Avapritinib is effective for treatment of minimal residual disease in acute myeloid leukemia with t (8;21) and kit mutation failing to immunotherapy after allogeneic hematopoietic stem cell transplantation

In patients with t(8;21) acute myeloid leukemia (AML) with recurrent measurable residual disease (MRD) after allogeneic hematopoietic stem cell transplantation (allo-HSCT), pre-emptive interferon-α therapy and donor lymphocyte infusion are noneffective in 30%–50% of patients. Avapritinib is a novel tyrosine kinase inhibitor targeting KIT mutations. We retrospectively report about 20 patients with t(8;21) AML and KIT mutations treated with avapritinib after allo-HSCT with MRD and most failing to respond to immunotherapy. Reduction of RUNX1-RUNX1T1 after 1 month of treatment was ≥1 log in 12 patients (60%), which became negative in 4 patients (20%). In 13 patients who received avapritinib for ≥3 months, the reduction was ≥1 log in all patients, which became negative in 7 patients (53.8%). The median follow-up time was 5.5 (2.0–10.0) months from avapritinib initiation to the last follow-up. Three patients underwent hematologic relapse and survived. Among all 20 patients, RUNX1-RUNX1T1 transcripts turned negative in 9 patients (45%). The efficacy did not differ significantly between D816 and non-D816 KIT mutation groups. The main adverse effect was hematological toxicity, which could generally be tolerated. In summary, avapritinib was effective for MRD treatment in patients with t(8;21) AML with KIT mutations failing to respond to immunotherapy after allo-HSCT.

Studies have demonstrated that the presence of KIT mutations is associated with decreased relapse-free survival (RFS) and overall survival (OS) in core-binding factor (CBF)-AML [10,11]. KIT is located on chromosome 4q12 and encodes a class III transmembrane receptor tyrosine kinase, which mediates downstream signal pathways involved in cell proliferation, differentiation, and survival. Mutations in KIT affecting its activation cause ligandindependent constitutive KIT activation, which contributes to leukemic transformation and has been reported in approximately one-third of adult patients with CBF-AML [10,12]. Therefore, in the National Comprehensive Cancer Network and European Leukemia Net recommendations, clinical trials for these molecular abnormalities are recommended [13].
Avapritinib is an oral tyrosine kinase inhibitor (TKI) that was designed to selectively target KIT and platelet derived growth factor receptor α (PDGFα) A-loop mutants [14]. The introduction of avapritinib has changed the prognostic outlook of advanced systemic mastocytosis and gastrointestinal stromal tumors verifying its efficacy in patients resistant to midostaurin [15]. These results highlighted the potential for avapritinib to be developed clinically for an AML population driven by constitutively active kinase conformations of KIT, especially the D816V mutation. There are few case reports on the efficacy and safety of avapritinib in AML patients with KIT mutations receiving allo-HSCT [16,17]. Therefore, we retrospectively analyzed the clinical data of 20 patients with t(8;21) AML treated with avapritinib after allo-HSCT at our center, most of whom also did not respond to immunotherapy, including IFN-α treatment and DLI. Additionally, it is not clear whether avapritinib has a consistent antagonistic effect against different KIT mutations; therefore, we also compared treatment responses in patients with different KIT mutations. To the best of our knowledge, this is the first clinical study to analyze the therapeutic efficacy of avapritinib in patients with t(8;21) AML carrying KIT mutations after allo-HSCT and failing immunotherapy in some patients.

SUBJECTS AND METHODS Patients and avapritinib treatment
Since 2021, at our center, avapritinib treatment was recommended for patients with t(8;21) AML and KIT mutations failing to respond to preemptive immunotherapy, including IFN-α treatment and DLI, after allo-HSCT. Avapritinib was also recommended for patients with RUNX1-RUNX1T1 transcript level of ≥0.1% within 60 days after allo-HSCT or who had uncontrolled GVHD. A total of 20 patients agreed and received avapritinib treatment. This retrospective study included these 20 consecutive patients who were treated with avapritinib for RUNX1-RUNX1T1-positive AML after allo-HSCT in the Peking University People's Hospital, Institute of Hematology, from October 2021 to July 2022. These 20 patients with t(8;21) AML had KIT mutations at diagnosis. The initial dose of avapritinib was 50 or 100 mg once daily according to patients' body weight of <50 or ≥50 kg, respectively. Avapritinib was administered in continuous 28-day cycles, and patients continued treatment until episodes of unacceptable toxicity, progressive disease, death, noncompliance, or physician decision. The study protocol was approved by the Ethics Committee of Peking University People's Hospital, and informed consent was obtained from all patients. Follow-up information was collected until September 30, 2022.

Transplantation protocol
All patients in this study received myeloablative conditioning regimens. Haploidentical HSCT and matched sibling donor transplantation were performed according to protocols reported previously by our institute [18][19][20]. The conditioning regimen for matched sibling donor transplantation patients was as follows: cytarabine 2 g/m 2 /day intravenous delivery for 1 day, cyclophosphamide 1.8 g/m 2 /day for 2 days, busulfan 0.8 mg/kg intravenous delivery, four times a day for 3 days, and nitrosourea (Semustine) 250 mg/kg for 1 day. The conditioning regimen for haploidentical HSCT patients was as follows: cytarabine 4 g/m 2 /day intravenous delivery for 2 days, cyclophosphamide 1.8 g/m 2 /day for 2 days, busulfan 0.8 mg/kg intravenous delivery, four times a day for 3 days, and nitrosourea 250 mg/kg for 1 day, and thymoglobulin (ATG, Sang Stat, Lyon, France) 2.5 mg/kg/day intravenous delivery for 4 days.

Measurement of RUNX1-RUNX1T1 transcript levels
Bone marrow samples were collected at the time of diagnosis, before every cycle of chemotherapy, and then at 3-month intervals for 2 years and at 6-month intervals for another 2 years. Bone marrow samples were collected at 1, 2, 3, 4.5, 6, 9, and 12 months after allo-HSCT and at 6-month intervals thereafter. Real-time quantitative reverse transcription polymerase chain reaction (RT-qPCR) was performed to quantitatively measure RUNX1/RUNX1T1 transcript levels. A direct sequencing method was performed to screen for c-KIT mutations.

Pre-emptive immunotherapy protocols (IFN-α therapy and DLI)
Some patients with RUNX1-RUNX1T1-positive AML received pre-emptive IFN-α therapy or DLI before hematologic relapse after allo-HSCT [21]. The therapeutic option was primarily based on donor availability and the intentions of physicians and patients. The patients received recombinant human IFN-α therapy by subcutaneous injection twice a week every 4 weeks [7]. IFN-α therapy was scheduled for six cycles or until the RUNX1-RUNX1T1 transcripts were negative in at least two consecutive tests. IFN-α therapy could be prolonged upon the request of patients. IFN-α therapy was discontinued in patients with grade ≥3 toxicity, severe infection, severe GVHD, non-relapse mortality (NRM), or relapse. Granulocyte colonystimulating factor-mobilized peripheral blood stem cells were administered instead of unstimulated donor blood lymphocytes [22]. The doses of mononuclear cells was 1×10 8 /kg. After DLI, subjects received immunosuppressive drugs such as CSA or MTX to prevent GVHD [22]. All patients received short-term immunosuppressive drugs after DLI. Patients could receive chemotherapy 48-72 h before DLI [22].

Statistical analysis
The primary study end point was the decrease in RUNX1/RUNX1T1 transcript levels ≥1 log compared with before treatment. The secondary end points were negativity of RUNX1/RUNX1T1, overall survival, and relapse-free survival. Summary statistics, such as proportions, median, and ranges, are used to describe patient characteristics and outcomes. The associations between RUNX1/RUNX1T1 expression and post-transplantation outcomes were analyzed by the Kaplan-Meier method. A two-sided P value of 0.05 was considered statistically significant. The independence of categorical parameters was calculated using the chi-square test or Fisher's exact test, and the distribution of continuous variables was calculated using the Mann-Whitney U test. All statistical analyses were performed using SPSS 23.0 (Chicago, IL, USA).

Avapritinib treatment
In this retrospective study, before starting avapritinib treatment, 7 patients (35%) received IFN-α therapy only, 6 (30%) received modified DLI only, and 3 (15%) received both DLI and IFN-α therapy. Avapritinib was also given for 4 patients with RUNX1-RUNX1T1 ≥ 0.1%, who were within 60 days after allo-HSCT, not suitable for IFN-α or DLI therapy. Avapritinib was initiated at a median of 4.5 (1.0-21.0) months after allo-HSCT. The starting dose was, in most cases (n = 18), 100 mg daily, except two child patients who received 50 mg daily. No patients required an increase in the dose of the drug during treatment. The median duration of treatment was 5.0 (1.0-9.0) months. The median RUNX1-RUNX1T1 transcript level before avapritinib treatment was 0.40% (0.01-46.90%) ( Table 2). All patients were tested for KIT mutations while testing the RUNX1-RUNX1 gene and were negative for KIT mutations in all the 20 patients.

Efficacy of avapritinib
The median RUNX1-RUNX1T1 transcript level after 1 month of avapritinib treatment was 0.012% (0.0-188.0%). Compared with the pretreatment RUNX1-RUNX1T1 level, the reduction of RUNX1-RUNX1T1 after 1 month of avapritinib treatment was ≥1 log in 12 patients (60%), ≥2 log in 5 patients (25%), and became negative in 4 patients (20%) (Fig. 1). In the <1 log reduction of RUNX1-RUNX1T1 group (n = 8), 4 patients stopped avapritinib after 1 month of treatment. Among these 4 patients, three patients stopped avapritinib due to ineffective treatment. The level of RUNX1-RUNX1 in these three patients increased by 2-4 times respectively after 1 month of treatment compared with that before treatment. All the three patients underwent hematologic relapse (HR, bone marrow blasts >5%). Two patients progressed to HR after one month of treatment, and the other patient progressed to HR 2 months after drug withdrawal. The fourth patient stopped avapritinib for financial reasons and received DLI treatment. In the ≥2 log reduction of RUNX1-RUNX1T1 group, 3 patients stopped avapritinib treatment owing to financial reasons (Table 2). Among these 3 patients, one patient chose to receive IFN-α therapy after discontinuation of avapritinib and developed chronic GVHD. After discontinuation of avapritinib for 4 months, RUNX1-RUNX1T1 remained negative. The other patient stopped avapritinib without other treatment, and RUNX1-RUNX1T1 remained negative for 4 months after avapritinib discontinuation. The RUNX1-RUNX1T1 of the third patient decreased by 2 log after 1 month of avapritinib treatment but did not turn negative, and the RUNX1-RUNX1T1 turned negative after 2 months of avapritinib withdrawal.
The remaining 13 patients were treated with avapritinib for >3 months. The median RUNX1-RUNX1T1 transcript level after 3 months of avapritinib treatment was 0.0% (0.0-0.051%). Compared with the pretreatment RUNX1-RUNX1T1 transcript level, the reduction of RUNX1-RUNX1T1 transcript level after 3 months of avapritinib treatment was ≥1 log in all 13 patients (100.0%), ≥2 log in 11 patients (84.6%), and became negative in 7 patients (53.8%) (Fig. 2). Among the remaining 13 patients, one patient achieved RUNX1-RUNX1T1 negative after 4 months of avapritinib treatment and stopped the treatment after 5 months owing to financial reasons, and RUNX1-RUNX1T1 transcripts remained negative for 9 months after the discontinuation of avapritinib (Table 2).  Discontinuation of avapritinib treatment owing to financial reasons, n (%)
The median follow-up time was 5.5 (2.0-10.0) months from initiation of avapritinib treatment to the last follow-up. All of the 20 patients were still alive.The RUNX1-RUNX1T1 turned negative in 4 patients after 1 month of avapritinib treatment, and in 7 patients after 3 months of avapritinib treatment. And with the prolongation of avapritinib treatment, two other patients achieved RUNX1-RUNX1T1 negative after 4 and 5 months of avapritinib treatment respectively. Therefore, of the 20 patients, RUNX1-RUNX1T1 transcripts turned negative in 9 patients (45%). Overall, the median time from the initiation avapritinib treatment to RUNX1-RUNX1T1 turning negative was 2.0 (1.0-5.0) months ( Table 2).

Effect of KIT mutations on the efficacy of avapritinib
In the 13 patients with D816 mutation (sole or combined), the reduction of RUNX1-RUNX1T1 transcript level after 1 month of avapritinib treatment was≥1 log in 8 patients (61.5%), and turned negative in 3 patients (23.1%). In the 7 patients with a non-D816 mutation, the reduction of RUNX1-RUNX1T1 transcript level after 1 month of avapritinib treatment, was ≥1 log in 4 patients (57.1%), and turned negative in 1 patient (14.3%).
Among the 13 patients receiving avapritinib for more than 3 months, 10 patients with the D816 mutation (sole or combined) showed reduction of RUNX1-RUNX1T1 transcript level after 3 months of avapritinib treatment, which turned negative in 5 patients (50.0%). In the 3 patients with a non-D816 mutation, the reduction of RUNX1-RUNX1T1 transcript level after 3 months of avapritinib treatment turned negative in 2 patients (66.7%). The efficacy of avapritinib did not differ significantly between the D816 mutation and non-D816 mutation groups ( Table 3).

Adverse events after avapritinib treatment
Three of the 20 patients developed grade ≥3 neutropenia, 2 developed grade ≥3 anemia, and 7 developed grade ≥3 thrombocytopenia independently of pretreatment status during avapritinib treatment, which led to treatment interruption. The common non-hematological side effects of avapritinib were grade 1-2 edema in 8 cases (40%) and grade 1-2 fatigue in 9 cases (45%). Two cases (10%) reported grade 1-2 acroanesthesia (Table 5). These adverse events were thought to possibly be related to avapritinib treatment. Avapritinib-induced myelosuppression could be reverted with dose interruption/dose reduction in all patients. In the two patients who stopped avapritinib due to hematological toxicity, RUNX1-RUNX1T1 turned negative after treatment, but RUNX1-RUNX1T1 turned positive 1 month after drug withdrawal due to hematological toxicity, and RUNX1-RUNX1T1 turned negative 1 month after restarting treatment.

DISCUSSION
To the best of our knowledge, this is the first clinical study involving the largest number of patients undergoing allo-HSCT that retrospectively analyzed the therapeutic efficacy of avapritinib in patients with t(8; 21) AML carrying KIT mutations after allo-HSCT most failing to immunotherapy. We found that in patients with t(8; 21) AML with KIT mutations who were RUNX1-RUNX1T1 positive after allo-HSCT most failing to immunotherapy, avapritinib could decrease RUNX1-RUNX1T1 transcript levels, and some patients even achieved RUNX1-RUNX1T1-negative conversion.
Avapritinib has rapid effects, which was demonstrated in our study, as 80% of patients experienced a decrease in RUNX1-RUNX1T1 transcript level after 1 month of administration, 60% experienced a decrease in RUNX1-RUNX1T1 transcript level by ≥1 log, and 20% achieved RUNX1-RUNX1T1-negative conversion. With the prolongation of treatment time, the efficacy rate in patients increased. In patients who received avapritinib for >3 months, RUNX1-RUNX1T1 transcript level decreased by more than 1 log in all patients, and 46.2% of patients achieved RUNX1-RUNX1T1-negative conversion. Avapritinib was ineffective after 1 month of treatment in four patients with RUNX1-RUNX1T1 transcript level of >1% before treatment. Xue et al. reported a case wherein a patient with t(8;21) AML and a KIT D816I mutation was treated with avapritinib (dose 200 mg/d) following relapse after allo-HSCT. After 1 month of treatment, RUNX1-RUNX1T1 transcript level decreased from 109.4% to 0.13% [17]. The patient achieved RUNX1-RUNX1T1-negative status after 1.5 months of treatment. Additionally, Yin et al. reported four cases of t(8;21) AML with a KIT D816 mutation being treated with avapritinib (dose 100-200 mg/ day) combined with other chemotherapy drugs [16]. Therefore, there is evidence to support increasing the drug dose or combining avapritinib with other drugs under the condition of patient tolerance to improve its efficacy.
In this study, avapritinib was given for 4 patients within 60 days after allo-HSCT. Since clearance of MRD after allo-HSCT can take some time paralleling the establishment of graft-versus-leukemia (GvL) effect and GvHD prophylaxis reduction, the observed MRD evolution could not be totally attributed to avapritinib in these patients but to to a GvL underlying reaction. Our previous study found that RUNX1-RUNX1T1 reduction <3-log at 1, 2, and 3 months after HSCT identified high relapse patients [8]. In the future prospective study, we will fix a threshold to start preemptive strategy with avapritinib, such as an increasing MRD load (>1 log) in two separate samples to distinguish MRD persistence/relapse from low level persisting MRD.
It is not clear how long avapritinib treatment should be maintained to achieve effective treatment for patients with t(8;21) AML. In our study, one patient achieved RUNX1-RUNX1T1-negative conversion after 4 months of treatment and stopped avapritinib due to a financial reason after 5 months. The patient had chronic GVHD, and RUNX1-RUNX1T1 continued to be negative after 5 months of drug withdrawal. In another patient, RUNX1-RUNX1T1 turned negative after 1 month of avapritinib treatment, and avapritinib was discontinued due to a financial reason. This patient selected IFN-α treatment and developed moderate chronic GVHD. After 4 months of avapritinib withdrawal, RUNX1-RUNX1T1 continued to be negative. One patient stopped avapritinib after 1 month treatment due to financial reason with RUNX1-RUNX1T1 turning negative, and RUNX1-RUNX1T1 remained negative for 4 months after avapritinib discontinuation. In the other two   patients, RUNX1-RUNX1T1 turned negative after treatment, but RUNX1-RUNX1T1 turned positive 1 month after drug withdrawal due to hematological toxicity, and RUNX1-RUNX1T1 turned negative 1 month after restarting treatment. It is not clear whether different KIT mutations affect the efficacy of avapritinib. We found that there was no significant difference in the overall response rate between patients with KIT-D816 mutations and patients with non-D816 KIT mutations. Our study showed that avapritinib could be used in patients with KIT mutations other than D816.
In patients receiving avapritinib treatment after allo-HSCT, the most serious adverse effect of avapritinib was myelosuppression, and thrombocytopenia had the highest incidence, which reached grade 3-4 in 35% of patients but could be recovered after drug withdrawal. The main non-hematological adverse reactions were edema and fatigue, which were relatively mild and tolerable. Xue et al. reported that grade 3 leukopenia and thrombocytopenia developed during avapritinib treatment but were resolved with supportive treatment [17]. Yin et al. reported that two out of four patients developed grade 3 febrile neutropenia, grade 3 anemia, and grade 4 thrombocytopenia during avapritinib treatment in combination with other chemotherapy drugs [16].
Our study has some limitations. First, our study is a retrospective study, and there may be some bias in the selection of patients for receiving avapritinib treatment. Second, the number of patients included in our study is currently the largest, it is relatively small to evaluate the clinical efficacy of avapritinib. Third, the observation time was not long enough to investigate the long-term efficacy and maintenance of treatment time with avapritinib. Consequently, further observational studies are needed to answer these questions.
In conclusion, our results indicate that avapritinib is effective for the treatment of MRD in AML with t(8;21) and KIT mutations after allo-HSCT, and the main adverse effect is hematological toxicity, which could generally be tolerated. Therefore, our study supports the use of avapritinib to treat patients with t(8;21) AML with KIT mutations. Most patients in our study received avapritinib treatment after failing of immunotherapy, however, avapritinib treatment may be more effective as first line treatment, which needs further research to confirm.

DATA AVAILABILITY
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.