Outcomes After Reirradiation for Recurrent Pediatric Intracranial Ependymoma

Accepted Manuscript Outcomes after re-irradiation for recurrent pediatric intracranial ependymoma Derek S. Tsang, MD, Elizabeth Burghen, RN, MSN, Paul Klimo, Jr., MD, MPH, Frederick A. Boop, MD, David W. Ellison, MD, PhD, Thomas E. Merchant, DO, PhD PII:

S0360-3016(17)33963-9

DOI:

10.1016/j.ijrobp.2017.10.002

Reference:

ROB 24529

To appear in:

International Journal of Radiation Oncology • Biology • Physics

Received Date: 9 May 2017 Revised Date:

31 August 2017

Accepted Date: 2 October 2017

Please cite this article as: Tsang DS, Burghen E, Klimo Jr P, Boop FA, Ellison DW, Merchant TE, Outcomes after re-irradiation for recurrent pediatric intracranial ependymoma, International Journal of Radiation Oncology • Biology • Physics (2017), doi: 10.1016/j.ijrobp.2017.10.002. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

Title Page Title: Outcomes after re-irradiation for recurrent pediatric intracranial ependymoma Authors: Derek S. Tsang, MD,1 Elizabeth Burghen, RN, MSN,1 Paul Klimo, Jr., MD, MPH,2,3 Frederick A. Boop, MD,2,3 David W. Ellison, MD, PhD,4 and Thomas E. Merchant, DO, PhD1

RI PT

Affiliations: 1 Department of Radiation Oncology, St. Jude Children’s Research Hospital, Memphis, Tennessee 2 Department of Surgery, St. Jude Children’s Research Hospital, Memphis, Tennessee 3 Department of Neurosurgery, University of Tennessee Health Science Center, Memphis, Tennessee 4 Department of Pathology, St. Jude Children’s Research Hospital, Memphis, Tennessee

M AN U

Corresponding author: Dr. Thomas E. Merchant St. Jude Children’s Research Hospital 262 Danny Thomas Place, MS 210 Memphis, TN 38105 Phone: (901) 595-3604 Fax: (901) 595-3113 E-mail: [email protected]

SC

Running title: Re-irradiation for ependymoma

TE D

Word count (body + summary + abstract + figure legends): 3989 Word count (abstract): 299 Figures: 3 Tables: 3

Keywords (MeSH): ependymoma, necrosis, pediatrics, recurrence, re-irradiation Conflict of interest: None.

AC C

EP

Acknowledgments: The authors would like to thank Keith A. Laycock, PhD, ELS, for editing the manuscript. This work was supported by the American Cancer Society (grant no. SPAMM-15-210-01COUN) and by American Lebanese Syrian Associated Charities (ALSAC). Neither organization was involved in the study design; in the collection, analysis, or interpretation of data; in the writing of the report; or in the decision to submit the article for publication.

ACCEPTED MANUSCRIPT

Summary Repeat radiation therapy (RT) for recurrent pediatric ependymoma results in long-term disease control in some patients. Five-year survival was 57% for all patients. Patients with distant-only failure treated with

RI PT

craniospinal irradiation had the best outcomes (5-year survival of 76%). The 10-year incidence of grade 3 or higher radiation necrosis was low at 7.9%. Repeat RT is a feasible, safe, and effective treatment for

AC C

EP

TE D

M AN U

SC

children with recurrent ependymoma.

ACCEPTED MANUSCRIPT

Blinded title page Title: Outcomes after re-irradiation for recurrent pediatric intracranial ependymoma

AC C

EP

TE D

M AN U

SC

RI PT

Running title: Re-irradiation for ependymoma

ACCEPTED MANUSCRIPT

Abstract Purpose: Re-irradiation for recurrent pediatric ependymoma is feasible and safe, but the long-term

RI PT

outcomes and the optimal dose and volume for re-treatment are unknown.

Methods and Materials: Patients with recurrent ependymoma treated with a second course of fractionated radiotherapy (RT2) were reviewed retrospectively. Eligible patients had localized, intracranial

SC

ependymoma at initial diagnosis that was treated with focal radiation (RT1) without craniospinal

irradiation (CSI-RT1), and were aged ≤21 years at the time of RT2. The median doses of RT1, focal RT2,

measured from the first day of RT2.

M AN U

and CSI-RT2 were 59.4, 54, and 39.6 Gy, respectively. The primary endpoint, overall survival (OS), was

Results: We included 101 patients in the study. The median interval between RT1 and RT2 was 26.8 months (interquartile range [IQR] 18.0–43.1). The median durations of OS and freedom from progression

TE D

(FFP) were 75.1 and 27.3 months, respectively. Male sex and anaplastic histology at recurrence were associated with decreased OS and FFP on multivariate analysis. Distant-only failure treated with CSIRT2 was independently associated with improved OS compared to individuals with local failure treated

EP

with focal RT2 (HR 0.37, 95% CI 0.16–0.87). Among individuals experiencing any distant failure after RT1, gain of chromosome 1q was adversely associated with poorer OS (HR 3.5, 95% CI 1.1–10.6). No

AC C

distant-only failures were observed in individuals with RT1 local failure who received CSI-RT2 (n = 10). The 10-year cumulative incidence of grade ≥3 radiation necrosis after RT2 was 7.9%.

Conclusions: Re-irradiation for relapsed pediatric ependymoma was well tolerated by most patients and resulted in long-term survival in a subset of patients. The best results were observed in patients who experienced distant-only failure after RT1 and were treated with CSI as part of RT2, without anaplasia at recurrence. The option of re-irradiation should be discussed with patients who develop recurrent ependymoma.

ACCEPTED MANUSCRIPT

Introduction Ependymoma is a central nervous system (CNS) tumor that exhibits variable ependymal differentiation and is the third most common CNS tumor in children [1]. For localized ependymomas, the standard

RI PT

treatment is surgery followed by focal irradiation (RT1) of the primary tumor site [2]. However, despite aggressive treatment, one-third of patients experience recurrence [3,4]. Retrospective studies have shown re-treatment of ependymoma with a second course of fractionated radiation therapy (RT2) to be effective

SC

[5-9], and this approach is the subject of an ongoing prospective phase II study (ClinicalTrials.gov

identifier: NCT02125786). Prior studies of re-irradiation for ependymoma have included only a limited

M AN U

number of patients without long-term follow-up. Furthermore, the role of molecular markers associated with prognosis in de novo ependymoma, such as gain of chromosome 1q and C11orf95–RELA fusion [1013], is unknown in patients selected for re-treatment.

The optimal treatment volume for re-irradiation is unclear. Focal re-irradiation and craniospinal

TE D

irradiation (CSI-RT2) may be used at the time of recurrence for local and distant failures, respectively. The purpose of CSI is to comprehensively treat disseminated disease in the neuraxis in the setting of distant (metastatic) recurrence. There is some indication that CSI-RT2 improves survival, even in patients

EP

with local-only recurrences (5, 7, 8), but CSI is associated with important late toxicities, especially

AC C

neurocognitive impairment.

This study aimed to evaluate overall survival (OS) in patients with recurrent pediatric ependymoma treated with a second course of irradiation. Secondary endpoints included freedom from progression (FFP), patterns of failure with focal RT versus CSI, clinicopathologic factors associated with OS and FFP, and cumulative incidence of radiation necrosis after RT2.

ACCEPTED MANUSCRIPT

Methods and Materials This was a retrospective cohort study of all patients with recurrent ependymoma treated with RT2 at a single institution between 1985 and 2017. Eligible patients had localized, intracranial ependymoma at

RI PT

initial diagnosis that was treated with focal radiation (RT1) without CSI and were aged 21 years or younger at the time of RT2. Patients treated with RT2 to the brain for secondary neoplasms without

SC

evidence of ependymoma were excluded. This study was approved by the institutional review board.

Data collection

M AN U

Data were collected from clinical charts, radiotherapy records, and the institutional picture archiving and communication system. Histologic specimens obtained at diagnosis and recurrence, if available, were centrally reviewed by institutional neuropathologists. 1q gain by infratentorial tumors was evaluated with a fluorescence in-situ hybridization (FISH) probe for EXO1 [13]. The rearrangement of C11orf95 and RELA in supratentorial tumors was evaluated with FISH probes for the respective targets [14]. Symptoms

TE D

of radiation necrosis were graded in accordance with the CNS necrosis subscale of the Common Terminology Criteria for Adverse Events version 4.0. Asymptomatic enhancement on contrast-enhanced,

Treatment

EP

T1-weighted MR imaging was classified as grade 1; patients with T2 signal change only were graded 0.

AC C

RT2 was delivered as focal irradiation of the primary site or as CSI (CSI-RT2). With CSI-RT2, foci of resected or gross residual disease at primary or metastatic sites were supplementally irradiated (boost RT). The gross tumor volume (GTV) included the tumor bed after any recent surgeries, as well as all sites of gross residual disease. The clinical target volume (CTV) margin was 0.5 cm or less, and modified for anatomic boundaries to tumor spread. For photon treatment, a planning target volume (PTV) margin of 0.3 cm was applied. The CSI-RT2 dose was 36.0 to 39.6 Gy. The primary and any metastatic sites, including those in the brain and spine, were treated to 54 Gy. Treatments were delivered in 1.8-Gy fractions on 5 days each week. CSI-RT2 was administered to 40 cases of infratentorial ependymoma.

ACCEPTED MANUSCRIPT

Through May 2007, including 15 of 17 cases, the CSI-RT2 dose to the upper cervical cord was entirely shielded in regions that previously received 41.4 Gy or higher. After May 2007, whereas the upper cervical spinal cord volume was entirely shielded in a single case, the upper cervical spinal cord in the

RI PT

remaining CSI-RT2 cases was treated to 27 Gy (n = 4), 36 Gy (n = 14), and 39.6 Gy (n = 4), regardless of the prior dosimetry. The decision to provide more comprehensive CSI-RT2 coverage after May 2007 was made after two of the early patients with metastatic disease developed disease progression in the shielded

SC

volume.

M AN U

Statistical analysis

Clinical factors and baseline characteristics were reported descriptively. Fisher’s exact test was used to compare proportions between groups. FFP was defined as the time from RT2 to ependymoma recurrence. OS and FFP were measured from the first day of RT2 and reported using the Kaplan-Meier method. Comparisons between groups were performed using the log-rank test. Cox regression was used to create

TE D

multivariate models to evaluate factors associated with OS and FFP, using stepwise variable selection. Patients who were alive at last follow-up or lost to follow-up were censored. The cumulative incidence of necrosis was reported while accounting for competing risks of tumor recurrence, death, or receipt of a

EP

third course of radiotherapy (RT3)[15]; patients with pre-existing grade 2 or higher radiation necrosis from RT1 were counted as having a competing event at time zero or were otherwise excluded from the

AC C

analysis. Analyses were completed using SAS 9.4 (Cary, NC).

Results

A total of 101 patients were eligible for inclusion. Their baseline characteristics and radiation details are listed in Table 1. The median follow-up was 81.5 months. The median time from the first day of RT1 to initial progression was 21.7 months (interquartile range [IQR] 12.8–35.0). The median time from initial progression to the first day of RT2 was 3.3 months (IQR 1.5–8.2). The median interval between the first

ACCEPTED MANUSCRIPT

day of RT1 and the first day of RT2 was 26.8 months (range, 4.1-138; IQR 18.0–43.1). Ninety-eight (97%) were treated with RT2 after 1996 and 66 (65%) were treated with RT2 after 2006.

RI PT

Figure 1 shows the patterns of initial RT1 failure. Forty-six patients (46%) received focal RT2 for postRT1 local failure (treatment group, “LF/focal”); and 55 patients (54%) received CSI as part of RT2 for post-RT1 local failure (LF/CSI, n = 10), distant-only failure (DF/CSI, n = 31), or combined failure

SC

(CF/CSI, n = 14). The pathologic characteristics of the tumors are listed in Table 2. Sixteen of 28 patients (57%) with infratentorial primaries and 1q gain had distant or combined failure after RT1, whereas 14 of

M AN U

42 patients (33%) without 1q gain had distant or combined failure (P = 0.084) after RT1.

The most commonly used RT2 dose was 54 Gy, in 69 of 101 patients (Table e1). 11 of 31 patients with RT1 distant failures received an RT2 dose greater than 54 Gy, whereas 11 of 56 patients with RT1 local failures received RT2 doses less than 54 Gy. No patient that received CSI-RT2 went on to receive

TE D

supplemental RT2 treatment to a total dose greater than 54 Gy. Primary site (supratentorial vs. infratentorial) was not associated with RT2 dose (P = 0.39).

EP

Progression and survival

Figure 2 shows the OS and FFP for all patients. The median OS time for all patients was 75.1 months

AC C

(95% confidence interval [CI] 52.5–159.4). The 1-, 2-, and 5-year estimates of OS were 95.5% (95% CI 88.4–98.3), 74.9% (95% CI 64.0–82.8), and 57.3% (95% CI 45.5–67.4). Forty-six patients died. Causes of death were ependymoma (38 patients [83%]), secondary neoplasm (4 patients [9%]), stroke (1 patient [2%]), sepsis with renal failure and thrombotic thrombocytopenic purpura (1 patient [2%]), chronic late complications from combined chemotherapy and radiation (1 patient [2%], who died 30.4 years after RT2), and unknown (1 patient [2%]). The median FFP was 27.3 months (95% CI 16.3–45.7). The estimates of FFP at 1, 2, and 5 years were 71.3% (95% CI 60.8–79.5), 53.3% (95% CI 42.3–63.2), and

ACCEPTED MANUSCRIPT

36.7% (95% CI 26.4–47.0), respectively. Among the 57 patients who experienced recurrence, the median time from RT2 to disease progression was 12.4 months (IQR 7.7–25.4).

RI PT

Figure 3 shows the OS and FFP stratified by various clinicopathologic factors. The subgroup of patients with DF/CSI had the longest 5-year OS and FFP (75.5% [95% CI 53.2–88.3] and 48.5% [95% CI 27.9– 66.3], respectively). Patients with anaplastic histology at recurrence experienced reduced OS (hazard ratio

SC

[HR] 4.7, 95% CI 1.9–11.4, P = 0.0007) and FFP (HR 2.2, 95% CI 1.1–4.1, P = 0.018). Those with an interval of 4 years or longer between RT1 and recurrence had improved OS and FFP. Survival stratified

M AN U

by 1q gain is presented in Figures e1 and e2. Among patients who experienced local failure after RT1 of an infratentorial primary tumor, 1q status was not prognostic after RT2 (OS HR 0.62, 95% CI 0.21–1.9, P = 0.41; FFP HR 1.4, 95% CI 0.61–3.4, P = 0.41). However, among individuals who experienced any distant failure after RT1 (e.g., distant-only or combined failure), 1q gain was adversely associated with OS (HR 3.5, 95% CI 1.1–10.6, P = 0.029) and FFP (HR 3.0, 95% CI 1.1–8.3, P = 0.037). Finally,

TE D

C11orf95–RELA fusion status was not associated with OS or FFP, although this analysis was limited by the small number of patients (Figure e3).

EP

Table 3 lists factors associated with OS and FFP on multivariate analysis. The initial site of disease, extent of pre-RT2 surgery, and receipt of pre-RT2 chemotherapy were not significantly associated with

AC C

OS or FFP, after adjusting for other factors. Variables associated with reduced OS and FFP included male sex, anaplastic histology at recurrence, treatment group, and a short interval between RT1 and first progression (OS only). Receiving DF/CSI was associated with improved OS and FFP, after adjusting for other variables. Neither 1q status nor C11orf95 status were included as model variables because information about these molecular markers was specific to infratentorial and supratentorial primary tumors, respectively.

Patterns of failure

ACCEPTED MANUSCRIPT

Fifty-seven patients experienced disease progression after RT2. Figure 1 shows the pattern of RT2 failure. No individuals with RT1 local failure who received CSI-RT2 (n = 10) had distant-only failures. In contrast, among those patients who received focal RT2 for RT1 local failure (n = 46), there were 11

RI PT

distant-only failures (P = 0.18). Among patients with infratentorial primary tumors, known 1q status, local failure after RT1, and subsequent progression after focal RT2, five of six patients (83%) with 1q gain

SC

developed distant-only failure, compared with four of 13 patients (31%) without 1q gain (P = 0.057).

RT3

M AN U

Fourteen patients received RT3 directed at a site in the neuraxis (four patients underwent stereotactic radiosurgery [SRS], eight received fractionated photon RT, one received photon stereotactic body RT, and one received proton RT). Two patients underwent SRS to a site of local failure within prior RT1 and RT2 treatment. The median time between RT2 and RT3 was 32.1 months (range, 12.3–133). Eleven patients experienced progression of their ependymoma (10 distant, 1 local) after RT3. Three patients

TE D

remained free of disease at last follow-up. The median OS after RT3 was 46.5 months (95% CI 19.0–NR [not reached]), and the median FFP was 15.5 months (95% CI 8.6–28.4, Figure e4).

EP

Six patients received a fourth course of RT (RT4) for distant failure after RT3. Four patients received RT4 SRS. The fifth patient received 30.6 Gy, in 1.8-Gy fractions, to a limited area of the spine, and the

AC C

sixth patient received 35 Gy, in 2.5-Gy fractions, to an infratentorial nodule.

Toxicity after RT2

Table e2 presents the distribution and treatments for radiation necrosis, and the cumulative incidence curves are shown in Figure e5. Radiation necrosis of grade 3 or higher after RT2 was observed in seven patients, with a 10-year cumulative incidence of 7.9% (95% CI 3.4–14.8%). Twenty-five patients developed any-grade radiation necrosis, including temporary, asymptomatic enhancement on MRI, with a 10-year cumulative incidence of 26.9% (95% CI 18.2–36.2). The median duration of abnormal MR

ACCEPTED MANUSCRIPT

enhancement in patients with grade 1–only necrosis was 5.6 months (IQR 3.9–7.6); for symptomatic patients with grade 2 or higher necrosis, the median duration from symptom onset to resolution was 4.2

RI PT

months (IQR 3.1–9.3).

There was no association between age at RT2 and necrosis of any grade; patients younger than 7 years and those aged 7 years or older had 10-year cumulative incidences of 22.7% and 31.0%, respectively

SC

(P = 0.48) (Figure e6). The median ages of patients with or without necrosis of any grade were similar (7.4 years and 7.5 years, respectively). The median time between RT1 and RT2 was similar between those

M AN U

with no necrosis (2.1 years), those with any-grade necrosis (2.5 years), and those with necrosis of grade 3 or higher (1.9 years). Patients with RT1-to-RT2 intervals of less than 2 years or 2 years or more had crude any-grade necrosis rates of 19.4% and 31.8%, respectively (P = 0.16) (Figure e6).

The seven observed cases of radiation necrosis of grade 3 and higher are listed in Table e3. Of those who

TE D

received CSI-RT2, 6 out of 53 (11.3%) developed severe radiation necrosis, as compared to one out of 44 who received focal RT2 (P = 0.12). The CSI dose, when given, was 39.6 Gy in 5 of 6 patients. Among patients who received CSI-RT2, 12.5% (5 of 35) received a CSI dose greater than 36 Gy and developed

EP

severe necrosis, as compared with 7.1% (1 of 13) who received a CSI dose of 36 Gy or less (P = 1.0). Four patients developed necrosis in-field to RT1 and RT2 (including boost RT); the remainder developed

AC C

necrosis in regions that received RT1 and CSI-RT2. The total nominal dose that resulted in necrosis ranged from 108–119 Gy.

Nine patients developed secondary cancers after RT2 (Table e4). Eight patients received chemotherapy for ependymoma before a secondary neoplasm was diagnosed. The median time to a second cancer was 162 months after RT1 (IQR 101–190) and 91 months after RT2 (IQR 69–153).

Discussion

ACCEPTED MANUSCRIPT

To our knowledge, this is the largest study yet of patients with recurrent ependymoma treated by reirradiation. We demonstrated that patients who received RT2 remained at substantial risk of further disease progression and that long-term survival without recurrence was possible for certain subgroups.

RI PT

The subgroup of patients with DF/CSI had good outcomes. In this subgroup of patients, local control at the primary site was successfully achieved with RT1, suggesting that their tumors had radiosensitive characteristics that might be effectively treated with surgery and re-irradiation. Patients with LF/CSI did

SC

not develop distant-only failures after RT2. This observation raises important questions about patient selection and the role of CSI-RT2 in treating locally recurrent ependymoma. Aside from the evaluation of

M AN U

RT1 distant failures, the overall sample size was insufficient to enable the identification of additional clinicopathologic factors that would predict a benefit from CSI-RT2.

The incidence of severe brain necrosis in our patients was low. Previous studies found that an interval between RT1 and RT2 of less than 6 months was associated with increased risk of necrosis [16,17]. Most

TE D

patients in this series had an interval of 9 months or longer between RT1 and RT2. With this knowledge driving clinical practice in the present study, we found no association between the length of the RT1-toRT2 interval and the development of necrosis. Nonetheless, continued meticulous attention to normal

AC C

1q gain

EP

tissue dose is required, particularly for eloquent structures.

This is the first report of the prevalence of 1q gain (40%) in a large series of patients with recurrent infratentorial ependymoma. Kilday et al. found 1q gain in two (33%) of six cases of recurrent ependymoma [18]. Not unexpectedly, our cohort was enriched for patients with 1q gain as compared to de novo ependymoma, given the association between 1q gain and tumor recurrence [13,19,20]. The prevalence of 1q gain in de novo infratentorial tumors ranged from 13% [13] to 19% [19].

ACCEPTED MANUSCRIPT

Chromosome 1q gain is a negative prognostic factor in de novo ependymoma [13,20]. In this study, 1q gain was associated with poorer prognosis in patients with any distant failure after RT1, but not in individuals who experienced local failure after RT1. One might speculate that 1q gain is prognostic in

RI PT

tumor clonogens that have not been previously irradiated, but the prognostic value of 1q gain is lost for patients with previously irradiated, locally recurrent ependymoma. Furthermore, CSI-RT2 may play an important role in treating locally recurrent ependymoma with 1q gain by reducing the risk of post-RT2

SC

distant-only failure.

M AN U

Comparison with other studies

Re-irradiation of recurrent pediatric intracranial ependymoma is established as a safe, effective treatment that may affect long-term survival in selected patients. An early study by XXXX et al. reviewed 38 patients who underwent re-irradiation [17]. No patient with post-RT1 distant failure who received CSIRT2 died. The 5-year survival for patients with post-RT1 local failure who received focal RT2 was 67%.

TE D

Messahel et al. reviewed 62 pediatric patients who received RT2 for relapsed ependymoma; receipt of CSI-RT2 may have been associated with improved OS in the subgroup of patients older than 3 years at diagnosis, but it was unclear whether this applied to all patients with post-RT1 local and distant

EP

recurrence or only to a subgroup thereof [8]. Receipt of RT2 was independently associated with improved OS. Zacharoulis et al. reviewed 82 patients with recurrent ependymoma. Patients with synchronous local

AC C

and distant relapse had worse survival (HR 3.6, 95% CI 1.7–7.6), and the 5-year OS for re-irradiated patients (n = 40) was 16% [21]. Bouffet et al. reported on 47 patients with relapsed ependymoma, 18 of whom received RT2 [9]. The 3-year OS for patients who received RT2 was 81%. Eaton et al. reviewed 20 patients, most of whom received focal-only proton RT2 [6]. The 3-year OS and progression-free survival (PFS) were 78.6% and 28.1%, respectively. Among those patients who received external beam RT, there were four cases of grade 2 or higher radiation necrosis. Finally, Lobón et al. reviewed 32 patients who were re-treated; for the entire cohort, the median PFS and OS were 1.2 years and 3.5 years, respectively

ACCEPTED MANUSCRIPT

[7]. Similar to the findings of the present study, the median PFS was highest for distant post-RT1 relapses treated with CSI-RT2 (6.8 years).

RI PT

In this study, male sex was associated with reduced OS and FFP. Male sex is a known negative prognostic factor in de novo pediatric ependymoma.[3,4] Males are also over-represented in patients with PF-EPN-A type ependymoma as assessed by DNA methylation; these tumors experience poorer

SC

outcomes.[11] The prognostic effect of sex on survival persists even after accounting for the molecular subgroup of ependymoma. [22] In the setting of recurrent ependymoma, Eaton et al. found a trend to

M AN U

worse progression-free survival among males. [6] It is not known why males with pediatric ependymoma experience poorer outcomes than females. One wonders whether unmeasured molecular features of ependymoma in boys make their tumors more aggressive.

RT3

TE D

RT3 may be helpful in selected cases and can provide effective local control and palliation. Long-term disease control may be possible, but comprehensive CSI is challenging in these patients, particularly those who have received RT1 and RT2 directed to the same site or have previously received CSI. In this cohort,

EP

RT3 was generally reserved for patients with a long disease-free interval after RT2.

AC C

The oldest patient in this study received RT2 in 1985, and the long-term results demonstrate the potential for long-term survival and disease control in children treated with a second course of high-dose, fractionated RT. The study also incorporated molecular information that is prognostic for de novo ependymoma; the roles of 1q gain and C11orf95 rearrangements in recurrent ependymoma were not previously known. The limitations of this study include the retrospective nature of the data collection and missing molecular data for some cases, particularly those with older pathology samples.

ACCEPTED MANUSCRIPT

A phase II study is prospectively enrolling patients with recurrent ependymoma for RT2 (RERTEP, NCT02125786). In this study, patients are eligible for RT2 if at least 9 months had elapsed after the initiation of RT1, to allow for repair of previously-irradiated tissues. The CSI-RT2 dose (if given) is 36

RI PT

Gy, while recurrent primary and metastatic sites are treated to 54 Gy. Decisions about cervical cord shielding after 27 Gy CSI-RT2 are made on an individualized basis. Boost radiotherapy is delivered to sites of resected or residual disease with a CTV margin of up to 0.5 cm and a PTV margin of 0.2-0.3 cm,

SC

with daily volumetric image guidance. RT2 dose to organs at risk (e.g. brainstem) are kept as low as possible, with maximum point brainstem dose kept at or below 54 Gy. It is anticipated that this study will

M AN U

provide the best-available evidence for the safety and efficacy of RT2 in patients with recurrent ependymoma, and the results are eagerly awaited.

Conclusions

In this study of 101 patients with recurrent ependymoma treated with second-course irradiation, the 5-

TE D

year OS and FFP were 57% and 37%, respectively. Long-term survival was observed in a subset of patients; those patients with distant-only failure treated with CSI-RT2 followed by boost RT2 had the best outcomes (5-year OS and FFP of 76% and 48%, respectively). Factors associated with improved survival

EP

were female sex, non-anaplastic histology at recurrence, a longer recurrence-free interval after RT1, and distant-only failure after RT1 (treated with CSI-RT2 followed by focal boost irradiation). CSI, as part of

AC C

RT2, is indicated for patients with any RT1 distant failure and might be helpful in treating RT1 local failure (particularly in individuals with 1q gain) by reducing the incidence of subsequent distant failures after RT2. The 10-year cumulative incidence of severe radiation necrosis (grade 3 or higher) after RT2 was low at 7.9%. Thus, re-irradiation for recurrent ependymoma was well-tolerated and safe for most patients, and it offered the possibility of long-term disease control in some patients. The option of reirradiation should be discussed with patients who develop recurrent ependymoma.

References

ACCEPTED MANUSCRIPT

Ostrom QT, et al. Cbtrus statistical report: Primary brain and other central nervous system tumors diagnosed in the united states in 2009-2013. Neuro Oncol 2016;18:v1-v75.

[2]

Merchant TE. Current clinical challenges in childhood ependymoma: A focused review. J Clin Oncol 2017;35:2364-2369.

[3]

Massimino M, et al. Final results of the second prospective aieop protocol for pediatric intracranial ependymoma. Neuro Oncol 2016;18:1451-1460.

[4]

Merchant TE, et al. Conformal radiotherapy after surgery for paediatric ependymoma: A prospective study. Lancet Oncol 2009;10:258-266.

[5]

Murray L, et al. Re-irradiation for relapsed paediatric ependymoma. ASCO Meeting Abstracts 2016;34:10565.

[6]

Eaton BR, et al. Use of proton therapy for re-irradiation in pediatric intracranial ependymoma. Radiotherapy and oncology : journal of the European Society for Therapeutic Radiology and Oncology 2015;116:301.

[7]

Lobón M, et al. Re-irradiation of recurrent pediatric ependymoma: Modalities and outcomes: A twenty-year survey. SpringerPlus 2016;5:1-9.

[8]

Messahel B, et al. Relapsed intracranial ependymoma in children in the uk: Patterns of relapse, survival and therapeutic outcome. European journal of cancer 2009;45:1815-1823.

[9]

Bouffet E, et al. Survival benefit for pediatric patients with recurrent ependymoma treated with reirradiation. Int J Radiat Oncol Biol Phys 2012;83:1541-1548.

[10]

Korshunov A, et al. Molecular staging of intracranial ependymoma in children and adults. J Clin Oncol 2010;28:3182-3190.

[11]

Pajtler KW, et al. Molecular classification of ependymal tumors across all cns compartments, histopathological grades, and age groups. Cancer Cell 2015;27:728-743.

[12]

Witt H, et al. Delineation of two clinically and molecularly distinct subgroups of posterior fossa ependymoma. Cancer cell 2011;20:143-157.

[13]

Godfraind C, et al. Distinct disease-risk groups in pediatric supratentorial and posterior fossa ependymomas. Acta Neuropathol 2012;124:247-257.

[14]

Parker M, et al. C11orf95-rela fusions drive oncogenic nf-kappab signalling in ependymoma. Nature 2014;506:451-455.

[15]

Gray RJ. A class of k-sample tests for comparing the cumulative incidence of a competing risk. The Annals of Statistics 1988;16:1141-1154.

[17] [18]

[19]

SC

M AN U

TE D

EP

AC C

[16]

RI PT

[1]

Nieder C, et al. Update of human spinal cord reirradiation tolerance based on additional data from 38 patients. Int J Radiat Oncol Biol Phys 2006;66:1446-1449.

XXXX

Kilday JP, et al. Copy number gain of 1q25 predicts poor progression-free survival for pediatric intracranial ependymomas and enables patient risk stratification: A prospective european clinical trial cohort analysis on behalf of the children's cancer leukaemia group (cclg), societe francaise d'oncologie pediatrique (sfop), and international society for pediatric oncology (siop). Clin Cancer Res 2012;18:2001-2011.

Araki A, et al. Chromosome 1q gain and tenascin-c expression are candidate markers to define different risk groups in pediatric posterior fossa ependymoma. Acta Neuropathol Commun 2016;4:88.

ACCEPTED MANUSCRIPT

Mendrzyk F, et al. Identification of gains on 1q and epidermal growth factor receptor overexpression as independent prognostic markers in intracranial ependymoma. Clin Cancer Res 2006;12:2070-2079.

[21]

Zacharoulis S, et al. Treatment and outcome of children with relapsed ependymoma: A multiinstitutional retrospective analysis. Child's nervous system : ChNS : official journal of the International Society for Pediatric Neurosurgery 2010;26:905-911.

[22]

Ramaswamy V, et al. Therapeutic impact of cytoreductive surgery and irradiation of posterior fossa ependymoma in the molecular era: A retrospective multicohort analysis. J Clin Oncol 2016;34:2468-2477.

AC C

EP

TE D

M AN U

SC

RI PT

[20]

ACCEPTED MANUSCRIPT

Figure Legends Figure 1. Patterns of failure after RT1 and RT2. Patients who underwent CSI-RT2 received sequential

RI PT

boost RT to sites of resected and/or gross disease. RT2 CSI = craniospinal irradiation as part of reirradiation; LF = local failure; DF = distant-only failure; CF = combined local and distant failure.

Figure 2. OS (a) and FFP (b) for all patients.

SC

Figure 3. OS and FFP for patients by treatment group (1 = LF/focal, 2 = DF/CSI, 3 = CF/CSI, 4 = LF/CSI; first row; a, b), extent of pre-RT2 surgery (second row; c, d), recurrent histology (third row; e, f), and

AC C

EP

TE D

ependymoma; GTR = gross total resection.

M AN U

time between RT1 and progression (fourth row; g, h). AEP = anaplastic ependymoma; EP = grade II

ACCEPTED MANUSCRIPT

Table 1. Baseline characteristics for all patients Characteristic

n = 101

Initial diagnosis

4.3 (0.62–17.8)

RT1 start

4.7 (1.1–17.9)

RT2 start

7.5 (2.5–21.97) 37 (36.6%)

Initial site of ependymoma 74 (73.3%)

M AN U

Infratentorial Supratentorial

27 (26.7%)

Pre-RT1 surgery GTR

72 (71.3%)

NTR

11 (10.9%) 18 (17.8%)

TE D

STR Median RT1 dose in Gy (range) RT1 modality

EP

Photon

59.4 (37.8–69.6)a

81 (80.2%) 1 (1.0%)

Proton

19 (18.8%)

AC C

Photon + SRS

Post-RT1 pattern of failure Local only

56 (55.5%)

Distant only

31 (30.7%)

Local and distant

14 (13.9%)

Number of lesions at post-RT1 failure 1

SC

Female sex

RI PT

Median age, years (range)

63 (62.4%)

ACCEPTED MANUSCRIPT

2

16 (15.8%)

≥3

22 (21.8%)

69 (68.3%)

NTR

15 (14.9%)

STR

16 (15.8%)

No surgery

1 (1.0%)

SC

GTR

RI PT

Pre-RT2 surgery

31 (30.7%)

Median RT2 CSI dose in Gy, if given (range)

39.6 (12.0b–44.4)

Median RT2 total dose in Gy (range) RT2 modality Photon

M AN U

Pre-RT2 chemotherapy for recurrence

54.0 (36.0–59.4)

87 (86.1%)

Photon + SRS

1 (1.0%)

13 (12.9%)

RT3 modality (n = 14) Fractionated photon

EP

Photon SRS

TE D

Proton

8 (57.1%) 4 (28.6%) 1 (7.1%)

Fractionated proton

1 (7.1%)

AC C

Photon SBRT

a

69.6 Gy was delivered in 1.2-Gy fractions twice daily (bid).

b

The one patient who received 12 Gy underwent total body irradiation for synchronously diagnosed acute

lymphoblastic leukemia, followed by focal RT2 for recurrent ependymoma. One other patient received 35.2 Gy CSI; all others received total CSI doses of 36 Gy or greater. GTR = gross total resection; Gy = gray; NTR = near-total resection; SBRT = stereotactic body radiotherapy; SRS = stereotactic radiosurgery; STR = subtotal resection

ACCEPTED MANUSCRIPT

Table 2. Pathology of ependymoma tumors n (%)

Present

28 (40.0%)

Absent

42 (60.0%)

RI PT

1q gain (IT only, n = 74)

Unknown 4

12 (63.2%)

Absent

7 (36.8%)

Unknown 8

Recurrence pathology

M AN U

Present

SC

C11orf95-RELA fusion (ST only, n = 27)

EP AEP NET No surgery

EP

15

15

AEP

9

59

TE D

Initial pathology

2

1

AEP = anaplastic ependymoma; EP = ependymoma; IT = infratentorial; NET = neuroepithelial tumor; ST

AC C

EP

= supratentorial

ACCEPTED MANUSCRIPT

Table 3. Multivariable analysis of clinicopathologic factors associated with OS and FFP Variable

Overall survival HR (95% CI)

P-value HR (95% CI)

P-value

0.35 (0.17–0.72) 0.004

0.35 (0.18–0.68)

Time from RT1 to recurrence (per year) 0.75 (0.59–0.96) 0.024

0.87 (0.74–1.03)

0.010

Less than GTR before RT2 (vs. GTR)

1.8 (0.99–3.2)

0.055

*

Reference

AEP

5.1 (1.9–13.5)

NET

**

Treatment group

Reference

0.001

4.5 (0.51–40.4)

Reference

LF/CSI

0.98 (0.38–2.6)

DF/CSI

0.37 (0.16–0.87) 0.023

TE D

LF/focal

CF/CSI

2.5 (1.2–5.1)

M AN U

EP

SC

Recurrent histology

2.0 (0.84–4.9)

0.002

RI PT

Female (vs. male)

Freedom from progression

0.013 0.18

Reference

0.97

0.12

0.83 (0.34–2.0)

0.68

0.34 (0.16–0.72)

0.004

1.5 (0.71–3.3)

0.28

* not significant; ** insufficient events

EP

AEP = anaplastic ependymoma; CF = combined failure; CSI = craniospinal irradiation; DF = distant

AC C

failure, EP = ependymoma; GTR = gross total resection; LF = local failure; NET = neuroepithelial tumor

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT

AC C

EP

TE D

M AN U

SC

RI PT

ACCEPTED MANUSCRIPT