Efficacy of tyrosine kinase inhibitors in Ph-like acute lymphoblastic leukemia harboring ABL-class rearrangements

Philadelphia(Ph)-like acute lymphoblastic leukemia (ALL) is a high-risk subtype of B-precursor ALL characterized by a gene expression profile (GEP) similar to Ph-positive ALL but lacking the specific BCR-ABL1 fusion gene.1,2 Several pediatric and adult groups have reported on the poor baseline features of Ph-like ALL including a high white blood cells count (WBC) at diagnosis and a frequent association with IKZF1 intragenic deletion. Patients with Ph-like ALL are at higher risk of induction failure or high post-induction minimal residual disease (MRD) levels.3–5 This poor early response translates into inferior outcome in terms of event-free survival (EFS) and overall survival (OS) as compared to other B-ALL patients. Ph-like ALL was recognized as a provisional entity by the 2016 WHO classification but strict and unequivocal diagnostic criteria have not been established yet.6 Ph-like ALL are characterized by a multitude of oncogenic events that lead to the aberrant activation of cytokine receptors or signaling factors, the most frequent beeing rearrangements of CRLF2, fusions and mutations of JAK kinases and fusions involving ABL-class kinases (ABL1, ABL2, CSF1R or PDGFRB).7 As expected, Ph-like cells harboring ABL-class rearrangements have shown sensitivity to tyrosine kinase inhibitors (TKIs) such as imatinib and dasatinib in vitro or in patient-derived xenograft (PDX).8–10 In addition, we and others have reported single experiences of clinical benefit of TKI treatment in early resistant patients.11,12 In an attempt to enable targeted therapy in B-ALL patients with poor response to chemotherapy, we have developped an integrative diagnostic strategy to identify Ph-like alterations in newly diagnosed or relapsing patients in a timely manner. Here, we report on the characteristics, response and outcome of 24 patients with B-ALL harboring ABL-class fusions who could be treated with a combination of TKI and chemotherapy, either during frontline treatment (N=19) or at relapse (N=5). Patients were initially enrolled in or treated according to pediatric FRALLE/CAALL/EORTC (NCT02716233; NCT01185886) or adult GRAALL/EWALL trials (NCT00327678; NCT02617004) in french clinical centers. Cytogenetic analyses were performed locally according to standard recommendations. Molecular analyses were performed centrally and evolved over time with the introduction of new techniques, up to the current algorithm shown in Supplementary Figure 1. Briefly, fusion transcripts are searched for using an in-house multiplexed method (derived from 13) and cases negative for classifying alterations are subjected to RNA-sequencing for detection of alternative fusions. Of the 24 patients reported here, 12 had fusions involving ABL1, including NUP214-ABL1 (n=6), ETV6-ABL1 (n=3) and other partners in single cases (RCSD1, RANBP2 and LSM14A) (Figure 1A). Nine cases had a PDGFRB rearrangement, including EBF1-PDGFRB (n=6) and other partners in single cases (NUMA1, ETV6 and ATF7IP). ZC3HAV1-ABL2 and MEF2D-CSF1R fusions were identified in single cases, and a patient with a ZMYM2-FGFR1 fusion (not strictly ABL-class) was also included in the present cohort. The resulting cohort (Table 1) included 16 males and 8 females and median age at diagnosis was 24 years (range 5-72). Two patients were previously reported (#1,24).11,14 As expected, patients had baseline characteristics and early response to treatment associated with a poor prognosis. Median WBC was 30x109/L (range 4-570x109/L). Intragenic IKZF1 deletions were detected in 11 out of 24 patients (46%). A poor response to prednisone prephase (≥1 G/L blasts in peripheral blood at day 8) was observed in 14 out of 22 evaluable patients (64%). After induction therapy, only 16 out of 24 (67%) of patients reached complete remission, all with detectable minimal residual disease (MRD), including 7 with MRD ≥10-2, and only one with MRD <10-4 (Table 1, Figure 1B). In 19 patients, the ABL-class fusion was identified at initial diagnosis work-up and TKI was introduced during frontline treatment, within the first month of consolidation or salvage for most of them (n=14). In 5 patients, the Ph-like status was diagnosed at relapse and TKI was introduced in association with salvage therapy. In this retrospective study, the choice of TKI, TKI dosage, and combination was up to the physician choice. Fourteen patients were exposed to imatinib, 9 to dasatinib, one to ponatinib. Three patients were switched from imatinib to dasatinib during frontline treatment (Table 1). Among the 19 patients treated with TKI frontline, 7 out of 8 primary refractory patients subsequently achieved CR, one after a switch to dasatinib (#10). One patient died early of sepsis in a context of uncontrolled disease (#16). One patient was lost of sight after salvage until she relapsed (#13). In 14 out of 18 patients (78%), a MRD level below 10-4 was achieved within a median time of 2.5 months (range 1.4-14.8) (Figure 1B). Allogeneic hematopoietic stem cell transplant (HSCT) was performed in 9 patients, of whom 3 with a sibling donor (SIB), 4 with a matched unrelated donor (MUD), and 2 with a haploidentical donor (Haplo). All patients were in CR before HSCT and 6/9 (67%) had undetectable MRD. One patient was additionally exposed to blinatumomab in combination to dasatinib in bridge to HSCT. After a median follow-up of 36 months (range 8-73), 12 patients were alive in first CR. Six patients relapsed, 3 patients received an alternative TKI, including one in association to inotuzumab. The median remission duration and OS were not reached. At 3 years, EFS was 55% (95%CI: 27-76) and OS was 77% (95%CI: 50-91) (Figure 1C-D). The 5 patients who were treated with TKI at relapse achieved CR, including two patients who had refractory disease to several lines of treatment (#22,24). A MRD level below 10-4 was achieved in 3 patients of whom two could proceed to HSCT (1 Haplo, 1 MUD) and remained alive in remission. The three other patients further relapsed and died of progressive disease after exposition to second-line TKIs (dasatinib, n=2; ponatinib, n=1). The efficacy of TKIs in Ph-like ALL has been suggested by several case reports.11,12,15–17 In a large pediatric and adult cohort, Roberts et al. mentioned 11 patients with Ph-like ALL and slow response or failure to induction who achieved rapid responses upon TKI or ruxolitinib therapy.3 In frontline patients, the 3-year OS of 77% we observe compares favorably with retrospective survival rates of Ph-like adult patients treated without TKI. In the GMALL group experience in 19 adult patients with Ph-like ALL and a median age of 31 years old (range 16-59), the 5-year OS was 22%.19 In the MDACC experience, the 5-year survival of 56 Ph-like adult patients with a median age of 33.5 years (range 15-71) treated with either Hyper-CVAD or augmented BFM was 23%.7 Of note, these series included all Ph-like ALL cases characterized by GEP, and no specific subgroup analysis was performed in patients with ABL-class fusion genes. In conclusion, we report the largest cohort of patients with ABL-class kinase rearrangement exposed to TKI frontline or at relapse, and show promising MRD response and outcome similar to those observed in early trials of imatinib combined with chemotherapy in Ph-positive ALL.18 Prospective screening strategies are feasible and should be generalized to identify these high-risk patients and to propose early TKI-based intervention. In future studies, several questions remain to be adressed including the choice of TKI according to the fusion transcript, whether these patients should benefit from recently approved blinatumomab19, and finally, the benefit of HSCT in patients who achieve good MRD response upon targeted therapy.