From April 2004 to December 2016, 29 patients with recurrent high-grade glioma underwent 38 cases of GKR. GKR was done with type C with/without an automatic positioning system. The male-to-female ratio was 10 : 19, and the median age was 53.8 years (range, 20–75). GKR was performed for 11 cases of pathologically diagnosed anaplastic oligodendrogliomas, five anaplastic astrocytomas, and 22 glioblastomas. The tumors were graded following the World Health Organization classification [15]. The tumor recurrence was diagnosed with contrast-enhanced brain magnetic resonance (MR) images including diffusion, MR spectroscopy and perfusion images [11]. C-methionine (MET) positron emission tomography (PET) images were checked to differentiate the recurred lesion from radiation necrosis, and delineate the target volume for GKR. Before GKR, 29 patients underwent treatments as follows : radiotherapy of 59.4 Gy plus concomitant and adjuvant temozolomide (n=12), radiotherapy of 40–59.4 Gy (n=17), temozolomide chemotherapy (n=6), procarbazine, lomustine, and vincristine (PCV) chemotherapy (n=4), nimustine and cisplatin (n=2), carmustine and cisplatin (n=1), and bevacizumab and irinotecan (n=1). The Eastern Cooperative Oncology Group (ECOG) performance status (PS) scales (grades 0–5) were used to assess disease progression and how the disease and treatments affected activities of daily living to determine patient prognosis and appropriate treatment [14]. An ECOG PS grade of 0 indicates that the patient is fully active and able to perform all predisease activities without restriction. A patient with an ECOG grade 1 PS is restricted in physically strenuous activity but is ambulatory and able to perform light or sedentary work, for example, light housework or office work). Grade 2 indicates that a patient is ambulatory and capable of all self-care but is unable to perform any work activities for more than 50% of waking hours. Grade 3 indicates that a patient is capable of only limited self-care and is confined to a bed or a chair for more than 50% of waking hours. When GKR was performed, 26 cases had an ECOG PS grade of 1, 10 had PS grade 2, and two had PS grade 3. The clinical characteristics are summarized in Table 1.

GKR was performed using Leksell Gamma Plan^Ⓡ (Elekta Instrument AB, Stockholm, Sweden) radiosurgery planning software. The median prescription dose was 16 Gy (range, 10–24), and the median target volume was 7.0 mL (range, 1.1–15.7). After GKR, the median follow-up duration was 11.5 months (range, 1.1–75.1). GKR was targeted on relatively wellcircumscribed recurred lesions on brain MR images and/or MET-PET images. Most target lesions were MR-enhanced, and MET-PET images was used to delineate the target lesion with/without contrast enhancement. A total of 13 of the 29 (44.8%) patients received concurrent chemotherapy consisting of temozolomide (n=9); PCV (n=3); and ifosfamide, carboplatin, and etoposide (ICE) (n=1).

After GKR with/without chemotherapy, the patterns of recurrence were categorized into local, distant, and cerebrospinal fluid (CSF) dissemination. Local recurrence was defined as recurrence at the primary tumor site, while distant recurrence was defined as a spatially separated recurrence within the brain. CSF dissemination was defined as recurrence at the ventricles, subarachnoid space, and spine as observed from imaging studies. The recurrence occurred in 29 cases out of 38. Local recurrence was noted in 14 patients. Of them, 10 underwent further treatment as follows : operation (n=3), temozolomide chemotherapy (n=4), PCV chemotherapy (n=2), and bevacizumab and irinotecan (n=1). Distant recurrence developed in 12 patients, including two combined with local recurrence. Of the 12 patients, 10 underwent further treatment : GKR (n=6), temozolomide (n=1), PCV (n=1), and nimustine and cisplatin (n=2). Meanwhile, CSF dissemination was noted in three patients. Of them, two were subjected to additional treatment: radiotherapy and ICE chemotherapy (n=1) and nexavar chemotherapy (n=1).

The median range of survival was defined as survival up to the last follow-up. Variables affecting progression-free survival (PFS) and overall survival (OS) after GKR were determined via univariate and multivariate analyses. The single variables were ECOG PS, pathology, concurrent chemotherapy, radiation dose, and target tumor volume. PFS was calculated from the date of surgery to the date of radiological progression or last follow-up visit, and OS was calculated from the date of surgery to the date of death or last follow-up. Survival probability was calculated using the Kaplan-Meier method, and comparisons were performed with the log-rank test. We examined variables using the Cox proportional hazard analysis model to identify the independent predictors of survival. All statistical analyses were performed with a significance level of p<0.05 using SPSS version 21.0 (SPSS, Chicago, IL, USA).

Starting from when the patients underwent GKR, the median PFS and OS were 5.0 months (range, 1.1–28.1) and 13.0 months (range, 1.1–75.1), respectively. The prognostic factors related to PFS and OS on univariate analysis are summarized in Table 2. The median PFS was significantly longer in patients with anaplastic oligodendroglioma (6.9 months [range, 4.4–18.4]) compared to those with glioblastoma (4.5 months [range, 1.1–28.1]) (p=0.001) (Fig. 1A). Moreover, survival of patients with an ECOG PS 1 and tumor volume less than 10 mL was statistically significantly different compared to those with an ECOG PS 3 and tumor volume over 10 mL (p=0.024 and 0.008, respectively). The median PFS of those with an ECOG PS 1 and PS 3 was 6.5 months (range, 2.6–28.1) and 1.1 months (range, 1.1–4.5), respectively (Fig. 1B). Meanwhile, the median PFS of patients with a tumor volume less than 10 mL was 5.8 months (range, 1.6–28.1), while for those with over 10 mL was 3.4 months (range, 1.1–7.7) (Fig. 1C). Differences in PFS based on radiation dose and combined chemotherapy were not statistically significant. Meanwhile, differences in OS in terms of ECOG PS and tumor volume were statistically significant (p=0.000 and 0.002, respectively). The median OS of patients with an ECOG PS 1, 2, and 3 was 16.8 months (range, 4.2–75.2), 7.7 months (range, 3.7–19.4), and 1.1 months (range, 1.1–4.5), respectively (Fig. 2A). The median OS of patients with tumor volume less than 10 mL was 14.7 months (range, 3.7–75.1), while for those with over 10 mL was 8.4 months (range, 1.1–13.0) (Fig. 2B).

Factors related to PFS and OS on multivariate analysis are listed in Table 3. Patients with tumor volume less than 10 mL had longer PFS (hazard ratio [HR] 0.273, 95% confidence interval [CI] 0.106–0.703, p=0.007) and OS (HR 0.239, 95% CI 0.075–0.761, p=0.015) than those with over 10 mL. Moreover, patients with ECOG PS 1 showed longer PFS (HR 0.190, 95% CI 4.710–8.290, p=0.043) than those with an ECOG PS 3. Further, patients with ECOG PS 1 and 2 (HR 0.018, 95% CI 0.002–0.161, p=0.000) had longer OS than those with ECOG PS 3 (HR 0.062, 95% CI 0.007–0.541, p=0.012).

We present a case of a 21-year-old woman noted to have newly developed lesions on the right parietal lobe on routine follow-up radiological imaging. Seven years ago, she underwent craniotomy and tumor resection. Pathologically, glioblastoma was diagnosed. She received concomitant radiotherapy of 59.4 Gy and 6 cycles of adjuvant chemotherapy using temozolomide. Brain magnetic resonance images (MRI) showed a homogenous enhanced lesion on the parietal lobe, and methionine-PET image showed a hot uptake lesion (Fig. 3A and B). GKR was performed for this lesion. The tumor volume was 2.05 mL, and the marginal radiation dose was 18 Gy with 50% isodose (Fig. 3C). A combined 12 cycles of temozolomide chemotherapy was then administered after GKR. Follow-up brain MRI 24 months after treatment showed decreased enhancing lesion without any neurologic symptoms (Fig. 3D).

We present a case of a 22-year-old man who underwent craniotomy and tumor resection. Pathologically, anaplastic oligodendroglioma was diagnosed. He received radiotherapy at a dose of 50.4 Gy. Within 4 years from diagnosis, re-operation, PCV chemotherapy, and temozolomide chemotherapy were performed for repeated recurrent lesion. Brain MRI showed a non-enhancing lesion adjacent to the right frontal horn, and methionine-PET image showed a hot uptake lesion (Fig. 4A and B). Consequently, GKR was performed. The tumor volume was 2.0 mL, and the marginal radiation dose was 18 Gy with 50% isodose (Fig. 4C). Follow-up brain MRI after 11.3 months showed a stable enhancing lesion without any neurologic symptoms (Fig. 4D).