From February 15, 2017, through March 18, 2020, 48 patients underwent screening and 31 were included in the intention-to-treat population at five sites in Germany (Fig. 1). All 31 patients suffered from de novo AML. The age ranged from 21 to 80 years (median 59 years) with 65% being female. A total of 55% of patients had received standard 3 + 7 induction therapy followed by consolidation therapy (for details, see Additional file 1: Table S13), and at baseline, all patients had CR with at least one detectable MRD marker. Twenty-eight, two and one participants were positive for mutated NPM1, mutated IDH2 and RUNX1-RUNX1T1, respectively. For NPM1, MRD values ranged from 5 to 656,344 mutated NPM1 copies/10,000 ABL1 copies. FLT3 mutations were detected in nine (29%) patients comprising FLT3-ITD (n = 5, 16% of all patients) and FLT3-TKD (n = 4, 13% of all patients). The clinical characteristics are summarized in Table 1.

Consort flow diagram. Seventeen patients did not meet the inclusion criteria at screening and accordingly were not enrolled in the trial. Three patients underwent screening procedures twice prior to enrollment. 31 enrolled patients received FLYSYN at the indicated dose level of the assigned cohort. Safety oversight to proceed to the next higher dose cohort was performed by an independent data and safety monitoring board after an interim safety analysis of the first three study patients included in a dose level, evaluated on day 15 (cohorts 1–5) or day 43 (cohort 6) after FLYSYN application. Assessment of DLT could be done for all dose levels, whereas single patients of each cohort were not assessable at the primary safety endpoint, and 15 patients left follow-up during the assessment of efficacy for the secondary endpoint analysis. In cohort 6, one patient (asterisk) dropped out prior to third application of FLYSYN due to hematologic relapse. One patient was lost to follow-up. n number

No DLT was reported for FLYSYN at any dose cohort. For the primary safety endpoint, treatment-emergent AEs were observed in 25 (81%) of the 31 patients, of which in 15 (48.5%) patients AEs were considered related to FLYSYN (Table 2 and Additional file 1: Table S1). The most common any-grade treatment-emergent AEs in all cohorts were laboratory abnormalities and hematologic events [neutrophil count decreased (22.6%), anemia (19.4%) and white blood cell decreased (19.4%)], of which most were grade 1–2 (Table 2). The only grade 3 treatment-emergent AEs was neutropenia (6.5%, one in cohort 4 and cohort 6, each) and back pain (3.2%, one in cohort 6) (Additional file 1: Tables S2 and S3). Only three patients (9.7%) developed fever after infusion, but none above grade 1 (Table 2). Most common treatment-related AEs of any grade were decreased neutrophil count (22.6%), anemia (19.4%) and decreased white blood cell count (19.4%, Additional file 1: Table S1). No support with G-CSF was required, and no neutropenic fever was observed. In the 18 patients receiving FLYSYN at ≤ 15 mg/m2 (cohorts 1–4), most common treatment-emergent AEs were anemia (11.1%), constipation (11.1%), hypokalemia (11.1%), hypertension (11.1%), decrease in neutrophil (11.1%) and platelet (11.1%) count. Of those, anemia (11.1%), decreased neutrophil count (5.6%) and decreased platelet count (11.1%) were considered related to FLYSYN (Additional file 1: Tables S4 and S5). In the 13 patients receiving > 15 mg/m2 of the drug (cohorts 5 and 6), most common treatment-emergent AEs were decreased neutrophil (46.2%) and white blood cell count (46.2%), anemia (30.8%), fatigue (30.8%), decreased lymphocyte count (30.8%) and pain in extremity (30.8%). Of those, all hematologic events were assessed as related to FLYSYN (Additional file 1: Tables S4 and S5).

Overall, no grade 4 toxicity or serious AEs were observed (Table 2 and Additional file 1: Table S1). There was no dose interruption due to treatment-related AEs; one patient in cohort 6 showed disease progression during study treatment and was discontinued after second dosing.

During long-term safety follow-up until visit 11, no additional safety issues were noticed (Additional file 1: Table S6).

Of the total of 31 patients, six patients (19.3%) had no disease progression at the end of study visit (day 545), whereas twenty-five patients (80.7%) discontinued study follow-up prior to the end of study visit (Additional file 1: Table S7). Median time to follow-up discontinuation for these 25 patients was 86 days (range 22–409) after study drug administration. The reasons for study follow-up discontinuation were evidence of disease progression (MRD progression or hematological relapse, n = 17), proceeding to allo-HCT (n = 6), lost to follow-up (n = 1) and proceeding to alternative treatment (n = 1; Additional file 1: Table S7). The six patients proceeding to allo-HCT had a median time on study follow-up of 79 days (range 56–246 days). No patient discontinued study treatment because of AEs related to study drug or death (Additional file 1: Table S7).

Pharmacokinetics were assessed in each dose cohort. Peak FLYSYN concentrations were documented 6 h (median) post dose for dose levels ≤ 15 mg/m2 (Additional file 1: Table S8). Maximum FLYSYN levels (Cmax) were dose proportional and reached 23.1 µg/ml upon single dosing of 45 mg/m2. Half-life of FLYSYN was about 7.2 days. Average FLYSYN exposures [area under the curve (AUC)] was dose proportional and similar between dose cohorts 5 and 6 (Additional file 1: Table S8).

Efficacy was determined as best response until visit 10 (day 90). In total, 20 (65%) of 31 patients experienced reduction in MRD in BM as defined as any reduction compared to baseline (Fig. 2 and Additional file 1: Table S9). MRD reduction differed for each cohort and was 100%, 67%, 33%, 44%, 67% and 80% for cohorts 1, 2, 3, 4, 5 and 6, respectively. In cohorts 1–4, MRD was reduced in 56% of patients compared to 77% in cohorts 5 and 6. 11/31 (35%) patients achieved an overall molecular response (defined as any > 1 log MRD reduction or negativity in BM) to treatment, with 67%, 33%, 0%, 22%, 67% and 40% in cohorts 1, 2, 3, 4, 5 and 6, respectively. Occurrence of response appeared to be dose-dependent: 28% responses were seen with lower doses compared to 46% in patients receiving 45 mg/m2 FLYSYN (single or repetitive dosing). MRD negativity was observed in 19% of all patients until visit 10, with slightly higher numbers in cohorts 1–4 compared to cohorts 5 and 6 (22% vs. 15%), and two patients maintaining documented MRD negativity at end of study (Additional file 1: Table S9). As our trial was designed prior to publication of the current ELN guidelines regarding MRD detection, we separated our patients into MRDlow and MRDhigh groups using the cut-off of 200 NPM1 copies/10,000 ABL copies (2%) according to ELN guidelines [2, 34]. A total of 32% (n = 10) of our patients had a NPM1 MRD level below 2%, the other 68% displayed MRD levels above the ELN threshold. Response to treatment was observed more frequently in patients with MRD levels below the cut-off (60% vs. 22%). Of note, the difference in response at least partly leveled out when patients received higher FLYSYN doses, with 60% versus 43% responders upon receiving a total dose of 45 mg/m2 compared to 60% vs 9% responders when receiving ≤ 15 mg/m2 FLYSYN. Of note, the patients that achieved MRD negativity had presented with median baseline values of 11 NPM1 copies/10,000 ABL copies (Additional file 1: Table S9).

Waterfall plot showing MRD change after FLYSYN treatment. Waterfall plot of the best response after baseline (maximum change of MRD level from baseline) until visit 10 (up to 90 days after first FYSYN application). All dose cohorts are displayed. Six patients were MRD negative, and five patients (asterisk) had a MRD reduction of ≥ 1 log compared to MRD levels prior to FLYSYN treatment. MRD was measured with qPCR or NGS, depending on MRD marker. Seven patients (dagger) did not respond to FLYSYN and had MRD increase > 100%. MRD, minimal residual disease

Across the study, median time to MRD response was 29 days (range 15–92). For cohorts 1–4, median time to MRD response was 29 days (range 15–43), whereas in cohorts 5 and 6, MRD response occurred 50.5 days (range 15–92) after FLYSYN treatment. Median time to progression for all treated patients was 6.9 months (95%, CI 3 months–not reached; 0 deaths) (Additional file 1: Fig. S2 and Table S10), with no significant difference observed between cohorts 1–4 and cohorts 5–6. Details are provided in Additional file 1: Appendix.

None of the patients developed ADAs (Additional file 1: Table S11). Analysis of potential effects of FLYSYN treatment on stem cell reserve revealed no profound decrease in colony-forming units (Additional file 1: Fig. S3). There was no substantial difference in baseline CD19+, CD3+, CD3+/CD4+, CD3+/CD8+, CD56+ and CD14+ cell counts with regard to response (Additional file 1: Fig. S4a–f), and neither was there a relevant difference between responders and non-responders regarding percentage and absolute number of NK cells in PB and BM (Additional file 1: Fig. S4g–h). Examination of baseline surface FLT3 expression revealed expression on AML cells in all patients, with no substantial difference according to response (Additional file 1: Fig. S5a–b). Patients not responding to treatment had substantially higher MRD levels at baseline (Additional file 1: Fig. S5c).