM-AML seems to be more frequent than commonly believed, b) SM-AML should be regarded as secondary AML evolving from multimutated SM-AHN, c) KITmutations are not restricted to CBF-AML, d) progression to SM-AML is often triggered by acquisition of new mutations and/or karyotype evolution, e) SM-AML has an aggressive phenotype and a poor prognosis, f) treatment of SM-AML should include a combination of intensive chemotherapy and allogeneic SCT for eligible patients and g) the disease frequency and poor prognosis warrant routine screening of all AML cases for serum tryptase and KITD816V.

The presence of multiple additional mutations was recently reported in >90% of patients with KITD816V+advanced systemic mastocytosis (SM) (Schwaab et al., Blood 2013). In the vast majority of cases, additional mutations occur prior to KITD816V indicating a multi-mutated stem cell disease with phenotype modification through KITD816V (Jawhar et al., Leukemia 2015). We now sought to clarify the prognostic impact of individual mutations and the number of mutated genes in 70 multi-mutated patients with ISM (n=26), aggressive SM (ASM, n=34) and mast cell leukemia (MCL, n=10). All patients had an associated clonal hematologic non-mast-cell disease (AHNMD), such as a myelodysplastic/myeloproliferative syndrome unclassified (MDS/MPNu, n=28), chronic myelomonocytic leukemia (CMML, n=21), chronic eosinophilia leukemia (CEL, n=9), acute myeloid leukemia (AML, n=8) or primary mylofibrosis (PMF, n=4). Mutational analysis was performed by next-generation sequencing in the following genes: ASXL1, CBL, ETV6, EZH2, IDH2, JAK2, KRAS, NRAS, RUNX1, SF3B1, SRSF2, TET2and U2AF1. The most frequently identified mutated genes were TET2(n=33, 47% of cases), SRSF2(n=30, 43%), ASXL1(n=20, 29%), RUNX1(n=16, 23%) and JAK2(n=11, 16%). The molecular data were correlated with overall survival (OS), clinical characteristics (e.g. B- and C-findings), bone marrow (BM) findings, serum tryptase levels and KITD816V allele burden in peripheral blood. With a median observation time of 3.7 years, 25 of 70 patients (36%) have died. Three-year OS probability was 71%. In univariate analysis, OS was adversely influenced by mutations in SRSF2([hazard ratio (HR) 5.9, 95% confidence interval (CI)] [2.2-15.4], p<0.0001), ASXL1(HR 3.4 [1.5-7.7], p=0.002) and RUNX1(HR 2.4 [1.1-5.4], p=0.03) but was not influenced by mutations in TET2(p=0.3) or JAK2(p=0.1). In multivariate analysis of the most frequently mutated genes, SRSF2remained as the best independent poor risk marker for OS. Furthermore, inferior OS was significantly associated with the number of mutated genes in SRSF2/ASXL1/RUNX1(S/A/R), with median OS not reached for 0 mutation (n=25), 3.9 years for 1 mutation (n=27) and 2.7 years for >=2 mutations (n=18). The 3-year OS was 90% in patients with 0 mutation, 73% in patients with 1 mutation and 42% in patients with >=2 mutations in the S/A/R gene panel. Pairwise significantly different OS probabilities were observed for the comparisons 0 vs. 1 mutation (HR 9.2 [2.0-42.3], p=0.001), 0 vs. >=2 mutations (HR 4.1 [1.9-8.8], p<0.0001) and 1 and >=2 mutations in the S/A/R gene panel (HR 2.5 [1.1-6.0], p=0.04) (Figure 1). Regarding clinical characteristics, the S/A/R mutation profile is significantly related to C-findings including hepatosplenomegaly, albumin <35g/L, alkaline phosphatase >150 IU/L, ascites and weight loss, and other important disease-related parameters including bone marrow mast cell infiltration, serum tryptase levels and KITD816V+allele burden. We therefore conclude that a) presence and number of mutations in the S/A/R gene panel are adversely associated with advanced disease and poor survival in KITD816V+SM patients and b) the inclusion of molecular markers should be considered in upcoming prognostic scoring systems for SM patients.

Systemic mastocytosis (SM) is characterized by abnormal proliferation and accumulation of mast cells (MC) in various tissues and organs, predominantly skin, bone marrow (BM) and visceral organs. The extent of organ infiltration and subsequent organ damage is the basis for the classification of SM into indolent SM (ISM), smoldering SM (SSM), SM with associated clonal hematologic non-MC lineage disease (SM-AHNMD), aggressive SM (ASM) and MC leukemia (MCL). Depending on the subtype of SM, cell source (BM or peripheral blood) and assay sensitivity, an acquired mutation in the receptor tyrosine kinase KIT, usually KITD816V, is detectable in 80-90% of patients. Next-Generation Deep Amplicon Sequencing (NGS) was performed to investigate 18 candidate genes at known mutational hotspot regions as previously described. Additional mutations in genes encoding for signaling molecules (JAK2, CBL, KRAS, NRAS), transcription factors (RUNX1), epigenetic regulators (ASXL1, DNMT3A, EZH2, TET2) or splicing factors (SRSF2, SF3B1, U2AF1) are detectable in the vast majority of patients KITD816V+ advanced SM. In order to gain more insight into clonal evolution of myeloid progenitors in indolent and advanced SM with or without AHNMD, we explored the mutation profile of single-cell-derived CFU-GM colonies - by using Sanger sequencing - in 19 KITD816V+ SM patients (investigated colonies, n=285; median per patient, n=15; range 10-30). The study included 7 patients with ISM/SSM/ASM (0 additional mutations), 4 patients with SM-AHNMD (median 1 additional mutation, range 1-4) and 8 patients with ASM-AHNMD (median 3 additional mutations, range 1-4). KITD816V+ CFU-GM could be identified in all 8 patients with ASM ± AHNMD but in only 20% (1/5) of patients with SM-AHNMD, while all CFU-GM colonies derived from ISM patients were completely KITD816V negative. On the other hand, CFU-GM colonies from individual ASM ± AHNMD patients were never entirely KITD816V+ (median 60%, range 0-95). In contrast to KITD816V, additional mutations were identified in CFU-GM colonies from all 12 multi-mutated (A)SM-AHNMD patients and many of these mutations were present in 100% of the CFU-GM-derived colonies analyzed. In 8 patients, different subclones with variable proportions of the number of mutated genes were identified and allowed to generate putative evolutionary trees. These mutations included TET2and SRSF2mutations in 6/6 and 4/4 patients with (A)SM-AHNMD. In contrast, ASXL1mutations were not identified in all TET2/SRSF2positive CFU-GM colonies suggesting that they are likely to occur later than TET2and SRSF2. When additional mutations and KITD816V were detected concomitantly in individual single-cell-derived CFU-GM colonies, the overall frequency of CFU-GM colonies positive for additional mutations was always higher than those that were KITD816V positive, indicating that the KITD816V mutation was a secondary event. In contrast to advanced SM, all ISM patients were negative for additional mutations. CFU-GM colonies and also microdissected CD15+ cells derived from BM biopsies were entirely KITD816V negative in these patients, highlighting that the KITD816V mutation may be restricted to other (probably later) stages of stem cell development and possibly only to the MC lineage. In general, the relative proportion of KITD816V+ progenitors correlated well with established parameters for quantification of disease burden, e.g. BM MC infiltration, serum tryptase levels and KITD816V allele burden in individual patients. In conclusion, the presence of multi-mutated myeloid non-MC lineage progenitors of the CFU-GM-type suggests an initial clonal expansion at an early stage of hematopoietic development due to mutations other than KITD816V with a subsequent phenotype modification towards SM due to a later acquisition of KITD816V. In contrast, ISM/SSM is not affected by mutations at the CFU-GM level which may at least in part explain its excellent prognosis.

Abstract 2859

Abstract 4100