A Pilot Study of Hypofractionated Stereotactic Radiation Therapy and Sunitinib in Previously Irradiated Patients With Recurrent High-Grade Glioma

Abstract

Purpose/Objective(s):

Angiogenic blockade with irradiation may enhance the therapeutic ratio of radiation therapy (RT) through vascular normalization. We sought to determine the safety and toxicity profile of continuous daily-dosed sunitinib when combined with hypofractionated stereotactic RT (fSRT) for recurrent high-grade gliomas (rHGG).

Methods and Materials:

Eligible patients had malignant high-grade glioma that recurred or progressed after primary surgery and RT. All patients received a minimum of a 10-day course of fSRT, had World Health Organization performance status of 0 to 1, and a life expectancy of >3 months. During fSRT, sunitinib was administered at 37.5 mg daily. The primary endpoint was acute toxicity, and response was assessed via serial magnetic resonance imaging.

Results:

Eleven patients with rHGG were enrolled. The fSRT doses delivered ranged from 30 to 42 Gy in 2.5- to 3.75-Gy fractions. The median follow-up time was 40 months. Common acute toxicities included hematologic disorders, fatigue, hypertension, and elevated liver transaminases. Sunitinib and fSRT were well tolerated. One grade 4 mucositis toxicity occurred, and no grade 4 or 5 hypertensive events or intracerebral hemorrhages occurred. One patient had a nearly complete response, and 4 patients had stable disease for >9 months. Two patients (18%) remain alive and progression-free >3 years from enrollment. The 6-month progression-free survival was 45%.

Conclusions:

Sunitinib at a daily dose of 37.5 mg given concurrently with hypofractionated stereotactic reirradiation for rHGG yields acceptable toxicities and an encouraging 6-month progression-free survival.

Introduction

Recurrent glioblastoma multiforme (rGBM) remains a devastating disease. Strategies using systemic antiangiogenic agents that target vascular endothelial growth factor (VEGF) or the signaling of its receptor have demonstrated clinical efficacy in high-grade glioma (HGG). One antiangiogenic strategy is the use of VEGF receptor tyrosine kinase inhibitors (TKI). Sunitinib is an orally active, multitargeted receptor TKI that selectively targets VEGF receptors (VEGFR1, VEGFR2, VEGFR3), platelet-derived growth factor receptors, and other kinases (1-3). The US Food and Drug Administration has approved sunitinib for use in metastatic renal cell carcinoma (RCC), imatinib-refractory gastrointestinal stromal tumor (GIST) and pancreatic neuroendocrine tumors (PNET) (4).

Clinically, sunitinib may have an enhanced radiosensitizing effect in glioma (5). This finding is consistent with numerous preclinical studies that report growth delay and tumor cell killing by combinations of radiation therapy (RT) and VEGF pathway—targeting antiangiogenic agents (6-8). In preclinical models of various tumor types, sunitinib and ionizing radiation had a more than additive effect on growth delay compared with either as a single agent (9, 10).

The mechanism underlying the enhanced therapeutic ratio when angiogenic blockade and RT are combined is not fully understood but has been hypothesized to relate to vascular normalization (11, 12). Antiangiogenic agents may cause aberrant tumor vessels to regress (10, 13) and may decrease interstitial fluid pressure. Both these tumor environment changes may improve oxygen delivery to hypoxic tissues (14, 15). Furthermore, antiangiogenic agents have other intracellular effects, which may have an impact on the tumor response to RT. One such important effect of antiangiogenic agents may be to disrupt the phosphoinositol-3 kinase (PI3K)-Akt-mTOR signaling pathway. Disruption of this pathway has been proposed to radiosensitize the vascular endothelium (16). The ability of endothelial progenitor stem cells to mobilize and recruit can be impaired by agents that disrupt the PI3K-Akt-mTOR pathway (17-19).

Reirradiation with conventional involved field RT is usually not feasible in patients with recurrent disease because the ability to administer tumoricidal doses is limited by treatment-related toxicity. However, recent advances in RT techniques, including fractionated stereotactic RT (fSRT), allow for highly conformal treatment with an improved toxicity profile (20).

We evaluated the safety and toxicity profile of 37.5 mg oral sunitinib administered daily to patients undergoing fSRT for recurrent high-grade glioma (rHGG).

Methods and Materials

Patient selection

The Thomas Jefferson University Hospital Cancer Center Review Committee and Institutional Review Board granted approval for this study (IRB#06C.549). Eligibility requirements included patients with a confirmed diagnosis of primary glioma who were shown to have a contrastenhancing recurrence based on multidisciplinary review of magnetic resonance imaging (MRI). All patients were previously treated with RT and were required to have either a resectable recurrence or a recurrence deemed by the treating physician to be amenable to reirradiation. Other eligibility requirements included age >18 years, World Health Organization performance status of 0 to 2 (21), and life expectancy greater than 3 months. Prior surgery, chemotherapy, and RT were allowed provided there was resolution of all acute toxic effects of prior therapies. Adequate hematologic, hepatic, and renal functions (defined as hemoglobin >9 g/dL, platelets >100,000, absolute neutrophil count >1500/μL, aspartate aminotransferase and alanine aminotransferase <2.5 times the laboratory upper limit of normal [ULN], total serum bilirubin <1.5 times ULN, serum calcium <12 mg/dL, and serum creatinine <1.5 times ULN) were required.

Exclusion criteria included major surgery or RT within 2 weeks, systemic anticancer therapy within 14 days before enrollment, concurrent chemotherapy use, grade 3 hemorrhage within 4 weeks, major cardiovascular events within 6 months, uncontrolled hypertension, prolonged QTc interval, cardiac dysrhythmias, known human immune deficiency virus infection, concurrent warfarin treatment, pregnancy, and breastfeeding.

Study design

The study consisted of a baseline assessment period, a treatment period, and an observation period. A complete clinical evaluation, including medical history and physical examination, performance assessment, laboratory evaluation, electrocardiograms, and brain MRI, was completed during the assessment period.

During the treatment period, 37.5 mg sunitinib was given starting on RT day 1 and then daily including weekends. The continuous lower daily-dosed regimen (CDDR) of sunitinib was chosen on the basis of the preclinical data to maximize the vascular normalization, maximize the radiosensitizing effect, and minimize associated fatigue. CDDR sunitinib has been used in previous studies in metastatic RCC (22), PNET (4), and GIST (23) with a well-tolerated toxicity profile. The total dose and fractionation of the fSRT varied according to the size of recurrence, extent of resection, size of resection cavity, prior RT doses, and physician preference. A Gill-Thomas-Cosman relocatable frame was placed, and then patients were transported to the MRI and computed tomography (CT) suite, where the Brown-Roberts-Well fiducial cage was placed on the frame before imaging. All patients were treated on a Varian 600SR 6-MV stereotactic linear accelerator modified to a Novalis unit with a minimultileaf collimator. Target volumes were contoured by CT-MRI fusion with Novalis Brain Lab software and coapproved by a neurosurgeon and a radiation oncologist. Gross tumor volume (GTV) was defined by the T1-weighted gadolinium-enhanced tumor edge. The planning target volume (PTV) was equal to the GTV with no margin. T2/ FLAIR edema was not included. Novalis Brain Lab software was used for treatment planning to design a single-isocenter, conformal dynamic arcs plan with minimultileaf collimation. Targets were treated to the 85% to 90% isodose line. Thirty days of additional treatment with sunitinib after completion of RT was allowed for patients with minimal toxicity. Weekly assessments with hematology and chemistry panels were performed. Corticosteroid use was retrospectively assessed before, during, and 1 month after treatment. The primary endpoint was to evaluate the toxicity and safety profile of combining sunitinib and RT. Secondary endpoints were to assess radiographic tumor response at 1 month, progression-free survival (PFS), and overall survival (OS).

Toxicity grading and evaluation of response

During the active treatment and 4-week posttreatment observation period, patients were assessed for toxicities according to the Common Toxicity Criteria, version 3.0 (24), and evaluated for potential dose-limiting toxicity (DLT) weekly. DLT was defined as grade 3 to 5 toxicity probably attributable to sunitinib, excluding grade 3 hematologic toxicity unless it was clearly dose limiting. The study was designed to be terminated early if the actual DLT rate was 50% or greater.

The posttreatment observation and follow-up period consisted of a history, physical examination, hematology and chemistry studies, and repeated MRI 1, 3, 6, and 9 months after treatment. No formal neurocognitive functional assessments were performed. Complete response (CR) was defined as the disappearance of all evidence of disease. Partial response (PR) was defined as 30% or greater decrease in the sum of the longest diameter of target lesions. Progressive disease (PD) was defined as 20% or greater increase in the sum of the longest diameter of target lesions. Stable disease (SD) was defined as less than a 30% decrease or less than a 20% increase in the longest diameter of target lesions.

Statistical methods

The neurologic PFS rate and OS rate were calculated from the time of enrollment by use of the Kaplan-Meier method (25). Confidence interval (CI) was calculated by use of the ratio of the prescription isodose volume to the PTV as calculated from the integral dose-volume histograms for the PTV and the normal tissue volume. For patients with a multifocal recurrence, a CI was evaluated for each lesion.

Results

Patient characteristics and treatment

From March 2007 to January 2008, 11 patients were enrolled and received protocol treatment. Seven patients had rGBM, 1 patient had a grade 2 astrocytoma that dedifferentiated into GBM, 2 patients had recurrent grade 3 anaplastic astrocytoma (AA), and 1 patient had a grade 2 astrocytoma that dedifferentiated into AA. Table 1 shows the study population characteristics.

Table 1.

Snapshot of patient demographics and overview of course of treatment

Patient	Age	Sex	PS	Histology	Time from definitive therapy (mo)	Prior therapy	Other prior therapy	fSRT dose (Gy)	Fraction size (Gy)	PFS (mo)	OS (mo)	Therapy at recurrence after fSRT
1	41	F	1	GBM	5.5	60Gy + TMZ	None	37.5	2.5	12	18.3	Bev/CPT-11
2	50	F	1	AA	6.9	59.4Gy + TMZ	None	35	3.5	2.6	5.8	Supportive care
3	37	M	1	Grade 2 —> AA	96	54 Gy	None	42	3	38.6	38.6	NA
4	47	M	0	Grade 2 —> GBM	68.8	55.8 Gy	TMZ then carbo	37.5	2.5	2.5	6.3	Bev
5	51	F	1	AA	78.7	60Gy + TMZ	BCNU then TMZ	37.5	2.5	1.8	11	Supportive care
6	52	M	0	GBM	22	60Gy + TMZ	None	30	3	6.9	9	Bev/CPT-11
7	51	M	1	GBM	19.5	60Gy + TMZ	None	39	3	21.8	27.5	Bev/CPT-11 —> fSRT
8	63	F	0	GBM	4.6	60Gy + TMZ	None	35	3.5	5.8	16.3	Bev/CPT-11
9	62	F	0	GBM	11.9	60Gy + TMZ	None	35	3.5	38.6	38.6	NA
10	49	M	1	GBM	23.9	60Gy + TMZ	Bev/CPT-11 pre	35	3.5	3.3	3.3	Supportive care
11	67	M	1	GBM	16.0	60Gy + TMZ + Bev	None	35	3.5	3.3	3.3	TMZ

Abbreviations: AA = anaplastic astrocytoma; BCNU = bis-chloroethylnitrosourea; Bev = bevacizumab; carbo = carboplatin; CPT-11 = irinotecan; GBM = glioblastoma multiforme; NA = Not applicable; PS = ECOG Performance Status; TMZ = temozolomide.

All patients were previously treated with partial brain 3-dimensional conformal RT. The mean prior dose was 59.0 Gy (range, 50.4-60 Gy). The total fSRT doses delivered ranged from 30 to 42 Gy. The fSRT doses per fraction ranged from 2.5 to 3.75 Gy. The most commonly used regimen was 35 Gy in 10 daily fractions. The tumor volumes ranged from 0.05 to 72.01 cm³ (mean, 16.75 cm³). The CI ranged from 1.33 to 2.36 (mean, 1.92). All 11 patients completed the prescribed treatment without delay. Six (55%) patients chose to continue daily sunitinib for 30 days after completion of fSRT.

Toxicity evaluation

Five patients (45%) experienced only grade 1 toxicities, including thrombocytopenia (n = 4), leukopenia (n = 3), and hyponatremia (n = 3). Five patients (45%) experienced grade 2 toxicity, including hematologic disorders, fatigue, hypertension, esophagitis, and elevated liver enzymes. Toxicities by grade are listed in Table 2.

Table 2.

Acute toxicity profile of sunitinib and fSRT

Toxicity	Grade 1	Grade 2	Grade 3	Grade 4	Grade 5	Any
Leukopenia	3	2				5
Anemia	2					2
Thrombocytopenia	4	1				5
Fatigue	2	2				4
Candidiasis		1				1
Nausea	2					2
Vomiting	1					1
Diarrhea	1					1
Hypocalcemia	2					2
Hyponatremia	3					3
Acid reflux	1					1
Xerostomia	1					1
Hypoproteinemia	1					1
Hypochloremia	1					1
Elevated ALT	2	1				3
Elevated AST	1	1				2
Pain		1				1
Cough		1				1
Alopecia	1					1
Aphasia	1					1
Stomatitis	1	1		1		3
Anorexia	1					1
Hyperglycemia	1					1
Elevated ALP	1					1
Muscle weakness		1				1
Esophagitis		2				2
Hypertension		1				1
	33	15	0	1	0
Highest grade toxicity by patient
1						5
2						5
3						0
4						1
5						0

Abbreviations: ALT = alanine aminotransferase; ALP = alkaline phosphatase; AST = aspartate aminotransferase.

One patient (9%) experienced a grade 4 toxicity that met established DLT criteria outside the radiation portal: a 49-year old man with GBM was previously treated to 60 Gy in 2-Gy daily fractions with concurrent and adjuvant temozolomide. After 1 year he experienced evidence of recurrence and was treated with bevacizumab and irinotecan. He had SD for approximately 7 months before further progression in the left frontal lobe. He was enrolled in the study and was treated with 35 Gy in 3.5-Gy daily fractions of fSRT to the left frontal lobe enhancing lesion, given concurrently with sunitinib. After 10 days of sunitinib, he experienced a nonhealing oral ulcer on the left lateral aspect of the tongue, which rapidly progressed. Sunitinib was discontinued, but the ulcer became more extensive, necessitating hospital admission for pain control, electrolyte imbalances, and fluid resuscitation. The ulcer improved over the course of 2 weeks. Shortly thereafter he chose to enter a hospice program.

No difference in toxicity was observed based on total dose of RT delivered or fraction size. No correlation was found between tumor size or dosimetry parameters (volume receiving 35 Gy [V35], volume receiving 30 Gy [V30], or CI) and toxicity. No differences were observed in skin toxicity, hair loss, or toxicity of other tissues within the radiation portal. No grade 4 or 5 hypertensive events or intracerebral hemorrhages occurred.

Treatment response

Ten patients had follow-up MRI imaging to assess 1-month tumor response. One patient (10%) demonstrated PR, 6 patients (60%) had SD, and 3 patients (30%) had disease progression (Table 3). Before the initiation of treatment, 7 (64%) patients were receiving high-dose corticosteroids. At the 1-month follow-up visit, 3 patients (27%) required corticosteroids. No correlation was found between tumor size or dosimetry parameters (V35, V30, or CI) and response.

Table 3.

Early treatment response as determined by gadolinium-enhanced MRI 1 month after treatment and 6-month PFS and median OS

Response	All patients	rGBM	rAA	Median PFS (mo)	Median OS (mo)
Partial response	1 (10%)	1 (14%)	0 (0%)	21.8	27.5
Stable disease	6 (60%)	4 (57%)	2 (66%)	9.5	13.7
Progressive disease	3 (30%)	2 (29%)	1 (33%)	2.5	6.3
6-month PFS	5 (45%)	4 (50%)	1 (33%)
Median OS (mo)	11.0	12.7	11.0

Abbreviations: MRI = magnetic resonance imaging; OS = overall survival; PFS = progression-free survival; rAA = recurrent anaplastic astrocytoma; rGBM = recurrent glioblastoma multiforme

Survival

With a median follow-up time of 11 months overall and 38.6 months among those remaining alive, 2 patients (18%) continue to have SD, and 9 (82%) have had disease progression or death. Four patients had progression within the field of reirradiation. Three patients had progression near the border of the reirradation field. One patient had a distant progression, and 1 patient died without imaging assessment of progression. The PFS at 6 months was 45% (Fig. 1). The median PFS (mPFS) was 5.8 months. Remarkably, 2 patients (18%) remain alive and without progression of disease. The median overall survival (mOS) is 11.0 months. Seven (64%) patients were alive 6 months after treatment, and 5 patients (45%) remained alive after 1 year, with a survival range for all patients of 3.3 to 38.6 months (Fig. 1). Of the 2 patients who are alive and free of progression 3 years after treatment, 1 had rGBM and the other recurrent anaplastic astrocytoma (rAA).

Fig. 1.

Kaplan-Meier curves of progression-free survival (PFS) (dashed line) and overall survival (OS) (solid line).

Among the 8 patients with rGBM, the 6-month PFS was 50%, and the mPFS was 6.4 months (range, 2.5-38.6 months). The 1-year OS among patients with rGBM was also 50% (Table 4), and the mOS was 12.7 months (range, 3.3-38.6 months).

Table 4.

Treatment response, progression-free survival, and overall survival compared with published treatment alternatives

	Sunitinib + fSRT	fSRT alone		Bevacizumab	Bevacizumab + fSRT
Results	Wuthrick et al (present study)	Fogh et al (20)	Combs et al (33)	Friedman et al (30)	Gutin et al (34)
No. of patients	11	147	172	167	25
Patient characteristics
Recurrent grade 2	0%	0%	41%	0%	0%
Recurrent grade 3	27%	29%	24%	1%	20%
Recurrent GBM	73%	71%	34%	99%	80%
Posttherapy MRI response*
PR/CR	10%	10%	NA	33%	52%
SD	60%	60%	NA	NA	48%
PD	30%	30%	NA	NA	NA
Progression-free survival (rGBM only)
Median (mo)	6.4	NA	5.0	4.9	7.3
6-month	50%	NA	NA	46.50%	64%
Overall survival (rGBM only)
Median (mo)	12.7	11.0	8.0	9.0	14.2
1-year	50%	42%	23%	25%	63%
2-year	25%	11%	29%	0%	20%

Abbreviations: CR = complete response; fSRT = hypofractionated stereotactic radiation therapy; MRI = magnetic resonance imaging; PD = progressive disease; PR = partial response; rGBM = recurrent glioblastoma multiforme; SD = stable disease.

^*1 patient without evaluable MRI.

Discussion

This trial sought to assess the safety and tolerability of combining fSRT and CDDR sunitinib in treating rHGG. Overall, the combination of sunitinib and fSRT was well tolerated. The vast majority of patients did not experience severe toxicities, and all patients were able to complete the course. Toxicity was limited to grades 1 to 2 in 91% of patients. No patients required craniotomy for symptomatic cerebral radionecrosis. The limitations are the selection for enrollment only of patients whose recurrences were deemed treatable by reirradiation, the small number of patients enrolled, the mix of grade 3 and 4 gliomas, and heterogeneity in the doses and fractionation schedules used. CDDR sunitinib has been combined with standard-fraction RT (26-28) and stereotactic body RT (29), with an acceptable toxicity profile; however, we present the first experience with reirradiation in any context.

We previously reported on 147 patients with rHGG treated by a similar hypofractionation schedule (20). Forty-two patients had rAA and 105 patients had rGBM. The rates of minimal response, SD, and PD were 10%, 60%, and 30%, respectively, at the 3-month follow-up MRI. mOS after reirradiation was 10 and 11 months for patients with rAA and rGBM, respectively. On multivariate analysis, we found that younger age, small PTV, and short interval between initial RT and fSRT positively affected postreirradiation survival.

Because of the limited efficacy of systemic cytotoxic agents, targeted agents have entered as a treatment for rHGG. Bevacizumab has been studied alone and in combination with irinotecan (30) with a response rate (RR) in rGBM of 33%. The 6-month PFS, mOS, and 1-year OS were 47%, 9.0 months, and 25%, respectively. The authors reported a 46% rate of grade 3 + toxicity with bevacizumab alone. Interestingly, monotherapy with 37.5 mg monotherapy had insufficient activity to warrant further investigation (31). In 21 patients with rHGG, the median time to progression and OS were 1.6 and 3.8 months, respectively. Cilengitide, an integrin-targeting peptide with effects on migration, proliferation, survival, and angiogenesis, has been examined in rGBM (32). Cilengitide was well tolerated in the 81 patients (5% non-hematologic and 10% hematologic grade 3 + toxicity), with a RR, 6-month PFS, and mOS of 9%, 15%, and 9.9 months, respectively, at the highest dose level.

Another report of combining hypofractionated RT with an antiangiogenic agent used the combination of bevacizumab and fSRT for rHGG (34). After an infusion of 10 mg/kg bevacizumab, 25 patients (20 with rGBM and 5 with rAA) were treated with 30 Gy in 6-Gy fractions over a 2.5-week period. The authors defined PTV as the contrast-enhancing lesion plus a 5 mm margin. Three patients discontinued treatment because of intratumoral hemorrhage, wound dehiscence, and bowel perforation. No clinically relevant radiation necrosis was observed. Patients with rGBM had RR, 6-month PFS, mOS, and 1-year survival of 50%, 65%, 12.5 months, and 54%, respectively. This treatment approach was well tolerated, with signs of efficacy.

Comparison of our experience with the PFS and OS of other treatment schemes in rGBM is possible in a limited way (Table 4). CDDR sunitinib combined with fSRT did not yield the same RR as regimens with bevacizumab; however, the 6-month PFS, mOS, 1-year OS, and 2-year OS are similar. When antiangiogenic drugs alter vascular permeability, they may change the apparent size of tumors without affecting the underlying tumor mass, compromising our ability to visualize tumor on MRI. Therefore, RR as an indication of treatment efficacy should be cautiously interpreted when antiangiogenic agents are used. A likely better surrogate for efficacy is PFS. It is unclear from this small pilot study whether combining fSRT reirradiation with antiangiogenic agents yields improved outcomes in comparison with fSRT alone; however, the mOS, 1-year OS, and 2-year OS are numerically favorable compared with retrospective series of fSRT alone.

However, we observed that the 2-month MRI response predicted PFS and OS. The patient who achieved a PR had 21.8-month PFS and 27.5-month OS. Among the patients with SD, 9.5-month mPFS and 13.7-month OS were observed. PD at the 2-month MRI was a poor prognostic sign, inasmuch as that cohort’s mOS was 6.3 months.

Inasmuch as bevacizumab is increasingly being used earlier in the treatment of GBM, the field of neurooncology has sought effective therapies for rGBM after bevacizumab. Controversy exists regarding whether to discontinue bevacizumab and initiate other agents or continue with bevacizumab therapy after PD. In this study, both patients treated after progression receiving bevacizumab did poorly. Patients with postbevacizumab progression remain in need of novel therapeutic strategies. Additionally, the 1 patient with DLT toxicity had been pretreated and experienced progression through bevacizumab. The toxicity experienced (nonhealing oral ulcer) may have been related to the longterm angiogenic blockade.

The promising outcome we observed is in accordance with preclinical studies demonstrating synergism between radiation and sunitinib, possibly by improved tumor perfusion. In conclusion, CDDR sunitinib combined with fSRT yielded acceptable toxicities and adverse events. This unique combination also produced several long-term PFS. This hypothesis-generating trial adds to the body of literature suggesting the safety of adding antiangiogenic agents to RT and suggests that a subgroup of rHGG patients can attain prolonged PFS with this treatment strategy.

Summary.

Preclinical studies suggest that angiogenic blockade given concurrently with radiation therapy can improve therapeutic ratio. This pilot study establishes the safety and tolerability of concurrent continuous daily dosed sunitinib, 37.5 mg, and hypofractionated stereotactic radiation therapy as reirradiation for patients with recurrent high-grade glioma. The progression-free survival rate with this concurrent strategy compares favorably with published rates obtained by reirradiation or antiangiogenic agents alone.