A phase 1 study of vinblastine in combination with carboplatin for children with low-grade gliomas: a Children's Oncology Group phase 1 consortium study

Abstract

Both carboplatin and vinblastine have demonstrated single-agent activity in children with low-grade gliomas. A phase 1 trial evaluating 2 different schedules of these 2 agents in combination was performed: (1) Schedule A = carboplatin (140 mg/m²) weekly × 3 + vinblastine (4.5 or 3.5 mg/m²) weekly × 6, every 6 weeks; (2) Schedule B = carboplatin (300, 400, or 500 mg/m²) on day 1 + vinblastine (4.0 mg/m²) weekly × 3, every 4 weeks. Twenty-six patients, median (range) age 4.4 (0.7–14.8) years, were enrolled. Four of 9 patients enrolled on Schedule A had recurrent grade 4 neutropenia, suggesting that this schedule was not feasible. Seventeen patients were enrolled on Schedule B. At the 500 mg/m² carboplatin dose level, 2 of 3 patients developed dose-limiting toxicity (elevated alkaline phosphatase, neutropenia). At the 400 mg/m² carboplatin dose level, none of the 6 patients had dose-limiting toxicity. Ten of 16 patients who received treatment on Schedule B completed the prescribed 12 courses. Of the 21 patients evaluable for response, central review confirmed 1 partial response and 6 minor responses. Eleven patients had stable disease (>3 months) and 3 developed progressive disease. Seven of 9 patients with visual pathway tumors and acute visual changes prior to enrollment had documented improvement. The recommended phase 2 dose and schedule is carboplatin, 400 mg/m²/dose on day 1, and vinblastine, 4 mg/m²/dose, weekly × 3, repeated every 4 weeks. Further study of this regimen in patients with low-grade glioma is warranted.

A number of treatment regimens for patients with unresectable low-grade gliomas have been evaluated over the last 2 decades, yet the long-term outcome remains suboptimal. A long-term follow-up study from St. Jude found that children with hypothalamic/chiasmatic low-grade gliomas treated with chemotherapy had a 6-year progression-free survival rate of only 12 ± 11%.¹ The combination of carboplatin plus vincristine has been considered by many to be the frontline treatment for young children with unresectable low-grade gliomas. Although carboplatin is known to be one of the most effective single agents against low-grade gliomas,^2–4 the activity of vincristine alone against low-grade gliomas has not been evaluated in the modern imaging era. Furthermore, the primary toxicity of vincristine is neurologic, often exacerbating preexisting deficits and confounding the clinical picture. As a result, the value of vincristine in childhood brain tumor regimens is being reconsidered. Vinblastine, another frequently utilized cytotoxic agent, also has potent anti-angiogenic properties. The doses of vinblastine needed to cause endothelial damage are substantially lower than cytotoxic doses,⁵ and picomolar concentrations of vinblastine can inhibit angiogenesis.⁶ A multi-institutional pilot study using single-agent vinblastine given weekly for a year in pediatric patients with recurrent and refractory low-grade gliomas showed significant efficacy.⁷ We therefore performed a phase 1 trial of the combination of carboplatin and vinblastine for children with unresectable low-grade gliomas.

Patients and methods

Eligibility

Patients <22 years of age with low-grade gliomas were eligible if they (1) had >1.5 cm² residual tumor following surgery or biopsy and were felt to be at high risk for neurologic and/or visual impairment if the tumor progressed or (2) had a low-grade glioma that had shown clear radiographic and/or clinical evidence of progression. Biopsy was not required for patients with neurofibromatosis type 1 (NF-1)–related visual pathway tumors (VPTs) or for non–NF-1 patients with VPTs involving the optic nerves and/or optic radiations. Patients with NF-1–related VPTs had to have evidence of clinical and/or radiographic progression. Newly diagnosed patients could not have received prior therapy except for steroids and/or surgery. Patients with recurrent tumors had to have recovered from the acute toxic effects of all prior treatment and could not have received myelosuppressive chemotherapy within 3 weeks of study entry, external beam radiation or gamma knife radiosurgery to a target lesion within 9 months, or biologic therapy or hematopoietic growth factor support within 7 days prior to enrollment. Patients who had previously received carboplatin and/or vinblastine were eligible as long as they had not progressed while on therapy and had not required dose reductions due to toxicity or developed an allergy to carboplatin. Other eligibility criteria included a KPS or Lansky score ≥50% and adequate bone marrow (absolute neutrophil count [ANC] ≥1000/µL, platelet count ≥100 000/µL, and hemoglobin ≥8 g/dL), renal (normal serum creatinine for age or a creatinine clearance rate [CrCl] or glomerular filtration rate [GFR] ≥70 mL/min/m²), and hepatic (total bilirubin <1.5 × normal and serum glutamic pyruvic transaminase <2.5 × normal) function. Exclusion criteria included pregnancy, uncontrolled infection, a diffuse intrinsic brainstem glioma, or the presence of any syndrome of inappropriate secretion of antidiuretic hormone or shunt-related ascites at the time of study entry. Informed consent was obtained from patients, parents, or guardians at the time of enrollment. The protocol was approved by the institutional review boards of the participating institutions.

Drug Administration and Study Design

Two different treatment schedules that involved administering carboplatin as a 30- to 60-min IV infusion and vinblastine as an IV push were evaluated (Table 1). Courses on both schedules were repeated when ANC was ≥1000/mm³ and platelets ≥100 000/mm³ immediately prior to the start of each course. Schedule A administered 6 weekly doses of vinblastine, at a starting dose of 4.5 mg/m²/week, in combination with 3 weekly doses of carboplatin at a starting dose of 140 mg/m²/week. When this schedule proved intolerable due to myelosuppression, Schedule B was then studied. Schedule B administered 3 weekly doses of vinblastine at a starting dose of 4 mg/m²/week, in combination with a single dose of carboplatin, at a starting dose of 400 mg/m² every 4 weeks. On the latter schedule, patients could receive a maximum of 12 four-week courses in the absence of tumor progression or unacceptable toxicity. The use of growth factors was not allowed.

Table 1.

Chemotherapy schedules

Dose Level	Carboplatin (mg/m2)	Schedule	Vinblastine (mg/m2)	Schedule
Schedule A: Repeated every 6 weeks
−1	140	Weekly × 3	3.5	Weekly × 6
1a	140	Weekly × 3	4.5	Weekly × 6
2	140	Weekly × 3	6	Weekly × 6
Schedule B: Repeated every 4 weeks
−1	300	Day 1	4.0	Weekly × 3
1a	400	Day 1	4.0	Weekly × 3
2	500	Day 1	4.0	Weekly × 3

^aStarting dose level.

Determination of the Maximum Tolerated Dose

A minimum of 3 patients were treated at each dose level. If no unacceptable toxicity was experienced, the dose was escalated in subsequent patients. If 1 patient experienced dose-limiting toxicity (DLT), 3 additional patients were accrued. If 2 or more patients experienced DLT, then the maximum tolerated dose (MTD) was exceeded, and 3 additional patients were treated at the next lower dose level. The MTD was defined as the maximum dose level at which no more than 1 of 6 patients experienced DLT and above which 2 or more patients encountered DLT.

DLT was evaluated during course 1 and was defined as grade 4 thrombocytopenia or neutropenia; a delay of >14 days between treatment courses resulting from drug-related myelosuppression or nonhematologic toxicity (NHT) for Schedule A or >7 days for Schedule B; or grade 3 or 4 NHT (excluding nausea and vomiting, infection, fever, or grade 3 hepatic toxicity that returned to grade 1 or less within 1 week). Toxicities were graded according to Common Terminology Criteria for Adverse Events v3.0 (http://ctep.cancer.gov/forms/CTCAEv3.pdf).

Toxicity Evaluations and Dose Modifications

Patients were evaluated at least weekly during the first course of chemotherapy and then prior to each course. Complete blood counts were performed at least weekly. Chemistries and liver function tests were performed weekly during the first course of chemotherapy and then prior to each course. Patients who experienced a hematologic DLT had chemotherapy held until recovery of ANC ≥1000/mm³ and platelet count ≥100 000/mm³ and then resumed at the next lower dose level. Patients were permitted up to 2 dose reductions. Patients with reversible grade 2 NHT that required dose reduction or grade 3 NHT felt to be related to vinblastine had further doses held until toxicity recovered to baseline, at which point vinblastine was resumed at the next lower dose level. If the serum creatinine increased to greater than the upper limit of normal for age, a CrCl or GFR was obtained. If the CrCl or GFR was <75% of normal for age, the carboplatin dose was held (with continuation of the vinblastine) and restarted with a 25% dose reduction during the next course once the CrCl or GFR recovered to ≥75% normal for age. Children with mild or questionable allergic reactions to carboplatin without respiratory compromise could continue on therapy with diphenhydramine and steroid pretreatment. If the allergic reaction included any degree of respiratory compromise, the patient was taken off protocol therapy.

Response Evaluation

Response was assessed by MRI scan after course 1 and every other course thereafter. Patient response required stability or decrease of steroid dose and confirmation by a second scan at least 4 weeks later. All reported responses were confirmed by central review. Complete response (CR) was defined as no evidence of tumor on MRI; partial response (PR) as a >50% decrease in the product of the greatest tumor diameter and its perpendicular diameter on MRI scan; and minor response (MR) as ≥25% but ≤50% reduction in the product of the greatest tumor diameter and its perpendicular diameter. Progressive disease was defined as a >25% increase in tumor size or the emergence of new lesions. All other patients were considered to have stable disease.

Results

Patient Characteristics

Between September 2006 and September 2008, 26 patients with a median age of 4.4 years (range, 0.7–14.8 years), 12 male and 14 female, were enrolled (Table 2). Seven patients (27%) had received prior chemotherapy; 2 patient had received prior radiation therapy. Of the 15 patients who underwent surgery or biopsy, 13 had juvenile pilocytic astrocytomas, 1 had a grade 2 astrocytoma, and 1 had a low-grade glioma “not otherwise specified.” Fifteen patients had VPTs, 6 of whom had NF-1.

Table 2.

Patient characteristics

#	Dose Level	NF-1	Prior Regimens	No. Courses	Age (y)	Diagnosis	Tumor Location	DLT	Reason off Study	Best Response
Schedule 1
1	1	No	0	4	1.9	JPA	Medulla	Gr 4 neutropenia	Recurrent Gr 4 neutropenia	SD
2	1	No	3	1	12.5	Low-grade	Third ventricle	Gr 4 neutropenia	Family choice	NE
3	1	No	0	9	4.6	JPA	Basal ganglia	No	Completed therapy	MR
4	2	No	0	4	1.5	JPA	Medulla	No	Recurrent neutropenia	SD
5	2	No	1	3	3.8	JPA	Hypothalamus/chiasm	No	Recurrent neutropenia	NE
6	2	No	4 + RT	2	10.6	JPA	Hypothalamus/chiasm	G 4 neutropenia	Recurrent neutropenia	NE
7	2	No	0	9	4.4	JPA	Midbrain (exophytic)	No	Completed therapy	SD
8	2	No	0	1	5.0	No	Cervicolthoracic spine	No	PD	PD
9	2	Yes	0	9	4.3	ND	Visual pathway	No	Completed therapy	MR
Schedule 2
10	0	Yes	0	8	1.9	ND	Visual pathway	G 4 neutropenia	Recurrent neutropenia
		Yes	0	12	1.1	ND	Visual pathway	Inevaluable	Completed therapy	MR
12	0	No	1	12	3.7	ND	Hypothalamus/chiasm + ventricular metastases	No	Completed therapy	MR
13	0	No	0	4	5.2	JPA	Visual pathway	No	Visual acuity worsening despite radiographic improvement	MR
14	0	No	0	11	11.3	JPA	Hypothalamus/chiasm	No	Family choice 11	SD
15	0	Yes	0	12	6.6	ND	Visual pathway	No	Completed therapy	SD
16	0	No	1	5	5.6	ND	Visual pathway	No	PD	PD
17	1	Yes	2	12	13.5	ND	Scattered/pons/temporal	No	Completed therapy	SD
18	1	No	0	0	0.8	JPA	Brainstem	Inevaluable	Never started treatment	NE
19	1	No	1	12	3.8	JPA	Thoraco-lumbar spine	No	Completed therapy	SD
20	1	Yes	0	12	2.6	ND	Visual pathway	No	Completed therapy	PR
21	1	No	0	3	11.9	Grade 2 astro	Thalamus	No	PD	PD
22	1	Yes	0	12	2.3	ND	Visual pathway	No	Completed therapy	SD
23	1	No	RT	12	14.8	JPA	Leptomeningeal metastases	No	Completed therapy	SD
24	2	No	0	12	3.0	ND	Visual pathway	Gr 3 alk phos	Completed therapy	MR
25	2	No	0	12	10.9	JPA	Hypothalamus/chiasm	No	Completed therapy	SD
26	2	No	0	3	0.7	ND	Visual pathway	Gr 4 neutropenia	Recurrent Gr 4 neutropenia, course 2	NE

RT, radiation therapy; ND, not done; NE, not evaluable; PR, partial response; MR, minor response; SD, stable disease; PD, progressive disease; JPA, juvenile pilocytic astrocytoma.

Toxicity

Nine patients were treated on Schedule A (Table 3). Two of the 3 patients enrolled on the starting dose level (4.5 mg/m²/week vinblastine) developed dose-limiting neutropenia during course 1, and the vinblastine dose was de-escalated. One of 6 patients enrolled at dose level –1 (3.5 mg/m²/week vinblastine) developed dose-limiting neutropenia during week 4. Moreover, 4 of the 9 patients came off study because of recurrent grade 4 neutropenia in subsequent courses of chemotherapy, suggesting that Schedule A was not feasible.

Table 3.

Summary of dose-limiting toxicities

Schedule	Dose Level	No. Entered	No. Evaluable	No.with DLT	DLT (n)
A	1	3	3	2	Neutropenia (2)
	−1	6	6	1	Neutropenia (1)
B	1	7	6	1	Neutropenia (1)
	2	7	6	0
	3	3	3	2	Alkaline phosphatase (1)
					Neutropenia (1)

Seventeen patients were subsequently enrolled on Schedule B, of whom 3 were not fully evaluable for toxicity. One patient never started treatment, 1 was <10 kg and received chemotherapy doses based on weight rather than body surface area, and 1 did not have the appropriate safety labs drawn during course 1. One of 6 patients at dose level 1 (300 mg/m² carboplatin) had dose-limiting neutropenia. At the 400 mg/m² carboplatin dose level, none of the 6 patients had DLT. At the 500 mg/m² carboplatin dose level, 2 of 3 patients developed DLT. One patient had grade 4 neutropenia and the other patient developed an asymptomatic grade 3 elevation of alkaline phosphatase during week 4 of course 1. The latter patient had no other liver function abnormalities and a normal liver ultrasound. Chemotherapy was held until the alkaline phosphatase returned to normal 4 weeks later and restarted at the next lower dose level without recurrence of the alkaline phosphatase elevation.

Complications from myelosuppression were more frequent on Schedule A. Out of 42 courses, there were 2 episodes of fever and neutropenia without documented infection, 1 episode of cellulitis at the mediport site with neutropenia requiring IV antibiotics, 1 episode of a toenail infection with neutropenia not requiring hospitalization, and 1 episode of Clostridia dificile colitis. There was only 1 episode of fever and neutropenia without documented infection out of 153 courses on Schedule B. There were no episodes of bacteremia on either schedule. Thrombocytopenia was minimal, with only 1 episode of grade 3 thrombocytopenia noted in 195 patient-courses of therapy. Two of the 5 patients who developed a hematologic DLT had previously received multiple regimens of chemotherapy. No patient on either regimen received either a red cell or platelet transfusion.

NHT was mild, consisting primarily of grade 1 or 2 nausea and vomiting (Table 4). One patient developed a grade 2 peripheral neuropathy requiring a brace on his right lower extremity during course 10. Five patients on Schedule B (monthly carboplatin) developed allergic reactions to carboplatin during courses 5 (n = 2), 8, 9 or 12. Two of these patients received no further carboplatin either because they had completed therapy or because they had progressive disease. The other 3 patients continued on therapy with pretreatment.

Table 4.

Nonhematologic toxicities occurring in >10% of patients

Toxicity Type	Maximum Toxicity Grade
	Course 1 (24 patient-courses)a				Courses 2–12 (160 patient-courses)
	Grade 1	Grade 2	Grade 3	Grade 4	Grade 1	Grade 2	Grade 3	Grade 4
Fatigue	3				4	1
Anorexia	1	2			2	1
Constipation	2	1			5
Diarrhea	4				3	3
Nausea	4	1			8	1
Vomiting	5	4	1		3	6	2
ALT, SGPT	4				7
AST, SGOT	4				6
Pain, stomach	3				1

ALT, alanine aminotransferase; AST, aspartate aminotransferase; SGPT, serum glutamic pyruvic transaminase; SGOT, serum glutamic oxalacetic transaminase. ^a23 patient-courses were deemed evaluable for nonhematologic toxicities.

Response

Twenty-one patients were evaluable for response. Five patients were not evaluable for the following reasons: 1 patient never started treatment, 3 were removed from treatment for toxicity, and 1 was removed by family choice. There was 1 PR and 6 MRs, all of which were confirmed by central radiology review. The patient with the PR improved to a CR, and 1 patient with an MR improved to meet criteria for a PR during the last course of treatment, which could only be confirmed on the first off-study scan. Eleven patients had stable disease lasting >3 months, and 3 developed progressive disease. Seven of 9 patients with VPTs who had acute ophthalmologic changes prior to enrollment showed documented improvement during the course of treatment, including resolution of optic nerve swelling and improvement in proptosis and/or visual acuity. Ten of 16 patients who received treatment on Schedule B completed the prescribed 12 courses. One patient stopped therapy after 11 courses at the family's request.

Discussion

Although chemotherapy has been shown to delay the need for radiation therapy in the majority of patients with unresectable low-grade gliomas, the published CR rates for patients treated with the most commonly used chemotherapy regimens range from 0% to 5%.^4,8–10 Long-term progression-free survival rates with chemotherapy alone are quite low and non–NF-1 patients with such tumors frequently receive more than 1 regimen of chemotherapy to delay the use of radiation for as long as possible. For patients with visual pathway gliomas, however, treatment with chemotherapy rarely results in significant recovery of vision, and many of these patients sustain significant visual impairment.¹¹

The combination of carboplatin and vincristine is generally considered the standard frontline treatment for unresectable low-grade gliomas, as evidenced by the fact that all of the patients who had received prior chemotherapy in this trial had received carboplatin and vincristine. A randomized Children's Oncology Group study using a “pick the winner” design compared carboplatin/vincristine with TPCV (6-thioguanine, procarbazine, CCNU, and vincristine) and found that the 5-year event-free survival with carboplatin/vincristine was 34.6% ± 4.8% compared with a 48.9% ± 4.8% 5-year event-free survival with TPCV, casting doubt on whether carboplatin/vincristine should truly be considered the best standard regimen.¹² Most oncologists do not utilize the TPCV regimen as frontline therapy largely because of the concern for adverse late effects in patients who are expected to have a reasonably long-term survival. Patients with NF-1, who frequently develop low-grade gliomas, are at particularly high risk for alkylator-related second malignancies,¹³ which limits the number of available regimens for these patients.

An alternative regimen used by many institutions in the United States and Europe consists of single-agent carboplatin administered monthly at a dose of 560 mg/m²/month.^3,4 Toxicity with this regimen is quite tolerable; 85% of patients experienced disease stabilization or better for a median of 22 months.³ No randomized studies comparing single-agent carboplatin with carboplatin/vincristine have been undertaken, although the published disease stabilization and event-free survival rates are similar.^3,8 Because the data for the efficacy of vincristine as a single agent against low-grade gliomas are obscure at best, it would seem rational to evaluate the combination of carboplatin with an agent with proven activity.

A recent clinical trial of single-agent vinblastine given weekly for a year demonstrated activity against recurrent and refractory low-grade gliomas.⁷ As of the most recent published follow-up, the study had enrolled 51 children with progressive low-grade gliomas following prior chemotherapy (n = 50) and/or radiation therapy (n = 9). Independent central radiology review of the first 31 patients confirmed 1 CR, 3 PRs, 9 MRs, 10 stable disease, and 10 progressive disease, with an overall response rate (CR + PR + MR) of 42%. Preclinical data showing that the antitumor activity of cytotoxic chemotherapeutic agents can be augmented by combining them with angiogenesis inhibitors¹⁴ also supports the rationale for this combination. Furthermore, vinblastine can be given at home with the assistance of personnel from a home care agency, decreasing the need for frequent clinic visits.

Our study found that the myelosuppressive effects of vinblastine can be limiting when this drug is combined with carboplatin. However, a reasonable alternative schedule proved tolerable, which allowed for 3 weekly doses of vinblastine combined with monthly carboplatin. The results were encouraging, with both radiographic and clinical responses seen. Objective responses were observed in both recurrent and newly diagnosed patients. However, given the small number of patients and the different dosing schedules, no conclusions can be drawn as to whether there was a higher response rate in newly diagnosed patients. A comparison of this regimen with others in terms of efficacy, toxicity, and convenience is warranted.

Funding

National Cancer Institute, National Institutes of Health [U01 CA97452, U10 CA 98543 to PA]; National Center for Research Resources, National Institutes of Health [MO1 RR00188 to SB].