In 2008, in response to the identification of patients with “chronic eosinophilic leukemia” or “hypereosinophilic syndrome” who carried recurrent tyrosine kinase fusion genes involving PDGFRA, PDGFRB, or FGFR1, the World Health Organization classification of myeloid neoplasms included a new category termed “Myeloid/lymphoid neoplasms (MLN) with eosinophilia and rearrangements of PDGFRA, PDGFRB, or FGFR1” [1]. This World Health Organization category was revised in 2016 with the addition of PCM1::JAK2 as a provisional entity [1]. In the recent fifth edition of the World Health Organization classification, similar to the recent update to the International Consensus Classification, the category was renamed to “myeloid/lymphoid neoplasms with eosinophilia and tyrosine kinase gene fusions” and both classifications added novel subtypes with new JAK2 rearrangements (e.g., BCR::JAK2, ETV6::JAK2) as well as fusions involving FLT3, and the ETV6::ABL1 fusion [2, 3]. Although eosinophilia (>0.5 × 109/l) or hypereosinophilia (>1.5 × 109/l) are characteristic of this subgroup, they are not universally present [4]. The clinical phenotype is largely influenced by the involved tyrosine kinase fusion gene and/or the fusion partner gene [5, 6]. For example, most patients with the FIP1L1::PDGFRA fusion gene present with a chronic myeloid neoplasm with eosinophilia; however, mixed lineage presentations are more common in patients with FGFR1 fusions [5,6,7,8]. Furthermore, in MLN with FGFR1 rearrangements, translocations involving ZMYM2 are more commonly associated with a T-lymphoblastic lymphoma phenotype, whereas translocations involving BCR tend to lead to a phenotype resembling BCR::ABL1 positive chronic myeloid leukemia [5, 6, 9].

Treatment of patients with MLN and rearrangements of PDGFRA, PDGFRB, or FGFR1 is dependent upon the involved tyrosine kinase fusion gene [5, 10, 11]. Imatinib is associated with complete hematologic, cytogenetic, and molecular responses in patients with PDGFRA- and PDGFRB-rearranged MLNs and is approved by the US Food and Drug Administration for these indications [4, 10,11,12]. Pemigatinib was approved for relapsed/refractory MLN with FGFR1 rearrangement in August 2022 [13].

Here we propose comprehensive response criteria based on the heterogenous clinical presentations of patients with MLN with eosinophilia and tyrosine kinase gene fusions. The MLN International Working Group (MLN IWG) was formed to adjudicate diagnoses and treatment responses in the FIGHT-203 study of pemigatinib in MLN with FGFR1 rearrangements.

MLN with FGFR1 rearrangements

MLN with FGFR1 rearrangement was previously known as 8p11 myeloproliferative syndrome [14]. The defining cytogenetic abnormality, a translocation at the 8p11 locus, was found to involve the FGFR1 gene [14]. Table 1 lists the diagnostic criteria for MLN with FGFR1 rearrangement. Clinical presentation can be in the form of chronic phase (CP) of a myeloid neoplasm detected in the bone marrow (BM)/peripheral blood (PB) (e.g., myeloproliferative neoplasm [MPN], myelodysplastic syndrome [MDS], or MDS/MPN), or blast-phase (BP) disease detected in the BM/PB (e.g., acute myeloid leukemia [AML], T- or B-cell acute lymphoblastic leukemia [ALL], mixed phenotype acute leukemia [MPAL]), and/or extramedullary disease (EMD) that is recognized as a BP component [4, 5, 7, 8]. Different phases and lineages of the disease can be seen in the same patient (e.g., chronic myeloid neoplasm in the bone marrow with concomitant T-cell acute lymphoblastic lymphoma in an EMD site) [4, 5]. Further, both primary BP disease and secondary BP disease as a consequence of rapid progression from CP, usually within 1–2 years, are reported in many patients [6].

Table 1 Diagnostic criteria for MLN with FGFR1 rearrangement [1,2,3].

Sixteen fusion partners to FGFR1 have been reported, with t(8;13)(p11;q12) involving ZMYM2 being the most common [6, 15, 16]. FGFR1-associated translocations or alterations can be identified with conventional cytogenetic analysis and confirmed with fluorescence in situ hybridization (FISH) using FGFR1 break-apart probes [5, 11]. Infrequently, FGFR1 rearrangements are cryptic by conventional cytogenetic analysis and can only be detected by FISH, reverse transcriptase-polymerase chain reaction (RT-PCR), and/or next-generation sequencing analysis [5, 17]. Fig. 1 summarizes the known gene fusion partners to FGFR1 and their respective translocations.

Fig. 1: FGFR1 fusion partners.
figure 1

Sixteen fusion partners of FGFR1 have been currently characterized. Chromosome breakpoints for the fusion partners are shown below each partner gene. ZMYM2 on chromosome 13q12 and BCR on chromosome 22q11 are the most common fusion partner genes of FGFR1.

Treatment with multikinase inhibitors with nonspecific anti-FGFR1 activity, including ponatinib and midostaurin, only provide short-term hematologic responses and rarely result in cytogenetic responses [6, 8, 18, 19]. Current treatment of patients with CP disease includes hydroxyurea or one of the multikinase inhibitors with nonspecific anti-FGFR1 activity [7, 8]. Treatment of patients with BP disease includes intensive induction chemotherapy followed by allogeneic hematopoietic stem cell transplantation in patients achieving disease control [5, 7, 8]. However, the option of allogeneic hematopoietic stem cell transplantation may be limited by patient age and comorbidities and lack of response to chemotherapy. In a review of 45 patients, 14 with CP and 31 with BP, the 1-year overall survival was 43.1%, and 46.2% of patients with CP disease progressed to BP at 1 year [7].

Fight-203 study

FIGHT-203 is a phase 2, open-label, multicenter study evaluating the efficacy and safety of pemigatinib (INCB054828) in adult patients with MLN with FGFR1 rearrangements [20]. Pemigatinib is a selective and potent inhibitor of FGFR 1–3 and is approved for the treatment of patients with advanced or metastatic cholangiocarcinoma with FGFR2 rearrangements [21, 22].

Patients enrolled in FIGHT-203 have a documented MLN with an 8p11 translocation on standard karyotyping and/or evidence of an FGFR1 rearrangement on break-apart FISH. Most patients enrolled in the study had at least one prior therapy; however, treatment-naive patients were also enrolled [20]. The primary endpoint of FIGHT-203 is complete clinical response (CR) rate. Secondary endpoints include overall response rate defined as the percentage of patients who achieved a best overall response of CR or partial response (PR), and cytogenetic response rates based on conventional cytogenetics or break-apart FISH (complete cytogenetic response and partial cytogenetic response) [20]. Primary and secondary endpoints were assessed by the investigators according to protocol-defined criteria [20]. In addition, a Central Review Committee (CRC), also known as the MLN IWG, consisting of hematopathologists and hematologists, convened regularly to retrospectively review and adjudicate diagnoses and responses. The CRC developed comprehensive response criteria based on the heterogeneous clinical presentations of patients enrolled in the trial. The criteria are a composite of previously published response criteria for MDS/MPN, acute leukemia, and lymphoma [23,24,25]. During their review, the CRC members discussed the histopathologic, laboratory, and radiologic results, and arrived at a consensus decision to assign patients to respective categories of clinical presentation. For the adjudication of each clinical and cytogenetic response, committee members reviewed and discussed the data, and assigned a response by consensus.

Categories of clinical presentation

During the CRC adjudication, it became evident that response criteria were needed that could address both the CP and BP presentations, as well as the potential presence of EMD.

The following five potential clinical presentation categories may be seen in MLN with FGFR1 rearrangement that are also applicable to other MLNs with tyrosine kinase fusion genes: (1) CP disease involving the BM/PB without EMD; (2) CP disease involving the BM/PB with concurrent EMD; (3) BP disease involving the BM/PB without EMD; (4) BP disease involving the BM/PB with EMD; and (5) EMD only (ie, without evidence of BM/PB involvement). In addition to these five presentation categories, and because the majority of patients had received therapy before enrollment in the study, the CRC recognized a sixth category to reflect those patients with evidence of persistent 8p11 cytogenetic abnormality/FGFR1 rearrangement but without morphologic and/or radiologic evidence of disease in the BM/PB or EMD.

Response criteria based on clinical presentation

Criteria for chronic-phase disease in the bone marrow and peripheral blood

The CP disease response criteria adopted by the CRC (Table 2) are modified from the IWG for MDS/MPN proposed criteria and require evaluation of the spleen/liver by palpation, BM, PB smear, and complete blood count with differential [23]. Criteria for CR, complete response with partial hematologic recovery (CRh), PR, stable disease, loss of response, and progressive disease (PD) are summarized in Table 2.

Table 2 Response criteria for chronic-phase disease in the bone marrow/peripheral blood.

A CRh category was included in the response criteria. CRh meets the criteria for CR except there is no requirement for normal age-adjusted cellularity and allows cytopenias defined as hemoglobin ≥8 g/dL, platelet count ≥50 × 109/L, and absolute neutrophil count ≥0.5 × 109/L. Historically, this category of “less than CR” was first introduced in AML response criteria with the progressive use of less intensive therapies in the treatment landscape [26], where patients receive treatment up to the day of response assessment. Consequently, myelosuppressive effects of therapy may confound response assessment by preventing full recovery of blood counts in the absence of morphologic evidence of AML [26]. Similarly, worsening cytopenias were observed with the use of KIT inhibitors in advanced systemic mastocytosis, and the CRh category was introduced to the modified IWG response criteria used in the evaluation of avapritinib in advanced systemic mastocytosis [27]. In the context of advanced systemic mastocytosis, the CRh category recognizes that in the absence of evidence of systemic mastocytosis due to successful treatment, persistently low blood counts may instead relate to treatment-associated myelosuppression or the presence of a concomitant-associated hematologic neoplasm.

Criteria for blast-phase disease in the bone marrow and peripheral blood

The response criteria for BP disease in the BM were largely based on the published response criteria for acute leukemia as summarized in Table 3 [25]. PR was modified to include partial hematologic recovery consisting of (1) an absolute neutrophil count >0.5 × 109/L and (2) platelet count >50 × 109/L.

Table 3 Response criteria for blast-phase disease in the bone marrow/peripheral blood.

Extramedullary disease

The response criteria pertaining to EMD (Table 4) are based on modified Lugano criteria [24]. The presence of splenic and/or liver enlargement was not considered EMD, but instead was evaluated under the CP disease response criteria. Consequently, components of the Lugano criteria pertaining to organ enlargement were not included. However, the presence of discrete splenic and/or hepatic lesions was considered EMD (extralymphatic lesions). Similarly, responses in the BM are addressed by the CP- and BP-specific criteria.

Table 4 Response criteria for EMD.

Overall clinical response based on phase of the disease and involved compartment(s)

The CRC developed composite response criteria for overall clinical response for (1) CP disease in the BM/PB with or without presence of EMD (Table 5); (2) BP disease in the BM/PB with or without the presence of EMD (Table 6); or (3) EMD only since this also represents BP disease (Table 6). For CP disease in the BM/PB (Table 5), six overall response categories are noted: CR, CRh, PR, stable disease, loss of response, and progressive disease. For BP disease, seven overall response categories are possible: CR, complete response with incomplete hematologic recovery, morphologic leukemia-free state, PR, stable disease, loss of response, and progressive disease. In both CP and BP diseases, a guiding principle is that overall clinical response is anchored to the lowest quality response among the BM/PB and the EMD disease components.

Table 5 Overall clinical responses in chronic-phase disease with or without EMD.
Table 6 Overall clinical responses in blast-phase disease with or without EMD, or EMD only.

Cytogenetic response (by cytogenetics and fish) and molecular responses

Cytogenetic responses were assessed separately from clinical responses. The criteria for cytogenetic response were developed based on the cytogenetic response criteria proposed for MDS/MPN and are summarized in Table 7 [23]. Molecular responses were defined based on the detection of FGFR1 fusion transcripts using either semiquantitative or quantitative RT-PCR assays (Table 7). In contrast to the use of international scale, which harmonizes quantitative RT-PCR evaluation of BCR::ABL1, no such standardization for molecular monitoring of FGFR1 and other fusion genes currently exists, but is a high priority for future development. For patients who were enrolled in the FIGHT-203 study with persistent FGFR1 rearrangement but no morphologic evidence of disease (due to receiving prior therapy), only cytogenetic and/or molecular responses were evaluated.

Table 7 Response criteria for cytogenetic and molecular responses.

Concluding remarks

Heterogeneous clinical presentations are observed in patients with MLNs with FGFR1 rearrangement and other tyrosine kinase fusion genes. This clinical variability presents a challenge for diagnosis and assessment of response. The FIGHT-203 study is the first prospective trial of targeted therapy in MLN with FGFR1 rearrangement and provided a unique opportunity to generate response criteria, which could adequately address the variable presentations of these diseases. This phenotypic diversity reflects differences in disease acuity (CP vs BP disease), lineage (myeloid vs lymphoid vs mixed phenotype disease), and the variable presence of EMD. We found that these response criteria permit adjudication of the manifold presentations of MLNs, including CP and BP disease with or without EMD, or EMD only. Although these criteria were generated in the context of the FIGHT-203 study of pemigatinib for MLN with FGFR1 rearrangement, they can also be used to assess therapies for other MLNs with tyrosine kinase fusion genes, including PDGFRA, PDGFRB, JAK2, FLT3, and ETV6::ABL1. In addition, the response criteria can be applied outside of trials because they incorporate commonly used histopathologic, cytogenetic/FISH, and imaging techniques.

The use of FISH testing is a key adjunct in the diagnosis and follow-up of these disorders, especially when banded metaphases cannot be obtained or are inadequate in number, but its use in response assessment is affected by the different normal cutoffs for different probes and lack of standardized definition of “cytogenetic FISH” response. Therefore, future studies are needed to confirm if there is a difference in FISH results between BM and PB samples and to confirm the correlation between karyotype and FISH results. Similarly, molecular analysis of FGFR1 fusion transcripts by RT-PCR using a semiquantitative or quantitative assay has not been standardized and is currently not widely available.

These newly proposed response criteria require evaluation in future prospective clinical trials, including whether the categories of response within CP and BP disease correlate with long-term endpoints such as progression-free survival and overall survival.