Re-irradiation of recurrent glioblastoma multiforme using ¹¹C-methionine PET/CT/MRI image fusion for hypofractionated stereotactic radiotherapy by intensity modulated radiation therapy

Patients were recruited from September 2007 to August 2011. Adult patients (aged ≥ 18 years) with histopathologic confirmation of GBM who had local recurrent tumor were eligible. Primary treatment consisted of subtotal surgical resection in all patients. All patients had a Karnofsky performance status (KPS) ≥ 60 and were previously treated with external postoperative radiotherapy to a mean and median dose of 60 Gy (range 54 Gy–68 Gy) with concomitant and adjuvant Temozolomide (TMZ) chemotherapy. Additional inclusion criteria consisted of the following: age 18 years or older; first macroscopical relapse at the original site; and adequate bone marrow, hepatic, and renal function. In all cases, a multidisciplinary panel judged the resectability of the lesion before inclusion in this study. Thus patients with non-resectable lesions were included in the study.

This prospective nonrandomized single-institution study was approved by the Department of Radiation Oncology of Kizawa Memorial Hospital Institutional Review Board. Informed consent was obtained from each subject after disclosing the potential risks of HS-IMRT and discussion of potential alternative treatments. Baseline evaluation included gadolinium-enhanced brain MRI and MET-PET, complete physical and neurological examination, and blood and urine tests within 2 weeks before treatment. After completion of HS-IMRT, patients underwent a physical and neurological examination and a repeat brain MRI and MET-PET. TMZ chemotherapy was continued in a manner consistent with standard clinical use after HS-IMRT course.

CT (matrix size: 512 × 512, FOV 50 × 50 cm) was performed using a helical CT instrument (Light Speed; General Electric, Waukesha, WI). The patient head was immobilized in a commercially available stereotactic mask, and scans were performed with a 2.5-mm slice thickness without a gap.

MRI (matrix size: 256 × 256, FOV 25 × 25 cm) for radiation treatment planning was performed using a 1.5-T instrument (Light Speed; General Electric). A standard head coil without rigid immobilization was used. An axial, three-dimensional gradient echo T1-weighted sequence with gadolinium and 2.0-mm slice thickness were acquired from the foramen magnum to the vertex, perpendicular to the main magnetic field.

An ADVANCE NXi Imaging System (General Electric Yokokawa Medical System, Hino-shi, Tokyo), which provides 35 transaxial images at 4.25-mm intervals, was used for PET scanning. A crystal width of 4.0 mm (transaxial) was used. The in-plane spatial resolution (full width at half-maximum) was 4.8 mm, and scans were performed in a standard two-dimensional mode. Before emission scans were performed, a 3-min transmission scan was performed to correct photon attenuation using a ring source containing 68Ge. A dose of 7.0 MBq/kg of MET was injected intravenously. Emission scans were acquired for 30 min, beginning 5 min after MET injection. During MET-PET data acquisition, head motion was continuously monitored using laser beams projected onto ink markers drawn over the forehead skin and corrected as necessary. PET/CT and MRI volumes were fused using commercially available software (Syntegra, Philips Medical System, Fitchburg, WI) using a combination of automatic and manual methods. Fusion accuracy was evaluated by a consensus of 3 expert radiologists with 15 years of experience using anatomical fiducials such as the eyeball, lacrimal glands, and lateral ventricles.

Target volumes were delineated on the registered PET/CT-MRI images and were used to plan HS-IMRT delivery. In this preliminary study, GTV was defined as using an area of high uptake on MET-PET (Figure 1). This high-uptake region was defined using a threshold value for the lesion versus normal tissue counts of radioisotope per pixel index of at least 1.3. A previous study with high-grade gliomas, comparing the exact local MET uptake with histology of stereotaxically guided biopsies, demonstrated a sensitivity of 87% and specificity of 89% for the detection of tumor tissue at the same threshold of 1.3-fold MET uptake relative to normal brain tissue[16]. Further, this threshold was validated in a study using diffusion tensor imaging in comparison with MET uptake[17]. Although the delineation of GTV was done by using the automatic contouring mode (Philips Pinnacle v9.0 treatment planning system), the final determination of GTV was confirmed by consensus among 3 observers based on the co-registered MRI and PET data. The PET/MRI fusion image was positioned properly by CT scans equipped with Helical TomoTherapy (TomoTherapy Inc., Madison, WI). Finally, the GTV was expanded uniformly by 3 mm to generate the planning target volume (PTV). The prescribed dose for re-irradiation was based on tumor volume, prior radiation dose, time since external postoperative radiotherapy, and location of the lesion with proximity to eloquent brain or radiosensitive structures. Radiosensitive structures, including the brainstem, optic chiasm, lens, optic nerves, and cerebral cortex, were outlined, and dose-volume histograms for each structure were obtained to ensure that doses delivered to these structures were tolerable. Considering these different conditions, the dose for PTV was prescribed using the 80-95% isodose line, and the total doses were arranged from 25 Gy to 35 Gy in each patient. The dose maps and dose-volume histograms of a representative case are illustrated in Figure 2.

After HS-IMRT course, a dose of 200 mg/m²/day for 5 days with TMZ chemotherapy was administered. Cycles were repeated every 28 days with 3 or more cycles after HS-IMRT. Treatment was discontinued when unequivocal progression or severe toxicity occurred. TMZ was not administrated in patients who refused treatment or did not meet inclusion criteria. Patient inclusion criteria for TMZ chemotherapy consisted of the following: KPS score 50 or higher; evidence of good maintenance of major organ function (bone marrow, liver, kidney, etc.) in routine laboratory studies; expected to survive more than 3 months.

Regular serial neurological and radiological examinations were initially performed at 1 month after completion of treatment and then every 2 months thereafter or in the event of neurological decay. Follow-up examinations included MRI and MET-PET imaging. If the MRI revealed further enlargement of the enhanced mass, the lesion was diagnosed as “local progression”, and the day on which MRI first revealed lesion enlargement was defined as the date of progression. However, in cases with low MET uptake on PET in the MRI-enhanced lesion, the diagnosis was changed to “radiation necrosis”. The criterion of MET-PET differential diagnosis between local progression and radiation necrosis was utilized in a previous MET-PET study by Takenaka et al.[18]. However, all patients couldn’t be applied with PET examination in the follow up study. In cases without PET examination, diagnosis of radiation necrosis was based on pathologic examination or clinical course. Cases in which lesions showed spontaneous shrinkage or decreased in size during corticosteroid treatment were also defined as “radiation necrosis”. A diagnosis of “distant failure” was defined as the appearance of a new enhanced lesion distant from the original tumor site. Either local progression or distant failure was defined as disease progression. Acute and late toxicities were determined based on the Common Terminology Criteria for Advance Events (version 4).

Survival events were defined as death from any cause for OS and as disease progression for progression-free survival (PFS). OS and PFS were analyzed from the date of HS-IMRT to the date of the documented event using the Kaplan-Meier method. Tumor- and therapy-related variables were tested for a possible correlation with survival, using the Log-rank test. Variables included age (≥50 vs. <50), KPS (≥70 vs. <70), and combined TMZ chemotherapy (“yes” vs. “no”). P-values of less than 0.05 were considered significant. Prognostic factors were further evaluated in a multivariate stepwise Cox regression analysis.

No patients demonstrated significant acute toxicity, and all patients were able to complete the prescribed radiation dose without interruption. In the late phase, Grade 2 radiation necrosis was observed in 1 patient, although the lesion was decreased in size during corticosteroid treatment. One patient who received a dose of radiation of 25.0 Gy, experienced Grade 4 radiation necrosis in the form of mental deterioration 4 months after HS-IMRT. In this case, a second necrotomy was required 5 months after HS-IMRT. Although after a second surgery, the patient’s clinical condition was stable for a long period, disease progression occurred 40 months after HS-IMRT, causing death. The actuarial reoperation rate was 4.8% for radiation necrosis.

With a median follow-up of 12 months, the median OS was 11 months from the date of HS-IMRT, with a 6-month and 1-year survival rate of 71.4% and 38.1%, respectively (Figure 3A). The survival rates by age, KPS, and TMZ chemotherapy are shown in Figure 4A-C. OS of patients with KPS 70 or over was 12 months versus 5 months for patients with KPS less than 70. KPS was a significant prognostic factor of OS as tested by univariate analysis (p < 0.001). OS of patients who received combined TMZ chemotherapy was 12 months versus 6 months for patients who did not receive chemotherapy. The differences between the two groups approached significance (p = 0.079). In the multivariate model, only KPS remained statistically significant (p = 0.005) (Table 2).

The median PFS was 6 months from the date of HS-IMRT treatment, with a 6-month and 1-year survival rate of 38.1% and 14.3%, respectively (Figure 3B). KPS was the significant prognostic factor for PFS as tested by univariate analysis (p = 0.016). On multivariate analysis, none of the variables were significantly predictive of PFS (Table 3).A representative case is shown in Figure 5, in which a follow-up MET-PET scan was performed to improve the diagnostic accuracy.