The Addition of Cixutumumab or Temozolomide to Intensive Multiagent Chemotherapy Is Feasible but Does Not Improve Outcome for Patients with Metastatic Rhabdomyosarcoma: A Report from the Children’s Oncology Group

Precis:

Addition of cixutumumab or temozolomide to intensive multi-agent chemotherapy for metastatic rhabdomyosarcoma is safe and feasible. Neither agent improved outcome compared to the chemotherapy backbone alone.

Abstract

Purpose

Outcome for patients with metastatic rhabdomyosarcoma (RMS) remains poor. A previous COG study (ARST0431) for patients with metastatic RMS showed no improvement in outcome using multiple cytotoxic agents in a dose intensive manner. We report the results of the subsequent COG study (ARST08P1), which evaluated the feasibility and efficacy of adding cixutumumab (insulin-like growth factor-1 monoclonal antibody) or temozolomide to the ARST0431 intensive chemotherapy backbone.

Patients and Methods

We conducted two non-randomized pilot studies in metastatic RMS, initially to determine feasibility, with both pilots expanded to assess efficacy. All patients received 54 weeks of chemotherapy including vincristine/irinotecan, interval-compressed vincristine/doxorubicin/cyclophosphamide alternating with ifosfamide/etoposide, and vincristine/dactinomycin/cyclophosphamide. In Pilot 1, patients received cixutumumab (3, 6, or 9 mg/kg) IV once weekly throughout therapy. In Pilot 2, patients received oral temozolomide (100 mg/m²) daily x 5 days with irinotecan. All patients received radiation to the primary tumor and to metastatic sites.

Results

One hundred and sixty-eight eligible patients were enrolled (97 on Pilot 1 and 71 on Pilot 2). Most patients were ≥ 10 years old (73%) with alveolar histology (70%) and bone and/or bone marrow metastases (59%). Toxicities observed in each pilot were similar to those seen on ARST0431. With a median follow-up of 2.9 years, the 3-year EFS was 16% (95% CI: 7–25%) with cixutumumab and 18% (95% CI: 2–35%) with temozolomide.

Conclusion

Addition of cixutumumab or temozolomide to intensive multi-agent chemotherapy for metastatic RMS was safe and feasible.Neither agent improved outcome as compared to the same chemotherapy on ARST0431.

Introduction

Approximately 15% of rhabdomyosarcoma (RMS) patients have distant metastases at diagnosis.¹ While the majority of patients with localized disease can be cured, outcome for patients with metastatic RMS is poor and has not improved over the past several decades.^1–4 Although small subsets of patients with metastatic RMS have a more favorable outcome,⁵ long-term event-free survival (EFS) for patients with metastatic RMS is less than 20%.^5–7

A recent Children’s Oncology Group (COG) study, COG ARST0431,⁷ evaluated an intensive chemotherapy regimen using the most active agents identified from previous phase 2 window studies.⁸ The ARST0431 regimen included six weeks of initial therapy with vincristine and irinotecan (VI), followed by interval-compressed cycles of vincristine, doxorubicin, and cyclophosphamide (VDC) alternating with ifosfamide and etoposide (IE), 2 subsequent courses of VI with radiation therapy to primary and metastatic sites, and 4 cycles of vincristine, dactinomycin and cyclophosphamide (VAC). Although the early EFS results were encouraging,⁹ more mature 3-year EFS was 38% for all patients, but was only 20% for patients with more than one risk factor as defined by Oberlin et al. (age > 10 years or < 1 year, unfavorable primary site, 3 or more metastatic sites, and bone or bone marrow involvement).^5,7 Hence, intensification of traditional treatment modalities did not improve outcome for the majority of patients with metastatic RMS.

While prognosis is significantly better for patients with localized RMS, outcome has not improved in recent years despite the evaluation of additional cytotoxic agents.^3,10 Therefore, novel treatment strategies are needed for both localized and metastatic RMS. Evaluating new agents added to an established chemotherapy backbone in patients with metastatic RMS may help to identify therapies that are worthy of further study in patients with localized RMS while also determining whether these agents may be safe and effective for patients with metastatic disease.

The insulin-like growth factor-1 receptor (IGF-1R) is a potentially important therapeutic target for RMS.^11,12 It is highly expressed in RMS and appears to be involved in the initiation and progression of the disease.^11,13 Cixutumumab (IMC-A12) is a human IgG1/λ monoclonal antibody against the IGF-1R that has shown strong in vitro and in vivo activity against RMS.^14,15 Cixutumumab has been evaluated in pediatric single-agent phase 1 and phase 2 studies,^16,17 but the addition of cixutumumab to a multiagent chemotherapy backbone has not previously been studied in pediatric patients or in patients with RMS.

Alkylating agents such as cyclophosphamide and ifosfamide are highly active in RMS and are a component of therapy for intermediate- and high-risk disease. Temozolomide is an alkylating agent that has demonstrated synergy with irinotecan in preclinical models of pediatric solid tumors,¹⁸ and is active in combination with irinotecan in patients with neuroblastoma and pediatric sarcomas.^19–21 The addition of temozolomide to irinotecan for patients with previously untreated metastatic RMS has not previously been evaluated.

We report the results of COG study ARST08P1, which included two pilot studies to evaluate the addition of cixutumumab or temozolomide to an intensive multi-agent chemotherapy backbone for the treatment of metastatic RMS. The primary aim of COG ARST08P1 was to evaluate the feasibility of combining cixutumumab or temozolomide with the chemotherapy regimen used in the predecessor study, COG ARST0431. An additional objective was to assess the efficacy of adding cixutumumab or temozolomide to the chemotherapy backbone.

Patients and Methods

Eligibility

Patients < 50 years old with newly diagnosed RMS metastatic at sites other than regional nodes were eligible. To establish safety of the combination regimens, we initially excluded patients younger than 10 years with metastatic embryonal RMS (ERMS) who have more favorable outcomes. After the initial 110 patients were enrolled and study treatment was determined to be safe and feasible, eligibility was expanded to all patients younger than 50 years with metastatic RMS regardless of histology. Patients had adequate renal, liver, and cardiac function. Patients with uncontrolled infections or known diabetes mellitus as well as women who were pregnant or breastfeeding were excluded. All enrolled patients had central pathology review of their tumors in order to confirm the diagnosis of RMS and to subclassify into histological types. Review was performed by two senior pediatric pathologists experienced in pediatric sarcoma diagnosis: sub-classification was based on guidelines promulgated by the Intergroup Rhabdomyosarcoma Study.²²

The trial was reviewed and approved by the National Cancer Institute’s Pediatric Central Institutional Review Board. Each participating COG institution also obtained approval from its local institutional review board. Informed consent was obtained from the patient or parent/guardian prior to enrollment, with patient assent when appropriate.

Study Design

The clinical trial consisted of two pilot studies, with block sequential enrollment of 20 patients, adding either cixutumumab (Pilot 1) or temozolomide (Pilot 2) to the backbone chemotherapy used in COG ARST0431.⁷ Patients received 54 weeks of chemotherapy, beginning with two cycles of VI (Weeks 1–6), followed by 6 cycles of alternating interval-compressed VDC and IE (Weeks 7–19) (Supplemental Table A1). VI cycles were repeated at Weeks 20–25 and Weeks 47–51. Interval-compressed VDC/IE cycles were administered again at Weeks 28–34 followed by 4 cycles of VAC during Weeks 35–46. Cumulative doses were 375 mg/m² of doxorubicin, 10.8 g/m² of cyclophosphamide, 45 g/m² of ifosfamide, and 2.5 g/ m² of etoposide. Dexrazoxane was administered with each doxorubicin dose as a cardioprotectant, and neutrophil growth factor was given after VDC, IE, and VAC cycles. Initiation of each cycle of chemotherapy required an absolute neutrophil count ≥ 750/µl and platelet count ≥ 75,000/µl.

Radiation therapy (50.4 Gy) to the primary tumor was administered for most patients starting at Week 20. Patients requiring emergent radiation and those with intracranial extension of parameningeal disease started primary site radiation beginning at Week 1. Radiation to metastatic sites was given at Week 20 and/or Week 47.

For patients on Pilot 1, cixutumumab was administered as a 1-hour intravenous infusion once weekly throughout treatment. If the start of a chemotherapy cycle was delayed due to myelosuppression, cixutumumab was also held until blood count recovery. Cixutumumab was also held during radiation therapy. Doses of chemotherapy agents were adjusted for hyperbilirubinemia and sinusoidal obstruction syndrome (SOS), for any SOS cixutumumab was permanently discontinued. The cixutumumab dose was escalated in sequential cohorts of 3, 6, and 9 mg/kg (the pediatric phase II dose), with escalation dependent upon tolerability at the previous dose level. After enrollment of the initial 20 patients to each cixutumumab cohort, accrual to Pilot 1 was suspended to monitor for toxicity. The subsequent 20 patients were enrolled on Pilot 2, followed by accrual of 20 patients to the next higher cixutumumab dose level (if the previous dose level was determined to be safe and tolerable). Patients on Pilot 2 received oral temozolomide 100 mg/m²/day x 5 days with each VI cycle at Weeks 1–6, 20–25, and 47–51.

Statistical Methods

The primary aim of this study was to assess the feasibility of adding each novel agent onto the chemotherapy backbone. Feasibility for Pilot 1 was defined as a rate of serious cardiac toxicity, defined as Grade 3 or higher by CTCAE v.4 attributable to cixutumumab, of 10% or lower. For Pilot 2, the addition of temozolomide was considered feasible if the rate of grade 4 non-hematologic toxicity throughout treatment was less then 15%. Adverse events were graded according to the NCI Common Toxicity Criteria (version 4.0).

A secondary objective was to evaluate the EFS and overall survival (OS) for patients treated with each pilot regimen. EFS was defined as the time from enrollment to first progression, recurrence after initial response, second malignancy, or death. OS was defined as the time from enrollment to death from any cause. EFS and OS were estimated using the method of Kaplan-Meier,²³ and comparisons were made between pilot studies using the log-rank test. We also analyzed EFS and OS for subsets of patients based on previously defined risk factors in metastatic RMS patients. Outcome was determined separately for patients with ERMS < 10 years old, ERMS ≥ 10 years old, and ARMS. In addition, outcome was analyzed by the number of risk factors as defined by Oberlin et al. Oberlin risk factors are age (< 1 or ≥ 10 years), unfavorable primary tumor site, bone or bone marrow metastases, and three or more metastatic sites, histology is not included in this risk model.⁵

Initial planned enrollment included expanding the more feasible and compelling pilot study to 75 patients, in order to allow comparison with ARST0431. This provided approximately 80% power at the 0.01 significance level (one-sided) to detect an increase in EFS associated with a reduction in failure risk to 59% of that with ARST0431 therapy (corresponding to an EFS of 56% at 3 years).

ARST08P1 was amended in September 2012 to address a limited supply of cixutumumab, which would no longer be available after March 2014. To ensure that any enrolled patient could complete therapy, a cutoff for enrollment on Pilot 1 at the cixutumumab dose level of 9 mg/kg was established at 60 patients or until January 15, 2013 (whichever occurred earlier). Thereafter, enrollment to Pilot 2 (temozolomide) was expanded to 75 patients.

Results

One hundred and sixty-eight eligible patients were enrolled on ARST08P1 between January 19, 2010 and July 19, 2013. Pilot 1 met its accrual goal in January 2013. Seven enrolled patients were found to be ineligible (3 assigned to cixutumumab and 4 assigned to temozolomide). Five patients had ineligible histology on central review and were removed from protocol therapy. For 2 patients, initial staging was incorrect.

Among 97 eligible patients who received cixutumumab, 19 were assigned to a dose of 3 mg/kg, 18 to 6 mg/kg, and 60 to 9 mg/kg. Seventy-one eligible patients were enrolled on Pilot 2 and received temozolomide. Table 1 shows the characteristics of eligible patients, which were similar between those treated with cixutumumab and those treated temozolomide. The majority of patients had alveolar histology (70%), were 10 years of age or older (73%), had primary tumors 5 cm or larger (80%), and had regional lymph node involvement (67%).

Table 1.

Patient and Tumor Characteristics (n=168)

Characteristic	Pilot 1	Pilot 2	All
Age (years)
0–9	22	23	45
10+	75	48	123
Sex
Male	46	42	88
Female	51	29	80
Histology
Alveolar	65	52	117
Embryonal	27	13	40
Other	5	6	11
Primary Site
Extremity	27	19	46
GU Bladder/Prostate	10	3	13
GU non-Blader/Prostate	10	8	18
Head and Neck	3	5	8
Intrathoracic	2	0	2
Parameningeal	10	8	18
Perineum	6	2	8
Retroperitoneum	12	13	25
Trunk	8	6	14
Other	9	7	16
Tumor Size, cm
≤ 5	18	15	33
> 5	79	56	135
Regional Lymph Node Involvement
No	42	34	76
Yes	50	36	86
Not evaluated	5	1	6
Number of Metastatic Sites
1	38	26	64
2	22	22	44
3	20	13	33
4	11	8	19
5	6	2	8
Bone Marrow Metastases
No	38	22	60
Yes	59	49	108

Toxicity

Grade 3 and higher non-hematologic toxicities are shown in Table 2. The most commonly observed Grade 3 or 4 toxicities on Pilot 1 (cixutumumab) were febrile neutropenia (42%), diarrhea (36%), and hyperglycemia (26%). Four patients (4%) had Grade 3 or 4 cardiotoxicity, below the 10% feasibility target. Three cases of severe hepatopathy were observed in the cohort receiving cixutumumab at 9 mg/kg, one of which was fatal. The study was amended on May 31, 2013 to reduce cixutumumab to 6 mg/kg for all patients initially assigned to the 9 mg/kg cohort and still receiving therapy, after which no further episodes of hepatopathy were observed. The most common Grade 3 or higher toxicities that occurred on Pilot 2 (temozolomide) were febrile neutropenia (46%) and infection (24%). The frequency and severity of toxicities with the addition of either cixutumumab or temozolomide were similar to those observed on ARST0431.⁷

Table 2.

Grade 3 and higher Non-hematologic Toxicities (Number of patients. (%))

Toxicity	Cixutumumab* (N=97)		Temozolomide (N=71)
	Grade 3	Grade 4	Grade 3	Grade 4
Increased ALT	16 (16)	5 (5)	12 (17)	1 (1)
Anorexia	13 (13)	0 (0)	7 (10)	0 (0)
Increased AST	11 (11)	1 (1)	7 (10)	0 (0)
Colitis	12 (12)	0 (0)	10 (14)	0 (0)
Dehydration	16 (16)	0 (0)	3 (4)	0 (0)
Diarrhea	34 (35)	1 (1)	11 (15)	1 (1)
Febrile Neutropenia	40 (41)	1 (1)	30 (42)	3 (4)
Hyperglycemia	22 (23)	3 (3)	2 (3)	1 (1)
Hypokalemia	13 (13)	4 (4)	8 (11)	3 (4)
Infections	18 (19)	0 (0)	15 (21)	2 (3)
Oral Mucositis	16 (16)	0 (0)	13 (18)	1 (1)
Nausea	12 (12)	0 (0)	7 (10)	0 (0)
Pain	10 (10)	0 (0)	2 (3)	0 (0)
Sepsis	0 (0)	10 (10)	0 (0)	5 (7)
Vomiting	14 (14)	0 (0)	9 (13)	0 (0)
Weight Loss	13 (13)	0 (0)	2 (3)	0 (0)

^*All dose levels of cixutumumab combined

Outcome

The median follow-up for all 168 eligible patients is 2.9 years. Estimated 3-year EFS and OS for all patients were 16% (95% CI: 8% to 23%) and 41% (95% CI: 32% to 50%), respectively. Patients receiving temozolomide had statistically significantly poorer EFS (p=0.0226) and OS (p=0.0297) than those receiving cixutumumab (Figure 1 and Figure 2). Median times from enrollment to an event were 21.5 months with cixutumumab and 17.5 months with temozolomide. Three-year EFS and OS were 16% (95% CI: 7% to 25%) and 47% (95% CI: 36% to 58%), respectively, with cixutumumab (all dose levels combined) and 18% (95% CI: 2% to 35%) and 33% (95% CI: 18% to 47%) with temozolomide. EFS by cixutumumab dose level compared to temozolomide is shown in Appendix Figure A1.

Figure 1:

Comparison of event-free survival by treatment arm. IMC-A12: Pilot 1 with cixutumumab; Temzolomide: Pilot 2 with temozolomide.

Figure 1A (online only): Comparison of event-free survival by cixutumumab dose level (1A, 3 mg/kg; 1B, 6 mg/kg; 1C, 9 mg/kg) and temozolomide.

Figure 2:

Comparison of overall survival by treatment arm. IMC-A12: Pilot 1 with cixutumumab; Temzolomide: Pilot 2 with temozolomide.

Outcomes for subgroups defined by histology/age and Oberlin risk factors are shown in Table 3. Patients with ARMS (3-year EFS 6%) had worse outcome than those with ERMS regardless of age. In addition, outcome was worse for patients with two or more Oberlin risk factors (3-year EFS 9%). On both treatment arms, EFS and OS were better for patients with ERMS less than 10 years of age compared to older ERMS patients and those with ARMS (Table 4).

Table 3.

3-year EFS and OS by risk-defined subgroups.

Outcome	ERMS < 10 Yrs	ERMS 10+ Yrs	ARMS	Oberlin 0–1	Oberlin 2+
n	12	25	117	38	130
EFS (95% CI)	64 (11 to 99)	48 (18 to 78)	6 (1 to 11)	38 (14 to 62)	9 (3 to 15)
OS (95% CI)	61 (18 to 99)	60 (34 to 87)	36 (26 to 46)	70 (51 to 88)	33 (24 to 43)

Table 4.

3-year EFS and OS for cixutumumab and temozolomide treatment arms by risk group.

	Outcome	ERMS < 10 Years	Others
Cixutumumab	n	6	91
	EFS (95% CI)	67 (1 to 99)	14 (5 to 22)
	OS (95% CI)	67 (13 to 99)	46 (34 to 57)
Temozolomide	n	6	65
	EFS (95% CI)	60 (1 to 99)	15 (1 to 31)
	OS (95% CI)	60 (1 to 99)	31 (17 to 46)

Discussion

Treatment of metastatic RMS remains a significant challenge as the majority of these patients are not cured. Several clinical trials over the past few decades have shown no benefit to intensification of cytotoxic therapy over standard VAC chemotherapy for metastatic RMS.^2,7,8,24,25 In addition, while outcome has improved over time for patients with localized RMS with improved staging, local control therapy, and supportive care, no systemic treatment has proved superior to VAC alone.^10,26

As the limit of what can be achieved with traditional cytotoxic agents in the treatment of RMS has likely been reached, focus has shifted to identification of novel biologic agents with different mechanisms of action. The majority of patients with metastatic RMS have alveolar histology, the molecular hallmark of which is PAX3 or PAX7 gene fusion with FOXO1. Unfortunately, efforts to target PAX/FOXO1 fusion oncoproteins directly are not yet clinically viable.^27,28 However, downstream pathways activated by PAX/FOXO1 may be amenable to targeted therapy.

One downstream effect of PAX-FOXO1 fusion is activation of the IGF-1R pathway.¹¹ In RMS, signaling through IGF-1R appears to promote resistance to cytotoxic therapy, and IGF-1R inhibition results in significant preclinical antitumor activity.^15,29,30 Therefore, while only limited single-agent activity was seen with the IGF-1R inhibitor, cixutumumab, in pediatric phase 1 and phase 2 studies,^16,17 we hypothesized that the addition of cixutumumab to a multi-agent chemotherapy backbone for metastatic RMS would be feasible and might improve outcome by enhancing the effects of cytotoxic therapy. By evaluating activity of cixutumumab in patients with metastatic RMS, we also aimed to determine whether clinical evaluation of IGF-1R inhibition for localized RMS was warranted.

In this study, the addition of cixutumumab to the intensive multiagent chemotherapy regimen used in ARST0431 proved to be feasible with similar toxicity to the chemotherapy backbone alone. However, we did not find an outcome benefit with the addition of this agent. Despite the dose de-escalation from 9 mg/kg to 6 mg/kg on this study, it is unlikely that the lack of efficacy was due to insufficient cixutumumab drug levels. Based on pediatric phase II data and better target trough concentrations sufficient to inhibit IGF-1R,³¹ the recommended phase 2 dose of cixutumumab in adults was 6 mg/kg, based on the trough concentrations. It is therefore more likely that mechanisms of resistance to IGF-1R inhibition developed or were unmasked. Activation of pathways to overcome IGF-1R inhibition, such as Her2, have been identified in models of RMS.³² Future evaluation of IGF-1R inhibitors in RMS will require a better understanding of this and other mechanisms of resistance.

In this study, the 3-year EFS with cixutumumab (16%) and with temozolomide (18%) were inferior to the overall 3-year EFS of 38% on ARST0431. While observed EFS was poor on both pilot studies, survival outcomes for patients receiving temozolomide were worse, with shorter time to progression compared to patients receiving cixutumumab (p-value 0.023). As patient characteristics and toxicity were similar between treatment arms, the reason for shorter median time to progression with temozolomide is not known. Outcomes were similarly worse for patients with ARMS than for those with ERMS on both treatment arms.

Because the initial ARST08P1 eligibility criteria excluded ERMS patients <10 year old, the inferior 3-year EFS observed on this study compared to ARST0431 may at least in part be attributable to the lower percentage of these patients on ARST08P1 (12/168 on ARST08P1 compared to 20/97 on ARST0431). The 3-year EFS for patients less than 10 years old with ERMS of 64% observed on this study is concordant with outcome for the same subset of patients on ARST0431. However, in this study, the 38 patients with < two Oberlin risk factors had a 3-year EFS of 38%, which was inferior to 43 similar patients on ARST0431 (69%), despite virtually identical backbone chemotherapy. Three year EFS was also somewhat inferior among patients with two or more Oberlin risk factors (n=130, 9%) compared to ARST0431 (n=66, 20%). Although the addition of cixutumumab or temozolomide could have antagonized the backbone chemotherapy, firm explanations for the inferior EFS are elusive. Assessing whether addition of either agent impacted dose intensity was not an aim of this trial and has not been analyzed.

Due to the lack of outcome improvement with cixutumumab or temozolomide, we do not recommend either agent for the treatment of patients with newly diagnosed metastatic RMS. While the administration of cixutumumab in combination with chemotherapy was generally feasible, the rare but potentially life-threatening liver toxicity observed is a concern. Monitoring for liver toxicity will be important if cixutumumab is evaluated in combination with chemotherapy in future clinical trials. As neither cixutumumab nor temozolomide improved outcomes, neither agent will be evaluated in the upfront setting for patients with localized RMS. In addition, as outcome for the majority of patients with metastatic RMS with the ARST0431 backbone is not superior to outcomes with less intensive backbones, future studies for metastatic RMS will evaluate novel agents on a less intensive backbone, such as VAC or VAC/VI and explore the utility of early novel endpoints (including early functional imaging responses) as part of smaller phase II screening or selection design studies to prioritize agents for testing in larger phase III studies.¹