Fifty-one adult patients (median [range] age of 67 [22–82] years) were enrolled in the study. Among enrolled patients, 27 (52.9%) had received 1 prior induction regimen; 11 (21.6%), 8 (15.7%), and 4 (7.8%) had received 2, 3, or > 3 prior induction regimens, respectively (see Supplementary Table 1 in the Online Resource). Five (9.8%) patients had received ≥ 1 allogeneic HSCT prior to GO treatment. There were 19 (37.3%) patients classified as adverse-risk according to the ELN 2017 recommendations [12]. Of 51 patients assigned to treatment, 50 (98.0%) received ≥ 1 dose of GO during Cycle 1 (safety analysis set), 46 (92.0%) patients received all doses in Cycle 1, and 9 (17.6%) received further doses in Cycle 2. After GO treatment, 18 (35.3%) patients went on to receive ≥ 1 further systemic treatment, and 2 (3.9%) patients went on to receive an allogeneic HSCT.

The upper limit of the 2-sided 90% confidence interval (CI) for least squares (LS) mean differences in QTc using Fridericia’s formula (QTcF) was < 10 ms for all time points during Cycle 1, including those on D4 and D7 assessed in the primary analysis (Fig. 1). The largest mean QTcF change from baseline was 5.10 ms (90% CI 2.2–8.1 ms). Most patients had a QTcF ≤ 450 ms (44/49, 89.8%) and a maximum increase from baseline of ≤ 30 ms (46/49, 93.9%). No patients had a post-baseline QTcF > 480 ms, and no patients had a change from baseline of > 60 ms.

QTcF change from baseline during C1. Baseline is defined as the last average of the planned triplicate 12-lead electrocardiogram measurements available prior to the first dose of GO. LS means estimated from a linear model with visit/time post-dose as fixed effect and unstructured variance/covariance matrix for the repeated measurements with CI using approximate t-distribution with degree of freedom by the Kenward–Roger method. The time points assessed in the primary analysis of QTcF were D4 (at 0 h) and D7 (at 0, 2, 4, and 6 h). C cycle, CI confidence interval, D day, GO gemtuzumab ozogamicin, LS least squares, QTcF QT interval corrected for heart rate using Fridericia’s formula

All-causality, treatment-emergent AEs (TEAEs) were reported in 49 (98.0%) patients, the most common being febrile neutropenia (40.0%) and thrombocytopenia (22.0%; Table 1). Serious TEAEs occurred in 34 (68.0%) patients. The most frequently reported were febrile neutropenia (22.0%), sepsis (14.0%), and disease progression (10.0%). Grade 3–4 TEAEs occurred in 27 (54.0%) patients, the most common being febrile neutropenia (36.0%), thrombocytopenia (18.0%), anemia (10.0%), and neutropenia (10.0%).

Seventeen (34.0%) patients had infections; 10 (20.0%) patients experienced grade 3–4 infections while 6 (12.0%) patients experienced grade 5 infections. Sixteen (32.0%) patients experienced hemorrhage; 2 (4.0%) patients experienced grade 3–4 hemorrhage, and no patients experienced grade 5 hemorrhage. The most-reported (≥ 10.0%) infectious TEAE was sepsis, occurring in 7 (14.0%) patients. The most-reported (≥ 5.0%) grade 3–4 infections were sepsis and pneumonia, in 3 (6.0%) patients each. The most-reported (≥ 5.0%) grade 5 infection was sepsis, in 4 (8.0%) patients. One (2.0%) patient experienced an infection (grade 3) that resulted in temporary discontinuation of GO and grade 4 sepsis that resulted in permanent discontinuation. Both infection and sepsis resulting in septicemia occurred simultaneously in the same patient; sepsis progressed, and the patient died. Both events were assessed as unrelated to GO by the investigator. The most-reported (≥ 10.0%) hemorrhage TEAE was epistaxis in 5 (10.0%) patients. One (2.0%) patient had grade 3 gastric hemorrhage and 1 (2.0%) patient had grade 4 traumatic intracranial hemorrhage. No hemorrhage resulted in permanent or temporary discontinuation of GO.

Three (6.0%) patients temporarily discontinued GO due to TEAEs, including increased alanine aminotransferase, increased transaminases, and infection, each in 1 (2.0%) patient. Two (4.0%) patients permanently discontinued GO due to TEAEs, including pyrexia and sepsis, each in 1 (2.0%) patient.

Grade 5 TEAEs occurred in 16 (32.0%) patients, of whom 5 (10.0%) died of disease progression and 4 (8.0%) died of sepsis. Multiple organ dysfunction, pyrexia, atypical pneumonia, COVID-19 pneumonia, AML (consistent with disease progression), respiratory failure, and capillary leak syndrome (CLS; detailed below) accounted for the deaths of the remaining patients.

One patient experienced grade 3 atrial fibrillation and supraventricular tachycardia, which were considered unrelated to GO. No patients experienced grade ≥ 4 cardiac conduction TEAEs. VOD/SOS was not reported; however, 1 patient experienced treatment-related grade 5 CLS associated with pleural effusion, ascites, hyperbilirubinemia, and endothelial syndrome.

A total of 43 (86.0%) patients experienced a shift from grade ≤ 2 at baseline to grade 3 or 4 post-baseline in hematology and coagulation laboratory parameters. The most common were decreased white blood cell count in 25/49 (51.0%) patients, decreased lymphocyte count in 24/49 (49.0%) patients, and anemia in 21/50 (42.0%) patients.

Following administration of multiple fractionated infusions of GO at 3 mg/m2 (Cycle 1, Day 7), exposures as measured by geometric mean area under the plasma concentration–time profile (AUC) from time zero to 336 h post-dose and maximum plasma concentration (Cmax) were 461,500 pg h/mL and 11,740 pg/mL; 1639 pg h/mL and 58.8 pg/mL; and 26,820 ng h/mL and 585.6 ng/mL for conjugated calicheamicin, unconjugated calicheamicin, and hP67.6 antibody, respectively (see Supplementary Table 2 in the Online Resource). In general, the concentration–time profiles of conjugated calicheamicin and unconjugated calicheamicin are similar to that of total hP67.6 antibody for Cycle 1, Day 1 and Cycle 1, Day 7.

Of 50 patients treated with GO, 12 (24.0%) had positive ADAs against GO at baseline. This was likely due to pre-existing host antibodies that were cross-reactive with GO. There was no treatment-boosted ADA response.

Treatment-induced ADA was detected in 6 (12.0%) patients. No patients experienced anaphylaxis, hypersensitivity, or other clinical sequelae related to ADA. Among the 6 patients positive for treatment-induced ADA, 2 (33.3%) experienced an infusion-related reaction, both pyrexia (grade 1 and grade 3). Among the 44 patients negative for treatment-induced ADA, 7 (15.9%) patients experienced infusion-related reactions. All were grade 1 or 2, except for 1 instance of grade 3 urticaria.

Of the 18 patients who had positive ADAs against GO, none were positive for NAbs at baseline; therefore, no patients had a treatment-boosted response. Treatment-induced NAbs were detected in 1 (2.0%) patient.

The best overall response was CR in 2 (3.9% [95% CI 0.5–13.5]) patients and CRi in 3 (5.9% [95% CI 1.2–16.2]) patients (see Supplementary Table 3 in the Online Resource), with an overall rate of 9.8% (95% CI 3.3–21.4). CR + CRi was achieved by 3 (6.0%) of the 50 patients in Cycle 1. Nine of the 50 patients receiving Cycle 1 proceeded to Cycle 2, of whom 3 (33.3%) patients achieved CR + CRi in Cycle 2. Among the 5 patients who achieved CR or CRi (see Supplementary Table 3 in the Online Resource), there were 2 and 3 patients stratified with ELN favorable or intermediate risk, respectively. One patient who achieved CR/CRi had previously received 2 HSCTs, 1 had previously been treated with 3 induction regimens, and the remaining patients had received 1 or 2 prior induction regimens.

The median OS was 2.8 (95% CI 1.7–4.2) months, with 45 deaths reported (88.2%; see Supplementary Fig. 1 in the Online Resource). Of the 45 deaths, disease progression was the most common cause in 35 (77.8%) patients.