Abstract
Aim: This retrospective analysis evaluates feasibility of wide-field re-irradiation using pulsed reduced dose rate (PRDR) technique in patients with recurrent ependymoma. PRDR employs a dose rate of 6 cGy/min, as opposed to 400-600 cGy/min for conventional radiation, allowing for enhanced normal tissue repair. Patients and Methods: Five patients with recurrent ependymoma having eight lesions (two brain, six spinal cord) were treated with PRDR. Progression-free survival (PFS) and overall survival (OS) were estimated by Kaplan Meier method. Results: The median interval between two radiation courses was 58 months (range: 32-212 months). The median PRDR dose was 40 Gy (range: 30.6-54 Gy) with a median cumulative lifetime dose of 105.2 Gy (range: 90-162.4 Gy). At a median post-PRDR follow-up of 64 months, estimated 4-year OS and PFS from PRDR was 60% and 35.7%, respectively. None of the patients developed necrosis on serial magnetic resonance imaging scans, and only one patient had progressive mild radiculopathy. Conclusion: In patients with large-volume recurrent ependymoma, re-irradiation with wide-field PRDR is a feasible option.
Ependymomas are rare neuro-epithelial tumors that arise from ependymal cells lining the ventricles and cerebro-spinal fluid (CSF) outflow track. Standard treatment recommendation includes maximal safe resection followed by consideration of adjuvant radiotherapy (RT) based on pathological findings. Treatment options for recurrent ependymoma, however, are limited and are defined by the treatment received at the time of primary diagnosis, interval from initial treatment and performance status. Median survival after relapse has been reported to be between 8.7 to 24 months, with 2-year progression-free survival (PFS) of 29% (1, 2). Surgical de-bulking is performed whenever feasible. Active chemotherapy agents that are used include cyclophosphamide, cisplatin, carboplatin, lomustine, etoposide and temozolomide, although overall benefit from chemotherapy is limited (3, 4). Lack of local control is the most important cause of death in patients with recurrent disease (5, 6). This stresses the importance of local therapy in the form of surgery or re-irradiation, when surgery is not feasible. Re-irradiation can be performed using highly conformal fractionated or stereotactic radiosurgery-based techniques. Preliminary reports from single-institution studies have demonstrated feasibility of re-irradiation with encouraging outcomes (7-11).
This approach of re-irradiation, however, is limited to a restricted population with limited volume disease due to the radiation tolerance of the neural structures. Conventional fractionated RT employs delivery of radiation at a dose rate of 4-6 Gy/min wherein biological repair of normal tissue DNA damage occurs in between consecutive fractions. Reducing the dose rate at which radiation is delivered can potentially allow enhanced DNA damage repair to occur intra-fraction, thereby improving normal tissue tolerance (12). Pulsed reduced dose rate (PRDR) RT is a technique employed at our institution to deliver RT in re-irradiation settings by dividing each radiotherapy fraction into a number of sub-fractions of 0.2 Gy delivered in a pulsed manner separated by a fixed time interval, creating an apprarent dose-rate of 0.0667 Gy/min (13). This also allows radiation to be delivered in the low-dose hyperradiosensitivity (LD-HRS) range that can potentiate radiation-induced biological tumor tissue damage (14). The combination of increased sub-lethal damage repair in normal tissue and increased tumor-tissue damage from LD-HRS may improve the therapeutic ratio.
Our clinical experience of re-irradiation using PRDR for recurrent glioma and recurrent breast cancer has been reported (13, 15-17). The median PRDR re-irradiation doses used in those patients were 50 and 54 Gy, respectively, with median cumulative doses of 106 and 110 Gy respectively. Our experience has demonstrated the feasibility of delivering wide-field re-irradiation to sensitive neural and chest wall tissues (13, 15-17). This is critical in patients with ependymoma who tend to have disease tracking along the meninges. In this case series, we present the results of using wide-field PRDR re-irradiation in pediatric and young adult patients with recurrent, large-volume, ependymomas.
Patients and Methods
After obtaining approval from the institutional review board (#M-2007-1448), five patients with eight lesions were identified in this retrospective chart review who were treated with re-irradiation using PRDR for recurrent ependymoma from January 2000 to December 2010. The median age at original diagnosis was 10 years (range: 3-31 years). This included three males and two females. For the original diagnosis, all the patients underwent primary surgery followed by adjuvant therapy with RT. One patient also received concurrent chemotherapy with RT (doublet of cisplatin and etoposide). Radiation was delivered to a median dose of 48.4 Gy (range: 36-55.8 Gy). Patient-, disease- and treatment-related variables were noted for original presentation and at recurrence. Disease was classified as intra-cranial or spinal, based on the site of origin.
The individual patient radiation treatment plans/port films were reviewed to identify the radiation dose-fractionation employed both at the time of original treatment and for PRDR re-irradiation. Radiation planning techniques were categorized as 2- or 3-dimensional (2-D or 3-D). PRDR treatment was delivered by using a series of 0.2 Gy pulses separated by 3-minute intervals, creating an apparent dose rate of 0.0667 Gy/min. The dose rate of the linear accelerator was reduced to 1 Gy/min during each 0.2 Gy pulse.
Pattern of failure from first treatment was noted and the recurrence sites were labeled as in-field or out-of-field with respect to the original radiation fields. In addition, the estimate of PRDR irradiated volume was obtained by calculating the volume of 100% and 50% isodose lines for patients that were planned with 3D-conformal RT technique. For patients planned with 2D-technique, since no volumetric information was available, the field sizes of the radiation fields were used as a surrogate of treatment volume.
Kaplan Meier estimates of overall survival (OS) and PFS were calculated, both for the original presentation and subsequent to recurrence. In patients with multiple lesions, analysis of post-PRDR PFS was carried out by considering each irradiated lesion separately. PFS was calculated from the time of PRDR completion to the development of subsequent recurrence, or date of last follow-up in the absence of recurrence. If PRDR re-irradiation was employed twice for the same site in a patient, PFS for each PRDR treatment was calculated separately by measuring the time interval from completion of PRDR to the diagnosis of subsequent recurrence. For OS, individual patients were assessed irrespective of the number of lesions irradiated in that individual. Log-rank test was used for univariate analysis. Statistical analysis was performed using SPSS v15.0 software (SPSS Inc., Chicago, USA).
Results
Pattern of first recurrence. All but one patient (#5), had recurrence within the radiation field. The patient who developed out-of-field recurrence had spinal disease at original presentation and developed a metachronous infra-tentorial lesion four years later. The estimated median PFS for all patients at the initial presentation was eight years (95% confidence interval (CI): 3.7-12.2 years). The 4-year and 8-year PFS were 80% and 40%. The median PFS for patients with intra-cranial disease and spinal disease was two years and 13 years, respectively (p=0.207).
Therapy at local recurrence and timing of PRDR. The primary therapy at the time of first local recurrence was surgery for all five patients. A gross total resection was attempted when feasible. Post-surgery, two patients (#1 and #3) received PRDR re-irradiation, one patient (#4) received chemotherapy alone and one patient (#2) was observed. Patient #2 had a subsequent additional recurrence seven-years from his first treatment, which was managed with postoperative conventionally fractionated RT. Patient #4 developed subsequent additional recurrence, which was re-irradiated with PRDR. Patient #5, who had thoracolumbar spinal disease at original presentation, developed a recurrence in the lumbar spine level about 15 years after first recurrence. This was managed with surgery followed by standard fractionation re-irradiation. He developed further recurrence at this lumbar level two years later. Simultaneously, at that time he developed recurrent disease even at the thoracic level. Both the lumbar and the thoracic recurrences were within the original RT field. These two sites of recurrences were then re-irradiated with PRDR using two separate fields (Figure 1A). Both these sites received a second course of PRDR re-irradiation for additional recurrence three years after the first PRDR course. Overall, this patient received four separate courses of RT to the lesion at the lumbar level and three courses of RT to the lesion at the thoracic level, over a period of 20 years. The PFS for each PRDR treatment course in this patient was calculated separately. Thus, overall five patients had PRDR treatment courses to eight sites.
Patient #5 with spinal ependymoma at presentation and recurrence. A: The isodose distribution (light green, 105%; red, 100%; pink, 90%; dark green, 50%, blue, 10%) from the treatment plan treating the thoracic and the lumbar spine with separate large-fields using PRDR technique. Cumulative life-time dose to the thoracic level was 120.6 Gy and to the lumbar level was 162.4 Gy. B and C: Findings on T2-weighted magnetic resonance images at the time of last follow-up showing lack of radiation necrosis at both sites, progressive disease at lumbar level and controlled disease at thoracic level.
The median age at PRDR was 25 years (range: 9-44 tomography (CT) scan was performed when planning with years). The median time interval between the original RT course and PRDR course was 58 months (range: 32-212 months).
PRDR re-irradiation volumes. Two of the eight PRDR treatments were directed to intracranial sites while remaining six targeted spinal locations (Table I). Information from diagnostic magnetic resonance imaging (MRI) was utilized when planning the radiation treatment. Co-registration of the MRI images with the radiation planning computerized 3D-techniques. 2D-Planning techniques with fluoroscopic simulation were used for four out of the eight PRDR treatments, all of which were spinal treatments. The largest dimensions of radiation portal for these individual spinal plans were 8 cm × 26 cm, 12 cm × 29 cm and 12 cm × 34 cm, with a mean portal area of 348 cm2. The entire spinal canal within the length specified above was included in the target volume. Four lesions were treated with 3D-conformal technique. For these patients, the median volume encompassed by the 100% isodose line was 82.9 cm3, while that encompassed by the 50% isodose line was 882 cm3. Figure 1A and 2A represent treatment plan/isodose lines demonstrating use of wide-field radiation.
PRDR re-irradiation dose. Median PRDR re-irradiation dose was 40 Gy (range: 30.6-54 Gy) (Table II). The cumulative radiation dose per site was a median of 105.2 Gy (range, 90-162.4 Gy). Prior to receiving PRDR re-irradiation, patients #2 and #5 had also received re-irradiation with standard doserate to a dose of 40 Gy and 45 Gy, respectively, for interval recurrences. Both these patients had subsequent recurrences, which were then treated with PRDR re-irradiation.
A: Treatment volume in patient #3 irradiated to a total dose of 54 Gy using PRDR technique. Cumulative dose to this region was 109.8 Gy. B: Lack of radiation necrosis seen on at the time of the last follow-up. The patient unfortunately developed progressive disease in the infratentorium resulting in hydrocephalous of the third and lateral ventricles and ultimately his death. The supratentorial re-irradiation site stayed controlled.
PRDR re-irradiation outcomes. At a median post-PRDR follow-up of 64 months (range: 8-86 months), three patients had died with progressive disease, one is alive with controlled disease, while one is alive with multiply-recurrent and progressive disease. Kaplan Meier estimates of 4- and 6-year OS post-PRDR was 60% and 40%, respectively, with an estimated median survival of 64 months (95% CI: 8-120 months) (Figure 3). Four out of the eight lesions that received PRDR were still controlled at the last follow-up. Estimated 2- and 4-year PFS post-PRDR was 53.6% and 35.7%, respectively, with a median estimated PFS of 34 months (95% CI: 11-57 months). The median post-PRDR PFS for patients with intracranial and spinal sites were 12 and 34 months, respectively (p=0.014).
PRDR toxicity. None of the eight sites that received PRDR re-irradiation developed any evidence of radionecrosis based on review of serial MRI studies. Figure 1B and C and Figure 2B show examples of MRI findings demonstrating lack of radiation necrosis. The only neurological toxicity noted was mild lumbo-sacral radiculopathy in patient #1, which developed about two years after PRDR re-irradiation. She received 50 Gy to the lumbo-sacral spine at the time of primary radiation and then received an additional 46 Gy PRDR for locally recurrent disease about 13 years after the first course. This patient was managed with six weeks of hyperbaric oxygen therapy with stabilization of deficits. The patient continues to have stable disease locally more than five years since PRDR with grade IV-V power in various muscle groups of bilateral lower extremities. Because of the retrospective nature of this study, formal neurocognitive outcomes were not performed for the patients with intracranial disease.
Kaplan Meier curves for OS (A) and PFS (B) post-PRDR.
Radiotherapy (RT) portals for primary treatment and at the time of pulsed reduced dose rate (PRDR) treatment.
Radiotherapy (RT) doses for primary treatment and at the time of pulsed reduced dose rate (PRDR) treatment.
Discussion
Recurrent ependymoma is a difficult local problem and often the ultimate cause of mortality. The usual treatment algorithm involves repeated attempts to surgically debulking, whenever feasible. This can be technically challenging due to scarring from prior surgery and radiation, concern of neural tissue toxicity or unresectable disease. RT has demonstrated efficacy in improving local control when used as an adjunct to surgery for primary diagnosis. Use of radiation in the recurrent setting is defined by the treatment received at the time of the primary diagnosis, the interval from initial radiation treatment and the patient's performance status.
Several series have described the use of stereotactic radiosurgery (SRS) for re-irradiation in patients with ependymoma having small-volume recurrent disease. Fractionated RT has also been used for re-irradiation when the volume of gross disease is not amenable to SRS. This is usually done with a stereotactic technique as fractionated stereotactic radiotherapy (FSRT) that allows for delivery of highly conformal standard fractionation RT with very tight planning margins. Table III summarizes literature regarding SRS and fractionated RT relative to treatment parameters and outcomes. The reported local control rates with SRS in these series ranges from 33%-100%. In the largest series of use of re-irradiation SRS in patients with ependymoma, Kano et al. at the University of Pittsburgh report of a 3-year PFS after SRS of 46%, with a median post-SRS PFS of 29.5 months (18). In a separate report of 21 pediatric patients (32 tumors), 3-year PFS was 42% with a median PFS of 26 months (10). Stauder et al. reported similar results at the Mayo clinic, Rochester, with use of Gamma Knife-based SRS in 26 patients (49 tumors) with recurrent intracranial ependymomas, although the median time to progression was 15 months (6). Merchant et al., at the St. Jude's Hospital, reported results of fractionated re-irradiation in 32 patients including 13 patients treated with FSRT and cranio-spinal irradiation (CSI) (7). Tumors in ten out of 13 patients treated with FSRT were controlled at a median follow-up of 30 months. Interestingly, PFS from the second course of radiation was longer than that from the first course in half of the patients. Three out of the 19 patients who received craniospinal irradiation (CSI) had some component of local progression, while five had a component of metastatic spread. Bouffet et al., at the Princess Margaret Hospital, Toronto, reported their experience using a re-irradiation approach similar to the one reported in the St. Jude's study (11). They noted that the survival outcomes of patients who received a full dose of re-irradiation was significantly superior to patients for whom re-irradiation was not offered and also for the patients in whom the first course of radiation was offered only when they developed recurrence. In our series of patients with large volume recurrent ependymoma, 2- and 4-year PFS post-PRDR was 53.6% and 35.7%, respectively with a median Kaplan Meier estimated PFS of 34 months. These results compare favorably with reported outcomes of highly focal re-irradiation series. Liu et al. treated six patients with recurrent ependymomas using hypofractionated RT to a dose of 24-30 Gy in three fractions of 8-10 Gy (9). At a median follow-up of 28 months, all six patients were alive without evidence of disease.
Although series using SRS have demonstrated high local control rates, recurrences along the adjacent regions of neuraxis are common. About 36% of the patients treated at the University of Pittsburgh developed neuraxis dissemination post-SRS (18). The cause of death was nearly equally distributed between neuraxis dissemination and local progression. Nearly 50% of the pediatric patients in a separate report also had distant intracranial or spinal recurrences (10). Three-quarter of the patient deaths were due to out-of-field neuraxis tumor relapse, with the remaining one-quarter dying from local progression. About one-quarter of the patients in the Mayo clinic series had distant tumor progression while 38% had a local or marginal tumor recurrence (6). When slightly larger volumes were re-irradiated with FSRT in the St. Jude Hospital series, three out of 13 patients (23%) had disseminated metastatic disease, with local control not assessable (7). Based on the trends seen in the pattern of recurrences between all these series, there may be a role for using wide-field RT techniques to prevent marginal recurrences and thereby prevent development of distant neuraxis disease. Our experience demonstrates that wide-field irradiation using the PRDR technique for large-volume recurrences may be feasible and can be considered as a treatment option.
Re-irradiation of neural tissues, however, is associated with a higher risk of neural toxicity. Reported rates of radiation necrosis with re-irradiation using fractionated RT for brain tumors in general range from 5-10% (19, 20). Stafford et al. noted evidence of radiation necrosis in two out of 12 (16.7%) patients treated with SRS (21). In the SRS series reported by Stauder et al., pathologically confirmed radiation necrosis was noted in about 10% of patients (6). In the Pittsburgh experience, 4.7% of the pediatric ependymomas and about 8% of all patients developed symptomatic adverse effects (10, 18). Using hypofractionated re-irradiation, 50% of patients in the University of Colorado series had radiographic radiation necrosis, although none of the patients required steroids or any other aggressive treatment (9). In the St. Jude experience, 83% of patients that were treated with re-irradiation SRS developed evidence of radiation necrosis on imaging or in pathological specimens (7). However, with conventionally fractionated radiotherapy at St. Jude, none of the 13 treated with FSRT and one out of 19 patients treated with focal RT plus CSI (99 Gy total dose) developed radiation necrosis. One case had transient myelopathy (54 Gy to cord). Even in the Toronto experience, only one out of 19 patients developed radiation necrosis (11).
Literature review of re-irradiation dose, volumes and outcome with stereotactic radiosurgery (SRS)/hypofractionated and conventionally fractionated radiotherapy (RT) in patients with recurrent ependymoma.
None of our patients developed radiation necrosis despite using wide-field RT. The largest dimensions of spinal fields planned with 2D-Techniques in our series were 26 cm, 29 cm and 34 cm with a mean portal area of 348 cm2. For patients planned with 3D-conformal/FSRT techniques, the median volume encompassed by the 100% isodose line was 82.9 cm3, while that encompassed by the 50% isodose line was 882 cm3. Even in our published experience of PRDR re-irradiation in patients with glioma, the mean treatment volume was 403.5 cm3 (16). For patients with recurrent breast cancer, the median volume encompassed by the 50% isodose line was 2,084 cm3 (15). Only one grade III and one grade IV late skin toxicity was noted from chest wall irradiations. Four out of 15 autopsies of patients with glioma showed evidence of microscopic necrosis, although the disease process itself could have been a contributing factor. Delivery of RT at reduced dose rates allows sub-lethal damage repair in the normal tissue, thereby improving radiation tolerance. This dose rate effect has been noted to be most prominent between 0.01 and 1 Gy/min (12). In addition, by pulsing and dividing each fraction into multiple small fractions of 0.2 Gy each, there is an added component of LD-HRS that can potentiate radiation-induced biological tumor tissue damage (14). The combined effect of these two phenomenoa could potentially improve the therapeutic index, which may allow use of wide-field RT. Based on our accumulated experience of re-irradiation with PRDR, the radiation dose currently employed for patients with brain tumors is 50-54 Gy in 2 Gy fractions. The technique for delivering PRDR has been described before and is easily adaptable to any radiation oncology facility with a conventional linear accelerator (13, 15-17).
Conclusion
Re-irradiation with PRDR is a feasible option for patients with large-volume recurrent ependymoma. In view of the rarity of this disease, clinical studies at multi-institutional level may need to be performed to assess if use of wide-field re-irradiation could potentially reduce marginal recurrences and improve outcomes. In addition, there may also be an opportunity to evaluate combined PRDR-based wide-field irradiation with lower-dose SRS boost to gross residual disease to limit the incidence of radiation necrosis otherwise seen with high-dose SRS alone.
Acknowledgements
Supported in part by Collaborative Ependymoma Research Network (CERN)
Footnotes
-
Abstract presented at the 16th Annual Scientific Meeting and Education Day of the Society for Neuro-Oncology, November 17-20, 2011, CA, U.S.A.
- Received April 11, 2013.
- Revision received May 4, 2013.
- Accepted May 8, 2013.
- Copyright© 2013 International Institute of Anticancer Research (Dr. John G. Delinassios), All rights reserved
