CyberKnife^® enhanced conventionally fractionated chemoradiation for high grade glioma in close proximity to critical structures

Patients with newly diagnosed resected unifocal high-grade gliomas (WHO Grade III and VI) in close proximity (<1 cm) to critical structures (Table 2) were evaluated. All patients were in RPA class 1 to 4 [19, 20]. Magnetic resonance imaging (MRI) was completed preoperatively and postoperatively. The Georgetown University Hospital institutional review board approved this study and all participants provided informed written consent.

The extent of surgical resection was documented as total tumor resection or subtotal tumor resection following review of operative reports and post operative MRI imaging (Table 2). Salvage surgery was routinely recommended for patients with good performance status and evidence of recurrence or radiation necrosis based on imaging studies.

Patients were placed in the supine treatment position with their heads resting on a standard support. A custom thermoplastic mask was crafted. Thin-sliced (1.25 mm) high-resolution CT images were obtained through the cranium for conventional and CyberKnife treatment planning. Treatment planning MRI imaging was completed selectively to enhance target and critical structure delineation when clinically indicated. Target volumes and critical structures were contoured by team neurosurgeons. Treatment volumes were generous including the contrast enhancing tumor volume when present and the surgical defect with a 3 cm margin. Critical structures in close proximity to the target volume were not excluded from the treatment volume during conventional radiation treatment. Forty to 50.4 Gy was delivered in 1.8 to 2.0 Gy fractions 5 days a week for a total of 4 to 5 1/2 weeks. Treatment was delivered using linear accelerators with nominal energies ≥ 6 MV. Intensity modulated radiation therapy (IMRT) technique was not permitted.

Following the completion of conventional radiation therapy, CyberKnife treatment was completed without a planned treatment break (Figure 1). The technical aspects of CyberKnife^® radiosurgical system for cranial tumors have been described in detail [17, 18]. The treatment volume for the radiosurgical boost included the contrast-enhancing lesion and the resection cavity as defined by the patient's neurosurgeon plus a 1 cm margin when clinically indicated (Figure 1A, B). Due to the submillimeter precision of CyberKnife treatment, no additional margin was added to correct for set-up inaccuracy. The treating neurosurgeon and radiation oncologist in consultation determined the prescription isodose line (Figure 1C). Twelve circular collimator ranging in diameter form 5 to 60 mm are available with the CyberKnife^® radiosurgical system. An inverse planning method with non-isocenteric technique was used. The treating physician and physicist input the specific treatment criteria, limiting the maximum dose to critical structures (Figure 1C). The planning software calculated the optimal solution for treatment. The DVH of each plan was evaluated until an acceptable plan was generated. Strict adherence to critical normal structure dose constraints was maintained (Table 1).

Image-guided radiosurgery was employed to eliminate the need for stereotactic frame fixation. Using computed tomography planning, target volume locations were related to cranial landmarks. With the assumption that the target position is fixed within the cranium, cranial tracking allows for anatomy based tracking relatively independent of patient's daily setup. Position verification was validated every third beam during treatment using paired, orthogonal, x-ray images [23, 24].

Patients received concurrent and/or adjuvant chemotherapy at the discretion of their medical oncologist. Typically, patients were administered Temozolomide with concurrent radiation at a dose of 75 mg/m2/d, given 7 d/wk from the first day of conventional irradiation until the last day of CyberKnife treatment. After a 4-week break, patients generally received 6 cycles or more of adjuvant Temozolomide on a 5-day schedule of 150 to 200 mg per square meter every 28 days.

Clinical evaluation and MRI imaging were performed at 3-6 month intervals following CyberKnife treatment for 5 years. Evaluation frequency beyond 5 years was determined by the medical oncologist. Throughout the follow-up period, a multidisciplinary team of neurosurgeons, radiation oncologists, medical oncologist and radiologists reviewed outcomes at a weekly central nervous system tumor board. Toxicity was scored according to the National Cancer Institute Common Terminology Criteria for Adverse Events, Version 3.0 [25]

The follow-up duration was defined as the time from the date of surgery to the last date of follow-up for surviving patients or to the date of death. Actuarial survival and local control was calculated using the Kaplan-Meier method.

Twenty four consecutive eligible patients were treated over a seven year period extending from January 2002 to January 2009 (Table 2) and were followed for a minimum of 12 months or until death. The mean age of the group was 52 years (range, 27-72). Tumors were evenly distributed between anaplastic glioma (WHO III) and glioblastoma multiformi (WHO IV). Ninety-two percent of the tumors involved the temporal and/or frontal lobes.

Thirteen tumors were completely resected; eleven were subtotaly resected. All patients received conventional radiation therapy following tumor resection, with a median dose of 50 Gy (range: 40 - 50.4 Gy). Upon completion of conventional treatment, an additional dose of 10 Gy was delivered in five successive 2 Gy daily fractions utilizing the CyberKnife^® image-guided radiosurgical system. Treatment plans were composed of hundreds of pencil beams shaped using a single circular collimator to generate highly conformal plans (mean new conformity index of 1.62, Table 3). Selected plans were inhomogeneous by design (mean homogeneity index of 1.38, Table 3) to minimize dose to adjacent critical structures. Radiation was delivered to a mean prescription isodose line of 73% (Table 3) in 5 approximately 1 hour long treatments. On average, 173 beams were employed to treat the mean prescription volume of 174 cc with a mean percent target coverage of 94%. An average of 58 verification images were taken during each treatment to account for intrafraction patient motion. Twenty-one patients received concurrent and/or adjuvant Temozolmide. Two patients received adjuvant procarbazine, lomustine, vincristine (PCV) alone and one patient declined chemotherapy.

The median follow-up was 23 months (range, 13-60 months) for glioblastoma multiforme patients and 63 months (range, 21-85 months) for anaplastic glioma patients (Table 4). No patients were lost to follow-up. Nine of twelve GBM patients (75%) experienced local progression, seven of which died during the follow-up period. Six of the twelve anaplastic patients (50%) experienced local progression, four deaths occurred during the clinical follow-up period. The median time to local progression was 16 months for the glioblastoma multiformi group and 33 months for the anaplastic glioma group. The median survival was 18 months for the glioblastoma multiforme group with a two-year survival rate of 37%. The median survival was not reached for the anaplastic glioma group and the 4-year survival rate was 71% (Figure 2). Of those who died in the glioblastoma multiforme group, 7 (89%) had local disease progression and of those who died in the anaplastic glioma group 4 (100%) had local disease progression (Figure 2). The median time to death was 18 months for the glioblastoma multiformi group and 36 months for the anaplastic glioma group. There were no severe (≥ grade 3) radiation complications per the National Cancer Institute Common Terminology Criteria for Adverse Events, Version 3.0 with this conservative treatment strategy.

Ultimately, 16 patients experienced local progression during follow-up (Table 5). Salvage surgery was clinically indicated and pursued in 10 patients, 4 with glioblastoma multiforme and 6 with anaplastic glioma. Each surgery confirmed recurrent glioma with treatment effect. Necrosis was not observed in the absence of tumor progression. Five patients completed salvage chemotherapy, 3 from the glioblastoma multiformi group and 2 from the anaplastic glioma group. A single glioblastoma multiforme patient survived 10 weeks following salvage CyberKnife radiosurgery.