Introduction
Seventy-eight thousand brain tumors are diagnosed in the United States per year with gliomas representing approximately one third [
1]. Virtually all gliomas eventually recur following treatment and in their natural history carry devastating neurological and psychological implications for those affected. Patients with a diagnosis of glioma may undergo multiple resections, radiation therapy and multiple lines of systemic treatment with diminishing treatment options as the disease progresses. Re-irradiation (re-RT) as a possible treatment option often enters the discussion and multiple retrospective studies have shown re-RT to be feasible and to improve outcome in selected patients [
2‐
13]. The benefits of re-RT include possible palliation by way of decreased steroid use, improvement in neurological symptoms, and in some patients an improvement in progression free survival and possibly overall survival. The timing of re-RT offering the best opportunity for benefit is unclear. Patient selection for re-RT varies according to the study, institution and available techniques and hence the type of patient who may benefit from its administration remains unclear. Prognostic scores using retrospective data have attempted to identify patient strata that may benefit from re-RT based on patient and histological factors [
14‐
17]. Significant heterogeneity in management has proven difficult with some studies reporting validation of existing scoring systems and others not. The radiation dose to be administered and its fractionation as well as the co-administration of chemotherapy suffer from lack of consensus and are as yet unclear. Re-RT is the subject of several ongoing trials, the results thereof, although highly relevant and anticipated, nonetheless are likely to leave us with unanswered questions. Many patients with recurrent glioma who may be considered for re-RT in the clinic would not necessarily have been eligible for the ongoing trials but according to existing data may nonetheless derive benefit from re-RT and may receive re-RT when no other options are available.
The intent of this publication was to obtain the opinion of radiation oncologists who 1) share an interest in the re-RT of patients with recurrent glioma defined as having published at least one paper on the subject and or having an ongoing practice or open protocol pertaining to re-RT or 2) have treated a significant number of cases. A simple but robust survey was employed as the springboard for expert discussion in an effort to gain a better understanding of patient selection and the philosophy underlying the re-RT dose and fractionation being offered.
Discussion
Re-irradiation is being offered for patients with recurrent glioma at both academic centers and in the community, on clinical protocols or outside of existing protocols. From retrospective data [
2‐
12] we understand that re-RT is feasible and that toxicity, if it occurs does not appear to be significant or clinically relevant. Despite numerous publications on re-RT in patients with recurrent glioma and several active protocols, we noted that there was no publication that specifically discussed the approach experts would adopt when faced with a potential re-RT case especially outside of a study protocol. Specifically, the question of our survey rested with the treatment volume, margins, dose and fractionation and more so than patient selection (seeing as this an evolving topic in and of itself with several available scoring systems [
14‐
18]. It is this information that we are frequently consulted on as re-RT experts and therefore, we felt that it was important to the recurrent glioma re-RT field to publish an approach other providers could follow and to provide a platform that others could contribute to (the live survey link below) in order to advance the field. To our knowledge this is in the only such survey and the only paper to discuss volumes, margins, dose and fractionation with respect to re-RT.
We obtained the opinion of radiation oncologists who share an interest in re-RT of patients with recurrent glioma using a simple survey involving 5 re-RT cases that could represent more common scenarios encountered in clinic as starting point for discussion and development of consensus. The responders share some common features including 1) all except one have treated more than 20 patients, although 3 of them have treated more than a 100, 2) they see at least 2 patients referred for re-RT per month and 3) treat at least 1 patient per month. With the data presented here on 13 responders, the numbers were insufficient to draw any conclusions with respect to difference in approach to patient selection or management based on the level of expertise of the responders although it is likely that such trends may emerge once the paper is published, readers take the survey and more data is available. We are aware of re-irradiation being practiced in the community as well as in academic centers and this too will be interesting to analyse. The volume seen in the community is likely far smaller than what the experts in the our paper have, which in a sense makes ours more homogenous of a cohort. Patient selection criteria, dose and fractionation to be employed are heterogeneous. Equally so is the perceived impact of previously administered dose to OAR with most responders rating it as extremely or very important while proceeding to re-RT with doses that are often unlikely to respect dose constraints. Despite the dose to the OAR having been reached dose limit in the first radiation course in all but one patient, greater than 50% of responders elected to offer re-RT in all but one case in which per responders’ comments, the reason for not offering re-RT was tumor size as opposed to dose to OAR. Perhaps this reflects a belief in normal tissue repair with increasing time from previous RT or a belief in lower dose per fraction decreasing the risk of late toxicity. These two considerations could be reflected in the minimal use of SRS (case 1 and 2).
Small volume “in field” recurrence in a young patient with a disease free interval of 2 years (case 1) prompted close agreement among responders all of whom perceived this case as an opportunity to elicit local control for the patient while considering both the risk of possible toxicity and the potential for longevity. Most responders chose a smaller margin and some employed SRS, although clearly hypofractionation was still favored by most responders irrespective of their geographical practice location. Three responders did choose to include T2 FLAIR and/or FDG-PET avid areas extending beyond the T1 gadolinium volume and some also pointed out that they would ultimately decide on their dose per fraction depending on dose previously administered to OAR. Offering clinical trials as well as considering concurrent chemotherapy with BEV and/or Lomustine was also mentioned.
The “out of field” BEV failure case (case 2) prompted the use of larger margins with concern for more extensive occult disease presence due to previous treatment with BEV. There was increased use of hypofractionation and conventional fractionation as opposed to SRS. Even when single fraction SRS was mentioned, the responders felt that hypofractionation should be considered and discussed with the patient. Several responders commented on this case representing an out of field recurrence, hence the willingness and ability to treat to full dose but also the concern that treatment of the T1 gadolinium enhancing tumor without addressing the T2 FLAIR or employing a larger margin, could result in marginal failure. In keeping with this view, 4 responders also chose to include the original resection cavity despite this being an out of field recurrence where the original resection cavity had remained stable. This too caused responders to favor hypofractionation or conventional fractionation over single fraction SRS.
A large volume recurrence with maximal previous dose to OAR (case 3), caused almost half the responders to not offer re-RT favoring either clinical trial or best supportive care. Responders also commented that re-RT could only be recommended if the OAR could be spared. Of the responders who did offer re-RT some rationalized that the long natural history and superior KPS did justify large volume re-RT but also stated that they would avoid hypofractionation.
In the older patient with a disease interval of less than a year from the original radiation (case 4) most responders would not offer re-RT, favoring systemic treatment or other options. The options included BEV plus CCNU, BEV alone or possible tumor treating fields (TTF). The responders who did offer re-RT in this case did so while expressing the concern of this representing possible pseudoprogression and the need to rule this out.
The case of recurrent anaplastic astrocytoma (case 5) was more split between resection or clinical trial and re-RT. Some of those who would re-RT, commented that they would consider a biopsy of the original cavity to help decide as to whether they would include it in their volume vs. treating the new enhancement alone. Some expressed that this was a limited volume out of field recurrence where optic nerve tolerance would play a role in re-RT planning and therefore they would treat the recurrent region only, given that the optic apparatus had previously received maximal dose. Two responders stated that they would treat with concurrent BEV and another that they would treat aggressively considering the long interval from previous RT. Some commented on foregoing treatment of the T2 FLAIR signal area as enveloping the chiasm and would especially consider doing so if systemic concurrent treatment were given. This case reflected the most heterogeneity in terms of volume to be treated and the use of systemic treatment as well as underlying rationale.
The question of clinical benefit as well as the concern over toxicity are not being addressed in this paper and remain a matter of debate in the absence of robust randomized trials that carry a best supportive care arm. Re-RT remains therefore remains controversial. We acknowledge that many patients are not referred for re-RT due to real or perceived lack of clinical benefit on the part of the provider. Therefore, the patients who do obtain re-RT reflect some level of selection bias. Cases attempting to elucidate the perception of clinical benefit on the part of the provider and hence the decision to offer re-RT were deliberately included in the questionnaire (especially so Case 3 and 4). We also acknowledge that this paper represents a reflection of the current practice in a number of centers who do practice re-RT. Prospective evidence is lacking.
Overriding patterns of practice/recommendations for management:
1
Offer re-RT to patients with smaller recurrences, especially if:
-
located in a favorable location.
-
ability to spare or minimize dose to OAR.
-
long interval since previous RT defined as greater than or equal to 6 months.
-
well defined area of recurrence ie. no previous use of BEV/BEV failure.
-
consider SRS or hypofractionated dose/fractionation in cases that meet size and location criteria above.
2
Consider clinical trial, systemic treatment or best supportive care in cases with:
3
Consider re-RT on a case by case basis in scenarios where:
-
OAR have received maximal dose previously and cannot be spared if further RT given.
-
Surgical resection possible.
-
If re-RT is proposed in cases where OAR toxicity is a concern, most responders would favor conventional fractionation over hypofractionation.
The lack of published toxicity following re-RT in the setting of doses that exceed published dose constraints reflect the limited life expectancy of patients treated as well as perhaps a significant tumor related neurologic deterioration which prevents adequate attribution of toxicity following re-RT. The authors recommend rigorous testing of visual fields, audiology, neurocognitive function prior to and following re-RT to better examine possible impact on toxicity as well as patient reported outcomes for quality of life metrics.
The cumulative dose administered to OAR as a result of re-RT and therefore the impact of those doses remains unclear. The limited toxicity observed should be examined further as per testing above but also should prompt the collection of OAR doses among the RT treatments for the purposes of designing a model of toxicity risk. Current guidelines [
19,
20] lack the data to make recommendations or offer risk estimation for doses above 60 Gy or with hypofractionated regimens.
The following were not addressed in this study and require further research: the most optimal timing of re-RT, the co-administration of systemic agents and the impact of re-RT on brain imaging specifically the interpretation of treatment failure versus tumor progression on MRI.
Seeing as the process underlying the decision to offer re-RT is more complex than the patient or disease specific factors deemed relevant in the questionnaire, we suspect that the decision to offer re-RT is likely based on a collection of case features as summarized in the “Overriding patterns of practice/recommendations for management” section. The need to understand this better in a larger sample is part of the rationale for the inclusion of a live link available for the duration of a year following publication. This survey will be available as live link (
https://www.surveymonkey.com/r/G7MWVZ8) for radiation oncologists willing to take it for one year from publication and results thereof will be published once available.