Does the interval from tumour surgery to radiotherapy influence survival in paediatric high grade glioma?

Prognosis of high grade glioma (HGG) in childhood is similarly poor as in adults. Standard of care in children with HGG older than three years of age includes maximal safe resection followed by focal radiotherapy (RTX) and variable adjuvant chemotherapy (CTX). Because HGGs are fast growing, aggressive malignancies neuro-oncology practice usually aims at minimising delay to initiation of RTX after surgery. More recently though, several distinct epigenetic and biological subgroups of HGG gliomas in children were described requiring determination of molecular markers for treatment stratification in future trials [1], thus most likely leading to longer delays in initiation of treatment. Yet no data exist in the paediatric HGG population whether the interval from surgery to initiation of radiotherapy (ISRT) influences survival. From a biological point of view and considering growth dynamics of malignant solid tumours such as high grade glioma an increased risk of relapse and a subsequent negative impact on survival may be postulated [1]. In adults delayed initiation of RTX was indeed proven to negatively impact outcome in head and neck, breast and other cancers [2]. In recent years an increasing number of studies investigated ISRT in adult patients with HGG, but results remain conflicting. Whereas some studies could identify the expected negative impact on survival [3‐7], others failed to show any correlation with survival [8‐12]. Some large studies even suggest a positive effect of moderately delayed RTX on survival in adult patients with HGG [13‐18].

The present institutional study investigates whether the interval between surgery and initiation of RTX influences survival in paediatric patients with HGG.

Patients with HGG treated at the Department of Paediatrics, Medical University of Vienna, Austria, from 1995 to 2015 were included in this retrospective single-centre analysis. The following eligibility criteria had to be met: histologically confirmed high-grade glioma (WHO grades 3 or 4), age 3–19 years at diagnosis, application of postoperative adjuvant radiation therapy and sufficient clinical data on primary treatment and follow-up. Patients with diffuse intrinsic brainstem glioma (DIPG) were excluded, as well as patients with a secondary tumour following prior cranial RTX. Patients younger than three years were excluded as RTX is generally withheld in these children [19]. Follow-up was conducted according to institutional standards (Table 1) and was continued at the Department of Paediatrics even after patients had reached adulthood.

This retrospective analysis was approved by the ethical review board of the Medical University of Vienna.

The medical histories of 59 patients with HGG were evaluated; 38 of them met the above mentioned criteria (Fig. 1). All patients received radiochemotherapy (RCTX) following surgery. Patient characteristics are listed in Table 2. According to a median ISRT of 18 days, the cohort was divided in two groups: group 1 started simultaneous RCTX within 17 days of surgery, group 2 after ≥18 days. An exploratory analysis also evaluated the impact of an ISRT of ±28 days (see below).

Like its adult counterpart paediatric high grade glioma (pHGG) has an unfavourable prognosis. In order to improve survival it is important to recognise potential influencing factors. Molecular subgrouping of pHGG [24] may influence treatment decisions in the future (e. g. presence of BRAF V600E mutations [25] or of a hypermutated phenotype [26] for which specific therapies are available). Since molecular subgrouping is often feasible in real time before starting adjuvant therapy [18], it is fundamental to identify a time limit that is safe to perform all necessary molecular testing. Prompted by a recent study involving a targeted drug that required a delay in start of radiotherapy for 4 weeks [27] we retrospectively analysed our consecutive series of patients above 3 years of age with HGG. As data on the best time point of radiotherapy in children with HGG are lacking, the present analysis evaluated a possible influence of the interval from surgery to radiotherapy (ISRT) in this patient population. Unlike those postulated from other tumour entities [2] and defying radiobiological theorems [1], numerous adult studies could not detect a presumed negative influence of ISRT on patient survival in HGG [8‐18]. A large study conducted by Blumenthal et al. [13] even suggested a favourable influence of prolonged waiting times until start of RTX. Similar to this observation ISRT could not be identified as a significant risk factor in our single-centre population of pHGG. Multiple factors may possibly influence such a finding like an uneven distribution of age, tumour grade or EOR. Whereas indeed our cohort of patients with the shortest ISRT comprised more patients with grade IV tumours, there was no significant difference in EOR or age. Studies involving adults found similar and even further bias: patients with unfavourable baseline characteristics and highly aggressive tumours seem to be over-represented in cohorts with shorter ISRT [6, 13‐16]. In the large study by Blumenthal et al. [13] the cohort with shortest delays had worse scores of performance status and fewer gross total resections. This is underlined by findings of Spratt et al. [6] and Wang et al. [16]. Lai et al. [14] demonstrated that patients with biopsy only were more likely to be irradiated sooner after surgery. In an analysis by Wehming et al. [15] patients with an ISRT <24 days suffered more often from a grade IV tumour, had a minor EOR and a higher median age. Irwin et al. described a direct correlation of a better performance status to a prolonged ISRT [3]. Clinicians thus tend to apply RTX early in patients with a large tumour load and poor prognosis. This bias may consequently mask the effects of late treatment initiation.

In the present analysis only nine of 38 children were treated with an ISRT of more than 28 days. The overall length of ISRT in our cohort was therefore shorter than in the published data in adult HGG [13, 15, 18]. But, while overall ISRT seems longer in the adult population, only few patients with HGG had very long delays above 6 weeks, which is in contrast to studies in breast or head and neck cancer that identified a detrimental effect of delayed RTX [2]. In cohorts including patients with very long delays survival was impaired and it was deduced that delaying RTX beyond 6 weeks after surgery was a negative predictive factor [4‐7]. Han et al. [17] even postulated an optimal time window of 30–34 days post-surgery, after which survival worsened. In contrast, studies postulating a positive effect of postponing RTX had a minority of patients with prolonged waiting times over 6 weeks, e. g. 0.7% [13] and 2.4% [14], respectively. Considering that in our study 29/38 patients started irradiation within 28 days after surgery this 4‑week interval may be too narrow to cause significant differences in outcome. Yet, when comparing patients receiving RTX within four weeks or thereafter we could not even identify a trend towards a significant difference.

In addition to ISRT we also analysed other factors possibly influencing outcome in pHGG. Previous findings identified a higher extent of resection (EOR) [28‐31] and lower tumour grade [32, 33] as major positive predictive factors for improved survival in pHGG. In accordance with the literature it was possible to reproduce the favourable influence of a greater EOR in univariate as well as multivariable analysis. Presence of a higher tumour grade could also be established to be associated with poorer PFS and OS. Young age as a predictor for favourable outcome [30, 32, 34] could not be evaluated in our cohort since patients under the age of 3 years (i. e. not receiving radiotherapy [19]) were excluded from the study.

Radiotherapy following surgery is considered the gold standard of HGG therapy in patients older than three years of age. Adjuvant chemotherapy was also used in all our patients but regimens varied over the past 2 decades. Interestingly and in contrast to previous findings [35], temozolomide (TMZ) seemed to be superior when compared to all other therapies in an exploratory analysis (the MGMT status being unknown in our retrospective cohort).

Our data therefore suggest the absence of a significant survival cost of moderate delays of simultaneous RCTX and waiting times of up to four weeks after surgery to be safely tolerable. The four week time frame may be used for molecular analysis needed for including patients with pHGG in future prospective clinical trials.