Prospective study to assess the survival outcomes of planned irradiation of ipsilateral subventricular and periventricular zones in glioblastoma

Purpose/objective(s) Retrospective evidence suggests that the irradiation of stem cells in the periventricular zone (PVZ), specifically the subventricular zone (SVZ), to higher doses may be associated with improved outcomes. Materials/methods This was a prospective study, done from 2012 to 2017 in glioblastoma patients to assess the efficacy of planned irradiation of ipsilateral PVZ and SVZ on survival outcomes. The clinical target volume included the tumour bed with a 1.5–2 cm margin, perilesional oedema and was expanded to encompass the ipsilateral PVZ (5 mm lateral expansion adjacent to the ventricles, including the SVZ, which was a 5 mm expansion lateral to lateral ventricle). The ipsilateral PVZ was planned to receive a dose of ≥50 Gy. Results 89 patients were recruited of which 74 patients were available for the analysis. Median age was 48 years. Mean doses to ipsilateral PVZ and SVZ were 56.2 and 55.1Gy, respectively. Median overall survival in the entire group was 13 months. There was no significant correlation between survival and doses to ipsilateral, contralateral, or bilateral PVZ and SVZ. Median survival was 16, 12 and 6 months for Eastern Cooperative Oncology Group (ECOG) PS 1, 2 and 3, respectively (p = 0.05). Conclusion Planned irradiation of potential stem cell niches in the ipsilateral cerebral hemisphere did not result in improved survival as suggested by retrospective studies. Doses to contralateral or bilateral PVZ or SVZ also did not influence survival.


Introduction
Glioblastoma is the most common malignant brain tumour [1]. Despite deeper insights into molecular biology and advances in therapeutics, the outcome remains poor for this aggressive tumour [2,3].The median overall survival (OS) is in the range of 14.6 to 21.1months [4,5].
Recent research in glioblastoma demonstrated that the heterogeneity in survival outcomes and recurrence patterns may be related to neuronal stem cells (NSC) [6,7]. NSCs reside in two areas in the adult mammalian brain: the subventricular zone (SVZ) and the subgranular zone (SGZ) [8] These have properties of self-renewal and multilineage potency along with tumour initiation and propagation [9,10]. Hence, eliminating these cancer stem cells may prove useful for cure in gliomas [10].
Several retrospective studies have been done to assess the impact of incidental radiation dose to SVZ during the course of standard chemoradiation (CRT) on survival in high grade gliomas but with conflicting results [11][12][13][14][15][16][17][18]. The aim of our study was to prospectively analyse the effect of SVZ and periventricular zone (PVZ) irradiation on survival in glioblastoma.

Materials and methods
Patients with histopathological diagnosis of primary glioblastoma, planned for adjuvant CRT at our institute from Jan 2012 to June 2017, were included in this prospective study. Following Institutional Review board approval patients of glioblastoma were prospectively recruited and informed consent was taken. All patients had undergone maximal safe resection. Patient demographics, clinical history and examination, radiological imaging and treatment details were documented. The details of extent of resection were collected from neurosurgeons' treatment records and post-operative imaging.

Radiotherapy planning
A thermoplastic immobilisation mask was done and CT simulation scans with 3 mm axial slices were acquired in treatment position. Pre and postoperative magnetic resonance imaging (MRI) images were co-registered with simulation scans. The gross tumour volume (GTV) was contoured as the contrast enhancing lesion on T1-weighted MRI or the post-operative cavity. The clinical target volume was a 1.5-2 cm margin around the GTV which was expanded to include the ipsilateral SVZ and PVZ, oedema and adjusted to anatomic barriers. The planning target volume was generated with a 5-mm isotropic margin expansion. The SVZ was contoured as a 5 mm lateral expansion of the lateral ventricles as per published guidelines [19]. The PVZ was contoured as 5 mm expansion all around the lateral ventricles. According to the spatial relationship with the tumour location SVZ and PVZ were classified as ipsilateral (i), contralateral (c) and bilateral (b). According to the spatial relationship with SVZ the tumours were classified as contacting SVZ and non-contacting SVZ tumours. Planned dose of adjuvant radiotherapy (RT) was 60 Gy in 30 fractions, 2 Gy per fraction, over 6 weeks with dynamic Multi-Leaf Collimator Intensity Modulated Radiotherapy (MLC IMRT) technique. A cone down boost volume was planned after 46 Gy if doses to critical structures exceeded tolerance levels. The ipsilateral SVZ and PVZ was planned to receive a mean dose of 50 Gy or more. All patients were planned for chemotherapy with TMZ, 75mg/m 2 PO daily throughout the course of radiation followed by 150-200 mg/m 2 PO for 5 days every 4 weeks for 6-12 cycles.

Dosimetry
The plans were generated with Eclipse TM version 8.6 (Varian Medical Systems) treatment planning system. Dose-volume parameters for each SVZ and PVZ volume were extracted from the treatment plans. Dose volume histograms were generated and mean dose to the SVZ and PVZ structures were documented.

Statistics
Data are presented as mean plus standard deviations (SD) and median for continuous variables. Categorical variables are presented as frequencies (%). The 25th, 50th and 75th percentiles were calculated for the dose to the iSVZ, cSVZ, bSVZ, iPVZ, cPVZ and bPVZ. Mean SVZ and PVZ doses were dichotomised at their median value at their median value into high dose and low dose arm and univariate analysis was done.
Survival times were calculated from the date of surgery and were estimated by the Kaplan-Meier methods with 95% confidence intervals (CI). Progression free survival (PFS) was defined as the time relapsed from date of surgery until the date of radiological or clinical evidence of progression or recurrence or patient death (if death occurred prior to radiological evidence of disease progression). Survival curves were compared using log rank test. Cox regression analysis was done for multivariate analysis.
Only variables with significance at or below the p = 0.05 level were considered as significant. All statistical analysis was done using SPSS v20 software (IBM, Chicago, IL, USA).

Results
89 patients with primary glioblastoma were recruited for the study. Patients who did not complete the prescribed treatment and those who were lost to follow up (n = 15) were excluded from the analysis. Thus, a total of 74 patients were included in the final analysis. Patient characteristics are tabulated in Table 1. The most common presenting complaint was headache (62%) followed by hemiparesis (24%). Other presentation symptoms were seizures, vomiting, fall, blurring vision, altered behaviour and memory disturbance. Mean duration of symptoms before surgery was 3.2 months. Tumours were grouped based on their relationship with SVZ as suggested by Jafri et al [6], and 62% of the cases were categorised as Group I. 70% of the tumours were contacting the SVZ. Other patient and treatment related characteristics are documented in Tables 1 and 2, respectively. The mean and median doses to iSVZ were 56.2 and 58 Gy, respectively. Dosimetric data are tabulated in Table 3.

Survival
Median OS was 13 months and median PFS was 11 months (Figure 1a and b). Univariate analysis of prognostic factors showed significantly higher in survival in patients with better performance status (p = 0.05) and patients who received adjuvant chemotherapy (p = 0.00). Patients were divided into subgroups based on whether they received lower or higher than the median dose to ipsilateral subventricular zone (iSVZ) and iPVZ. Univariate analysis of prognostic factors in the higher and lower dose groups are given in Table 4. On comparing the lower and higher dose groups based on the median dose received by the ipsilateral, contralateral and bilateral SVZ and PVZ dose there was no significant difference in survival outcomes (Table 5).
On multivariate analysis, Eastern Cooperative Oncology Group (ECOG) performance status and concurrent and adjuvant temozolomide showed significant survival benefit (Table 6).

Discussion
Glioblastoma is an aggressive and highly invasive primary CNS malignancy. Recent studies demonstrated that Glioblastomas contacting the SVZ are associated with earlier recurrence and poor survival [6,20,21].
Given the potential clinical significance of the stem cell niches in the SVZ, the concept of irradiation of SVZ seems logical to eliminate the CSCs. Many studies attempted to correlate incidental dose delivered to the SVZ with survival outcomes. These studies divided the patient population into a low SVZ dose and a high SVZ dose group and compared survival [22].

Evidence supporting improved survival with higher SVZ and PVZ doses
The first study testing this hypothesis [11] was a retrospective study of 55 patients with high-grade glial tumours treated uniformly with surgery, conformal radiotherapy and chemotherapy. Patients were divided into groups based on the median bilateral periventricular dose of 43 Gy into a high dose (n = 28) and a low-dose group (n = 27 Iuchi et al [23] reported on a prospective phase II trial of 46 patients with Glioblastoma treated with hypofractionated radiotherapy and concomitant chemotherapy with temozolomide. Median OS was 20.0 months and the results showed that radiation necrosis in the SVZ strongly correlated with improved survival but caused deterioration in the performance status of long-term survivors. Median OS was 36.2 months in patients with SVZ necrosis versus 13.3 months without necrosis (HR 4.08; 95 % CI 1.97-9.10, p = 0.0001). In multivariate analysis only SVZ necrosis was significantly associated with prolonged survival.
Another retrospective study of 43 patients with glioblastoma [24] was conducted to identify SVZ related prognostic factors for survival and recurrence patterns. Results on multivariate analysis showed contact to SVZ, as well as reduced bSVZ radiation dose coverage (V20Gy <84%), might be independent poor prognostic factors for time to progression (7 months versus 5.2 months, p = 0.017).
Interim analysis of our study data (54 patients) investigating planned neural stem cell niche irradiation in glioblastoma [25] showed mean dose of 58 Gy or greater to the iSVZ correlated positively with improved OS (16 months versus 14 months, p = 0.03).
A recent retrospective study [26], including 47 patients with glioblastoma, showed a non-significant trend towards improved survival at a median follow up of 19 months with mean dose higher than 56 Gy to the ipsilateral SVZ.

Evidence showing no correlation of survival with SVZ or PVZ doses
Researchers from the Netherlands [12], however, could not replicate these results. They analysed 87 patients with newly diagnosed glioblastoma treated with standard trimodality therapy. Using a cutoff of 43 Gy (threshold from Evers et al), there was no correlation between ipsilateral, contralateral or bilateral SVZ dose with PFS or OS. Threshold values of 30, 40 or 50 Gy also could not demonstrate any correlation between SVZ dose and survival. However, in the subgroup with complete resection (n = 32), using a cutoff 30 Gy mean dose for bSVZ, there was a statistically significant correlation between dose and OS (p = 0.015). In our study, we did not find any significant PFS benefit with dose higher than 43 Gy or 30 Gy to bSVZ.
A larger study at Johns Hopkins University, by Chen et al, [15]

SVZ and PVZ doses and neurocognitive effects
In the scenario of longer survival neurocognitive toxicity of radiation is one of the important concerns which can result due to damage to neural progenitor cells (NPCs). Redmond et al [32] in their study in mice and human models demonstrated that NPC sparing clinical plans may reduce neurocognitive effects. A prospective study in 30 patients showed inverse relationship between RT dose to the NPC niches and neurocognitive effects in patients with no effects on recurrence or survival [33]. A recent retrospective analysis comparing the impact of sparing stem cell with IMRT and non IMRT plans showed no significant difference in the OS, concluding that there was no impact on sparing stem cells on survival.
Many of the studies which demonstrated a positive correlation of doses to the SVZ and PVZ with survival were retrospective and analysed the incidental coverage of SVZ in treatment fields. Heterogeneity in the patient characteristics and dose cut offs at which statistically significant relationships are observed limit the validity of the conclusions drawn. There is conflicting data regarding the benefit of higher dose to SVZ on survival. Prospective data including ours failed to demonstrate any improvement in survival outcomes. Recent data has shown worse outcomes with higher SVZ doses. The detrimental effect on neuro-cognition with higher doses to NSCs is also of concern. Thus, there is no strong evidence to suggest higher dose to SVZ or PVZ improves survival it will be prudent to spare these areas and reduce neurocognitive decline.
Limitations he study did not evaluate or stratify for other established prognostic factors in Glioblastoma, including MGMT methylation, IDH1 mutation, etc. Neurocognitive function was not assessed in this study.

Conclusion
Increasing dose to the iSVZ and PVZ did not translate into improved survival outcomes in patients of Glioblastoma. Future studies should focus on sparing of these areas to preserve neurocognitive function.

Funding declaration
No funding was received for the study.