Final report of the efficacy and safety of gemtuzumab ozogamicin (Mylotarg) in patients with CD33-positive acute myeloid leukemia in first recurrence
Abstract
BACKGROUND
In this study, the authors analyzed the efficacy and safety of gemtuzumab ozogamicin (GO) (Mylotarg®), an antibody-targeted chemotherapy for CD33-positive acute myeloid leukemia (AML).
METHODS
Patients with CD33-positive AML in first recurrence were entered in 3 open-label, single-arm, Phase II studies. Patients received monotherapy with GO 9 mg/m2 as a 2-hour intravenous infusion in 2 doses separated by 2 weeks. Patients were evaluated for remission, survival, and treatment-emergent adverse events.
RESULTS
Two hundred seventy-seven patients (median age, 61 yrs) were treated with GO, and 71 patients (26%) achieved remission, which was defined as ≤ 5% blasts in the bone marrow without leukemic blasts in the peripheral blood, neutrophil recovery to ≥ 1500/μL, hemoglobin ≥ 9 g/dL, and independence from red blood cell and platelet transfusions. Complete remission (CR) with platelet recovery (≥ 100,000/μL) or without full platelet recovery (< 100,000/μL) (CRp) was observed in 35 patients (13%) and 36 patients (13%), respectively. The median recurrence-free survival was 6.4 months for patients who achieved CR and 4.5 months for patients who achieved CRp. Although expected incidences of Grade 3 or 4 neutropenia (98%) and thrombocytopenia (99%) were observed, the incidence of Grade 3 or 4 sepsis (17%) and pneumonia (8%) was relatively low. Grade 3 or 4 hyperbilirubinemia and hepatic aspartate aminotransferase and alanine aminotransferase elevations were reported in 29%, 18%, and 9% of patients, respectively; 0.9% of patients who did not undergo prior or subsequent hematopoietic stem cell transplantation developed hepatic venoocclusive disease after GO treatment.
CONCLUSIONS
When it was administered to patients with CD33-positive AML in first recurrence, single-agent GO induced a 26% remission rate with a generally acceptable safety profile. Cancer 2005. © 2005 American Cancer Society.
Acute myeloid leukemia (AML) has an annual incidence of 2.4 per 100,000 adults in the United States.1 The incidence is higher in adults age 65 years or older (12.6 per 100,000), and the median age of onset is approximately 62–64 years.1, 2 The prognosis for patients with AML generally is poor. Although initial complete remission (CR1) rates of 50–70% usually are obtained in patients with newly diagnosed AML, the majority of patients eventually will develop recurrent disease, often within the first year.1, 3 For patients with recurrent or refractory AML, there is a lower likelihood of obtaining a second remission (CR2).4 Less than 30% of patients with AML survive beyond 3 years.5
For patients with recurrent AML, the objectives of chemotherapy vary from the induction of long-term remission, to providing a “bridge” to hematopoietic stem cell transplantation (HSCT), or, in some patients, temporary prolongation of life and palliation of symptoms. Most current therapies cause substantial toxicity and induce CR2 in only approximately 25% of patients, depending on individual patient factors, such as age, duration of CR1, and cytogenetic characteristics.4, 6-15
The survival of patients in CR2 improves with HSCT; without it, the median duration of CR2 is short, with long-term survival rates of 2–5%. In patients who are able to undergo allogeneic HSCT from human leukocyte antigen-matched siblings or other donors, long-term survival rates range from 25% to 40%.16 However, only a minority of patients benefit from HSCT because of other factors, such as older age and comorbidities, substantial morbidity associated with reinduction therapy, limited availability of appropriate donors, and treatment-associated toxicity.
Gemtuzumab ozogamicin (GO) (Mylotarg®; Wyeth Pharmaceuticals, Philadelphia, PA) is a new antibody-targeted chemotherapy agent that targets the CD33 surface antigen of leukemic cells by means of a humanized anti-CD33 monoclonal antibody (hP67.6) conjugated to a derivative of the antitumor antibiotic calicheamicin.17 The antibody-antigen complex is internalized into target cells, and the cytotoxic calicheamicin is released intracellularly through hydrolysis.18-20 Because approximately 90% of patients with AML have myeloid blast cells that express the CD33 surface antigen,21 GO is a particularly attractive therapeutic approach.
Three Phase II studies evaluating GO in the treatment of patients with AML at first recurrence have been completed, and the interim results from those studies (n = 142 patients) have been reported.22 These results led to approval by the United States Food and Drug Administration of GO for the treatment of patients with CD33-positive AML in first recurrence age 60 years or older who are not candidates for treatment with other cytotoxic chemotherapy.23 A more detailed interim report on the outcomes and toxicities observed in 101 patients age 60 years or older in these Phase II trials also has been published.24 This is the final report on the efficacy and safety analyses for which data from 277 patients are now available.
MATERIALS AND METHODS
Patients
The current analysis includes data from 3 multicenter, open-label, single-arm, Phase II studies that were conducted to evaluate the efficacy and safety of GO as monotherapy. At study entry, all patients had CD33-positive AML in untreated first recurrence, as determined by analysis of bone marrow aspirates and immunophenotyping.25 In addition, all patients were required to have an Eastern Cooperative Oncology Group performance status of 0–2, the absence of hematologic disorders (such as myelodysplasia) before their diagnosis of AML, adequate renal and hepatic function, and a peripheral white blood cell count < 30,000/μL. Patients who developed AML secondary to chemotherapy or toxin exposure were excluded from the study.
Although all three studies were of similar design, inclusion criteria differed between studies. Study 1 (n = 84 patients in the United States and Canada) and Study 2 (n = 95 patients in Europe) enrolled patients who age 18 years or older who had a prior CR1 of at least 6 months' duration. In Study 2, patients who underwent prior HSCT were permitted (2 patients underwent prior allogeneic HSCT, and 25 patients underwent prior autologous HSCT). Patients enrolled in Study 3 (n = 98 patients in the United States and Europe) had to be age 60 years or older with a CR1 of ≥ 3 months' duration. Patients were enrolled from May 12, 1997 through August 28, 2000.
Cytogenetic analyses were performed on 181 patients (65%) at first recurrence. Cytogenetic data that were obtained at initial diagnosis also were available from 200 patients (72%). Three risk groups were defined as follows: The favorable-risk group included patients with t(8;21), t(15;17), t(16;16), inv(16) or +14 chromosomal abnormalities; the poor-risk group included patients with t(9;22), −5/del(5q), −7/del(7q), inv(3q), i(17q), del(20q), +13, 11q, or 17p chromosomal abnormalities or complex karyotypes (≥ 3 abnormalities); and the intermediate-risk group included patients with all other chromosomal abnormalities or with normal karyotypes.26, 27 Patients for whom cytogenetic analysis was not performed or was inconclusive were categorized into an unknown-risk category.
All protocols were approved and monitored by central or local institutional review boards, and the studies were conducted in a manner consistent with the Declaration of Helsinki and Good Clinical Practice guidelines. All patients provided written informed consent before treatment.
Study Design
Based on results obtained in the Phase I dose-escalation study, 277 patients were scheduled to receive 2 doses of monotherapy with GO 9 mg/m2 as a 2-hour intravenous infusion with 14–28 days between doses. Treatment with hydroxyurea was permitted to reduce initial peripheral white blood cell counts to < 30,000/μL before treatment. Acetaminophen and antihistamines were administered before GO infusion, and two additional doses of acetaminophen were permitted after GO infusion. Corticosteroids were not given. Before they received the second dose of GO, patients must have recovered from all reversible, nonhematologic toxicities and have no evidence of uncontrolled infection, disease progression, or detectable formation of antibodies to GO.
Patients were evaluated during the treatment period (from the initial dose of GO until 28 days after the last dose). After treatment, patients received the AML therapy that was most appropriate, as determined by their individual physicians. Survival data were collected until February 28, 2002.
Efficacy Measures
Because the study design and treatment regimens were similar in all 3 Phase II studies, data were pooled for the purpose of analyzing a greater number of patients. The primary efficacy measure was CR rate. CR was defined as 1) the absence of leukemic blasts in peripheral blood; 2) ≤ 5% leukemic blasts in the bone marrow, as measured in bone marrow aspirates or biopsy samples; 3) peripheral blood counts with hemoglobin ≥ 9 mg/dL, absolute neutrophil count (ANC) ≥ 1500/μL, and platelets ≥ 100,000/μL; and 4) red blood cell transfusion independence for ≥ 2 weeks and platelet transfusion independence for ≥ 1 week. Determination of remission status was evaluated approximately 28 days after the last dose of GO. All bone marrow slides were evaluated by a central reviewer (J.M.B.), who was blinded to treatment outcomes.
Some patients had a CR with incomplete platelet recovery (CRp). These patients met all the criteria for CR (including freedom from platelet transfusion for ≥ 1 wk), but platelet counts were < 100,000/μL. CRp was included as a secondary efficacy measure in these studies, and the overall remission rate was based on the combined total of patients who had CR or CRp. Patients who did not meet the criteria for CR or CRp were categorized as no remission (NR).22
To assess the durability of response, recurrence-free survival (RFS) and overall survival rates also were evaluated. RFS was measured from the first documentation of CR or CRp to the date of recurrence, death, or data cut-off. Overall survival was measured from the initial GO administration to death or data cut-off. All survival data were analyzed using Kaplan–Meier estimates.
Safety Measures
The incidence of treatment-emergent adverse events (TEAEs) was used to assess safety. A TEAE was any event that was not present at baseline or that was present at baseline and worsened during the treatment period. The United States National Cancer Institute Common Toxicity Criteria (version 1) was used to assess the severity of all TEAEs.
Analysis for Antibodies to GO
All patients were screened for antibodies to the hP67.6 monoclonal antibody and the calicheamicin-linker portion of GO using enzyme-linked immunosorbent assays.22
RESULTS
Patient Demographics and Baseline Clinical Characteristics
In total, 277 patients enrolled in the 3 phase II studies had follow-up data available, including 157 patients (57%) age 60 years or older. Baseline clinical characteristics are provided in Table 1. The median age was 61 years, and the median duration of CR1 was 10.6 months. Most patients (93%) had received intensive treatment in CR1 to prevent recurrence after initial induction therapy. Cytoreduction therapy with hydroxyurea within 7 days of GO treatment was administered to 68 of 277 patients (25%). Of 181 patients who had cytogenetic studies performed at the time of recurrence, 6 patients (3%) were in the favorable-risk category, and 63 patients (35%) were in the poor-risk category.
| Characteristic | No. of patients (%) | ||
|---|---|---|---|
| Age < 60 yrs (n = 120) | Age ≥ 60 yrs (n = 157) | All patients (n = 277) | |
| Age (yrs) | |||
| Median (range) | 47 | 68 | 61 |
| Range | 20–59 | 60–87 | 20–87 |
| Gender | |||
| Female | 60 (50) | 66 (42) | 126 (45) |
| Make | 60 (50) | 91 (58) | 151 (55) |
| Ethnic origin | |||
| White | 110 (92) | 154 (98) | 264 (95) |
| Black | 4 (3) | 2 (1) | 6 (2) |
| Asian | 2 (2) | — | 2 (< 1) |
| Other | 4 (3) | 1 (<1) | 5 (2) |
| Duration of CR1 (mos) | |||
| Median | 10.9 | 9.9 | 10.6 |
| Range | 4–117 | 2–74 | 2–117 |
| < 6 mos | 5 (4) | 32 (20) | 37 (13) |
| 6–12 mos | 62 (52) | 62 (39) | 124 (45) |
| ≥ 12 mos | 53 (44) | 63 (40) | 116 (42) |
| Cytogenetics at recurrence | |||
| Known | 87 (73) | 94 (60) | 181 (65) |
| Favorable risk | 5 (6) | 1 (1) | 6 (3) |
| Intermediate risk | 58 (67) | 54 (57) | 112 (62) |
| Poor risk | 24 (28) | 39 (41) | 63 (35) |
| Unknown | 33 | 63 | 96 |
- CR1: first complete remission.
Drug Exposure and Treatment Response
In all, 277 patients received the first dose of GO, 203 patients received 2 doses, and 7 patients received 3 doses. For patients who received 2 doses, there was a median of 16 days (range, 14–31 days) between doses.
Of the 277 patients who initiated a course of treatment with GO, 20 patients later received > 1 course (1 course was defined as 2–3 doses of GO). These patients achieved remission, subsequently developed recurrent disease, and were treated further with repeated courses of GO. Nineteen patients received 2 courses of GO, and 1 patient received 4 courses of GO.28 In total, 299 courses of GO were administered. Sixty-seven patients received only 1 dose of GO because of disease progression (n = 36 patients); adverse events (n = 26 patients; infection [n = 11 patients], bleeding [n = 4 patients], venoocclusive disease [VOD; n = 4 patients], and other events [n = 7 patients], including 1 patient each with acute renal failure/hepatic injury, bradycardia/hyperkalemia, diffuse fluid retention, fever, refractoriness to platelet transfusions, tonic/clonic seizure, and an unspecified adverse event); or early death (n = 4 patients; due to sepsis [n = 2 patients], infection [n = 1 patient], and fungemia and bacteremia [n = 1 patient]). In addition, one patient received only one dose before proceeding directly to HSCT.
Seventy-one patients (26%) achieved overall remission, including 35 patients (13%) who achieved CR and 36 patients (13%) who achieved CRp. There were no significant differences in overall remission rates between patients younger than 60 years (28%) and patients age 60 years or older (24%) (Table 2). The median time to remission for the 35 patients who achieved CR was 66 days (95% confidence interval [95%CI], 52–79 days; range, 38–239 days). The median time to remission for the 36 patients who achieved CRp was 65.5 days (95% CI, 51–75 days; range, 43–150 days).
| Remission type | Age < 60 yrs (n = 120) | Age ≥ 60 yrs (n = 157) | All patients (n = 277) | |||
|---|---|---|---|---|---|---|
| No. (%) | 95%CI | No. (%) | 95%CI | No. (%) | 95%CI | |
| CR | 16 (13) | 8–21 | 19 (12) | 7–18 | 35 (13) | 9–17 |
| CRp | 17 (14) | 8–22 | 19 (12) | 7–18 | 36 (13) | 9–18 |
| ORa | 33 (28) | 20–36 | 38 (24) | 18–32 | 71 (26) | 21–31 |
- 95%CI: 95% confidence interval; CR: complete remission; CRp: complete remission with incomplete platelet recovery; OR: overall remission.
- a OR = CR + CRp. Percentages do not add up to 100% because of rounding.
Forty percent (111 of 277 patients) of the patients had ≤ 5% leukemic blast cells in bone marrow aspirates that were obtained 5–11 days after the first dose of GO based on analysis by an experienced central reviewer. These 111 patients included 56 of 66 patients who achieved overall remission (85%) who had bone marrow analysis performed and 55 of 169 NR patients (32%) who had bone marrow analysis performed (P < 0.0001). Ten of 66 patients who achieved overall remission (15%) had > 5% blast cells in bone marrow aspirates after the first dose of GO but subsequently achieved remission by the end of treatment.
Survival Characteristics of Patients Treated with GO
The median overall survival was 4.9 months (5.3 mos and 4.5 mos for patients younger than 60 yrs and age 60 yrs and older, respectively) (Fig. 1). RFS was similar in both patients who achieved CR and patients who achieved CRp (P = 0.719) (Fig. 2). The median RFS was 6.4 months for patients with CR, 4.5 months for patients with CRp, and 5.2 months for the combined group (CR plus CRp). A significant difference in RFS was observed between patients younger than 60 years and patients age 60 years and older (P = 0.008). This difference may have been influenced by postremission treatment options, especially HSCT.
This graph illustrates overall survival among all patients who received gemtuzumab ozogamicin (n = 277 patients).
This chart shows that recurrence-free survival for 35 patients who achieved complete remission (CR) (median, 6.4 mos) and for 36 patients who achieved complete remission with incomplete platelet recovery (CRp) (median, 4.5 mos) did not differ significantly (P = 0.719; log-rank test).
For those patients who received ≥ 2 doses of GO and completed the treatment period, the median overall survival was 12.2 months for patients in the CR group, 12.9 months for patients in the CRp group, and 4.2 months for patients who did not enter remission (CR or CRp vs. NR; P < 0.001) (Fig. 3). For all patients in remission (CR and CRp), the median overall survival was 12.6 months. The median overall survival for patients younger than 60 years in the CR group and in the CRp group was 17.2 months and 18.4 months, respectively. For patients age 60 years or older in the CR group and in the CRp group, the median overall survival was 11.7 months and 11.4 months, respectively.
For patients who received ≥ 2 doses of gemtuzumab ozogamicin (GO) and remained in the study for > 28 days after the last dose of GO, overall survival differed significantly among patients who achieved either complete remission (CR) (n = 35 patients; median, 12.2 mos) or complete remission with incomplete platelet recovery (CRp) (n = 36 patients; median, 12.9 mos) and patients who had no remission (NR) (n = 124 patients; median, 4.2 mos; P < 0.001; log-rank test). Eighty-two treated patients did not complete the treatment period.
Effect of Pretreatment Factors on Clinical Outcome
Remission rates were similar for all cytogenetic risk groups, although there were only a small number of patients in the favorable-risk group (Table 3). Response to GO was more likely in patients with a longer duration of CR1 (Table 4). In patients who had a CR1 that lasted < 12 months (n = 161 patients), the overall remission rate was 19%, with a median overall survival of 4.0 months. By comparison, patients who had a CR1 ≥ 12 months (n = 116 patients) had an overall remission rate of 34%, with a median overall survival of 7.9 months.
| Age and risk group | No. of patients | OR: No. of patients (%) | 95%CI |
|---|---|---|---|
| Age < 60 yrs | |||
| Favorable risk | 5 | 2 (40) | 5–85 |
| Intermediate risk | 58 | 17 (29) | 18–43 |
| Poor risk | 24 | 6 (25) | 10–47 |
| Unknown | 33 | 8 (24) | 11–42 |
| Age ≥ 60 yrs | |||
| Favorable risk | 1 | 0 | 0–97 |
| Intermediate risk | 54 | 16 (30) | 18–44 |
| Poor risk | 39 | 9 (23) | 11–39 |
| Unknown | 63 | 13 (21) | 11–33 |
| All ages | |||
| Favorable risk | 6 | 2 (33) | 4–78 |
| Intermediate risk | 112 | 33 (29) | 21–39 |
| Poor risk | 63 | 15 (24) | 14–36 |
| Unknown | 96 | 21 (22) | 14–31 |
- 95%CI: 95% confidence interval; OR: overall remission.
| Age group | No. of patients | OR: No. of patients (%) | 95%CI |
|---|---|---|---|
| Age < 60 yrs | |||
| All | 120 | 33 (28) | 20–26 |
| CR1 ≤ 6 mos | 5 | 1 (20) | 1–72 |
| CR1 > 6 mos to < 12 mos | 62 | 14 (23) | 13–35 |
| CR1 ≥ 12 mos | 53 | 18 (34) | 22–48 |
| Age ≥ 60 yrs | |||
| All | 157 | 38 (24) | 18–32 |
| CR1 ≤ 6 mos | 32 | 3 (9) | 2–25 |
| CR1 > 6 mos to < 12 mos | 62 | 13 (21) | 12–33 |
| CR1 ≥ 12 mos | 63 | 22 (35) | 23–48 |
| All ages | |||
| All | 277 | 71 (26) | 21–31 |
| CR1 ≤ 6 mos | 37 | 4 (11) | 3–25 |
| CR1 > 6 mos to < 12 mos | 124 | 27 (22) | 15–30 |
| CR1 ≥ 12 mos | 116 | 40 (34) | 26–44 |
- 95%CI: 95% confidence interval; CR1: first complete remission; OR: overall remission.
Twenty-seven patients underwent HSCT before GO treatment (median time from HSCT to first dose of GO, 11.7 mos; range, 4.3–112.4 mos). Two patients had undergone allogeneic HSCT, had no remission after GO treatment, and survived for 5.8 months and 3.1 months. Of the remaining 25 patients who had undergone autologous HSCT, 1 patient achieved CR after GO treatment and survived for at least 41.2 months, 2 patients achieved CRp and survived for 10.8 months and 23.0 months, and the remaining 22 patients had NR and a median overall survival of 3 months (range, from 0.6 mos to ≥ 41.1 mos).
Therapy after GO Treatment
Forty-six patients underwent HSCT after treatment with GO and before AML recurrence. Fourteen patients who achieved overall remission underwent allogeneic HSCT, and 11 underwent autologous HSCT. Fifteen patients in the NR group underwent allogeneic HSCT, and 6 underwent autologous HSCT. The 100-day survival rate after allogeneic and autologous HSCT was 79% (11 of 14 patients) and 91% (10 of 11 patients), respectively, for the patients who achieved remission, and 33% (5 of 15 patients) and 100% (6 of 6 patients), respectively, for patients in the NR group. At last follow-up, 8 of 14 patients (57%) who achieved remission and underwent allogeneic HSCT and 3 of 11 patients (27%) who underwent autologous HSCT remained alive. Among the NR patients, 3 of 15 patients (20%) who underwent allogeneic HSCT and 4 of 6 patients (67%) who underwent autologous HSCT survived.
Of the 71 patients in the CR plus CRp group who achieved remission, 25 patients (35%) underwent allogeneic or autologous HSCT, 11 patients received additional chemotherapy, and 35 patients received no additional therapy; for the purposes of the current analysis, therapy after second recurrence was not counted. The median overall survival was > 18.3 months, 16.5 months, 12.2 months, and 11.2 months for patients who received allogeneic HSCT, autologous HSCT, additional chemotherapy, or no additional therapy, respectively (P = 0.007) (Fig. 4). When patients were stratified by age, the median survival of patients aged 60 years or older after HSCT was 7.5 months for those who achieved CR (n = 4 patients), 11.9 months for those who achieved CRp (n = 3 patients), and > 3.1 months for patients who had NR (n = 2 patients). For patients younger than 60 years, the median survival was 21.2 months, > 15.6 months, and 3.3 months for the CR group (n = 7 patients), the CRp group (n = 11 patients), and the NR group (n = 19 patients), respectively. RFS and overall survival were similar for both the CR group and the CRp group after HSCT. However, among the patients in the CR and CRp groups who received neither HSCT nor additional chemotherapy between GO treatment and subsequent recurrence, the overall RFS rates for the CR and CRp groups differed significantly (P = 0.030) (Fig. 5).
Overall survival differed significantly among patients who achieved remission after treatment with gemtuzumab ozogamicin (n = 71 patients) and subsequently received allogeneic hematopoietic stem cell transplantation (HSCT) (n = 14 patients; median survival, > 18.3 mos), autologous HSCT (n = 11 patients; median survival, 16.5 mos), additional chemotherapy (n = 11 patients; median survival, 12.2 mos), or no additional therapy (n = 35 patients; median survival, 11.2 mos; P = 0.007; log-rank test).
Recurrence-free survival differed significantly between 17 patients who achieved complete remission (CR) (median, 3.8 mos) and 18 patients who achieved complete remission with incomplete platelet recovery (CRp) (median, 2.4 mos) and received no further therapy after gemtuzumab ozogamicin treatment (P = 0.030; log-rank test) and before disease recurrence.
TEAEs
TEAEs were categorized as either infusion related (those that occurred on the day of GO administration) or those that occurred during the remainder of the treatment period. Hematologic adverse events were reported in virtually all patients and were expected, because CD33 is expressed on normal myeloid cells beyond the progenitor cell stage.
The incidence of Grade 3 or 4 infusion-related adverse events included chills (8%), fever (6%), hypotension (4%), nausea (3%), and hypertension (2%). Hypotension occurred several hours after completion of infusion and was transient and reversible with intravenous fluid administration in most patients. The incidence of infusion-related symptoms was significantly lower on repeat administration: 30% of patients experienced Grade 3 or 4 infusion-related events after the first dose, whereas only 10% experienced such events after the second dose (P < 0.0001).
During the treatment period, Grade 3 or 4 TEAEs that occurred in ≥ 5% of patients included sepsis (17%), fever (13%), chills (9%), nausea or emesis (10%), pneumonia (8%), dyspnea (8%), hypertension (8%), hypotension (8%), asthenia (6%), increased lactate dehydrogenase (6%), and neutropenic fever (6%). Grade 3–4 mucositis occurred in 9 of 277 patients (3%) after the first dose; 2 of those patients had hydroxyurea treatment for 2 days before the first dose of GO and developed mucositis 8 days and 15 days later.
Myelosuppression and Hematologic Adverse Events
Profound neutropenia (Grade 3 or 4) occurred in 267 of 272 patients (98%). However, CR, CRp, and NR patients had recovery of their ANC to 500/μL in a median of 40 days, 43 days, and 51 days, respectively, from the first dose of GO (Fig. 6). There was no significant difference in the median time to ANC recovery to 500/μL between younger and older patients. For CR, CRp, and NR patients, the median time to the recovery of ANC to 1500/μL was 48 days, 56 days, and 97 days, respectively, from the first dose of GO.
The time to absolute neutrophil count (ANC) recovery to 500/μL is illustrated in patients who achieved complete remission (CR) (n = 35 patients; median, 40 days), complete remission with incomplete platelet recovery (CRp) (n = 36 patients; median, 43 days), and no remission (NR) (n = 205 patients; median, 51 days) after the first dose of gemtuzumab ozogamicin (P < 0.001). The time to ANC recovery for patients who achieved CR and CRp did not differ significantly (P = 0.075).
The incidence of profound (Grade 3 or 4) thrombocytopenia was 99% (272 of 276 patients). Patients who achieved CR or CRp had recovery of platelet counts to 25,000/μL in a median of 36 days and 51 days, respectively, from the first dose of GO (Fig. 7). A small subset of patients in the NR group (n = 61 patients) also had recovery of platelets counts; for this subset, the median time to recovery was 32 days. Younger and older patients with CR had recovery of platelet counts to 25,000/μL in a median of 35 days and 38 days, respectively, from the first dose of GO. For younger and older patients with CRp, the median times to recover platelets to 25,000/μL were 39 days and 72 days, respectively. For patients who achieved CR, the median time to recovery of platelet counts to 100,000/μL was 50 days from the first dose of drug.
The time to platelet count recovery to 25,000/μL is illustrated in patients who achieved complete remission (CR) (n = 35 patients; median, 36 days), complete remission with incomplete platelet recovery (CRp) (n = 35 patients; median, 51 days), and no remission (NR) (n = 61 patients who recovered platelets; median, 32 days) from the first dose of gemtuzumab ozogamicin (P = 0.008). The time to platelet recovery for patients who achieved CR and CRp differed significantly (P < 0.001).
Despite the nearly universal occurrence of severe thrombocytopenia, the overall incidence of Grade 3 or 4 bleeding events was only 13% (36 of 277 patients). Epistaxis (3%) and intracranial hemorrhage (3%) occurred most frequently. Eight of nine patients with intracranial hemorrhage died from this condition. The incidence of severe bleeding events was 1 of 35 patients who achieved CR and 4 of 36 patients who achieved CRp.
Liver Function Abnormalities
Grade 3 or 4 hyperbilirubinemia (≥ 1.5 × the upper limit of normal [ULN]) was reported in 80 of 274 patients (29%) after treatment with GO. Grade 3 or 4 elevation of aspartate aminotransferase or alanine aminotransferase levels (≥ 5 × ULN) were observed in 49 of 274 patients (18%) and in 26 of 274 patients (9%), respectively, after GO treatment. Most abnormalities of liver function were transient and reversible and required no medical intervention.
Eight patients (3%) had moderate ascites after GO treatment. In 4 of these patients, the ascites resolved in a median of 42 days (range, 20–130 days). One patient age 24 years had ascites for 33 days and died from multiorgan failure 35 days after the first dose of GO.
Of the 299 courses of GO administered, 16 episodes (5%) of VOD occurred. Among 200 patients who received GO treatment without undergoing prior or subsequent HSCT, 215 courses were given, and 2 patients (0.9%) developed VOD. Signs of VOD were noted at 6 days and 28 days after the first dose of GO.
The incidence of VOD also was examined in the 52 patients who underwent HSCT after GO treatment (46 patients before and 6 patients after second recurrence). Among the 46 patients who underwent HSCT before second recurrence, 8 patients (17%) developed VOD after HSCT (7 allogeneic HSCT, 1 autologous HSCT). Five patients (9%) died with VOD after undergoing HSCT. One patient underwent HSCT after the first course of GO, developed VOD that resolved, received a second course of the drug, and developed a second and fatal case of VOD. The remaining 2 patients recovered from VOD and had overall survivals of at least 3.3 years and 3.5 years. None of the six patients who underwent HSCT after second recurrence experienced VOD.
Of the 27 patients who underwent HSCT prior to GO treatment, 5 patients (19%) developed VOD after GO treatment. Two patients died with VOD, and 3 patients recovered from VOD and had overall survival of 4.7 months, 8.9 months, and 10.8 months.
Hospitalization of Patients Treated with GO
The median total duration of hospitalization during the treatment period was 27 days (range, 0–723 days). Fifteen percent (n = 42) of 277 patients had brief hospitalizations, defined as ≤ 7 days during the treatment period, including 2% of patients with no hospitalization at any time during the treatment period. The 71 patients who achieved remission had a median of 18 days in the hospital (range, 0–71 days) versus 30 median days for the 206 NR patients (range, 0–723 days).
Early Treatment Mortality
Forty-four patients (16%) died within 28 days of the last dose of GO. Patients younger than 60 years had a mortality rate of 14% (17 of 120 patients), and patients age 60 years or older had a mortality rate of 17% (27 of 157 patients). Causes of death included disease progression (n = 13 patients), infection (n = 13 patients), intracranial hemorrhage (n = 8 patients), multiorgan failure (n = 4 patients), respiratory failure (n = 3 patients), amphotericin anaphylaxis (n = 1 patient), VOD (n = 1 patient), and unknown cause (n = 1 patient).
Immunogenicity of GO
None of the 277 patients developed antibody responses to the hP67.6 monoclonal antibody or to the calicheamicin-linker portion of the GO molecule. These results included 20 patients who received multiple courses of the drug without developing detectable antibody responses.
DISCUSSION
The current results indicate that GO is an effective monotherapy for patients with AML in recurrence and confirm the Phase II data from the first 142 patients reported by Sievers and colleagues.22 GO therapy resulted in a 26% overall remission rate in 277 patients with CD33-positive AML in first recurrence and demonstrated a favorable safety profile. These data supplement the currently approved indication in the United States, which is limited to patients with CD33-positive acute myeloid leukemia in first recurrence age 60 years or older who are not considered candidates for other cytotoxic chemotherapy.
Patients with CRp met criteria that were identical to those for CR, with the exception that platelet recovery to ≥ 100,000/μL was not observed. In some patients, incomplete platelet recovery was associated with the rapid administration of other antileukemic therapies before the end of the treatment period. RFS and overall survival after HSCT were similar between patients with CR and CRp, suggesting that these 2 groups are comparable clinically when other antileukemic therapies are given after GO therapy. However, RFS differed significantly for patients who achieved CR and CRp and who received no further therapy in the period after GO treatment and before they developed recurrent disease. The clinical significance of this finding is unclear. The significantly slower recovery of thrombopoiesis in patients who achieved CRp versus patients who achieved CR may have been due to larger amounts of residual (occult) AML or to a specific effect of GO on platelet precursors. Alternatively, biases may not have been detected in these Phase II studies because of the small numbers of patients who achieved CR and CRp.
Most patients tolerated GO well. The incidence of infusion-related symptom complex was similar to that seen with other antibody-based therapies.29 In most patients, symptoms resolved within 2–4 hours with supportive therapy, which included acetaminophen, diphenhydramine, and intravenous fluids. The early treatment-related mortality rate (16%) was similar to that reported with other intensive chemotherapeutic regimens.30, 31 However, it should be noted that this trial was not designed to compare GO with intensive chemotherapy. A comparison between treatment modalities warrants a direct head-to-head trial in patients with AML in first recurrence.
The rates of Grade 3 or 4 sepsis and pneumonia were only 17% and 8%, respectively, although Grade 3 or 4 myelosuppression occurred in nearly all patients. These are low rates of life-threatening infections for patients receiving reinduction therapy for AML and may be related to the low incidence of severe mucositis, which would provide fewer opportunities for invasion by microflora of the gastrointestinal tract. The incidence of severe nausea or emesis also was low (10%). Thus, GO may be of particular benefit in patients for whom these adverse events are treatment limiting.
This favorable safety profile may allow use of the drug in combination with other agents, such as anthracyclines and/or cytarabine. Studies currently are underway to evaluate combination regimens. Such studies may be particularly important for patients with a high load of CD33 antigen in the peripheral blood, in as much as this variable may be an adverse prognostic factor during single-agent treatment with GO.32
The most clinically significant drug-related toxicity observed was hepatic and was reflected by elevations in bilirubin and serum aminotransferases. These signs of liver injury occurred within a few days of drug exposure and were transient in most patients. However, along with others, we have described VOD in patients with AML after treatment with GO.33-35 Hepatic VOD occurs frequently after HSCT but is a rare occurrence after nonmyeloablative therapy. Among 200 patients who received 215 courses of GO, only 2 patients (0.9%) developed a VOD-like syndrome after monotherapy without prior or subsequent HSCT procedures. In contrast, of the 27 patients who underwent HSCT before GO treatment, 5 patients (19%) experienced VOD. Among the 52 patients who received 54 courses of GO with subsequent HSCT procedures, 8 patients (15%) developed VOD. The current trial was not designed to assess a relation among HSCT, VOD, and GO, and additional studies will be required to identify clinical factors that affect the risk of VOD when patients undergo HSCT after they receive GO treatment.
In conclusion, GO, which is an antibody-targeted chemotherapy, provides an effective and relatively well tolerated treatment option for patients with CD33-positive AML in recurrence. It is useful for cytoreduction of the leukemia tumor burden, allowing recovery of normal hematopoiesis; however, unfortunately, nonmaintained remissions are short. The best outcomes were observed in patients who were able to proceed with HSCT for postremission therapy. Further studies will be required to evaluate its usefulness as primary therapy or in combination with other antileukemic agents and to determine the mechanisms of leukemia cell resistance.
Acknowledgements
The authors thank Vincent H. J. van der Velden for CD33/Mylotarg laboratory analysis, Nadine Dodge for clinical programming, Robert Herbertson for biostatistical analysis, and Susan Leinbach for assistance in article preparation. They also thank all of the patients and clinical personnel who participated in these studies.
