- Email: [email protected]
Correlation of Acute and Late Brainstem Toxicities With Dose-Volume Data for Pediatric Patients With Posterior Fossa Malignancies
Accepted Manuscript Correlation of acute and late brainstem toxicities with dose-volume data for pediatric patients with posterior fossa malignancies Ronica H. Nanda, MD, Rohit G. Ganju, MD, Edward Schreibmann, PhD, Zhengjia Chen, PhD, Chao Zhang, PhD, Naresh Jegadeesh, MD, Richard Cassidy, MD, Claudia Deng, BA, Bree Eaton, MD, Natia Esiashvili, MD PII:
S0360-3016(17)30477-7
DOI:
10.1016/j.ijrobp.2017.02.092
Reference:
ROB 24125
To appear in:
International Journal of Radiation Oncology • Biology • Physics
Received Date: 9 September 2016 Revised Date:
14 February 2017
Accepted Date: 21 February 2017
Please cite this article as: Nanda RH, Ganju RG, Schreibmann E, Chen Z, Zhang C, Jegadeesh N, Cassidy R, Deng C, Eaton B, Esiashvili N, Correlation of acute and late brainstem toxicities with dosevolume data for pediatric patients with posterior fossa malignancies, International Journal of Radiation Oncology • Biology • Physics (2017), doi: 10.1016/j.ijrobp.2017.02.092. 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
Correlation of acute and late brainstem toxicities with dose-volume data for pediatric patients with posterior fossa malignancies Short time: Brainstem necrosis from posterior fossa radiation
RI PT
Ronica H. Nanda, MD1; Rohit G. Ganju, MD1; Edward Schreibmann, PhD1; Zhengjia Chen, PhD2; Chao Zhang, PhD2; Naresh Jegadeesh, MD1; Richard Cassidy, MD1; Claudia Deng, BA1; Bree Eaton, MD1; Natia Esiashvili, MD1 1
Department of Radiation Oncology, Winship Cancer Institute, Emory University College of Medicine; Department of Biostatistics and Bioinformatics Shared Resource, Winship Cancer Institute, Emory University Rollins School of Public Health
SC
2
M AN U
Corresponding author: Ronica H. Nanda, MD [email protected] 352.792.5218
AC C
EP
TE D
The authors have no conflicts of interest to disclose.
ACCEPTED MANUSCRIPT
Summary
AC C
EP
TE D
M AN U
SC
RI PT
The rates of brainstem toxicity and dosimetric factors for children treated with intensity-modulated radiation therapy for posterior fossa CNS malignancies remain unclear. Our series demonstrates a very low rate of severe brainstem toxicity in these patients, with severe toxicities limited to children irradiated to large volumes of the posterior fossa. Limiting the volume of the brainstem and posterior fossa irradiated may limit significant brainstem toxicities.
ACCEPTED MANUSCRIPT
Abstract: Purpose: Radiation-induced brainstem toxicity after treatment for pediatric posterior fossa (PF) malignancies is incompletely understood, especially in the era of intensity-modulated radiation therapy (IMRT). The rates of and predictive factors for brainstem toxicity after photon radiation therapy (RT) for PF tumors were examined.
SC
RI PT
Materials/Methods: After IRB approval, 60 pediatric patients treated at our institution for non-metastatic infratentorial ependymoma and medulloblastoma with IMRT were included in this analysis. Dosimetric variables, including mean and maximum dose to the brainstem, dose to 10-90% of the brainstem (in 10% increments), and the volume of brainstem receiving 40, 45, 50, and 55 Gy were recorded for each patient. Acute (onset within 3 months) and late (beyond 3 months of RT completion) RT-induced brainstem toxicities with clinical and radiographic correlates were scored using CTCAE version 4.0.
M AN U
Results: Patients from ages 1.4-21.8 years were treated using IMRT or VMAT postoperatively to the posterior fossa or tumor bed. At a median clinical follow up time of 2.8 years, fourteen patients developed symptomatic brainstem toxicity (crude incidence 23.3%). There was no correlation between the dosimetric variables examined and brainstem toxicity. Vascular injury or ischemia strongly trended toward predicting brainstem toxicity (P=0.054). Patients with grade 3-5 brainstem toxicity were treated to significant volumes of the posterior fossa.
AC C
EP
TE D
Conclusion: This series demonstrates a low but not negligible risk of brainstem radiation necrosis for pediatric patients with posterior fossa malignancies treated with intensity-modulated radiation therapy. No specific dose-volume correlations were identified but modern treatment volumes may help limit severe toxicity. Additional work investigating inherent biologic sensitivity may also provide further insight into this clinical problem.
ACCEPTED MANUSCRIPT
Introduction:
M AN U
SC
RI PT
Posterior fossa tumors in pediatric patients, including ependymomas and medulloblastomas, were conventionally treated with 3D-conformal techniques, which resulted in significant long-term toxicity. The adoption of intensity modulated radiation therapy (IMRT) has allowed for substantial dose sparing of the cochlea, optic apparatus, and brainstem. There was initial concern that the rapid dose fall-off would potentially compromise disease control outcomes by reducing dose to tumor volumes, and that the increased treatment time may increase repopulation, but a number of studies have shown that disease control with IMRT is equivalent to historical controls [1]; [2]; [3]. Changes in target volume recommendations have further limited adverse effects without compromising outcomes [4]; [5]. Despite these advances, dose constraints to the brainstem are not yet defined for this pediatric population. These tumors have a tendency to adhere to or invade the brainstem, which not only can limit resectability, but also requires that clinical target volumes (CTVs) include significant volumes of the brainstem. However, published cooperative group studies have not specified strict dose constraints for the brainstem in these patients and institutional practices have yet to be validated.
TE D
The rates of clinically significant brainstem toxicity using IMRT or volumetric arc therapy (VMAT) for patients with posterior fossa malignancies are not well-described in the literature. The rates of severe brainstem toxicity range from 2.5-18.5% with 3D-conformal therapy, but definitions of brainstem toxicity in these series are heterogeneous[6-8]. These reports suggest some risk factors for brainstem toxicity, including higher doses to larger volumes of infratentorial brain, postoperative ischemia, and extent of tumor [6], [4]. Multiple proton series have evaluated this risk as well. The rates range from 3.8% [9] in a series that was inclusive of both supratentorial and infratentorial tumors, to 16% [10] in AT/RT patients. Although the proton literature has started to specify certain dose-volume constraints, given differences in RBE and LET between proton therapy and photon therapy [11], one cannot directly extrapolate between the two modalities.
EP
We evaluated pediatric patients treated with intensity modulated radiation therapy at our institution for posterior fossa malignancies to determine the rates of and dose-volume predictors of brainstem toxicity.
AC C
Methods and Materials: Institutional review board approval was obtained for collection of patient data. Institutional patient databases were queried for patients ≤21 years old with primary CNS malignancies (N=170). Patients with medulloblastoma and ependymoma of the posterior fossa who were treated post-operatively after gross total or subtotal resection with IMRT or volumetric arc therapy (VMAT) were included (N=67) (there were no patients with biopsy only). Seven patients with missing dosimetric data (N=4) or insufficient follow up (less than one year, N=3) were excluded, allowing 60 patients for the final analysis. The gross tumor volume (GTV) was defined using rigid CT simulation fused with preoperative and postoperative high resolution MRI (T1-post contrast and when needed T2-weighted images were used). The GTV was defined as the tumor bed plus any residual disease if present. For medulloblastomas treated in the early period of follow up clinical target volume (CTV)1 was the entire craniospinal axis. CTV2 was the entire posterior fossa, including the tumor bed. The posterior fossa
ACCEPTED MANUSCRIPT
RI PT
CTV was anatomically defined as the space between the tentorium cerebelli and foramen magnum, bounded posteriorly by the occipital bone and laterally by the temporal bones. If there was disease into the cervical canal there would be a cone down to respect tolerance to the cord which would necessitate a third CTV (CTV3). Later, when a posterior fossa boost was abandoned in favor of tumor bed boost, CTV2 would be the tumor bed and residual disease plus 5-10 mm margin and CTV 3, if needed, would be the GTV plus a more limited margin respecting cord tolerance (0-3 mm).
SC
For ependymomas, CTV1 was defined as the GTV plus a 5-10 mm margin (depending on period when patient was treated), and CTV2 would be the GTV with a limited margin (0-3). Most ependymomas were treated in a two-phase plan due to the higher dose delivered in order to respect brainstem and spinal cord tolerance. PTV margin was 3-5 mm depending on whether IGRT (image-guided radiation therapy) was available.
M AN U
Medulloblastomas were treated with craniospinal irradiation to 18 to 36 Gy, depending on riskstratification and protocol enrollment, followed by a posterior fossa boost or tumor bed boost to 54-55.8 Gy (CTV2 to 50.4-54 Gy; CTV3, if needed, to 54-55.8 Gy). All medulloblastoma patients received concurrent and adjuvant chemotherapy. Ependymomas were treated with adjuvant involved field radiation therapy alone to a total dose of 55.8-59.4 Gy (CTV1: 50.4-54 Gy; CTV2: 55.8-59.4 Gy). Patients younger than age 3 were treated to 54-55.8 Gy; the remaining patients were treated to 59.4 Gy. Seven of sixteen ependymoma patients received chemotherapy either before radiation therapy. Constraints used for treatment planning are listed in Table 1.
AC C
EP
TE D
Brainstem volumes were determined by rigid registration with high-resolution thin-cut planning MRI sequences, typically T1 weighted with contrast. Brainstems were contoured from the superior cerebral peduncles through the medulla to the level of the foramen magnum. Dose-volume histograms were used to determine the following values: maximum and mean doses to the brainstem, as well as the doses to 1090% of the brainstem volume in 10% increments (i.e., D10, D20, D30, D40, D50, D60, D70, D80, and D90. In addition, the percent volumes of brainstem receiving more than 40 Gy (V40), 45 Gy (V45), 50 Gy (V50), and 55 Gy (V55) were determined. Additional data collected included extended or permanent postoperative shunting, posterior fossa syndrome, and postoperative vascular or cranial nerve injury. Presence of radiographic changes consistent with radiation necrosis based on institutional expert radiology interpretation was also determined, defined as T1 pre- or post-contrast enhancement within the treatment field. Acute brainstem toxicity was defined as new or progressive neurologic changes involving cranial nerves V-VII or IX-XII, or motor weakness or ataxia occurring during radiation or within 3 months of treatment completion. Late brainstem toxicity included these same symptoms with onset more than 3 months after treatment completion, or acute symptoms worsening beyond 3 months after radiation. Symptom onset after reirradiation, if given, was not counted toward the incidence of radiation necrosis. Any imaging findings suggestive of radiation necrosis, with or without corresponding clinical findings, were also documented. Toxicities were scored using NCI Common Terminology Criteria for Adverse Events version 4.0 [12]. The parametric p-value was calculated by t-test for numerical covariates and chi-square test for categorical variables. The non-parametric p-value is calculated by the Kruskal-Wallis test for numerical covariables and Fisher’s exact test for categorical covariates.
ACCEPTED MANUSCRIPT
SC
RI PT
Results: A total of 60 patients were included in the final analysis, 36 male patients and 24 female patients. The mean age at diagnosis was 8.2 years (median age 6.7 years; range 1.4-21.8 years) and the median follow up was 33.8 months (2.8 years; range 3.0-120.2 months). Median follow up for imaging studies (MRI brain with contrast) was 24.1 months (2 years; range 3.0-96.2 months). Sixteen patients had ependymoma and 44 patients had medulloblastoma. All patients were treated between 2005 and 2015. The last date of follow up was 4/28/2016 for the series. The majority of patients had a gross total or near total resection; however, 17 patients had a subtotal resection based on post-operative MRI findings. Only one patient had postoperative MRI changes in the brainstem suggestive of ischemia or iatrogenic injury, and three patients had evidence of posterior fossa syndrome. Four patients required extended or permanent shunting even after resection of the primary mass. Additional patient characteristics are listed in Supplemental Table 1.
M AN U
Mean and median brainstem D10-D90 and V40-V55 values (as detailed above) are in table 2. In this series, 14 patients (23.3%) experienced brainstem toxicity attributable to radiation. Table 3 lists the details of these patients. There are three grade 1 cases, nine Grade 2 cases, and two grade 3 or higher cases. The cumulative incidences of toxicity by grade are in table 4. The specific grade 1 and 2 toxicities included ataxia (n=3), dysmetria (n=2), and dysarthria (n=4). The patients with grade 2 toxicity were treated with a dexamethasone taper with improvement in symptoms. Hyperbaric oxygen or bevacizumab were not indicated or needed for these cases.
EP
TE D
Three of the 14 cases were in the acute phase (grade 1: two cases; grade 2: one case). None of these worsened beyond the acute phase to be subsequently classified as a late toxicity. The patient with acute grade 2 toxicity did have radiographic changes of radiation necrosis, but these findings occurred after resolution of symptoms (prior imaging did not show any radiographic changes), and she did not develop additional symptoms after developing radiographic necrosis. The patients with acute grade 1 toxicity did not have radiographic changes identified during follow up. The remaining 11 toxicities were categorized as late toxicities (grade 2: nine cases: grade 3: one case; grade 5: one case). These patients with late grade 2 effects did not have radiographic correlates.
AC C
The grade three and grade five toxicity cases are described as follows. Both cases were in patients treated for high risk medulloblastoma after subtotal resection with high-dose craniospinal irradiation (36 Gy); one patient received conformal tumor bed boost with a 10 mm CTV margin (and 3 mm PTV margin) while the other patient received whole posterior fossa boost. Both patients developed radiographic changes consistent with radiation necrosis. The patient with grade 3 toxicity developed a right facial palsy and numbness 118 days after radiation therapy that did not improve on a prolonged steroid taper. Surgical intervention (gold weight in eyelid) was required to correct corneal abrasions secondary to cranial nerve deficits. He has been followed closely since that time and has not developed evidence of recurrence over a 29-month period of follow up. The second patient with severe brainstem toxicity, which developed 378 days after radiation therapy, was scored as grade 5 toxicity as she developed PEG-tube dependence and respiratory failure, requiring permanent tracheostomy. She suffered a prolonged course with clinical deterioration before succumbing to sepsis and multiorgan failure. She did have a biopsy consistent with
ACCEPTED MANUSCRIPT
radiation necrosis and not recurrence prior to her clinical deterioration. Hyperbaric oxygen was not offered in either case, and bevacizumab was not routinely used for radiation necrosis in the time period of these patients’ symptoms. She had no evidence of disease recurrence or progression at time of death. Figure 1 shows the area of brainstem necrosis in the axial, sagittal, and coronal planes in the patient with grade 3 toxicity with overlying isodose lines. The second patient had similar radiographic findings.
SC
RI PT
There were two additional patients, not detailed above, with radiographic alterations in the brainstem on surveillance MRIs that were consistent with radiation change. The first patient with radiographic changes in the brainstem remained asymptomatic through the period of follow up (i.e., no acute or late toxicities). The second patient developed symptomatic radiation necrosis (left hemiparesis, ataxia) but only after reirradiation for tumor bed recurrence and was therefore not coded as a toxicity in this dataset. Therefore, a total of 5 patients in this series had radiographic changes consistent with radiation necrosis (grade 0: 2/5, 3.3% of cohort; grade 1: 0/5; grade 2: 1/5, 1.7%; grade 3: 1/5, 1.7%;; grade 4: 0/5; grade 5: 1/5, 1.7%;).
M AN U
There was no correlation with acute brainstem toxicity and any of the clinical covariates examined. None of the dosimetric variables predicted for acute toxicity. Univariate logistic regression analysis did not yield any significant predictors for acute toxicity (supplemental table A2).
Discussion:
TE D
Radiographic necrosis correlated with symptomatic brainstem toxicity (p=0.029). Vascular or cranial nerve injury from surgery was marginally associated with incidence of late toxicity (p=0.054). However, there were no dosimetry endpoints that predicted for late brainstem toxicity (including necrosis) in this series (supplemental table A3). None of the patients with late toxicities had experienced acute brainstem toxicities. Figure 2 demonstrates the cumulative dose-volume histograms of all patients in the series, with the DVH of patients with grade 3 and 5 radiation necrosis highlighted in red. The patients with grade 1-2 toxicity are highlighted in blue.
AC C
EP
To our knowledge, no prior studies have correlated dose-volume data with risks of brainstem toxicity in patients treated with IMRT or VMAT. Constraints from the proton beam therapy cannot be directly extrapolated to the photon setting due to differences in treatment planning, dose distribution, and inherent differences in Bragg peak, RBE and LET, which are also not yet fully understood [11]. Only 10% of children (up to 33% for ependymoma/medulloblastoma patients) requiring radiation therapy in the United States receive radiation therapy at a proton therapy center each year [13], which may be due to lack of resources or ability to travel to proton therapy centers. Therefore, it is still critical to determine risks and constraints for patients being treated with x-ray-based therapy. Our study demonstrated that the crude risk of any brainstem toxicity was 23.3% but that the 3-year rate of severe toxicity was 3.3%, consistent with prior studies. Radiographic radiation necrosis and postoperative vascular or ischemic injury suggested a higher risk for symptomatic brainstem toxicity, also consistent with the literature [14]. We did not detect any correlation with age, gender, GTV size, presence of shunting, or development of posterior fossa syndrome.
ACCEPTED MANUSCRIPT
RI PT
The rates of any symptomatic brainstem toxicity in patients treated with 3D-conformal therapy range from 2.5% to 18.5%, depending on the definitions [8]; [15], [6], and from 3.8 to 16% in the proton literature. Grade 3 or higher toxicity, however, appears to be less than 3-4% in both settings [9], [11]. At the very least, the three modalities, now based on our data, appear to be associated with equally low rates of severe brainstem toxicity despite concerns in each setting that the variable physical properties are more or less risky to normal organs at risk. However, the rates of grade 2 toxicity may be slightly higher in our series than in the proton literature, although these comparisons should be made with caution.
M AN U
SC
We were unable to correlate risk of late toxicity or radiographic brainstem necrosis with any dosimetric variables. Our mean D10 and mean D50 were lower than the D10 and D50 proposed as safe in the series by Indelicato, et al [9]. However, in the two patients with grade 3 or higher brainstem necrosis, the D10 and D50 exceeded these constraints. Our mean V40, V45, V50, V55, and V60 also exceeded the recommended limits in the aforementioned proton series. It is possible that the higher V40-V55, representing higher volumes of the brainstem receiving moderate to high doses, are responsible for the higher rate of grade 2 toxicity in our series compared with the Indelicato series. Other data demonstrates that intensity-modulated radiation therapy is associated with higher volumes of brainstem receiving higher doses for similar tumors, compared with conventional and intensity-modulated proton therapy [16]. However, it is unclear however how the dosimetric variables in the proton setting translate into patients treated with conformal photon-based techniques as no clear relationship with RBE, LET, and brainstem injury has been identified [11].
EP
TE D
Five patients had radiographic changes consistent with radiation necrosis in this series (8.3%) and three additional patients had (asymptomatic) T2-weighted sequence radiation induced changes (data not shown), for an 13.3% rate of radiation related imaging changes, compared with 17% rate of similar changes in a series of posterior fossa patients treated with IMRT (and compared with 43% in patients treated with protons in that same series) [17]. The rate of symptomatic changes was, however, similar in both series (5% in ours, 8.5% in the Gunther, et al series) amongst IMRT patients. Using standard definitions of imaging changes in terms of relationship of radiographic changes to the treatment fields, the MRI sequences used to the define them, and the period of time after radiation therapy for which these findings should be investigated, will further serve to develop clinically relevant guidelines.
AC C
Given that these patients were treated at a single institution by only two physician providers, the data is fairly homogeneous in terms of registration software and technique, contouring of normal structures and tumor volumes, treatment planning, and delivery. We were able to have full and complete dosimetry data for all included patients included in this study which allowed for a robust dosimetric analysis. Due to close collaboration with our colleagues at the children’s hospital, we were able to have careful and frequent documentation of clinical and radiographic events in follow up. With rare exception, patients had imaging and clinical follow up done at the primary center, allowing for a consistent basis for comparison of symptoms and imaging findngs. Our data is of course limited inherently by retrospective collection of data and relatively small cohort size. In addition, the intervals between imaging studies were in some cases prolonged due to inability for patients to come as instructed for follow up, as we are a tertiary referral center with patients occasionally traveling significant distances. Therefore, it is possible that some symptomatic toxicities and radiographic changes were missed due to occasionally infrequent follow ups. Furthermore, including both medulloblastoma and ependymoma patients lends heterogeneity, as these
ACCEPTED MANUSCRIPT
patients are treated to different doses, and often have variable extents of disease and biologic behavior. Chemotherapy was sometimes given prior to radiation; it is unclear what interaction chemotherapy may have had with radiation even when not delivered concurrently. There is also relatively limited follow up and additional late toxicities affecting the brainstem may be yet to be seen. However, in our series, almost all events manifested within the first two years after treatment.
EP
Conclusions:
TE D
M AN U
SC
RI PT
We propose that certain measures be taken in order to reduce the risk of radiation necrosis. Avoiding excess volume of uninvolved brainstem within the CTV can reduce the volume of brainstem receiving high doses, and careful attention should be given to avoiding “hot spots” or heterogeneity within the brainstem. The shift in practice from posterior fossa boost to conformal tumor bed boost likely has resulted in lower radiation necrosis rates than previously observed. Approximately three-quarters of patients observed to develop radiation necrosis in the Massachusetts General Hospital series received a posterior fossa boost [11], supporting the hypothesis that high volumes receiving high doses of radiation predispose patients to radiation necrosis. The patients with grade 3 and 5 brainstem toxicity had either generous CTV margins or the entire posterior fossa treated. The use of 5 mm margins, as opposed to 10 mm margins, as are now employed in current practice may help achieve the goal of reducing the volume of brainstem irradiated. Although severe toxicities were not seen in patients with ependymomas in our series, they have been otherwise reported. It is unclear whether dose escalation for ependymomas beyond 54-55.8 improves local control. There is some data suggesting improved outcomes with doses beyond 50 Gy, but not necessarily to 59.4 Gy [18]. Additional data is needed to determine the optimal dose for local control while limiting risk of toxicity. Despite overall low rates, radiation-induced damage to the brainstem can be significantly morbid, requiring expensive interventions when at all available, impairing quality of life, and posing a threat to life in rare but tragic cases. Finally, it is possible that there are biologic factors that we have not yet considered that predispose certain patients to toxicity despite the most meticulous, conservative, treatment planning. Studying tissue and blood samples from patients enrolled on prospective national studies for molecular markers may provide insight into these toxicity profiles.
AC C
Our data demonstrates that the risk of clinically significant brainstem injury is low but not negligible using conformal, inverse-planned techniques. Prospective collection of dose-volume data in a large trial may help define more precise constraints that can be employed for pediatric patients with posterior fossa malignancies to further minimize risk of radiation necrosis while maintaining tumor control. Additional factors, including molecular markers to identify patients with inherent radiosensitivity, may also be useful in caring for this population.
Figure Captions: Figure 1: Isodose lines from this patient’s IMRT plan are overlaid the MRI T1-weighted post-contrast image illustrating the area of radiation necrosis, in the axial, sagittal, and coronal planes.
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
Figure 2: The cumulative brainstem dose-volume histograms are displayed on this graph. The DVHs for patients with grade 3 and 5 brainstem toxicity are highlighted in red. The DVHs for patients with grade 12 acute and late toxicity are in blue. The brainstem DVHs for the remaining patients without toxicities are represented in gray.
ACCEPTED MANUSCRIPT
References
AC C
EP
TE D
M AN U
SC
RI PT
1. Schroeder T. M., Chintagumpala M., Okcu M. F., et al. Intensity-modulated radiation therapy in childhood ependymoma. Int J Radiat Oncol Biol Phys 2008;71:987-993. 2. Polkinghorn W. R., Dunkel I. J., Souweidane M. M., et al. Disease control and ototoxicity using intensity-modulated radiation therapy tumor-bed boost for medulloblastoma. Int J Radiat Oncol Biol Phys 2011;81:e15-20. 3. Paulino A. C., Mazloom A., Teh B. S., et al. Local control after craniospinal irradiation, intensitymodulated radiotherapy boost, and chemotherapy in childhood medulloblastoma. Cancer 2011;117:635-641. 4. Merchant T. E., Li C., Xiong X., et al. Conformal radiotherapy after surgery for paediatric ependymoma: A prospective study. Lancet Oncol 2009;10:258-266. 5. Ibrahim N. Y., Abdel Aal H. H., Abdel Kader M. S., et al. Reducing late effects of radiotherapy in average risk medulloblastoma. Chin Clin Oncol 2014;3:4. 6. Murphy E. S., Merchant T. E., Wu S., et al. Necrosis after craniospinal irradiation: Results from a prospective series of children with central nervous system embryonal tumors. Int J Radiat Oncol Biol Phys 2012;83:e655-660. 7. Mayo C., Yorke E., Merchant T. E. Radiation associated brainstem injury. Int J Radiat Oncol Biol Phys 2010;76:S36-41. 8. Shaw E., Arusell R., Scheithauer B., et al. Prospective randomized trial of low- versus high-dose radiation therapy in adults with supratentorial low-grade glioma: Initial report of a north central cancer treatment group/radiation therapy oncology group/eastern cooperative oncology group study. J Clin Oncol 2002;20:2267-2276. 9. Indelicato D. J., Flampouri S., Rotondo R. L., et al. Incidence and dosimetric parameters of pediatric brainstem toxicity following proton therapy. Acta Oncol 2014;53:1298-1304. 10. McGovern S. L., Okcu M. F., Munsell M. F., et al. Outcomes and acute toxicities of proton therapy for pediatric atypical teratoid/rhabdoid tumor of the central nervous system. Int J Radiat Oncol Biol Phys 2014;90:1143-1152. 11. Giantsoudi D., Sethi R. V., Yeap B. Y., et al. Incidence of cns injury for a cohort of 111 patients treated with proton therapy for medulloblastoma: Let and rbe associations for areas of injury. Int J Radiat Oncol Biol Phys 2015. 12. Institute National Cancer. Common terminology criteria for adverse events v4.0. NCI, NIH, DHHS 13. Chang AL, Yock TI, Mahajan A, et al. Pediatric proton therapy: Patterns of care across the united states. Int J Particle Ther 2014;1. 14. Merchant T. E., Chitti R. M., Li C., et al. Factors associated with neurological recovery of brainstem function following postoperative conformal radiation therapy for infratentorial ependymoma. Int J Radiat Oncol Biol Phys 2010;76:496-503. 15. Spreafico F., Gandola L., Marchiano A., et al. Brain magnetic resonance imaging after high-dose chemotherapy and radiotherapy for childhood brain tumors. Int J Radiat Oncol Biol Phys 2008;70:1011-1019. 16. MacDonald S. M., Safai S., Trofimov A., et al. Proton radiotherapy for childhood ependymoma: Initial clinical outcomes and dose comparisons. Int J Radiat Oncol Biol Phys 2008;71:979-986. 17. Gunther J. R., Sato M., Chintagumpala M., et al. Imaging changes in pediatric intracranial ependymoma patients treated with proton beam radiation therapy compared to intensity modulated radiation therapy. Int J Radiat Oncol Biol Phys 2015;93:54-63. 18. Taylor R. E. Review of radiotherapy dose and volume for intracranial ependymoma. Pediatr Blood Cancer 2004;42:457-460.
ACCEPTED MANUSCRIPT
Table 1: Treatment planning constraints for IMRT Volume
Dose Limit
Brainstem Globe Left Globe Right Lens Left Lens Right Optic Chiasm Optic Nerve Left Optic Nerve Right Cochlea Left Cochlea Right Spinal Cord C1-2 Spinal Cord Hypothalamic-pituitary axis
50% 50% 50% 100% 100% 50% 50% 50% 50% 50% 50% 50% 50%
≤ 60 Gy ≤ 10Gy ≤ 10 Gy ≤ 5 Gy ≤ 5 Gy ≤ 54 Gy ≤ 54 Gy ≤ 54 Gy ≤ 35 Gy ≤ 35 Gy ≤ 50 Gy ≤ 54 Gy ≤ 35 Gy
AC C
EP
TE D
M AN U
SC
RI PT
Organ At Risk
ACCEPTED MANUSCRIPT
RI PT
Table 2. Mean and median dose-volume data for brainstem organ at risk.
Mean Dose (Gy)
Median Dose (Gy)
D10 D20 D30 D40 D50 D60 D70 D80 D90
55.1 54.7 54.2 53.5 52.6 51.5 50.1 48.2 45.8
56.8 56.6 56.2 55.7 55.5 55.0 54.2 53.2 50.9
Dose to Brainstem (Gy)
Mean Volume (%) Median Volume (%)
V40 V45 V50 V55 Maximum
87.8 85.2 79.1 53.8 56.1 (Gy)
M AN U
SC
Volume of Brainstem (%)
AC C
EP
TE D
100 99 92.4 59.6 57.3 (Gy)
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
Table 3. Acute and late toxicities
ACCEPTED MANUSCRIPT
Time (Months)
Grade 1
6 12 24 36
Estimated cumulative incidence (%) (95% CI) 1.7 (0, 5.1) 3.4 (0, 8.2) 3.4 (0, 8.2) 3.4 (0, 8.2)
Grade 2
6 12 24 36
3.4 (0, 8.2) 8.6 (1.3, 1.6) 12.1 (3.6, 20.5) 14.4 (5.0, 23.9)
Grade 3-5
6 12 24 36
0 (0,0) 0 (0,0) 3.6 (0, 8.6) 3.6 (0, 8.6)
AC C
EP
TE D
M AN U
SC
Event
RI PT
Table 4. Cumulative incidence of toxicity by grade.
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
S0360-3016(17)30477-7
DOI:
10.1016/j.ijrobp.2017.02.092
Reference:
ROB 24125
To appear in:
International Journal of Radiation Oncology • Biology • Physics
Received Date: 9 September 2016 Revised Date:
14 February 2017
Accepted Date: 21 February 2017
Please cite this article as: Nanda RH, Ganju RG, Schreibmann E, Chen Z, Zhang C, Jegadeesh N, Cassidy R, Deng C, Eaton B, Esiashvili N, Correlation of acute and late brainstem toxicities with dosevolume data for pediatric patients with posterior fossa malignancies, International Journal of Radiation Oncology • Biology • Physics (2017), doi: 10.1016/j.ijrobp.2017.02.092. 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
Correlation of acute and late brainstem toxicities with dose-volume data for pediatric patients with posterior fossa malignancies Short time: Brainstem necrosis from posterior fossa radiation
RI PT
Ronica H. Nanda, MD1; Rohit G. Ganju, MD1; Edward Schreibmann, PhD1; Zhengjia Chen, PhD2; Chao Zhang, PhD2; Naresh Jegadeesh, MD1; Richard Cassidy, MD1; Claudia Deng, BA1; Bree Eaton, MD1; Natia Esiashvili, MD1 1
Department of Radiation Oncology, Winship Cancer Institute, Emory University College of Medicine; Department of Biostatistics and Bioinformatics Shared Resource, Winship Cancer Institute, Emory University Rollins School of Public Health
SC
2
M AN U
Corresponding author: Ronica H. Nanda, MD [email protected] 352.792.5218
AC C
EP
TE D
The authors have no conflicts of interest to disclose.
ACCEPTED MANUSCRIPT
Summary
AC C
EP
TE D
M AN U
SC
RI PT
The rates of brainstem toxicity and dosimetric factors for children treated with intensity-modulated radiation therapy for posterior fossa CNS malignancies remain unclear. Our series demonstrates a very low rate of severe brainstem toxicity in these patients, with severe toxicities limited to children irradiated to large volumes of the posterior fossa. Limiting the volume of the brainstem and posterior fossa irradiated may limit significant brainstem toxicities.
ACCEPTED MANUSCRIPT
Abstract: Purpose: Radiation-induced brainstem toxicity after treatment for pediatric posterior fossa (PF) malignancies is incompletely understood, especially in the era of intensity-modulated radiation therapy (IMRT). The rates of and predictive factors for brainstem toxicity after photon radiation therapy (RT) for PF tumors were examined.
SC
RI PT
Materials/Methods: After IRB approval, 60 pediatric patients treated at our institution for non-metastatic infratentorial ependymoma and medulloblastoma with IMRT were included in this analysis. Dosimetric variables, including mean and maximum dose to the brainstem, dose to 10-90% of the brainstem (in 10% increments), and the volume of brainstem receiving 40, 45, 50, and 55 Gy were recorded for each patient. Acute (onset within 3 months) and late (beyond 3 months of RT completion) RT-induced brainstem toxicities with clinical and radiographic correlates were scored using CTCAE version 4.0.
M AN U
Results: Patients from ages 1.4-21.8 years were treated using IMRT or VMAT postoperatively to the posterior fossa or tumor bed. At a median clinical follow up time of 2.8 years, fourteen patients developed symptomatic brainstem toxicity (crude incidence 23.3%). There was no correlation between the dosimetric variables examined and brainstem toxicity. Vascular injury or ischemia strongly trended toward predicting brainstem toxicity (P=0.054). Patients with grade 3-5 brainstem toxicity were treated to significant volumes of the posterior fossa.
AC C
EP
TE D
Conclusion: This series demonstrates a low but not negligible risk of brainstem radiation necrosis for pediatric patients with posterior fossa malignancies treated with intensity-modulated radiation therapy. No specific dose-volume correlations were identified but modern treatment volumes may help limit severe toxicity. Additional work investigating inherent biologic sensitivity may also provide further insight into this clinical problem.
ACCEPTED MANUSCRIPT
Introduction:
M AN U
SC
RI PT
Posterior fossa tumors in pediatric patients, including ependymomas and medulloblastomas, were conventionally treated with 3D-conformal techniques, which resulted in significant long-term toxicity. The adoption of intensity modulated radiation therapy (IMRT) has allowed for substantial dose sparing of the cochlea, optic apparatus, and brainstem. There was initial concern that the rapid dose fall-off would potentially compromise disease control outcomes by reducing dose to tumor volumes, and that the increased treatment time may increase repopulation, but a number of studies have shown that disease control with IMRT is equivalent to historical controls [1]; [2]; [3]. Changes in target volume recommendations have further limited adverse effects without compromising outcomes [4]; [5]. Despite these advances, dose constraints to the brainstem are not yet defined for this pediatric population. These tumors have a tendency to adhere to or invade the brainstem, which not only can limit resectability, but also requires that clinical target volumes (CTVs) include significant volumes of the brainstem. However, published cooperative group studies have not specified strict dose constraints for the brainstem in these patients and institutional practices have yet to be validated.
TE D
The rates of clinically significant brainstem toxicity using IMRT or volumetric arc therapy (VMAT) for patients with posterior fossa malignancies are not well-described in the literature. The rates of severe brainstem toxicity range from 2.5-18.5% with 3D-conformal therapy, but definitions of brainstem toxicity in these series are heterogeneous[6-8]. These reports suggest some risk factors for brainstem toxicity, including higher doses to larger volumes of infratentorial brain, postoperative ischemia, and extent of tumor [6], [4]. Multiple proton series have evaluated this risk as well. The rates range from 3.8% [9] in a series that was inclusive of both supratentorial and infratentorial tumors, to 16% [10] in AT/RT patients. Although the proton literature has started to specify certain dose-volume constraints, given differences in RBE and LET between proton therapy and photon therapy [11], one cannot directly extrapolate between the two modalities.
EP
We evaluated pediatric patients treated with intensity modulated radiation therapy at our institution for posterior fossa malignancies to determine the rates of and dose-volume predictors of brainstem toxicity.
AC C
Methods and Materials: Institutional review board approval was obtained for collection of patient data. Institutional patient databases were queried for patients ≤21 years old with primary CNS malignancies (N=170). Patients with medulloblastoma and ependymoma of the posterior fossa who were treated post-operatively after gross total or subtotal resection with IMRT or volumetric arc therapy (VMAT) were included (N=67) (there were no patients with biopsy only). Seven patients with missing dosimetric data (N=4) or insufficient follow up (less than one year, N=3) were excluded, allowing 60 patients for the final analysis. The gross tumor volume (GTV) was defined using rigid CT simulation fused with preoperative and postoperative high resolution MRI (T1-post contrast and when needed T2-weighted images were used). The GTV was defined as the tumor bed plus any residual disease if present. For medulloblastomas treated in the early period of follow up clinical target volume (CTV)1 was the entire craniospinal axis. CTV2 was the entire posterior fossa, including the tumor bed. The posterior fossa
ACCEPTED MANUSCRIPT
RI PT
CTV was anatomically defined as the space between the tentorium cerebelli and foramen magnum, bounded posteriorly by the occipital bone and laterally by the temporal bones. If there was disease into the cervical canal there would be a cone down to respect tolerance to the cord which would necessitate a third CTV (CTV3). Later, when a posterior fossa boost was abandoned in favor of tumor bed boost, CTV2 would be the tumor bed and residual disease plus 5-10 mm margin and CTV 3, if needed, would be the GTV plus a more limited margin respecting cord tolerance (0-3 mm).
SC
For ependymomas, CTV1 was defined as the GTV plus a 5-10 mm margin (depending on period when patient was treated), and CTV2 would be the GTV with a limited margin (0-3). Most ependymomas were treated in a two-phase plan due to the higher dose delivered in order to respect brainstem and spinal cord tolerance. PTV margin was 3-5 mm depending on whether IGRT (image-guided radiation therapy) was available.
M AN U
Medulloblastomas were treated with craniospinal irradiation to 18 to 36 Gy, depending on riskstratification and protocol enrollment, followed by a posterior fossa boost or tumor bed boost to 54-55.8 Gy (CTV2 to 50.4-54 Gy; CTV3, if needed, to 54-55.8 Gy). All medulloblastoma patients received concurrent and adjuvant chemotherapy. Ependymomas were treated with adjuvant involved field radiation therapy alone to a total dose of 55.8-59.4 Gy (CTV1: 50.4-54 Gy; CTV2: 55.8-59.4 Gy). Patients younger than age 3 were treated to 54-55.8 Gy; the remaining patients were treated to 59.4 Gy. Seven of sixteen ependymoma patients received chemotherapy either before radiation therapy. Constraints used for treatment planning are listed in Table 1.
AC C
EP
TE D
Brainstem volumes were determined by rigid registration with high-resolution thin-cut planning MRI sequences, typically T1 weighted with contrast. Brainstems were contoured from the superior cerebral peduncles through the medulla to the level of the foramen magnum. Dose-volume histograms were used to determine the following values: maximum and mean doses to the brainstem, as well as the doses to 1090% of the brainstem volume in 10% increments (i.e., D10, D20, D30, D40, D50, D60, D70, D80, and D90. In addition, the percent volumes of brainstem receiving more than 40 Gy (V40), 45 Gy (V45), 50 Gy (V50), and 55 Gy (V55) were determined. Additional data collected included extended or permanent postoperative shunting, posterior fossa syndrome, and postoperative vascular or cranial nerve injury. Presence of radiographic changes consistent with radiation necrosis based on institutional expert radiology interpretation was also determined, defined as T1 pre- or post-contrast enhancement within the treatment field. Acute brainstem toxicity was defined as new or progressive neurologic changes involving cranial nerves V-VII or IX-XII, or motor weakness or ataxia occurring during radiation or within 3 months of treatment completion. Late brainstem toxicity included these same symptoms with onset more than 3 months after treatment completion, or acute symptoms worsening beyond 3 months after radiation. Symptom onset after reirradiation, if given, was not counted toward the incidence of radiation necrosis. Any imaging findings suggestive of radiation necrosis, with or without corresponding clinical findings, were also documented. Toxicities were scored using NCI Common Terminology Criteria for Adverse Events version 4.0 [12]. The parametric p-value was calculated by t-test for numerical covariates and chi-square test for categorical variables. The non-parametric p-value is calculated by the Kruskal-Wallis test for numerical covariables and Fisher’s exact test for categorical covariates.
ACCEPTED MANUSCRIPT
SC
RI PT
Results: A total of 60 patients were included in the final analysis, 36 male patients and 24 female patients. The mean age at diagnosis was 8.2 years (median age 6.7 years; range 1.4-21.8 years) and the median follow up was 33.8 months (2.8 years; range 3.0-120.2 months). Median follow up for imaging studies (MRI brain with contrast) was 24.1 months (2 years; range 3.0-96.2 months). Sixteen patients had ependymoma and 44 patients had medulloblastoma. All patients were treated between 2005 and 2015. The last date of follow up was 4/28/2016 for the series. The majority of patients had a gross total or near total resection; however, 17 patients had a subtotal resection based on post-operative MRI findings. Only one patient had postoperative MRI changes in the brainstem suggestive of ischemia or iatrogenic injury, and three patients had evidence of posterior fossa syndrome. Four patients required extended or permanent shunting even after resection of the primary mass. Additional patient characteristics are listed in Supplemental Table 1.
M AN U
Mean and median brainstem D10-D90 and V40-V55 values (as detailed above) are in table 2. In this series, 14 patients (23.3%) experienced brainstem toxicity attributable to radiation. Table 3 lists the details of these patients. There are three grade 1 cases, nine Grade 2 cases, and two grade 3 or higher cases. The cumulative incidences of toxicity by grade are in table 4. The specific grade 1 and 2 toxicities included ataxia (n=3), dysmetria (n=2), and dysarthria (n=4). The patients with grade 2 toxicity were treated with a dexamethasone taper with improvement in symptoms. Hyperbaric oxygen or bevacizumab were not indicated or needed for these cases.
EP
TE D
Three of the 14 cases were in the acute phase (grade 1: two cases; grade 2: one case). None of these worsened beyond the acute phase to be subsequently classified as a late toxicity. The patient with acute grade 2 toxicity did have radiographic changes of radiation necrosis, but these findings occurred after resolution of symptoms (prior imaging did not show any radiographic changes), and she did not develop additional symptoms after developing radiographic necrosis. The patients with acute grade 1 toxicity did not have radiographic changes identified during follow up. The remaining 11 toxicities were categorized as late toxicities (grade 2: nine cases: grade 3: one case; grade 5: one case). These patients with late grade 2 effects did not have radiographic correlates.
AC C
The grade three and grade five toxicity cases are described as follows. Both cases were in patients treated for high risk medulloblastoma after subtotal resection with high-dose craniospinal irradiation (36 Gy); one patient received conformal tumor bed boost with a 10 mm CTV margin (and 3 mm PTV margin) while the other patient received whole posterior fossa boost. Both patients developed radiographic changes consistent with radiation necrosis. The patient with grade 3 toxicity developed a right facial palsy and numbness 118 days after radiation therapy that did not improve on a prolonged steroid taper. Surgical intervention (gold weight in eyelid) was required to correct corneal abrasions secondary to cranial nerve deficits. He has been followed closely since that time and has not developed evidence of recurrence over a 29-month period of follow up. The second patient with severe brainstem toxicity, which developed 378 days after radiation therapy, was scored as grade 5 toxicity as she developed PEG-tube dependence and respiratory failure, requiring permanent tracheostomy. She suffered a prolonged course with clinical deterioration before succumbing to sepsis and multiorgan failure. She did have a biopsy consistent with
ACCEPTED MANUSCRIPT
radiation necrosis and not recurrence prior to her clinical deterioration. Hyperbaric oxygen was not offered in either case, and bevacizumab was not routinely used for radiation necrosis in the time period of these patients’ symptoms. She had no evidence of disease recurrence or progression at time of death. Figure 1 shows the area of brainstem necrosis in the axial, sagittal, and coronal planes in the patient with grade 3 toxicity with overlying isodose lines. The second patient had similar radiographic findings.
SC
RI PT
There were two additional patients, not detailed above, with radiographic alterations in the brainstem on surveillance MRIs that were consistent with radiation change. The first patient with radiographic changes in the brainstem remained asymptomatic through the period of follow up (i.e., no acute or late toxicities). The second patient developed symptomatic radiation necrosis (left hemiparesis, ataxia) but only after reirradiation for tumor bed recurrence and was therefore not coded as a toxicity in this dataset. Therefore, a total of 5 patients in this series had radiographic changes consistent with radiation necrosis (grade 0: 2/5, 3.3% of cohort; grade 1: 0/5; grade 2: 1/5, 1.7%; grade 3: 1/5, 1.7%;; grade 4: 0/5; grade 5: 1/5, 1.7%;).
M AN U
There was no correlation with acute brainstem toxicity and any of the clinical covariates examined. None of the dosimetric variables predicted for acute toxicity. Univariate logistic regression analysis did not yield any significant predictors for acute toxicity (supplemental table A2).
Discussion:
TE D
Radiographic necrosis correlated with symptomatic brainstem toxicity (p=0.029). Vascular or cranial nerve injury from surgery was marginally associated with incidence of late toxicity (p=0.054). However, there were no dosimetry endpoints that predicted for late brainstem toxicity (including necrosis) in this series (supplemental table A3). None of the patients with late toxicities had experienced acute brainstem toxicities. Figure 2 demonstrates the cumulative dose-volume histograms of all patients in the series, with the DVH of patients with grade 3 and 5 radiation necrosis highlighted in red. The patients with grade 1-2 toxicity are highlighted in blue.
AC C
EP
To our knowledge, no prior studies have correlated dose-volume data with risks of brainstem toxicity in patients treated with IMRT or VMAT. Constraints from the proton beam therapy cannot be directly extrapolated to the photon setting due to differences in treatment planning, dose distribution, and inherent differences in Bragg peak, RBE and LET, which are also not yet fully understood [11]. Only 10% of children (up to 33% for ependymoma/medulloblastoma patients) requiring radiation therapy in the United States receive radiation therapy at a proton therapy center each year [13], which may be due to lack of resources or ability to travel to proton therapy centers. Therefore, it is still critical to determine risks and constraints for patients being treated with x-ray-based therapy. Our study demonstrated that the crude risk of any brainstem toxicity was 23.3% but that the 3-year rate of severe toxicity was 3.3%, consistent with prior studies. Radiographic radiation necrosis and postoperative vascular or ischemic injury suggested a higher risk for symptomatic brainstem toxicity, also consistent with the literature [14]. We did not detect any correlation with age, gender, GTV size, presence of shunting, or development of posterior fossa syndrome.
ACCEPTED MANUSCRIPT
RI PT
The rates of any symptomatic brainstem toxicity in patients treated with 3D-conformal therapy range from 2.5% to 18.5%, depending on the definitions [8]; [15], [6], and from 3.8 to 16% in the proton literature. Grade 3 or higher toxicity, however, appears to be less than 3-4% in both settings [9], [11]. At the very least, the three modalities, now based on our data, appear to be associated with equally low rates of severe brainstem toxicity despite concerns in each setting that the variable physical properties are more or less risky to normal organs at risk. However, the rates of grade 2 toxicity may be slightly higher in our series than in the proton literature, although these comparisons should be made with caution.
M AN U
SC
We were unable to correlate risk of late toxicity or radiographic brainstem necrosis with any dosimetric variables. Our mean D10 and mean D50 were lower than the D10 and D50 proposed as safe in the series by Indelicato, et al [9]. However, in the two patients with grade 3 or higher brainstem necrosis, the D10 and D50 exceeded these constraints. Our mean V40, V45, V50, V55, and V60 also exceeded the recommended limits in the aforementioned proton series. It is possible that the higher V40-V55, representing higher volumes of the brainstem receiving moderate to high doses, are responsible for the higher rate of grade 2 toxicity in our series compared with the Indelicato series. Other data demonstrates that intensity-modulated radiation therapy is associated with higher volumes of brainstem receiving higher doses for similar tumors, compared with conventional and intensity-modulated proton therapy [16]. However, it is unclear however how the dosimetric variables in the proton setting translate into patients treated with conformal photon-based techniques as no clear relationship with RBE, LET, and brainstem injury has been identified [11].
EP
TE D
Five patients had radiographic changes consistent with radiation necrosis in this series (8.3%) and three additional patients had (asymptomatic) T2-weighted sequence radiation induced changes (data not shown), for an 13.3% rate of radiation related imaging changes, compared with 17% rate of similar changes in a series of posterior fossa patients treated with IMRT (and compared with 43% in patients treated with protons in that same series) [17]. The rate of symptomatic changes was, however, similar in both series (5% in ours, 8.5% in the Gunther, et al series) amongst IMRT patients. Using standard definitions of imaging changes in terms of relationship of radiographic changes to the treatment fields, the MRI sequences used to the define them, and the period of time after radiation therapy for which these findings should be investigated, will further serve to develop clinically relevant guidelines.
AC C
Given that these patients were treated at a single institution by only two physician providers, the data is fairly homogeneous in terms of registration software and technique, contouring of normal structures and tumor volumes, treatment planning, and delivery. We were able to have full and complete dosimetry data for all included patients included in this study which allowed for a robust dosimetric analysis. Due to close collaboration with our colleagues at the children’s hospital, we were able to have careful and frequent documentation of clinical and radiographic events in follow up. With rare exception, patients had imaging and clinical follow up done at the primary center, allowing for a consistent basis for comparison of symptoms and imaging findngs. Our data is of course limited inherently by retrospective collection of data and relatively small cohort size. In addition, the intervals between imaging studies were in some cases prolonged due to inability for patients to come as instructed for follow up, as we are a tertiary referral center with patients occasionally traveling significant distances. Therefore, it is possible that some symptomatic toxicities and radiographic changes were missed due to occasionally infrequent follow ups. Furthermore, including both medulloblastoma and ependymoma patients lends heterogeneity, as these
ACCEPTED MANUSCRIPT
patients are treated to different doses, and often have variable extents of disease and biologic behavior. Chemotherapy was sometimes given prior to radiation; it is unclear what interaction chemotherapy may have had with radiation even when not delivered concurrently. There is also relatively limited follow up and additional late toxicities affecting the brainstem may be yet to be seen. However, in our series, almost all events manifested within the first two years after treatment.
EP
Conclusions:
TE D
M AN U
SC
RI PT
We propose that certain measures be taken in order to reduce the risk of radiation necrosis. Avoiding excess volume of uninvolved brainstem within the CTV can reduce the volume of brainstem receiving high doses, and careful attention should be given to avoiding “hot spots” or heterogeneity within the brainstem. The shift in practice from posterior fossa boost to conformal tumor bed boost likely has resulted in lower radiation necrosis rates than previously observed. Approximately three-quarters of patients observed to develop radiation necrosis in the Massachusetts General Hospital series received a posterior fossa boost [11], supporting the hypothesis that high volumes receiving high doses of radiation predispose patients to radiation necrosis. The patients with grade 3 and 5 brainstem toxicity had either generous CTV margins or the entire posterior fossa treated. The use of 5 mm margins, as opposed to 10 mm margins, as are now employed in current practice may help achieve the goal of reducing the volume of brainstem irradiated. Although severe toxicities were not seen in patients with ependymomas in our series, they have been otherwise reported. It is unclear whether dose escalation for ependymomas beyond 54-55.8 improves local control. There is some data suggesting improved outcomes with doses beyond 50 Gy, but not necessarily to 59.4 Gy [18]. Additional data is needed to determine the optimal dose for local control while limiting risk of toxicity. Despite overall low rates, radiation-induced damage to the brainstem can be significantly morbid, requiring expensive interventions when at all available, impairing quality of life, and posing a threat to life in rare but tragic cases. Finally, it is possible that there are biologic factors that we have not yet considered that predispose certain patients to toxicity despite the most meticulous, conservative, treatment planning. Studying tissue and blood samples from patients enrolled on prospective national studies for molecular markers may provide insight into these toxicity profiles.
AC C
Our data demonstrates that the risk of clinically significant brainstem injury is low but not negligible using conformal, inverse-planned techniques. Prospective collection of dose-volume data in a large trial may help define more precise constraints that can be employed for pediatric patients with posterior fossa malignancies to further minimize risk of radiation necrosis while maintaining tumor control. Additional factors, including molecular markers to identify patients with inherent radiosensitivity, may also be useful in caring for this population.
Figure Captions: Figure 1: Isodose lines from this patient’s IMRT plan are overlaid the MRI T1-weighted post-contrast image illustrating the area of radiation necrosis, in the axial, sagittal, and coronal planes.
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
Figure 2: The cumulative brainstem dose-volume histograms are displayed on this graph. The DVHs for patients with grade 3 and 5 brainstem toxicity are highlighted in red. The DVHs for patients with grade 12 acute and late toxicity are in blue. The brainstem DVHs for the remaining patients without toxicities are represented in gray.
ACCEPTED MANUSCRIPT
References
AC C
EP
TE D
M AN U
SC
RI PT
1. Schroeder T. M., Chintagumpala M., Okcu M. F., et al. Intensity-modulated radiation therapy in childhood ependymoma. Int J Radiat Oncol Biol Phys 2008;71:987-993. 2. Polkinghorn W. R., Dunkel I. J., Souweidane M. M., et al. Disease control and ototoxicity using intensity-modulated radiation therapy tumor-bed boost for medulloblastoma. Int J Radiat Oncol Biol Phys 2011;81:e15-20. 3. Paulino A. C., Mazloom A., Teh B. S., et al. Local control after craniospinal irradiation, intensitymodulated radiotherapy boost, and chemotherapy in childhood medulloblastoma. Cancer 2011;117:635-641. 4. Merchant T. E., Li C., Xiong X., et al. Conformal radiotherapy after surgery for paediatric ependymoma: A prospective study. Lancet Oncol 2009;10:258-266. 5. Ibrahim N. Y., Abdel Aal H. H., Abdel Kader M. S., et al. Reducing late effects of radiotherapy in average risk medulloblastoma. Chin Clin Oncol 2014;3:4. 6. Murphy E. S., Merchant T. E., Wu S., et al. Necrosis after craniospinal irradiation: Results from a prospective series of children with central nervous system embryonal tumors. Int J Radiat Oncol Biol Phys 2012;83:e655-660. 7. Mayo C., Yorke E., Merchant T. E. Radiation associated brainstem injury. Int J Radiat Oncol Biol Phys 2010;76:S36-41. 8. Shaw E., Arusell R., Scheithauer B., et al. Prospective randomized trial of low- versus high-dose radiation therapy in adults with supratentorial low-grade glioma: Initial report of a north central cancer treatment group/radiation therapy oncology group/eastern cooperative oncology group study. J Clin Oncol 2002;20:2267-2276. 9. Indelicato D. J., Flampouri S., Rotondo R. L., et al. Incidence and dosimetric parameters of pediatric brainstem toxicity following proton therapy. Acta Oncol 2014;53:1298-1304. 10. McGovern S. L., Okcu M. F., Munsell M. F., et al. Outcomes and acute toxicities of proton therapy for pediatric atypical teratoid/rhabdoid tumor of the central nervous system. Int J Radiat Oncol Biol Phys 2014;90:1143-1152. 11. Giantsoudi D., Sethi R. V., Yeap B. Y., et al. Incidence of cns injury for a cohort of 111 patients treated with proton therapy for medulloblastoma: Let and rbe associations for areas of injury. Int J Radiat Oncol Biol Phys 2015. 12. Institute National Cancer. Common terminology criteria for adverse events v4.0. NCI, NIH, DHHS 13. Chang AL, Yock TI, Mahajan A, et al. Pediatric proton therapy: Patterns of care across the united states. Int J Particle Ther 2014;1. 14. Merchant T. E., Chitti R. M., Li C., et al. Factors associated with neurological recovery of brainstem function following postoperative conformal radiation therapy for infratentorial ependymoma. Int J Radiat Oncol Biol Phys 2010;76:496-503. 15. Spreafico F., Gandola L., Marchiano A., et al. Brain magnetic resonance imaging after high-dose chemotherapy and radiotherapy for childhood brain tumors. Int J Radiat Oncol Biol Phys 2008;70:1011-1019. 16. MacDonald S. M., Safai S., Trofimov A., et al. Proton radiotherapy for childhood ependymoma: Initial clinical outcomes and dose comparisons. Int J Radiat Oncol Biol Phys 2008;71:979-986. 17. Gunther J. R., Sato M., Chintagumpala M., et al. Imaging changes in pediatric intracranial ependymoma patients treated with proton beam radiation therapy compared to intensity modulated radiation therapy. Int J Radiat Oncol Biol Phys 2015;93:54-63. 18. Taylor R. E. Review of radiotherapy dose and volume for intracranial ependymoma. Pediatr Blood Cancer 2004;42:457-460.
ACCEPTED MANUSCRIPT
Table 1: Treatment planning constraints for IMRT Volume
Dose Limit
Brainstem Globe Left Globe Right Lens Left Lens Right Optic Chiasm Optic Nerve Left Optic Nerve Right Cochlea Left Cochlea Right Spinal Cord C1-2 Spinal Cord Hypothalamic-pituitary axis
50% 50% 50% 100% 100% 50% 50% 50% 50% 50% 50% 50% 50%
≤ 60 Gy ≤ 10Gy ≤ 10 Gy ≤ 5 Gy ≤ 5 Gy ≤ 54 Gy ≤ 54 Gy ≤ 54 Gy ≤ 35 Gy ≤ 35 Gy ≤ 50 Gy ≤ 54 Gy ≤ 35 Gy
AC C
EP
TE D
M AN U
SC
RI PT
Organ At Risk
ACCEPTED MANUSCRIPT
RI PT
Table 2. Mean and median dose-volume data for brainstem organ at risk.
Mean Dose (Gy)
Median Dose (Gy)
D10 D20 D30 D40 D50 D60 D70 D80 D90
55.1 54.7 54.2 53.5 52.6 51.5 50.1 48.2 45.8
56.8 56.6 56.2 55.7 55.5 55.0 54.2 53.2 50.9
Dose to Brainstem (Gy)
Mean Volume (%) Median Volume (%)
V40 V45 V50 V55 Maximum
87.8 85.2 79.1 53.8 56.1 (Gy)
M AN U
SC
Volume of Brainstem (%)
AC C
EP
TE D
100 99 92.4 59.6 57.3 (Gy)
ACCEPTED MANUSCRIPT
AC C
EP
TE D
M AN U
SC
RI PT
Table 3. Acute and late toxicities
ACCEPTED MANUSCRIPT
Time (Months)
Grade 1
6 12 24 36
Estimated cumulative incidence (%) (95% CI) 1.7 (0, 5.1) 3.4 (0, 8.2) 3.4 (0, 8.2) 3.4 (0, 8.2)
Grade 2
6 12 24 36
3.4 (0, 8.2) 8.6 (1.3, 1.6) 12.1 (3.6, 20.5) 14.4 (5.0, 23.9)
Grade 3-5
6 12 24 36
0 (0,0) 0 (0,0) 3.6 (0, 8.6) 3.6 (0, 8.6)
AC C
EP
TE D
M AN U
SC
Event
RI PT
Table 4. Cumulative incidence of toxicity by grade.
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