Phase I and Pharmacokinetic Study of Sunitinib in Pediatric Patients with Refractory Solid Tumors: A Children's Oncology Group Study

Sunitinib is an oral small molecule multitargeted receptor tyrosine kinase inhibitor that inhibits VEGF receptors (VEGFR), platelet-derived growth factor receptors (PDGFR), c-KIT, Flt3, colony stimulating factor-1 receptor, and RET (1–3). Sunitinib has been studied in adults with a range of malignancies, particularly gastrointestinal stromal tumor (GIST) and renal cell carcinoma (4–6). The recommended adult dose is 50 mg/d for 28 days followed by a 14-day break (5, 7), though some groups have evaluated daily dosing without a 14-day break (8). In adults, common toxicities were fatigue, gastrointestinal symptoms, and myelosuppression (4–7, 9). Pharmacodynamic effects of sunitinib observed in studies in adults include: increased plasma VEGF; increased circulating endothelial cells (CEC); decreased soluble VEGFR2; and decreased monocyte counts (10–13).

In preclinical model systems, sunitinib has activity against a number of pediatric solid tumors, including neuroblastoma and sarcoma models with shown growth inhibition during sunitinib exposure (14, 15). Sunitinib has not previously been evaluated in children with the exception of 11 children with GIST (16, 17) who were treated with doses ranging from 25 to 50 mg/d. The toxicity profile of sunitinib in this small number of children was similar to the adult experience.

The current report describes the results of a pediatric phase I study of sunitinib in children with refractory or recurrent solid tumors. The primary aims of the study were: to define the maximum tolerated dose (MTD) and toxicities associated with sunitinib; and to characterize the pharmacokinetics of sunitinib in children. Secondary endpoints included an assessment of antitumor activity and pharmacodynamic biomarker studies.

Patients were eligible for participation if they were 2 to 21 years of age, had histologic diagnosis of solid malignancy with measurable or evaluable disease, and had no known curative options. Patients were required to have a Karnofsky (age >10 years) or Lansky (age ≤10 years) performance score 50 or more and to have recovered from previous therapy. Patients were required to have adequate baseline bone marrow, renal, hepatic, pancreatic, and cardiac function according to defined protocol criteria. Patients with preexisting hyper- or hypothyroidism were required to have stable thyroid function. Patients could not be receiving concomitant antihypertensive medications and had to have blood pressure less than 95th percentile for age, height, and gender.

Exclusion criteria included: inability to swallow capsules; body surface area less than 0.5 m² (due to available capsule sizes); concurrent use of strong CYP3A4 inducers or inhibitors; treatment with agents that might increase the risk of bleeding complications; presence of pleural-based tumors; or uncontrolled infection.

An amendment to the eligibility criteria was made after enrollment of the first 12 patients (part A) because of cardiac and hematologic toxicity. In the amended protocol (part B), patients with previous anthracycline treatment or cardiac radiation exposure were excluded. Patients with known bone marrow metastases were also excluded from part B to provide a clear assessment of the hematologic toxicity of sunitinib after a patient in part A developed dose-limiting neutropenia. This amendment also allowed enrollment of patients with central nervous system tumors. These patients were excluded from part A because experience with the use of antiangiogenic agents in children with central nervous system tumors was limited at the time of study initiation.

The protocol was approved by each Institutional Review Board. Written informed consent (and assent when applicable) was obtained for all patients.

Patients received sunitinib orally once daily for 28 days followed by a 14-day break, such that the duration of 1 cycle of therapy was 42 days. The starting dose level was 20 mg/m²/dose (approximately 80% of the approved adult dose) in part A and 15 mg/m²/dose in part B. All doses were given as capsules and were rounded to the nearest 12.5 mg using a dosing nomogram. Patients without disease progression or unacceptable toxicity could receive up to 9 cycles of therapy.

Patients had routine physical examinations and surveillance laboratory testing to evaluate for toxicity. Electrocardiograms and echocardiograms were obtained at the end of cycles 1 and 2 and thyroid stimulating hormone (TSH) levels were monitored at the end of cycle 3 and then every other cycle. On the basis of the toxicities observed in part A, patients in part B were required to have additional electrocardiograms and echocardiograms at the end of cycles 3, 6, and 9 as well as a TSH levels on day 28 of cycle 1, days 1 and 28 of cycle 2, and day 28 of every subsequent odd number cycle. All patients had baseline plain radiographs of the proximal tibial growth plate. Patients with open tibial growth plates had repeat radiographs at the end of cycle 3 and then every other cycle.

Toxicities were graded according to the National Cancer Institute (NCI) Common Terminology Criteria, version 3.0. Dose-limiting toxicity (DLT) was defined as any of the following that were attributed as at least possibly related to sunitinib: grade 4 neutropenia; grade 4 thrombocytopenia; any grade 4 nonhematologic toxicity; grade 2 cardiac systolic dysfunction; blood pressure more than 25 mmHg above the 95th percentile for age, height, and gender; any grade 2 nonhematologic toxicity for 7 or more days that required drug interruption; any nonhematologic toxicity that required drug interruption for more than 7 days; or any grade 3 nonhematologic toxicity with the exception of nausea and vomiting of less than 3 days duration, alanine aminotransferase (ALT) elevation that returned to baseline prior to the next cycle, aspartate aminotransferase (AST) elevation, fever of less than 5 days duration, electrolyte abnormalities responsive to oral supplementation, asymptomatic elevations of amylase or lipase resolving to less than grade 1 within 7 days of drug interruption.

Patients underwent disease reevaluation at the end of cycle 1 and then every other cycle. For patients with measurable disease, tumor response was evaluated by Response Evaluation Criteria in Solid Tumors (RECIST; ref. 18).

All patients were required to have trough plasma samples obtained on days 7, 14, 21, and 28 of cycle 1. Patients were also asked to participate in extended pharmacokinetic sampling after the first dose with samples obtained prior to sunitinib and 1, 2, 4, 6, 8, 24, and 48 hours after the first dose. Plasma concentrations of sunitinib and its main active metabolite, SU12662, were measured by using liquid chromatography/tandem mass spectrometry by Bioanalytical Systems, Inc. (BASi) as previously described (19). Sunitinib and SU12662 plasma concentration–time data were analyzed by standard noncompartmental methods using WinNonlin Pro (Pharsight Corp.).

Plasma was obtained at baseline and day 28 of cycle 1 in consenting patients to assess the impact of sunitinib on angiogenic biomarkers. VEGF, VEGFR1, VEGFR2, placental growth factor (PGF1), stromal cell-derived factor-1 (SDF-1), and endoglin levels were measured by ELISA using commercially available kits (R&D Systems, Inc.). CECs and endothelial progenitor cells (CEP) were measured using a multiparameter FACSCalibur flow cytometer (BD Biosciences) as previously described (20). CECs were defined as CD31+/CD45−/CD133− and CPCs were defined as CD31+/CD45−/CD133+. Additional testing was done to determine whether CECs and CEPs were apoptotic, expressed endoglin, or expressed VEGFR2 (21, 22). Monocyte counts were obtained at each institution by using standard methods.

The primary endpoint for determining dose escalation was DLT in cycle 1. In part A, a standard 3 + 3 phase I dose escalation design was utilized. Briefly, patients were enrolled in cohorts of 3 patients each and, if no first-cycle DLT was observed, the dose was escalated. If 1 patient had a DLT, then up to 3 additional patients were treated at that same dose level. If 2 or more patients in a given cohort of 6 patients had DLT, then the MTD was exceeded and the sunitinib dose was deescalated. In part B, a more conservative dose escalation strategy was adopted such that an additional 3 patients were enrolled at a dose level if 0 or 1 of the first 3 patients had a DLT. The dose was then only escalated if 1 or more patient in the cohort of 6 had a DLT. Therefore, a minimum of 6 patients were to be evaluated per dose level in part B before the dose could be escalated.

Changes in biomarkers of angiogenesis obtained at baseline and at the end of cycle 1 were assessed using the paired t test or Wilcoxon signed rank test for nonnormal data. Differences between dose level and clinical response (progressive disease vs. stable disease) based on changes in laboratory markers from baseline to end of cycle 1 were assessed by using the Wilcoxon rank-sum test.

Characteristics of the 23 patients, who received a median (range) of 1 (1 to 9) cycle of sunitinib, are shown in Table 1. Eight patients were younger than 12 years of age at study entry. One patient had received previous therapy with another VEGFR tyrosine kinase inhibitor, sorafenib.

In part A, 1 of 3 patients treated at the starting dose level of 20 mg/m² developed grade 2 systolic dysfunction. This dose level was expanded to 6 patients and 2 additional patients had DLTs consisting of grade 3 anorexia and grade 4 neutropenia. The dose was then deescalated to 15 mg/m². In the first 3 patients treated at this dose level, 1 patient had a prolonged grade 3 ALT elevation that was dose limiting. Three more patients were enrolled, one of whom developed dose-limiting cardiac toxicity with grade 3 diastolic and grade 2 systolic dysfunction (Table 2). Enrollment to part A was suspended and, following a detailed evaluation of both cases of reversible cardiac dysfunction, the protocol was amended to exclude patients with previous anthracycline or cardiac radiation exposure.

Eight patients were enrolled on part B of the study at a starting dose level of 15 mg/m². This starting dose was chosen for part B because it was anticipated that the risk of cardiac toxicity would be less in this cohort of patients who had not received previous anthracycline therapy or cardiac radiation. Two patients had early disease progression during cycle 1 and were therefore not fully evaluable for DLT. None of the 6 evaluable patients at this dose level experienced DLT in cycle 1. Three patients were treated at 20 mg/m², 2 of whom had DLTs in cycle 1. One patient developed grade 4 hyperuricemia in the setting of grade 2 diarrhea and dehydration. The second patient developed grade 5 aspiration pneumonia in the setting of grade 4 punctate hemorrhage into a progressive diffuse pontine glioma. Although this episode may have been related to the natural history of her disease, (23) a possible attribution to study therapy could not be excluded. These results established 15 mg/m²/d for 28 days followed by a 14-day break as the MTD for patients without previous cardiac radiation or anthracycline exposure. No cardiac dysfunction was observed in part B of the trial.

Because of dose rounding to accommodate capsule size, the 14 patients treated at the 15 mg/m² dose level received a mean actual dose of 14.8 ± 2.1 mg/m². The mean dose for the 6 evaluable patients treated at 15 mg/m² in part B was 14.5 mg/m². The 9 patients treated at the 20 mg/m² dose level received a mean actual dose of 18.8 ± 2.8 mg/m². Patients with DLT were not more likely to receive an increased dose due to rounding compared with patients without DLT.

Details of toxicity reported in more than 10% of patients in cycle 1 are shown in Table 3. Neutropenia, thrombocytopenia, and transaminase elevation were the most commonly reported toxicities. Gastrointestinal symptoms and fatigue were the most commonly reported symptoms. The incidence of neutropenia and thrombocytopenia did not seem to increase in subsequent cycles. Nonhematologic toxicity was uncommon in subsequent cycles of therapy. The main differences in toxicity profiles between parts A and B were the absence of reported cardiac toxicity and thrombocytopenia in part B.

Three of 23 patients developed hypothyroidism (≤ grade 2) after 1 cycle of therapy. Two patients developed hypothyroidism (≤ grade 2) during subsequent cycles of therapy. Hypertension was rarely reported. One patient had grade 1 and another patient had grade 2 QT_c interval prolongation, with maximal QT_c of 490 m/sec. Of the 15 patients with open tibial growth plates at study entry, 1 patient received 9 cycles of therapy and did not show any growth plate thickening. The remaining 14 patients were removed from protocol therapy before undergoing a second growth plate evaluation. Changes in hair color and texture observed in adult studies of sunitinib (7) were not reported.

All 23 patients submitted at least 1 sample for evaluation of steady-state trough levels (Table 4). Steady state concentrations of sunitinib, the active metabolite SU12662, and total drug (sunitinib + SU12662) were reached by day 7. Five of 14 patients (36%) at the 15 mg/m² dose level and 5 of 9 patients (56%) at the 20 mg/m² dose level had steady-state total drug trough concentrations more than 50 ng/mL, the target total drug concentration for kinase inhibition based on preclinical studies (2). Nine patients participated in the extended sampling study (Table 5). A median peak plasma concentration of 16.8 (range 9.5–61.4) ng/mL sunitinib was achieved 7 hours after the first dose. Median values of the sunitinib and SU12662 half-life, based on the accumulation ratios calculated from the steady-state data were 38.7 (range 23.9–61.7) and 93.4 (range 47.4–176) hours, respectively. Steady-state trough concentrations did not seem to correlate with occurrence of DLT (data not shown).

No objective responses were observed in the 17 patients evaluable for response. Four patients had a best response of stable disease, including patients with diffuse pontine glioma (2 cycles), osteosarcoma (4 cycles), epithelioid sarcoma (9 cycles), and ganglioglioma (9 cycles). The latter 2 patients did not progress on study therapy, but were removed from study after receiving the maximum of 9 cycles. Thirteen patients had a best response of progressive disease. Of the 6 patients not evaluable for best response, 2 patients did not complete cycle 1 due to toxicity. The remaining 4 patients had stable disease at the completion of cycle 1, but were removed from protocol therapy for toxicity after cycle 1. Of these, 1 patient with neuroblastoma who received 1 cycle at the 20 mg/m² dose level had resolution of the disease noted at baseline on a ¹²³I-meta-iodobenzylguanidine scan but had persistent bone marrow disease. The patient was removed from study due to toxicity and a subsequent ¹²³I-MIBG scan was not obtained.

Twelve patients had paired baseline and day 28 plasma samples available for evaluation of angiogenic factor biomarkers by ELISA. Plasma VEGF and PGF concentrations increased significantly over the course of the first cycle of therapy (P = 0.002 and 0.0003, respectively; Fig. 1A and B). Plasma soluble VEGFR2 and endoglin concentrations decreased significantly over the course of the first cycle of therapy (P = 0.0009 and 0.01, respectively; Fig. 1C and D). Changes in VEGF, PGF, soluble VEGFR2, and endoglin did not vary by sunitinib dose level or by response to therapy over the course of cycle 1. Plasma soluble VEGFR1 and SDF-1 concentrations did not change in a consistent manner with therapy.

Totals and subsets of CECs and CEPs in 11 patients with paired baseline and day 28 samples did not reliably change with therapy. Paired day 1 and day 28 absolute monocyte counts (AMC) were available for 18 patients. One patient had an increase in AMC. The remaining 17 patients had decreases in AMC (median decrease 117 μL; range 18–1253 μL).

The MTD of sunitinib for pediatric patients without previous anthracycline exposure or previous cardiac radiation was 15 mg/m²/d. This dose is lower than the approved adult dose of 50 mg/d (approximately 28 mg/m²/d based on an average adult body surface area of 1.8 m²). Previous case series have indicated that children with GIST are able to tolerate sunitinib doses of 25 to 50 mg (16, 17). The lower doses tolerated in this formal dose escalation study may reflect more intensive previous therapy in the patient population in the current study. In addition, adults treated with sunitinib require dose reductions in 11% to 50% of cases (5, 6, 24, 25). The typical reduced adult dose of 37.5 mg/d (approximately 21 mg/m²/d based on an average adult body surface area of 1.8 m²) is still greater than the pediatric dose determined in the current study. In addition, the toxicity profile of sunitinib in children may differ from the toxicity profile in adults. In the current study, myelosuppression and transaminase elevations were the most commonly reported toxicities compared with fatigue and gastrointestinal symptoms in adults (4–6).

The pharmacokinetic data of the current study found that the apparent clearance of sunitinib was similar between children and adults (7, 26–29). These results are consistent with a recent population pharmacokinetic analysis of sunitinib in adults (median age 55 years; range 18–87) that showed that clearance did not correlate with age (27). As a consequence of the lower doses tolerated in this study, steady-state trough total drug concentrations fell below target concentrations (50 ng/mL) in many of the patients. Although peak concentrations would be predicted to exceed target concentrations, these data suggest that kinase inhibition may be intermittent for some patients treated at the MTD in this study. It is possible that plasma drug concentrations may not reflect intratumoral drug concentrations and that there may be different target exposures required for antitumor activity in different model tumor types.

Due to cardiac toxicity, a recommended phase 2 dose was not determined in part A of the study, which included patients with previous anthracycline exposure or previous cardiac radiation. Reports from adult patients treated with sunitinib show an incidence of grade 3 or symptomatic congestive heart failure of 7% to 15% (30–32). As in this current study, the onset of cardiac toxicity is often early in the course of therapy with sunitinib (33, 34). Although both patients in the current study recovered normal cardiac function after discontinuing sunitinib, not all adults have had reversible cardiac dysfunction (32, 33). The mechanism of cardiac toxicity with sunitinib remains unclear, though some preclinical studies have shown an increase in myocyte apoptosis in response to sunitinib (30, 35). Because cardiac toxicity precluded the determination of a recommended sunitinib dose in patients with previous anthracycline exposure, future evaluation of this agent is likely to be limited to specific pediatric oncology populations who do not routinely receive therapy with anthracyclines, for example, children with primary central nervous system tumors.

There were no objectives antitumor responses in this trial. However, several patients with a variety of different tumors, including sarcoma and glioma, may have benefited as evidenced by stable disease for multiple cycles. Of note, the 2 patients with stable disease for 9 cycles had histologies which may show indolent growth: epithelioid sarcoma and ganglioglioma. In addition to recognized activity in GIST, sunitinib and other antiangiogenic tyrosine kinase inhibitors (sorafenib and pazopanib) have shown clinical activity in several other subtypes of sarcoma, including desmoplastic small round cell tumor, alveolar soft part sarcoma, synovial sarcoma, and angiosarcoma (8, 36, 37). Sunitinib has been effective in the management of adult patients with renal cell carcinoma and central nervous system metastases, indicating that this agent is distributed to central nervous system tumors (38–40). Experience with sunitinib for primary central nervous system tumors is limited to studies showing preclinical activity in glioma models (41, 42). Given a reported relationship between drug exposure and clinical benefit in adults (43), it is possible that the lack of objective responses in the current study was a result of inadequate drug exposure given the lower doses tolerated in children versus adults. Intra-patient dose escalation may be possible in a less heavily pretreated phase 2 population, which would result in increased sunitinib exposure, thereby maximizing the potential for antitumor activity.

As with previous studies, sunitinib resulted in consistent increases in plasma VEGF and PGF along with decreases in soluble VEGFR2 and monocyte counts (10, 11, 13). In contrast, changes in plasma SDF-1, total CECs, and CEC subsets noted in other studies could not be confirmed in this study. Plasma endoglin concentrations decreased significantly with sunitinib treatment. Previous published studies have not described the effects of sunitinib on plasma endoglin levels. Since endoglin is expressed on tumor-associated endothelial cells and increased levels of plasma endoglin have been reported in patients with malignant tumors (44), plasma endoglin levels may provide another pharmacodynamic tool for evaluating patients treated with sunitinib.

In conclusion, cardiac toxicity precluded a determination of a recommended phase 2 dose of sunitinib in children with previous therapy with a potential for cardiac toxicity. The MTD of sunitinib in children without previous anthracycline exposure or radiation therapy involving a cardiac field is lower than in adults (15 mg/m² vs. 28 mg/m²), despite similar apparent drug clearance. As a result, trough plasma concentrations at the MTD may be below “target” concentrations, as defined in preclinical studies. The implications of these findings are that further development of sunitinib in children with refractory malignancies is currently limited to a patient population without previous cardiotoxic therapy, specifically children with primary central nervous system tumors. In an attempt to maximize the potential for benefit from this agent in a planned phase II study in this population, a single intra-patient dose escalation is planned if the first cycle of therapy at 15 mg/m² is well tolerated. This strategy has successfully been employed for other anticancer agents (45) and may allow specific patients to safely receive higher doses of sunitinib than those identified in the current study.

Pfizer provided partial support for the analyses of sunitinib pharmacokinetics.

This work was supported by the Campini Foundation, Pfizer, NIH/NCRR/OD UCSF-CTSI grant number KL2 RR024130, and NCI U01 CA97452.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.