Experimental treatment group (group 1):

Since prospective phase III research on the optimal fractionation reirradiation regimen is sparse, a radiobiological theoretical optimal regimen is chosen from retrospective studies and phase I/II studies [8,23,24,29,30]. The linear-quadratic model describes cell killing, both for tumor control and for normal tissue complications following exposure to a varying amount of radiation. The most modern method of comparing fractionation schedules is to use the linear quadratic equation to calculated the equivalent dose in 2 Gy fractions (EQD2).

For GTV with a 5 mm margin (CTV_22Gy), the EQD2 may be up to 40 Gy based on the maximal cumulative dose of 100 Gy advised by Sminia and Mayer [15,31]. For this reason, the dose was selected to be 4 fractions of 5.5 Gy (EQD2_α/β=2 = 41 Gy).

For the tumor (CE on CE-T1 MRI, GTV), the maximal dose should not exceed 8 fractions of 5 Gy (EQD2_α/β=2 = 70 Gy). In a prospective phase I trial, hypofractionated stereotactic radiotherapy was given with 5 Gy per fraction to doses ranging from 20 to 50 Gy on a dose escalation program, in patients with recurrent glioma with a median target volume of 24cc. Exceeding 8 fractions of 5 Gy resulted in 6.4 times more radiation damage [32]. Therefore, four fractions of 7.5 Gy were considered safe with a comparable EQD2 _α/β=2 of 71 Gy.

Experimental treatment schedule therefore consists of 4 fractions of 7.5 Gy SIB to the CE on CE-T1 MRI and 4 fractions of 5.5 Gy to the GTV with a 5 mm margin.

Standard treatment group (historic controls, group 2):

In 1999, a phase I dose escalation of hypofractionated stereotactic radiotherapy in 20 patients with recurrent or persistent malignant gliomas was conducted. Three different total dose levels were sequentially evaluated: 24.0 Gy in 8 fractions of 3.0 Gy, 30.0 Gy in 10 fractions of 3.0 Gy and 35.0 Gy in 10 fractions of 3.5 Gy. No grade 3 toxicities were observed and the response rate was 78% [33]. From that time, 10 fractions of 3.5 Gy is a generally accepted regimen. Until now, this schedule has shown to be safe and effective in practice [34]. EQD2_α/β=2 of this schedule is 48 Gy.

Equipment and planning:

Radiotherapy will be planned and delivered with modern techniques including at least intensity modulated radiation therapy (IMRT) and volumetric modulated arc therapy (VMAT).

Immobilization:

All participants will be immobilized in a customized thermoplastic shell or relocatable stereotactic frame in a supine position according to institutional standards.

Data acquisition and definitions of target volumes and organs at risk:

The definition of volumes will be in accordance with the International Commission on Radiation Units and Measurements (ICRU) Report 50, ICRU Report 62. Volumes will be defined based on a volume MRI scan with contrast (T1 + Gadolinium, CE-T1) and co-registered to a computerized tomography (CT) scan. The MRI scans must have a maximum 2 mm and isotropic voxels, the CT should have 1–3 mm slice thickness. Amino-Acid-PET or SPECT-Examinations may be used in addition to contrast-enhancement on MRI for target volume definition but are not mandatory.

Gross Tumour Volume (GTV):

The GTV is defined as the contrast enhancing lesion on CE-T1 MRI sequence for both treatment groups. GTV can be expanded using the other clinically acquired MRI sequences according to the expertise of the local radiation oncologist and the medical team.

Clinical Target Volume (CTV):

The CTV of both, the experimental treatment group and the standard treatment group, is defined as the GTV with addition of a 3-dimensional margin of 5 mm limited by anatomic borders. The stereotactic CTV_boost of the experimental treatment group is equivalent to the GTV. CTV can be expanded using the other clinically acquired MRI sequences according to the expertise of the local radiation oncologist and the medical team.

Planning Target Volume (PTV):

The PTV and PTV_boost are defined as the CTV and stereotactic CTV_boost plus an isotropic 2 or 3 mm margin depending on radiotherapy center, to account for day to day setup variation related to the ability to immobilize the participant. The PTV should not extend outside the body contour.

Organs at Risk (OARs):

OARs include the following normal tissue volumes: the lenses, optic nerves and chiasm and the eyeball (includes anterior and posterior chambers); brainstem; pituitary; cochlea; skin (a 3 mm thick rim) and the uninvolved brain (= brain – CTV). OARs will be delineated following the local guidelines of the participating centers and will typically follow atlases such as CT-based OAR atlas [35] or the European Partical Therapy Network (EPTN)-consensus based atlas [36]. The complete atlas is provided as supplemental material of the published article, available at https://ars.els-cdn.com/content/image/1-s2.0-S0167814015004016-mmc1.pdf.

Recovery potential:

With regard to the recovery potential in the brain after prior irradiation, the following schedule will be utilized [37]:

• No recovery within 1 year.
• 30% recovery after 1 year.
• 50% recovery after 2 years.

Treatment planning:

IMRT or VMAT may be planned using fixed fields or arcs. The treatment plan to be used for each participant is based on an analysis of volumetric doses including the dose volume histogram (DVH) of the PTV and the OARs. Treatment planning should conform to ICRU 50, 62 and 83 rules for coverage of GTV, CTV and PTV. OAR planning risk volumes (PRV) should be defined by applying a margin equal to the PTV margin.

Prescribed dose and fractionation:

The dose prescribed to the PTV in the experimental group is 4 fractions of 5.5 Gy. The prescribed dose to the PTV_boost is 4 fractions of 7.5 Gy. All fractions should be given within 2 weeks. The maximum dose in the PTV_boost of the experimental group should not exceed 107% of the prescription dose.

The maximum dose in the PTV (i.e., PTV minus PTV_boost) should not exceed 130% of the prescription dose according to ICRU rapport 50. The dose calculation grid should be 3 mm or less.

Dose limitation to organs at risk:

It is important to minimize the dose to the OARs whenever possible. This must be weighed against the possibility of sub-optimal treatment of the target volume. The dose constraints for the Organs at Risk are given in Table 2.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Dose constraints for OARs in Gy. Radiobiological calculations performed with α/β = 2Gy. To calculate the dose constraint at reirradiation, the initial dose should be subtracted from the cumulative dose constraint for OARs. Subsequently, the reirradiation EQD2 should be converted to the dose constraint for the OARs according to the treatment group. The reirradiation dose of OARs should never exceed the EQD2 and total dose constraint given for initial treatment (shown in column ‘Constraint at reirradiation’).

https://doi.org/10.1371/journal.pone.0342337.t002

Timing of radiotherapy:

The interval between the planning-MRI and actual first fraction of reirradiation is 3 weeks at most.

Criteria for discontinuing or modifying allocated interventions:

Subjects can leave the study treatment group at any time, for any reason, if they wish to do so, without any consequences. The investigator can decide to withdraw a subject from the study for urgent medical reasons. If a patient is going off protocol treatment, the reason should be documented. Patients are not obligated to specify the reason of withdrawal.

Strategies to improve adherence to interventions:

Participants are actively encouraged to attend all scheduled radiation therapy sessions to ensure adherence to the study protocol.

Attendance will be monitored, and participants will be contacted to reschedule any missed appointments in a timely manner. Participants are encouraged to complete and return any questionnaires sent to them within the specified timeframe. If a questionnaire has not been completed within two weeks, a reminder will be sent to encourage timely submission.

Relevant concomitant care permitted or prohibited during the trial:

Concomitant dexamethasone and anticonvulsant therapy is allowed, in line with routine clinical care and local/national guidelines. Salvage surgery prior to start radiation is allowed. Concurrent systemic therapy including chemotherapy, immunotherapy or targeted therapy during reirradiation is not allowed, including the use of bevacizumab.

Adverse events (AEs):

Adverse events are defined as any undesirable experience occurring to a subject during the study, whether or not considered related to reirradiation treatment. All adverse events reported spontaneously by the subject or observed by the investigator or his staff will be recorded.

Serious adverse events (SAEs):

A serious adverse event is any untoward medical occurrence or effect that:

• results in death.
• is life threatening (at the time of the event).
• requires hospitalization or prolongation of existing inpatients’ hospitalization.
• results in persistent or significant disability or incapacity.
• is a congenital anomaly or birth defect.
• any other important medical event that did not result in any of the outcomes listed above due to medical or surgical intervention but could have been based upon appropriate judgment by the investigator. An elective hospital admission will not be considered as a serious adverse event.

The investigator will report all SAEs to the sponsor without undue delay after obtaining knowledge of the events. However, patients with recurrent glioma have a poor life expectancy. Therefore, a significant proportion of patients will die within the study duration of 1 year after radiotherapy. The SAE sensitive period will be up until 30 days after last reirradiation fraction to account for treatment related effects. Expected SAEs are hospitalization, persistent or significant disability or incapacity, or death due to glioma progression. These expected SAEs will be reported every year in a line listing.

The sponsor will report the SAEs through the web portal ToetsingOnline to the accredited METC that approved the protocol, within 7 days of first knowledge for SAEs that result in death or are life threatening followed by a period of maximum of 8 days to complete the initial preliminary report. All other SAEs will be reported within a period of maximum 15 days after the sponsor has first knowledge of the serious adverse events.

Follow-up of adverse events:

All AEs will be followed until they have abated, or until a stable situation has been reached. Depending on the event, follow up may require additional tests or medical procedures as indicated, and/or referral to the general physician or a medical specialist.

Plans for assessment and collection of outcomes:

The clinical data from the experimental arm will be collected from medical charts, pathology reports, radiology reports and web-based surveys (PROFILES web-based database). The data will be entered and managed using Castor EDC (Castor, Amsterdam, The Netherlands), a secure electronic data capture system.

Health related quality of life (HRQoL)

Patients will complete HRQoL questionnaires and the study team will record toxicity and anti-edema treatment by phone call at baseline and at 2, 4, 6, 13, 26, 39, and 52 weeks from the start of reirradiation or until disease progression (whichever is sooner). During the first 6 weeks HRQoL will be assessed with use of the EQ-5D-5L. At subsequent time points QLQ-C15-PAL and BN20 + 2 will be completed digitally or on paper.

EQ-5D-5L is a validated measure of health status developed by the EuroQol Group in order to provide a simple, generic measure of health for clinical and economic appraisal [40,41]. This tool is used for the HRQoL measurement during the first 6 weeks after therapy start, since it takes only a few minutes to complete.

Subsequently, HRQoL will be determined using the EORTC QLQ-C15 PAL questionnaire [42] and QLQ-BN20 + 2 questionnaire [43] every 3 months after finishing radiotherapy treatment. The QLQ-C15 PAL is a validated general core QoL assessment tool which is developed for (near) palliative cancer care patients. The QLQ-BN20 + 2 includes the QLQ-BN20 questionnaire that has been validated for use in patients with primary brain malignancies. In addition, two scales for concentration and memory were added which were omitted in the abbreviated QLQ-C15-PAL.

The QLQBN20 + 2 was first described by Caissie et al. and validated by Nguyen et al. in patients with brain metastases [43,44]. These questionnaires consist in total of 37 assessments instead of 50 assessments when using the QLQ-C30. This aims to minimize patients’ burden and increase the chance of maintaining follow-up as patients progress through their disease course [43].

Neurocognitive assessment (NCA, RISinG trial A only)

NCA will be measured by a core set of cognitive tests proposed by RANO working groups and International Cognition and Cancer Task Force [44,45]. The tests include measures of memory, processing speed, executive function and verbal fluency. The test battery takes about 30–45 minutes. This core assessment will take place at baseline and after 6 months, and will be administered by trained employees.

Plans to promote participant retention and complete follow-up:

Follow-up data will be collected from routine clinical data wherever possible, and do not require specific efforts from patients. If a study-specific questionnaire has not been completed within two weeks, a reminder will be sent to encourage timely submission.

Statistical methods for primary and secondary outcomes:

Demographic and baseline characteristics

Data on demographic and baseline characteristics will be summarized by means of descriptive statistics. For continuous variables, in case of normal distribution by means and standard deviations, and in case of non-normal distribution by medians and interquartile ranges. For discrete variables data will be summarized by frequencies and proportions.

OS and PFS

All statistical tests for the efficacy endpoint OS and PFS will be performed according to intention-to-treat (ITT) principle, which is the most unbiased method. Considering the trial being a non-inferiority trial, which is more sensitive to bias, a secondary per-protocol (PP) analysis will be done for OS [46].

The PP analyses will be performed on a “subset of patients who complied sufficiently with the protocol, such as exposure to treatment, availability of measures and absence of major protocol violations” according to the ICH E9 guidelines. The ITT analysis includes all participants that meet the inclusion criteria. Start date of reirradiation will be used as t = 0. Patients not known to have died during the study will be censored for OS at the day they were last known to be alive. Patients free from progression and alive at the last visit will be censored for PFS at the day of last assessment. Endpoints are measured from the first date of reirradiation.

Distributions of OS and PFS will be estimated by the Kaplan-Meier method. The OS rates at one year, PFS rates at six months and medians of OS and PFS will be presented with two-sided 95% confidence interval(CI) computed using the log-log transformation.

To minimize the effect of confounding when estimating the difference between the treatment and retrospective control group, stabilized inverse probability of treatment weighting (SIPTW) using propensity scores will be applied when comparing the groups on OS and PFS.

Propensity scores will be estimated using logistic regression including important covariates for OS and PFS: primary histology, GTV, KPS and age.

OS

The hazard rate for the experimental versus the retrospective control group and its two-sided 80 and 95% CIs will be estimated by means of a Cox proportional hazards model. We will adjust for confounding using propensity scores. We will conclude that 4 fractions is non-inferior to 10 fractions of RT if the higher 90% CI of the hazard ratio (HR) of 4 fractions versus 10 fractions was not above 1.4.

Only when non-inferiority has been demonstrated, we will test for superiority of the experimental treatment in a hierarchical fashion, using Kaplan-Meier method and two-sided log-rank test at a significance level α = 10% in ITT analysis.

PFS

The primary comparison of the time-to-event distributions between the two treatment groups will consist of the two-sided log-rank test at significance level α = 5% The hazard rates for the experimental versus the retrospective control group and its two-sided 95% CI will be estimated with a Cox proportional hazards model. We will adjust for confounding using propensity scores.

Recurrence patterns

Recurrence patterns will be compared between the treatment groups with logistic regression, also adjusted for confounding using propensity scores. Recurrences are defined as “central” if more than 95% of the tumor recurrence resides within the prescription 95% isodose surface (D95) of the PTV; “in-field” if more than 80% of recurrent lesion is inside D95 of the PTV, and “marginal” if 20% to 80% of the lesion is inside D95 of the PTV. In all other cases, recurrences are defined as outside the radiation field (“ex-field”) according to the study of Lee et al [47].

HRQoL in treatment group 1

HRQoL will only be assessed in the experimental group. Reasons for missing baseline and follow-up questionnaires will be assessed. Linear Mixed Effects models will be used for regression analyses of HRQoL. In addition, the repeated measures of the EQ-5D-5L, QLQ-C15-PAL and BN20 + 2 functional and symptom scales will be assessed by time to deterioration (TTD) analyses. The time until definitive HRQoL score deterioration (TUDD) is defined as time from inclusion to a first occurrence of a deterioration of the score that exceeds the minimal clinically important difference (MCID) of at least 10 points compared to the baseline score. This deterioration is considered definitive if there is no subsequent improvement exceeding the MCID relative to the baseline, or if the patient drops out after the deterioration, resulting in missing data. According to the construction of the HRQoL scores, deterioration corresponds to an increase (e.g., for symptomatic scales of the EORTC questionnaires and EQ-5D-5L) or a decrease (e.g., for functional scales) of the score.

Patients with no score available are excluded from the time-to-deterioration analyses. Patients with no baseline score are usually censored at baseline and those with no follow up scores but with a baseline score are censored one day after baseline. Patients with no deterioration before drop-out from the study, including death, are censored at the time of last HRQoL questionnaire completion or last follow-up.

The TUDD estimation will be calculated using the Kaplan-Meier method and described using median and 95% CI.

Neurocognitive Function (RISinG trial A only)

Analyses on both frequency and extent of cognitive decline in both arms will be performed using logistic and linear regression.

Toxicity

Toxicity will only be assessed in the experimental group. Analyses of radiotherapy toxicity will be performed in all patients who received at least one fraction. Acute events occurring from start of radiotherapy until 90 days and late events occurring from 91 days after end of radiotherapy will be presented. Further analyses of treatment toxicity will present the grade 3–5 acute or late side effect by treatment group. The time to occurrence of any severe late side effects, and radionecrosis separately (distinguishing symptomatic and asymptomatic), will be estimated by cumulative incidence. The time to severe late side effects will be calculated from the time of start of radiation treatment to the first evidence of any grade 3–5 late side effects. Patients alive without grade 3–5 late toxicity will be censored at the date of last follow-up, patients who died without experiencing late grade 3–5 side effects will be assessed as competing risk at the time of death.

Prognostic and predictive factors

Additional exploratory analyses using Cox regression analyses will study the prognostic impact of factors (other than treatment, age, gender, WHO-grade primary tumor, time interval in between first radiation and reirradiation and MGMT-status), including baseline KPS, number of recurrence, tumor volume and dexamethasone use at baseline. In addition, the independent predictive value for the treatment effect of the 4 and 10 fraction regimens of these factors will be investigated using Cox Regression analyses.

Handling of missing data:

An inclusion of 66 subjects in the experimental arm and 99 in the control group are aimed to generate a total of 165 patients If a subject in arm 1 discontinues from the study prematurely the reason, if given, must be fully evaluated and recorded appropriately.

After informed consent, patients who did not receive any reirradiation (e.g., withdrawal of consent or due to findings on planning-MRI) will be substituted by additional patients. Patients who received at least 1 fraction, withdrawing because of any reason will not be substituted by additional patients.

By only including patients who received at least one fraction of reirradiation in the experimental arm, we aim to optimize comparability with the retrospective control-arm (wherein all patients received reirradiation).

Patients withdrawn from the trial retain their identification codes, if already given. New patients must always be allotted a new identification code. Patients withdrawn from treatment will be asked if they will continue completing the questionnaires and if their medical record may be checked for reported toxicity and progression. Patients will be asked for oral and written informed consent.

Patients withdrawn from treatment will be asked if they will continue completing the questionnaires and if their medical record may be checked for reported toxicity and progression. Patients will be asked for oral and written informed consent.

Due to the possibility of checking the date of death in the National population registry in the Netherlands and due to the fact that reirradiation schedule is one of the inclusion criteria, we do not expect missing data for our primary analysis.

For HRQoL we use a linear mixed model. This method is robust, even in the case of missing data [48].

Composition of the coordinating center and trial steering committee:

The main team of the coordinating center is composed of the project controller, one principal investigator and one coordinating investigator. They meet biweekly and discuss day to day issues, concerning trial management, communication with the centers and protocol implementation. The steering committee consists of the main team of the coordinating center, the other members of the writing committee and the trial agency of the division. The principal investigators of the study sites coordinate the execution of the trial in their own center. Together with the steering committee, they are the investigator group. For independent patient support, an independent advisory physician is available.

The monitoring of the study is performed by a qualified and trained monitor to assure the quality and validity of the research data. Due to the well-known toxicity profile of reirradiation for high-grade gliomas and the low risk of unexpected adverse events, a data safety monitoring board (DSMB) will not be established.

Study monitoring:

The study will be monitored, by using a minimal, risk-based, primarily on-site monitoring approach, performed by a qualified and trained monitor to assure the quality and validity of the research data. Monitoring activities include review of patient flow, essential documents, informed consent, eligibility and targeted source data verification (minimum 5%), as well as verification of Serious Adverse Event Reporting. An initiation visit will be performed before enrolment, the first routine visit after the first three participants, followed by annual visits per site, with frequency adjusted based on recruitment. A close-out visit will be conducted at study end. Further details are provided in the Monitoring plan.