KIT D816 mutated/CBF-negative acute myeloid leukemia: a poor-risk subtype associated with systemic mastocytosis

KIT D816 mutations (KIT D816mut) are strongly associated with systemic mastocytosis (SM) but are also detectable in acute myeloid leukemia (AML), where they represent an adverse prognostic factor in combination with core binding factor (CBF) fusion genes. Here, we evaluated the clinical and molecular features of KIT D816mut/CBF-negative (CBFneg) AML, a previously uncharacterized combination. All KIT D816mut/CBFneg cases (n = 40) had histologically proven SM with associated AML (SM-AML). Molecular analyses revealed at least one additional somatic mutation (median, n = 3) beside KIT D816 (e.g., SRSF2, 38%; ASXL1, 31%; RUNX1, 34%) in 32/32 (100%) patients. Secondary AML evolved in 29/40 (73%) patients from SM ± associated myeloid neoplasm. Longitudinal molecular and cytogenetic analyses revealed the acquisition of new mutations and/or karyotype evolution in 15/16 (94%) patients at the time of SM-AML. Median overall survival (OS) was 5.4 months. A screen of two independent AML databases (AMLdatabases) revealed remarkable similarities between KIT D816mut/CBFneg SM-AML and KIT D816mut/CBFneg AMLdatabases (n = 69) with regard to KIT D816mut variant allele frequency, mutation profile, aberrant karyotype, and OS suggesting underlying SM in a significant proportion of AMLdatabases patients. Bone marrow histology and reclassification as SM-AML has important clinical implications regarding prognosis and potential inclusion of KIT inhibitors in treatment concepts.

In general, acquired mutations in KIT (usually KIT D816V) are detectable in >90% of patients with SM, acknowledged to be most relevant for disease pathogenesis [5]. In advSM, multi-lineage involvement (including nonmast-cell-lineage cells, e.g., monocytes, eosinophils, and others) of KIT mutations is frequently observed and the basis for the phenotype of SM-AHN [6][7][8]. Recent data have, however, also highlighted that the molecular pathogenesis of advSM is much more complex with the presence of one or more additional somatic mutations, e.g., in SRSF2, ASXL1, RUNX1, JAK2, TET2 [9][10][11]. These additional mutations are often acquired by neoplastic (stem) cells prior to KIT D816V thereby indicating a multi-mutated stem cell disease and a step-wise process of oncogenesis [12].
Core binding factor (CBF) positive AML (CBF pos AML) represents 5-8% of all AMLs and is defined by the presence of a t(8;21)(q22;q22) and the associated RUNX1-RUNX1T1 fusion gene, or an inv(16)(p13.1q22)/t(16;16)(p13.1;q22) with the resulting CBFB-MYH11 fusion gene. CBF pos AML is categorized to the genetically favorable risk group. However, KIT . mutations, most frequently at position D816 (KIT D816 mut ), are detectable in up to 45% of CBF pos patients and associated with adverse prognosis [13,14]. The potential association of KIT D816 mut /CBF pos AML with underlying SM has been described in various case reports, case-series, and/or literature reviews [15][16][17][18][19], however, there is little information available on KIT D816 mut /CBF neg AML [20]. We therefore evaluated (a) clinical and molecular genetic characteristics, (b) response to treatment, and (c) survival and prognostic factors in 40 patients with KIT D816 mut /CBF neg AML collected at 4 centers of the European Competence Network on Mastocytosis (ECNM). To further investigate whether KIT D816 mut /CBF neg defines a distinct AML subtype associated with SM and poor prognosis, two independent AML databases (AML databases , German/Austrian AML Study Group, Munich Leukemia Lab) were retrospectively screened for KIT D816 mut /CBF neg AML patients (selection criteria were all AML patients with available status on CBF and KIT D816 mut ).

Diagnosis of SM-AML
The diagnosis of SM-AML was established according to the WHO classification [2,[21][22][23]. Bone marrow biopsies and smears were evaluated by reference pathologists of the ECNM (H-PH and K Sotlar). A total of 48 CBF neg SM-AML patients, diagnosed in 4 ECNM centers between 2003 and 2018, were included in this retrospective analysis. Eight patients negative for KIT D816 mutations (n = 5) or with unknown KIT D816 mutation status (n = 3) were excluded. Among all SM-AML patient from the 4 ECNM centers, one patient was KIT D816 mut /CBF pos . The study design adhered to the tenets of the Declaration of Helsinki and was approved by the institutional review board of the Medical Faculty of Mannheim, Heidelberg University, as part of the "German Registry on Disorders of Eosinophils and Mast Cells". All patients gave written informed consent.
Subsequent to bcl2fastq and demultiplexing, alignment and variant calling were performed using JSI SeqNext v4.4.0 (JSI Medical Systems, Kippenheim, Germany) software with default parameters. Only basecalls with quality score of 30 or above were considered for further processing. In median~1800 reads were aligned to the target region. All regions below the minimal coverage of 400 reads were rejected and resequenced for higher depth. Variants were called with a variant allele frequency (VAF) cutoff of 3% and each assessed manually for pathogenicity. Mutation assessment was performed using COSMIC (v78), dbSNP (v150), ClinVar (2018-07), gnomAD (r2.0.2 and dbNSFP v3.5).
Qualitative and quantitative assessments of KIT D816V and KIT D816V expressed allele burden, respectively, was performed using allele-specific quantitative real-time reverse transcriptase polymerase chain reaction analyses (qRT-PCR) as previously described [24]. Molecular analyses were performed at diagnosis of SM ± AHN and at diagnosis of SM-AML.

Conventional cytogenetic analysis and fluorescence in situ hybridization
Cytogenetic analyses of at least 20 Giemsa-banded bone marrow metaphases (24 h and/or 48 h culture) was performed and interpreted according to the International System for Human Cytogenetic Nomenclature [25]. If necessary, chromosome banding analysis was combined with fluorescence in situ hybridization according to the manufacturer's instructions (Metasystems, Altlussheim, Germany) [26].

Statistical analyses
Statistical analyses considered clinical, laboratory, or molecular parameters obtained at the time of diagnosis. Overall survival (OS) analysis was determined as time from date of diagnosis to date of death or last follow up. Pearson correlation analysis was performed for the correlation between two parameters. Differences in the distribution of continuous variables between categories were analyzed by Mann-Whitney test (for comparison of two groups). For categorical variables, Fisher's exact test was used. OS probabilities were estimated with the Kaplan-Meier method and compared by the log-rank test in univariate analysis.
For the estimation of hazard ratios (HRs) and multivariate analysis, the Cox proportional hazard regression model was used. P-values < 0.05 (two-sided) were considered significant. There was no adjustment for multiple testing as all analyses were explorative. SPSS version 22.0.0 (IBM Corporation, Armonk, NY, USA) was used for statistical analysis.

De novo SM-AML and secondary SM-AML
De novo SM-AML was diagnosed in 11/40 (28%) patients. Secondary SM-AML evolving from indolent SM (n = 5) or SM-AHN (n = 24) was observed in 29/40 (73%) patients with a median time to progression of 24 months (range 2-116). The 24 patients with AHN were classified as MDS/ MPN-u (n = 8), CMML (n = 6), MDS (n = 5), or MPN associated with eosinophilia (MPN-eo) (n = 5) (Table 1a). The comparison between de novo and secondary AML revealed that patients with secondary AML were older, had a higher monocyte count, a higher AP level, and a lower serum tryptase level. However, there were no significant differences regarding OS (P = 0.2).
To further evaluate whether KIT D816 mut occurred in hematopoietic progenitor cells, we performed molecular analyses on DNA derived from CD34+ cells from 6 KIT D816V positive patients. KIT D816V was found in 1/6 (17%) patients while additional somatic mutations were detected in all 6 patients.
Comparison of KIT D816 mut /CBF neg SM-AML with KIT D816 mut /CBF neg AML from two independent databases To further investigate whether KIT D816 mut /CBF neg AML represents a distinct subtype which is associated with SM and poor prognosis, two independent AML databases (AML databases ) were retrospectively screened for KIT D816 mut / CBF neg AML patients. Overall, 69 KIT D816 mut /CBF neg AML databases patients could be identified. Mutation profile and karyotype were available from all patients, detailed clinical characteristics from 17/69 patients (Tables 1b and 3).  Fig. 2 KIT D816 variant allele frequency (VAF), somatic mutations, and aberrant karyotype in KIT D816 mut /CBF neg SM-AML in comparison to KIT D816 mut /CBF neg AML from the two databases (AML databases ). a KIT D816 VAF, b relative frequency distribution of mutated genes, and c aberrant karyotype. Gray columns: KIT D816 mut / CBF neg SM-AML and blue columns: KIT D816 mut /CBF neg AML databases . Asterisk represents targeted next-generation sequencing was performed in 32/40 SM-AML patients  Table 3): (a) The median KIT D816 VAF was 34% (range 3-54) and 29% (range 3-93), respectively, (b) with the exception of SRSF2 (38% vs. 18%), the frequency of the most frequently somatic mutations (RUNX1, TET2, ASXL1, NPM1, DNMT3A, IDH1/2) was highly similar between the two groups, (c) in contrast to de novo AML, the frequency of FLT3 aberrations was very low (3% and 7%, respectively), and (d) the frequency of an aberrant karyotype was 52% and 42, respectively, with a comparable rate of intermediate-risk and poor-risk karyotype.
In univariate analyses (including age, hemoglobin, platelets, AML subtype, treatment modalities [non-allogeneic vs. allogeneic SCT], somatic mutations, and aberrant karyotype), only age >60 years, at least one additional somatic mutation in the S/A/R gene panel (S/A/R pos ) and a poor-risk karyotype were identified as poor prognostic variables regarding OS. In multivariate analysis, S/A/R pos and a poor-risk karyotype remained the only independent adverse factors with regard to OS. Accordingly, a weighted score (based on the HR) of 1 was assigned to S/A/R pos and poor-risk karyotype. Significantly different OS probabilities were observed for the comparisons S/A/R neg + normal-/ intermediate-risk karyotype (0 point, n = 14), S/A/R pos or poor-risk karyotype (1 point, n = 23), and S/A/R pos + poorrisk karyotype (2 points, n = 10) with median OS not reached vs. 14.0 [6.2-21.8] vs. 7.0 months [4.5-9.6] (P = 0.001). These results were independent of treatment modalities (Fig. 4a, b).

Discussion
We report here on a large series of 40 patients with morphologically proven KIT D816 mut /CBF neg SM-AML. Approximately 65% of patients evolved from other advSM subtypes. Similar to previous reports concerning the molecular profile of advSM, all patients with SM-AML had at least one additional somatic mutation, most frequently affecting TET2, SRSF2, ASXL1, RUNX1, and NPM1. In contrast to de novo AML, only one patient had a FLT3 mutation. The overall molecular profile of SM-AML therefore was more similar to the profile of advSM than to that of de novo AML [28].
Using CFU-GM-colonies and microdissected cells, we have previously shown that mast cells and AHN cells are not only positive for KIT D816V but also for additional somatic mutations, indicating that both derive from a common progenitor [12]. However, a significant proportion of colonies were positive for additional somatic mutations but negative for KIT D816V [12]. In line with this and other  Data on patients treated with intensive chemotherapy ± allogeneic stem cell transplantation only data demonstrating the absence of KIT D816V in myeloid blasts of 50% of SM-AML cases [20], we confirmed the absence of KIT D816V but the presence of additional somatic mutations in CD34+ cells in 5 of 6 SM-AML cases, indicating that the additional somatic mutations rather than KIT D816V are the driving force for progression to secondary AML. In addition, serial molecular genetic analyses revealed the acquisition of new somatic mutations, e.g., in NPM1, IDH1/2, RUNX1, with or without karyotype evolution in >90% of patients as further underlying mechanisms for progression to secondary SM-AML. This data is reminiscent of reports on progression in other myeloid neoplasms such as MDS or MDS/MPN, and our previous reports on progression of SM to advSM or progression within advSM subtypes, e.g., to secondary mast cell leukemia, in which somatic mutations in NPM1, IDH2, or RUNX1 were also identified as late events and drivers for disease progression [29][30][31][32][33][34][35]. SM-AHN is the most common subtype of advSM but the diagnosis is challenging because the mast cell infiltrate may obscure the AHN and vice versa [20,[36][37][38]. This is particularly true for AML where the morphological but not the histological examination of bone marrow has been established as a standard diagnostic tool. Recently reported data collected from deep targeted sequencing indicated that KIT D816 mutations can be identified in 1-6% of patients with various subtypes of myeloid neoplasms, e.g., MDS, MDS/ MPN-u, CMML, polycythemia vera, essential thrombocythemia, or myelofibrosis [39][40][41][42][43][44]. However, many of these cases have not routinely been screened by histopathology for the presence of co-existing SM. Within our registry, all KIT D816V mut patients, who had initially been diagnosed as myeloid neoplasms such as CMML, triplenegative MF, and others, in fact fulfilled the WHO-criteria for a diagnosis of SM-AHN.
We therefore sought to investigate the incidence of KIT D816 mut /CBF neg in retrospective screens of two independent AML databases. Rather unexpectedly, 69 patients were identified which revealed remarkable similarities concerning the high KIT D816 VAF, the mutation profile and the aberrant karyotype ( Table 2), suggesting that the vast majority of these AML databases patients are likely to have SM-AML. Unfortunately, the lack of bone marrow trephine biopsies at initial diagnosis of AML has not allowed a definite re-evaluation of these cases and formal reclassification as SM-AML. However, based on our data, which are in line with previously published results, an underlying or concomitant SM can be diagnosed in most cases of KIT D816V mut AML, when the bone marrow is investigated using standard histopathological and molecular studies.
The median OS of the 40 SM-AML patients was 5.4 months and thus even worse as compared to patients with mast cell leukemia, which is defined by the presence of ≥20% mast cells in a bone marrow smear [29]. No patient achieved a CR on treatment with hypomethylating agents C B A Fig. 3 Kaplan-Meier estimates of overall survival (OS) of KIT D816 mut /CBF neg SM-AML and AML from the databases (AML databases ). a OS of all KIT D816 mut /CBF neg patients, b OS comparing the KIT D816 mut /CBF neg SM-AML cohort with intensive chemotherapy (ICT) ± allogeneic stem cell transplantation (SCT) (yellow), the KIT D816 mut /CBF neg AML databases cohort with ICT ± allogeneic SCT (green), and the KIT D816 mut /CBF neg SM-AML with non-intensive therapy (NIT)/best supportive care (BSC) (red), c OS of all KIT D816 mut /CBF neg patients treated with ICT only (blue) or with allogeneic SCT (gray). CI confidence interval, n.s. non-significant. Asterisk refers to included patients with SM-AML and AML databases n.s. non-significant, S/A/R pos at least one mutation in SRSF2, ASXL1, and/or RUNX1 a Included patients with SM-AML and AML from the two databases and none of the patients was treated with midostaurin. Following intensive induction chemotherapy in eligible patients, the CR rate of 40% was significantly inferior as compared to the general CR rate of de novo AML (70-80%) [45] and median survival following intensive chemotherapy with or without allogeneic SCT was 17 months. In addition to the aforementioned similarities regarding the molecular genetic characteristics (KIT D816V VAF, additional somatic mutations, and aberrant karyotype), the poor median OS of 26 months in 17 KIT D816 mut /CBF neg AML patients from the two independent AML databases adds further evidence that KIT D816 mut /CBF neg AML may in fact represent SM-AML in the vast majority, if not all patients. Independently of treatment modalities and consistent with previous reports on other advSM subtypes, e.g., mast cell leukemia, mutations in S/A/R and a poor-risk karyotype conferred an adverse impact on response to treatment, disease progression, and OS [10,29,30]. Midostaurin, an orally administered multi-kinase/FLT3-/ KIT-inhibitor improves survival in FLT3 pos AML and achieves overall response rates of 60% in patients with advSM [46,47]. Better survival is observed in advSM patients without additional somatic mutations in the S/A/R gene panel and a >25% reduction of the KIT D816V VAF at month 6 [30,[46][47][48][49]. If the presence of SM can be proven in KIT D816 mut /CBF neg AML by bone marrow histology and elevated serum tryptase, KIT inhibitors (e.g., midostaurin, potentially avapritinib [BLU-285, Blueprint Medicines, Cambridge, MA, USA]) in combination with intensive chemotherapy and allogeneic SCT may help to improve the poor prognosis of this distinct AML subtype [50,51].
We conclude that (a) progression to secondary AML from a preceding KIT D816 mut SM-AHN is frequently observed and may be triggered by the acquisition of additional somatic mutations with or without karyotype evolution, (b) KIT D816 mut /CBF neg AML is a distinct subtype with remarkable similarities compared to SM-AML cases concerning KIT D816 VAF mutation profile, aberrant karyotype, and poor prognosis, suggesting that a significant proportion of these AML patients may in fact have SM-AML, which is a strong argument to propose a new evaluation, (c) with its very high positive and negative predictive value, serum tryptase is an excellent screening marker for SM and should therefore be part of the diagnostic workflow in all AML patients. Cases with an elevated serum tryptase level should subsequently be screened for KIT D816 mut , and (d) bone marrow histology is mandatory in KIT D816 mut patients. This simple diagnostic procedure will allow reclassification to SM-AML and thus allow inclusion of KIT inhibitors in established treatment modalities of AML. Fig. 4 Kaplan-Meier estimates of overall survival (OS) of KIT D816 mut /CBF neg SM-AML and AML databases . OS of KIT D816 mut / CBF neg patients treated with a non-intensive therapy (NIT)/best supportive care (BSC) or intensive chemotherapy (ICT) ± allogeneic stem cell transplantation (SCT) or b ICT ± allogeneic SCT. Depending on the SRSF2/ASXL1/RUNX1 (S/A/R) mutation status and karyotype, three different cohorts were identified: S/A/R neg + normal-/intermediate-risk karyotype (green), S/A/R pos or poor-risk karyotype (yellow), and S/A/R pos + poor-risk karyotype (red). CI confidence interval, n.s. non-significant, S/A/R pos at least one mutation in the S/A/ R gene panel. Asterisk refers to included patients with SM-AML and AML databases Shoumariyeh, MM, LLFS, SF, BK, NvB, K Spiekermann, MH, GM, SK, HCK-N, TH, HD, WRS, PV, and AR provided patient material and information. H-PH and K Sotlar reviewed the bone marrow biopsies. MJ, KD, SK, JS, AF, NCPC, W-KH, PV, and AR wrote the paper.

B A
Conflict of interest H-PH, K Sotlar, PV, and AR served as consultants in a global phase-II-study examining the effects of midostaurin in advanced systemic mastocytosis. H-PH, K Sotlar, K Shoumariyeh, PV, NCPC, and AR received honoraria and/or travel support from Novartis Pharmaceuticals. MJ and JS received travel support from Novartis Pharmaceuticals. TH has equity ownership of the MLL Munich Leukemia Laboratory. MM is employed by the MLL Munich Leukemia Laboratory. The remaining authors declare that they have no conflict of interest.
Publisher's note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons. org/licenses/by/4.0/. 1