Response to tyrosine kinase inhibitors in myeloid neoplasms associated with PCM1‐JAK2, BCR‐JAK2 and ETV6‐ABL1 fusion genes

We report on 18 patients with myeloid neoplasms and associated tyrosine kinase (TK) fusion genes on treatment with the TK inhibitors (TKI) ruxolitinib (PCM1‐JAK2, n = 8; BCR‐JAK2, n = 1) and imatinib, nilotinib or dasatinib (ETV6‐ABL1, n = 9). On ruxolitinib (median 24 months, range 2‐36 months), a complete hematologic response (CHR) and complete cytogenetic response (CCR) was achieved by five of nine and two of nine patients, respectively. However, ruxolitinib was stopped in eight of nine patients because of primary resistance (n = 3), progression (n = 3) or planned allogeneic stem cell transplantation (allo SCT, n = 2). At a median of 36 months (range 4‐78 months) from diagnosis, five of nine patients are alive: four of six patients after allo SCT and one patient who remains on ruxolitinib. In ETV6‐ABL1 positive patients, a durable CHR was achieved by four of nine patients (imatinib with one of five, nilotinib with two of three, dasatinib with one of one). Because of inadequate efficacy (lack of hematological and/or cytogenetic/molecular response), six of nine patients (imatinib, n = 5; nilotinib, n = 1) were switched to nilotinib or dasatinib. At a median of 23 months (range 3‐60 months) from diagnosis, five of nine patients are in CCR or complete molecular response (nilotinib, n = 2; dasatinib, n = 2; allo SCT, n = 1) while two of nine patients have died. We conclude that (a) responses on ruxolitinib may only be transient in the majority of JAK2 fusion gene positive patients with allo SCT being an important early treatment option, and (b) nilotinib or dasatinib may be more effective than imatinib to induce durable complete remissions in ETV6‐ABL1 positive patients.

4-78 months) from diagnosis, five of nine patients are alive: four of six patients after allo SCT and one patient who remains on ruxolitinib. In ETV6-ABL1 positive patients, a durable CHR was achieved by four of nine patients (imatinib with one of five, nilotinib with two of three, dasatinib with one of one). Because of inadequate efficacy (lack of hematological and/or cytogenetic/molecular response), six of nine patients (imatinib, n = 5; nilotinib, n = 1) were switched to nilotinib or dasatinib. At a median of 23 months (range 3-60 months) from diagnosis, five of nine patients are in CCR or complete molecular response (nilotinib, n = 2; dasatinib, n = 2; allo SCT, n = 1) while two of nine patients have died. We conclude that (a) responses on ruxolitinib may only be transient in the majority of JAK2 fusion gene positive patients with allo SCT being an important early treatment option, and (b) nilotinib or dasatinib may be more effective than imatinib to induce durable complete remissions in ETV6-ABL1 positive patients.

| INTRODUCTION
More than 70 different tyrosine kinase (TK) fusion genes with recurrent involvement of at least 6 TK genes (PDGFRA, PDGFRB, FGFR1, JAK2, ABL1, FLT3) have been identified in clinically and morphologically distinct myeloid neoplasms with or without eosinophilia. 1 Patients may present in chronic phase (CP) or blast phase (BP) of myeloid or lymphoid origin. 2 The WHO 2017 classification defines some but not all of these fusions within a distinct subgroup as "myeloid/ lymphoid neoplasms with eosinophilia and rearrangement of PDGFRA, PDGFRB, FGFR1 or PCM1-JAK2 fusion gene" (MLN-eo). 3 Targeted treatment with TK inhibitors (TKI) is highly effective in many cases with rapid and durable complete remissions in patients with PDGFR fusion genes, for example, FIP1L1-PDGFRA or ETV6-PDGFRB. [4][5][6][7] In contrast, patients with fusions involving other TK, for example, FGFR1 or JAK2, present with a more aggressive phenotype and a highly variable sensitivity to currently available TKI. 1,[8][9][10] Long-term disease-free survival may therefore only be achieved through allogeneic stem cell transplantation (allo SCT).

| Patients
Eighteen patients with PCM1-JAK2 (n = 8), BCR-JAK2 (n = 1) [patients 6 and 7 were previously published, 9 ] and ETV6-ABL1 (n = 9) fusion genes were identified within the "German Registry for Disorders of Eosinophils and Mast Cells "and in cooperation with hematology centers in the UK (n = 3), Switzerland (n = 1) and USA (n = 1). All patients were treated with one or more TKIs for at least part of their clinical course. Patient demographics, disease characteristics, therapies, responses and follow-up data

| Cytogenetic and molecular analysis
Cytogenetic analysis and fluorescence in situ hybridization (FISH) were performed on bone marrow (BM) according to standard procedures. Specific nested reverse transcription polymerase chain reaction (RT-PCR) was performed for confirmation of PCM1-JAK2, BCR-JAK2 and ETV6-ABL1 fusion transcripts 8,19,55 and for detection of residual disease. Complete cytogenetic response (CCR) was defined by a normal karyotype and complete molecular response (CMR) was defined by an undetectable fusion transcript by nested RT-PCR.

| DISCUSSION
We sought to evaluate disease characteristics and response to treatment with TKI in 18 patients with myeloid neoplasms and associated JAK2 and ETV6-ABL1 fusion genes. Moreover, we integrated our data into available cases from the literature to gain a more thorough insight into phenotype, treatment efficacy and prognosis of these distinct myeloid neoplasms.
In our series, the most notable clinical and morphological features of patients with JAK2 fusion genes included a marked male predominance, lack of hypereosinophilia (≥1.5 × 10 9 /L) in the majority of patients, pathognomonic giant paratrabecular islets of predominantly immature proerythroblasts, 36,37,57 and primary BP or relatively rapid progression to secondary BP in a significant proportion of patients.
Patients can achieve CHR and CCR on ruxolitinib but it had to be stopped in all but one patient within the first 3 years, most frequently because of resistance, relapse or progression. In the extended analysis, three further patients were treated with ruxolitinib, with two patients achieving a CCR for more than 30 months. [12][13][14][15] There was, however, no long-term beneficial effect upon ruxolitinib ( Figure 2). After a median of 2 years, durable complete remissions were only observed on nilotinib, dasatinib or after allo SCT. We identified an additional 14 cases in the literature [16][17][18][19][20][21][22][23][24][25][26][27]53,54] with adequate data on response on TKI and follow-up. In concordance with our findings, imatinib can induce but not maintain long-term remissions. 25