- Email: [email protected]
Dose-Response Modeling of the Visual Pathway Tolerance to Single-Fraction and Hypofractionated Stereotactic Radiosurgery
Dose-Response Modeling of the Visual Pathway Tolerance to Single-Fraction and Hypofractionated Stereotactic Radiosurgery Susan M. Hiniker, MD,* Leslie A. Modlin, BA,* Clara Y. Choi, MD, PhD,* Banu Atalar, MD,† Kira Seiger, BA,* Michael S. Binkley, BA,* Jeremy P. Harris, MD, MPhil,* Yaping Joyce Liao, MD,‡ Nancy Fischbein, MD,§ Lei Wang, PhD,* Anthony Ho, PhD,* Anthony Lo, MS,* Steven D. Chang, MD,║ Griffith R. Harsh, MD,║ Iris C. Gibbs, MD,* Steven L. Hancock, MD,* Gordon Li, MD,║ John R. Adler, MD,║ and Scott G. Soltys, MD* Patients with tumors adjacent to the optic nerves and chiasm are frequently not candidates for single-fraction stereotactic radiosurgery (SRS) due to concern for radiation-induced optic neuropathy. However, these patients have been successfully treated with hypofractionated SRS over 2-5 days, though dose constraints have not yet been well defined. We reviewed the literature on optic tolerance to radiation and constructed a dose-response model for visual pathway tolerance to SRS delivered in 1-5 fractions. We analyzed optic nerve and chiasm dose-volume histogram (DVH) data from perioptic tumors, defined as those within 3 mm of the optic nerves or chiasm, treated with SRS from 2000-2013 at our institution. Tumors with subsequent local progression were excluded from the primary analysis of vision outcome. A total of 262 evaluable cases (26 with malignant and 236 with benign tumors) with visual field and clinical outcomes were analyzed. Median patient follow-up was 37 months (range: 2-142 months). The median number of fractions was 3 (1 fraction n ¼ 47, 2 fraction n ¼ 28, 3 fraction n ¼ 111, 4 fraction n ¼ 10, and 5 fraction n ¼ 66); doses were converted to 3-fraction equivalent doses with the linear quadratic model using α/β ¼ 2 Gy prior to modeling. Optic structure dose parameters analyzed included Dmin, Dmedian, Dmean, Dmax, V30 Gy, V25 Gy, V20 Gy, V15 Gy, V10 Gy, V5 Gy, D50%, D10%, D5%, D1%, D1 cc, D0.50 cc, D0.25 cc, D0.20 cc, D0.10 cc, D0.05 cc, D0.03 cc. From the plan DVHs, a maximum-likelihood parameter fitting of the probit doseresponse model was performed using DVH Evaluator software. The 68% CIs, corresponding to one standard deviation, were calculated using the profile likelihood method. Of the 262 analyzed, 2 (0.8%) patients experienced common terminology criteria for adverse events grade 4 vision loss in one eye, defined as vision of 20/200 or worse in the affected eye. One of these patients had received 2 previous courses of radiotherapy to the optic structures. Both cases were meningiomas treated with 25 Gy in 5 fractions, with a 3-fraction equivalent optic nerve Dmax of 19.2 and 22.2 Gy. Fitting these data to a probit dose-response model enabled risk estimates to be made for these previously unvalidated optic pathway constraints: the Dmax limits of 12 Gy in 1 fraction from QUANTEC, 19.5 Gy in 3 fractions from Timmerman 2008, and
*Department of Radiation Oncology, Stanford University, Stanford, CA. †Department of Radiation Oncology, Acibadem University, Istanbul, Turkey. ‡Department of Ophthalmology, Stanford University, Stanford, CA. §Department of Radiology, Stanford University, Stanford, CA. ║Department of Neurosurgery, Stanford University, Stanford, CA. Conflict of interest: none. Presented in part at the Radiosurgery Society SRS/SBRT Annual Meeting, May 2014. Address reprint requests to Scott G. Soltys, MD, Department of Radiation Oncology, Stanford Cancer Center, 875 Blake Wilbur Drive, Stanford, CA 94305-5847. E-mail: [email protected]
http://dx.doi.org/10.1016/j.semradonc.2015.11.008 1053-4296/& 2016 Elsevier Inc. All rights reserved.
97
S.M. Hiniker et al
98 25 Gy in 5 fractions from AAPM Task Group 101 all had less than 1% risk. In 262 patients with perioptic tumors treated with SRS, we found a risk of optic complications of less than 1%. These data support previously unvalidated estimates as safe guidelines, which may in fact underestimate the tolerance of the optic structures, particularly in patients without prior radiation. Further investigation would refine the estimated normal tissue complication probability for SRS near the optic apparatus. Semin Radiat Oncol 26:97-104 C 2016 Elsevier Inc. All rights reserved.
S
tereotactic radiosurgery (SRS) is a noninvasive, highly accurate form of radiation therapy that is increasingly used to treat benign and malignant intracranial tumors with high rates of local control. Recent large studies of patients treated with SRS for pituitary adenomas and meningiomas have shown long-term disease control of over 95% with reduced toxicity as compared to traditional external beam radiation therapy (EBRT).1,2 However, patients with tumors located near the optic nerves or optic chiasm (anterior optic apparatus) have been frequently excluded from SRS owing to concerns for visual toxicity. For tumors within 3 mm of the optic structures (“perioptic tumors”), single-fraction SRS risks vision loss either from radiation-induced optic neuropathy (RION) if the optic apparatus dose is too high, or from tumor progression if the tumor dose is too low to yield control.3,4 Historically, the SRS maximum dose (Dmax) limit to the optic pathway has been cited as 8 Gy in a single fraction. For single-fraction doses necessary to control benign tumors (13-16 Gy), a risk of blindness as high as 27% has been reported.5 Given these potential risks of singlefraction SRS (“SRS”), we have typically treated these perioptic tumors with hypofractionated SRS (“fSRS”) over 2-5 days, although evidence-based dose constraints for fSRS have not yet been reported. Here we summarize the literature on dose constraints for the optic structures and report our institutional experience with over 250 patients treated with singlefraction and hypofractionated SRS for perioptic tumors.
Optic Structure Tolerance to Radiation Optic apparatus dose constraints have been best defined for conventionally fractionated EBRT. The Emami data for EBRT dose causing a 5% risk of toxicity at 5 years (TD 5/5) for the chiasm and optic nerves is 50 Gy, and the 50% risk tolerance dose (TD 50/5) is 65 Gy.6 Of note, these estimates are for whole organ irradiation, and reports suggest they may be conservative, particularly in the setting of partial-organ dose. More recently, QUANTEC data estimates the risk of toxicity for optic nerve or chiasm maximum point dose (Dmax) of o55 Gy at o3%, for 55-60 Gy of 3%-7%, and for 460 Gy of 47%20%.7 Among 131 patients treated with EBRT at the University of Florida, no RION occurred at maximum doses at or below 59 Gy. The 15-year risk of RION was 11% with doses above 60 Gy with dose fractions of less than 1.9 Gy, and 47% with
doses above 60 Gy when fraction size was greater than or equal to 1.9 Gy.8 Other studies provide normal tissue complication probability (NTCP) calculations for optic structures, including a report of 39 patients treated for advanced paranasal sinus tumors, of which 13 patients were treated without optic nerve or chiasm sparing. Of these 13 patients, 3 patients experienced moderate or severe vision loss, and all 3 with vision loss received maximum dose to the optic apparatus of at least 64 Gy.9 Although hyperfractionation may reduce the risk of vision injury with EBRT, maximum dose to the optic apparatus best correlates with optic neuropathy.10
Single-Fraction SRS for Perioptic Tumors Early Studies Early studies of RION after SRS supported strict dose constraints for the optic structures. Among 62 patients treated with single-fraction SRS for cavernous sinus meningiomas, 17 who received an optic apparatus Dmax of 48 Gy developed visual complications; a single-fraction optic apparatus limit of 8 Gy was proposed.11 Another early report described 4 patients with RION after an SRS Dmax ranging from 7-14 Gy.12 Leber et al5 reported no cases of RION for a single-fraction dose of less than 10 Gy, but they reported a risk of 27% for 10-15 Gy and 78% for 415 Gy. Given these early reported complications, initial single-fraction SRS dose constraints for the optic apparatus were conservative at a Dmax of less than 8 Gy.
Dose Escalation Despite an early recommended optic apparatus Dmax of 8 Gy, multiple groups subsequently reported favorable vision outcomes with higher single-fraction doses. At the University of Maryland, 20 patients received an average Dmax of 9 Gy, with none experiencing RION with serial visual field testing.13 In a report from Norway of 100 patients treated for cavernous sinus meningiomas, 1 patient, with a visual pathway Dmax of 8.6 Gy, experienced RION, consistent with a 1% risk.14 Mayo Clinic reported on 88 patients treated with SRS for skull base meningiomas, with no RION at a median Dmax of 10 Gy (range: 1-16 Gy).15 These investigators later reported a rate of 1.1% for patients receiving up to 12 Gy to the optic apparatus.16 This series was later updated with a reported risk of RION of o1% at a maximum dose of o12 Gy.17 In this
Tolerance of the visual pathway to radiosurgery largest report to date, of 222 patients treated with SRS for perioptic tumors, 1 developed unilateral blindness after receiving a Dmax of 12.8 Gy. The above reports were primarily in patients treated with SRS for benign meningiomas or nonfunctioning pituitary adenomas. SRS doses are typically higher when treating functioning pituitary adenomas, yielding additional data for RION. A report from Yale described 31 patients treated with a marginal tumor dose of 35 Gy for functional pituitary adenomas with an optic apparatus Dmax of up to 13.8 Gy. One patient developed unilateral optic nerve pallor and temporal field defect 3 years after radiation, with a nerve Dmax of 7.4 Gy.18 The University of Virginia reported outcomes in 217 patients treated with GammaKnife radiosurgery (GKS) for recurrent pituitary adenoma, 131 of which were secretory, treated with median tumor dose of 23 Gy. In all, 3 patients (2%) had RION, although the specific optic apparatus Dmax for these patients was not provided. Although an increased number of isocenters was statistically related to RION, Dmax was not correlated.19 Other reports that suggest safety of single-fraction optic apparatus SRS doses greater than 12 Gy are limited by incomplete dosimetric reporting and patient heterogeneity. In a series of 100 patients treated with SRS for craniopharyngioma, Hasegawa et al reported 3 patients with RION. Of the 3 patients, 2 received 15 Gy to the optic apparatus, with the third treated with multiple courses of irradiation. The authors argue that their data potentially supports maximum doses of up to 14 Gy to a partial segment of the optic apparatus.20 Conversely, unilateral blindness was seen in a patient treated for chordoma following 14.8 Gy in a single fraction to the optic apparatus.21 The most detailed dose-volume analysis data for singlefraction SRS to date was reported in 2014 by Pollock et al, in which the risk of RION for single-fraction SRS for pituitary adenomas was calculated. Among 133 treated patients who had not received previous irradiation, with a total of 266 anterior visual sides analyzed, 29 anterior visual pathway sides (11%) received a maximum dose of greater than 12 Gy, and no RION occurred, with a 95% CI risk of optic neuropathy of 0%13.9% at 12 Gy.22 In a recent review of radiation dose-volume effects of the optic structures, QUANTEC threshold limits for increased toxicity are 60 Gy in 1.8 Gy/fraction, and 12 Gy for single-fraction SRS.4
Multiple Fraction SRS for Perioptic Tumors Dose constraints for hypofractionated SRS over 2-5 days for perioptic tumors have not been well described, though it is recognized that hypofractionation may reduce the risk of normal tissue toxicity. In our and other institutions, if the single-fraction optic apparatus Dmax is greater than the 1012 Gy reported above, fSRS over 2-5 days has been investigated. A prior series of 49 patients with perioptic tumors treated with fSRS at our institution showed excellent tumor control and visual field preservation at mean follow-up of 49 months, with 94% tumor control and 1 patient (2%)
99 experiencing RION.23 However, detailed optic apparatus dosimetry was not reported. Kim et al reported a Korean experience evaluating multisession GKS for perioptic tumors with tumor control in 21 of 22 patients. Vision deteriorated in the one patient with tumor progression, with no other complications.24 However, no detailed dosimetric analysis was reported. In a series of 23 patients treated with multisession GKS for perioptic tumors, Jo et al reported 12 with stable vision and 11 with improvement in visual fields or acuity or both. Patients were treated to a marginal dose of 22 Gy in 4 fractions, with optic nerve maximum dose ranging from 12.8-20.8 Gy.25 In another series of 20 patients treated with SRS for perichiasmatic pituitary adenomas who received 25 Gy in 5 fractions, vision remained intact in all patients and improved in 3 patients.26 Median maximum chiasm dose was 23.3 Gy (range: 18.3-25.1 Gy). Similarly, excellent results were reported in a series of patients treated for pituitary adenomas close to the optic apparatus to 21 Gy in 3 fractions, with dose to the optic apparatus reported as mean dose of 16.7 Gy to the nerve and 14.6 Gy to the chiasm.27 In a report from Japan of the treatment of craniopharyngioma with singlefraction or hypofractionated SRS, 43 patients received a median single-fraction marginal dose of 14.3 Gy, and median multiple fraction marginal dose of 21 Gy in 3 fractions or 25 Gy in 5 fractions. No patient experienced RION, though optic apparatus dosimetry was not reported.28 This group has also reported similar excellent outcomes using hypofractionated SRS for nonfunctioning pituitary adenomas with minimal vision toxicity, with median optic nerve maximum dose of 19.9 Gy in 5 fractions, and optic chiasm maximum dose of 20.3 Gy (range: 1.4-25.0 Gy).29 However, detailed dosimetric analysis has not yet been reported for hypofractionated SRS.
Dose-Response Model for Hypofractionated SRS As above, appropriate dose constraints for fSRS remain poorly defined. There have been no reports to date detailing the dosimetric parameters for hypofractionated SRS that yield long-term tumor control as well as acceptable toxicity to the optic pathway. In the present study, we report the tumor control and vision outcomes of 262 patients treated with single and hypofractionated SRS for perioptic tumors, specifically identifying dose-volume parameters for the optic structures in patients treated with single-fraction and hypofractionated SRS. We sought to construct a dose-response model for visual pathway tolerance to SRS delivered in 1-5 fractions.
Methods and Materials Patients We conducted a retrospective review of a prospectively maintained list of patients, identifying those with perioptic tumors within 3 mm of the optic nerves or chiasm, treated with radiosurgery at Stanford University from 2000-2013. This study was approved by the Stanford University Institutional
S.M. Hiniker et al
100 Review Board. Eligible patients included those having tumors within 3 mm of the optic apparatus treated with SRS inclusive of all histologies including benign and malignant disease. We identified 326 patients treated with SRS for perioptic tumors. We excluded patients for whom clinical follow-up was not available or for whom dose-volume histogram (DVH) data could not be retrieved because of software incompatibility between machine upgrades, yielding 262 evaluable patients. The median age was 53 years (range: 14-86 years). Among all, 62% of patients were women. Metastases or malignant disease comprised 26 cases, the remaining were benign tumors: meningiomas (132), secreting (25) and nonsecreting (43) pituitary adenomas, craniopharyngiomas (15), schwannomas (11); arteriovenous malformations, hemangioblastomas, hemangiopericytomas, chordomas (2 each); choroid plexus papilloma and hamartoma (1 each). A total of 154 patients (59%) had undergone at least 1 surgical resection prior to SRS, including craniotomy or trans-sphenoidal resection. A total of 34 patients (13%) had been treated previously with radiation therapy to the lesion in question, and of these, 27 patients were treated with EBRT, and 7 patients with SRS.
Treatment Details Radiosurgical planning was performed using thin-slice, highresolution computed tomography (CT) scans, with intravenous contrast unless there was a contraindication. Thinsection magnetic resonance imaging (MRI) scans were also performed in the majority of patients, and MRI-CT fusion was performed. The treating physician contoured both the anterior visual pathways and target tumor. As described previously,23 the standard inverse planning CyberKnife method was used for planning treatment. Dose selection was determined by the attending neurosurgeon and radiation oncologist and was based on tumor volume, histology, proximity to optic structures, and previous history of radiation. Radiosurgery was delivered over 1-5 fractions, separated by at least 12 hours. The median prescribed tumor dose was 18 Gy (range: 12-25 Gy) in 1 fraction (n ¼ 47), 24 Gy (range: 18-33 Gy) in 3 fractions (n ¼ 111), and 25 Gy (range: 18-40 Gy) in 5 fractions (n ¼ 66). The median Dmax to the optic nerve was 7.6 Gy (5-95 percentile range: 1.9-12.4 Gy) in 1 fraction, 13.4 Gy (2.7-23.3 Gy) in 3 fractions, and 19.6 Gy (3.8-29.4 Gy) in 5 fractions. Median values for other parameters in all fractionations are given in Tables e1-e2 of the Appendix A.
(CTCAE) grade 4 vision loss, corresponding to vision of 20/ 200 or worse in the affected eye. Response rates were classified as the presence or absence of local relapse. Local control was defined as decreased size or stability of the treated tumor. Local failure was classified as tumor growth or the need for subsequent treatment with surgery or repeat irradiation.
Analysis Time to event was measured from the date of SRS or fSRS. From the plan DVHs, a maximum-likelihood parameter fitting of the probit dose-response model was performed using DVH Evaluator software (DiversiLabs, LLC, Huntingdon Valley, PA). The 68% CIs were calculated using the profile likelihood method. The median number of fractions was 3; all doses were converted to 3-fraction equivalent doses with the linear quadratic model using α/β ¼ 2 Gy for the optic structures prior to the modeling. The following dose descriptors of the optic apparatus were analyzed: Dmin, Dmedian, Dmean, Dmax, V30 Gy, V25 Gy, V20 Gy, V15 Gy, V10 Gy, V5 Gy, D50%, D10%, D5%, D1%, D1 cc, D0.50 cc, D0.25 cc, D0.20 cc, D0.10 cc, D0.05 cc, D0.03 cc. Dose-response modeling was accomplished according to the flowchart in Figure 1. DVH data were loaded into the DVH Evaluator software and numerous dose descriptors Dx were extracted, where x is a specified volume and Dx is the DVH dosevolume point corresponding to the volume x. To ensure a relevant metric, it was required that 95% or more of the cases had a nonzero value for the selected Dx. For example, fewer than 95% of the cases had optic nerves exceeding a volume of 0.5 cc, so the D0.5 cc was not used. All Dx values were then converted to 3fraction equivalent doses using the linear quadratic model with α/ β ¼ 2 Gy for normal tissue tolerance. The probit dose-response model was used,30 with maximum-likelihood parameter fitting,31 with CIs via the profile likelihood method.32 Our model was compared to the previously unvalidated Dmax constraints of 12 Gy in 1 fraction from QUANTEC, 19.5 Gy in 3 fractions from Timmermann33, and 25 Gy in 5 fractions from AAPM Task Group 101, and all constraints had less than 1% risk.3,7,33
Results Tumor Control and Vision Outcomes Median follow-up was 36.8 months (range: 2-142 months). Among 262 patients, 19 experienced local progression of
Patient Follow-Up Patients were followed with clinical examination, brain imaging, and annual visual field testing. For benign tumors, followup is at 6 and 12 months post-SRS, then annually. For malignant tumors, follow-up is typically every 3 months for 1 year, then every 4 months. Evaluation of treatment response was assessed by clinical examination and radiographic studies. Evaluation of visual outcomes was assessed by formal visual field testing (performed by an ophthalmologist approximately yearly, when available) and clinical vision examination. RION was defined as common terminology criteria for adverse events
Figure 1 Flowchart of dose-response modeling process. (Color version of figure is available online.)
Tolerance of the visual pathway to radiosurgery
101 appeared stable, the patient underwent surgical decompression of the optic apparatus, without improvement, and the patient lost vision in that eye. In the second case of RION, the patient had 2 courses of irradiation previously: EBRT and SRS (20 Gy in 1 fraction), and surgery to a right parasellar grade I meningioma, measuring 14.33 cc. The patient was retreated to 25 Gy in 5 fractions, prescribed to the 78% isodose line; the maximum dose to the optic pathway was 27.7 Gy for the third course. Seven months after radiosurgery, the patient experienced gradual onset of vision loss in the right eye that became complete, attributed to optic neuropathy. For each patient, 3 critical structures could potentially cause the complication: the 2 optic nerves and the chiasm. Rather than combining all into a single optic pathway structure, an attempt was made to analyze the components separately, as the vision loss in both cases was unilateral. We used the structure receiving the highest dose as the likeliest to be responsible for vision loss, which was the optic nerve in the majority of patients. In the patient who experienced RION who had been previously treated with 2 courses of irradiation, the exact composite treatment plan for all courses was not available. The chiasm dose was the highest optic structure dose in the final treatment plan prior to the development of RION, so this was the dose used in the analysis. We noted that given the 2 courses of previous treatment, the dose used is likely to underestimate the total dose responsible for causing RION in this patient.
Normal Tissue Complication Probability
Figure 2 Optic nerve dose-response for small volumes: (A) D0.2 cc and (B) Dmax. AE, adverse event; MLE, maximum likelihood estimate; NfxED, N-fraction equivalent dose. (Color version of figure is available online.)
disease after treatment (7.3%). Of 26 patients with malignant disease, 6 (23%) experienced local failure, and 13 of 236 (5.5%) patients with benign tumors had local failure. Among the 86 patients with vision deficits prior to radiosurgery (33%), we found that 13% had improvement in vision after treatment. In all, 7 (2.7%) patients had worsening of vision following treatment; this was attributed to unequivocal tumor growth causing optic apparatus compression in 5 (1.9%) patients. A total of 2 patients (0.8%) experienced unilateral vision loss in the absence of progressive disease. The first patient was treated with 25 Gy in 5 fractions for a 7.84 cm3 sellar meningioma encasing the left optic nerve. The maximum dose to the optic nerve was 23.9 Gy. The patient experienced gradual onset of vision loss in the left eye 3 months after radiosurgery, and MRI revealed enhancement along the left optic nerve suspicious for RION. Though the tumor size
Graphs of the resulting NTCP curves for the optic nerve are shown in Figure 2, for D0.2 cc and Dmax, results for D50% and D10% are shown in the Appendix A, and model parameters are in Table 1. In the figures, the red squares represent the 2 cases with optic neuropathy, and the blue dots represent the doses for the non-complication cases. CIs are shown as the dashed green lines, which are reasonably tight in the clinical region where data exists, and diverge above these doses. Figure 2 (B) shows that for Dmax at 40 Gy in 3 fractions, the risk ranges from 5%-65%, whereas at 20 Gy, the CIs only range ⫾ 1% from the model estimate. This method of CI shows where the data are sparse and where the estimates mature. Dotted red lines intersect the NTCP curves at 1%, 2%, and 5% risk levels for visualization of the corresponding Dx values. The risk level of previously published dose tolerance limits was evaluated and is denoted in the second number in each cell of the DVH Risk Map (Fig. 3). Specifically, the QUANTEC single-fraction Dmax limit of 12 Gy has a risk level of 0.7% from this dataset, as shown in the upper-right cell of the table in Table 1 Optic Nerve Probit Model Parameters With 68% CI Dose Descriptor
TD50, Gy (68% CI)
m (68% CI)
D50% D10% D0.2cc D0.05cc Dmax
46.81 (29.56-242.84) 45.19 (32.98-93.99) 55.66 (34.78-434.24) 45.04 (33.25-91.00) 48.91 (37.12-94.44)
0.3387 (0.2950-0.3858) 0.2830 (0.2296-0.3450) 0.3361 (0.2907-0.3870) 0.2789 (0.2250-0.3416) 0.2459 (0.1878-0.3216)
S.M. Hiniker et al
102
Figure 3 DVH Risk Map showing published dose tolerance limits organized into high-risk (solid red line) and low-risk (dashed green line) categories. Estimates of risk for each selected limit are shown to the right of each cell in the table. Bold limits in the table indicate published limits, and italicized limits indicate new limits from the trendlines and from modeled estimates of risk. (Color version of figure is available online.)
model estimated a o1% risk of RION at 12 Gy in 1, 19.5 Gy in 3, and 25 Gy in 5 fractions.
Figure 3 and depicted graphically in the rightmost subplot. In the DVH Risk Map, each dose descriptor (D50%, D10%, D0.2cc, D0.05cc, and Dmax) is shown in a subplot, and all corresponding published dose tolerance limits are plotted as blue diamonds. For each dose descriptor in each fractionation, a reasonably high set of published limits were selected as high-risk limits and are labeled and plotted with a red trend line. At a lower dose, another set of low-risk limits were selected and labeled and were plotted with a dotted green trend line. From the model of the clinical dataset, the estimated risk levels for every dose tolerance limit along both high-risk and low-risk trend lines are displayed quantitatively in the tabular section of Figure 3. Estimated risk levels for optic nerve Dmax levels in 1-5 fractions are shown in Tables 2 and 3. To summarize, the
Discussion In this study of tolerance of the optic apparatus to singlefraction and hypofractionated SRS, we sought to review the literature and to better define the dose-volume parameters for the risk of RION. We report one of the largest studies to date examining vision outcomes after single-fraction and hypofractionated SRS for perioptic tumors. In this cohort of 262 perioptic tumors treated with SRS, there was a low rate of optic complications. Of the 262
Table 2 Estimated RION Risk Level of Common Dmax Levels for Optic Nerve One Fraction
Three Fractions
Five Fractions
Dmax (Gy)
Estimated Risk (%)
Dmax (Gy)
Estimated Risk (%)
Dmax (Gy)
Estimated Risk (%)
8 10 12 14 15
0.1 0.3 0.7 1.6 2.3
16 18 20 22 24
0.3 0.5 0.8 1.3 1.9
20 22.5 25 27.5 30
0.3 0.5 0.8 1.3 1.9
Tolerance of the visual pathway to radiosurgery
103
Table 3 Optic Nerve Dmax Values Corresponding to 1%, 2%, 3%, and 5% Risk of RION Number of Fractions
Dmax for Dmax for Dmax for Dmax for 1% Risk 2% Risk 3% Risk 5% Risk (Gy) (Gy) (Gy) (Gy)
1 2 3 4 5
12.7 17.5 20.9 23.7 26.1
14.6 20.2 24.2 27.5 30.3
15.9 21.9 26.3 29.9 32.9
17.5 24.2 29.1 33.1 36.6
analyzed patients, 2 (0.8%) experienced common terminology criteria for adverse events grade 4 vision loss in one eye, defined as vision of 20/200 or worse in the affected eye. Both cases were meningiomas treated with 25 Gy in 5 fractions, with optic pathway Dmax of 19.2 Gy and 22.2 Gy, respectively, in 3fraction equivalent doses. In one of these cases, the patient had received 2 previous courses of radiotherapy to this area. A model of these complications estimated the risk of vision injury for the previously unvalidated estimates of optic pathway tolerance, including a 0.2% risk for the estimated 3-fraction optic pathway constraint of Dmax r 15 Gy, and 0.7% risk for the estimated constraint of Dmax r 19.5 Gy.33,34 These data support the previously unvalidated estimates as safe guidelines, which may in fact underestimate the tolerance of the optic structures, particularly in patients without prior radiation. Patients were twice as likely to experience vision loss related to local progression of disease as compared to RION, and therefore it may be riskier to vision to adhere to dose constraints that may in fact be overly conservative rather than adopting more aggressive constraints to achieve tumor control.35 Our data suggest that perioptic tumors can be safely treated with SRS delivered over 1-5 fractions with a low risk of RION even with optic apparatus doses above the previously estimated constraints. We included patients with metastatic and malignant disease as long as they lived for at least 6 months, whereas most reports of RION for perioptic tumors focus on benign tumors. Minniti et al36 also report outcomes after fSRS for patients with skull base metastases involving the anterior visual pathway, and found a 2-year local control of 72%, with no RION and 51% of patients having improvement of preexisting cranial nerve deficits. In addition, we included patients with recurrent disease and 34 (13%) of our patients had received previous irradiation to the lesion in question, putting them at higher risk for RION. The predictors of optic neuropathy after radiation have not been well described. RION typically occurs within 3 years of treatment.4,12 The pretreatment condition of the optic nerve, even in those patients who have not received previous radiation, is likely to affect posttreatment vision outcomes. Animal studies revealed that a compressed optic nerve is more sensitive to SRS-induced RION than are noncompressed nerves. In cats, the minimum dose causing optic neuropathy is 12 Gy in noncompressed nerves, and 11 Gy in compressed nerves.37 Others have found that prior surgery or anterior visual pathway dysfunction may increase the risk of RION.38
One of our 2 patients with RION had surgery before, and this patient also had anterior visual pathway dysfunction prior to SRS. Other factors that may increase the risk of SRS toxicity include diabetes, hypertension, and collagen vascular disorders. Similar to others and not surprisingly, we found a higher rate of vision toxicity in patients undergoing retreatment with radiation (SRS or EBRT), with 1 of 34 retreated patients developing optic neuropathy, vs 1 of 228 previously untreated patients.16,19 Contouring the optic chiasm can be a challenge as it is a very small structure not well seen on CT scans. MRIs are beneficial in delineating the chiasm, but the distinction between chiasm and optic nerves may be difficult to determine and to some extent is arbitrary. Although absolute volumes are preferable in most situations, the small size of the optic structures and variability in contouring argues for a role of relative volume use in this setting. Similar to others, we found that the maximum dose to the optic apparatus was the most important predictor of toxicity, so volumetric constraints may be less important. The retrospective nature of our study is a notable limitation. Nevertheless, these results support the use of hypofractionated SRS for perioptic tumors, similar to other studies, and provide dose-volume constraints in a large cohort of patients. Further investigation is needed to refine the estimated NTCP for SRS near the optic apparatus.
Conclusions In summary, our data suggest that perioptic tumors can be safely treated with single-fraction and hypofractionated SRS with excellent local control and a low rate of vision complications. Our data provide support for the previously unvalidated estimates of hypofractionated dose constraints for the optic structures as safe guidelines, which may underestimate the actual tolerance of the optic apparatus. Further investigation and delineation of the dose-volume parameters of SRS for tumors in close proximity to the optic structures would be important, particularly in the setting of reirradiation. Our model suggests a less than 1% incidence of RION in our group of patients treated with an optic apparatus pathway maximum point dose of 12 Gy in 1, 19.5 Gy in 3, and 25 Gy in 5 fractions.
Appendix. Supplementary data Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.semradonc. 2015.11.008.
References 1. Sheehan JP, Niranjan A, Sheehan JM, et al: Stereotactic radiosurgery for pituitary adenomas: An intermediate review of its safety, efficacy, and role in the neurosurgical treatment armamentarium. J Neurosurg 102 (4):678-691, 2005 2. Pollock BE, Stafford SL, Link MJ, et al: Single-fraction radiosurgery of benign intracranial meningiomas. Neurosurgery 71(3):604-612, 2012; [discussion 613]
S.M. Hiniker et al
104 3. Benedict SH, Yenice KM, Followill D, et al: Stereotactic body radiation therapy: The report of AAPM Task Group 101. Med Phys 37 (8):4078-4101, 2010 4. Mayo C, Martel MK, Marks LB, et al: Radiation dose-volume effects of optic nerves and chiasm. Int J Radiat Oncol Biol Phys 76(suppl 3): S28-S35, 2010 5. Leber KA, Bergloff J, Pendl G: Dose-response tolerance of the visual pathways and cranial nerves of the cavernous sinus to stereotactic radiosurgery. J Neurosurg 88(1):43-50, 1998 6. Emami B, Lyman J, Brown A, et al: Tolerance of normal tissue to therapeutic irradiation. Int J Radiat Oncol Biol Phys 21(1):109-122, 1991 7. Marks LB, Yorke ED, Jackson A, et al: Use of normal tissue complication probability models in the clinic. Int J Radiat Oncol Biol Phys 76(suppl 3): S10-S19, 2010 8. Parsons JT, Bova FJ, Fitzgerald CR, et al: Radiation optic neuropathy after megavoltage external-beam irradiation: Analysis of time-dose factors. Int J Radiat Oncol Biol Phys 30(4):755-763, 1994 9. Martel MK, Sandler HM, Cornblath WT, et al: Dose-volume complication analysis for visual pathway structures of patients with advanced paranasal sinus tumors. Int J Radiat Oncol Biol Phys 38(2):273-284, 1997 10. Bhandare N, Monroe AT, Morris CG, et al: Does altered fractionation influence the risk of radiation-induced optic neuropathy? Int J Radiat Oncol Biol Phys 62(4):1070-1077, 2005 11. Tishler RB, Loeffler JS, Lunsford LD, et al: Tolerance of cranial nerves of the cavernous sinus to radiosurgery. Int J Radiat Oncol Biol Phys 27 (2):215-221, 1993 12. Girkin CA, Comey CH, Lunsford LD, et al: Radiation optic neuropathy after stereotactic radiosurgery. Ophthalmology 104(10):1634-1643, 1997 13. Ove R, Kelman S, Amin PP, et al: Preservation of visual fields after perisellar gamma-knife radiosurgery. Int J Cancer 90(6):343-350, 2000 14. Skeie BS, Enger PO, Skeie GO, et al: Gamma knife surgery of meningiomas involving the cavernous sinus: Long-term follow-up of 100 patients. Neurosurgery 66(4):661-668, 2010; [discussion 668-669] 15. Morita A, Coffey RJ, Foote RL, et al: Risk of injury to cranial nerves after gamma knife radiosurgery for skull base meningiomas: Experience in 88 patients. J Neurosurg 90(1):42-49, 1999 16. Stafford SL, Pollock BE, Leavitt JA, et al: A study on the radiation tolerance of the optic nerves and chiasm after stereotactic radiosurgery. Int J Radiat Oncol Biol Phys 55(5):1177-1181, 2003 17. Leavitt JA, Stafford SL, Link MJ, et al: Long-term evaluation of radiationinduced optic neuropathy after single-fraction stereotactic radiosurgery. Int J Radiat Oncol Biol Phys 87(3):524-527, 2013 18. Grant RA, Whicker M, Lleva R, et al: Efficacy and safety of higher dose stereotactic radiosurgery for functional pituitary adenomas: A preliminary report. World Neurosurg 82(1-2):195-201, 2014 19. Cifarelli CP, Schlesinger DJ, Sheehan JP: Cranial nerve dysfunction following Gamma Knife surgery for pituitary adenomas: Long-term incidence and risk factors. J Neurosurg 116(6):1304-1310, 2012 20. Hasegawa T, Kobayashi T, Kida Y: Tolerance of the optic apparatus in single-fraction irradiation using stereotactic radiosurgery: Evaluation in 100 patients with craniopharyngioma. Neurosurgery 66(4):688-694, 2010; [discussion 694-685] 21. Hauptman JS, Barkhoudarian G, Safaee M, et al: Challenges in linear accelerator radiotherapy for chordomas and chondrosarcomas of the skull
22.
23.
24.
25. 26.
27.
28.
29.
30. 31.
32.
33.
34. 35.
36.
37.
38.
base: Focus on complications. Int J Radiat Oncol Biol Phys 83(2):542-551, 2012 Pollock BE, Link MJ, Leavitt JA, et al: Dose-volume analysis of radiationinduced optic neuropathy after single-fraction stereotactic radiosurgery. Neurosurgery 75(4):456-460, 2014; [discussion 460] Adler Jr JR, Gibbs IC, Puataweepong P, et al: Visual field preservation after multisession cyberknife radiosurgery for perioptic lesions. Neurosurgery 59(2):244-254, 2006; [discussion 244-254] Kim JW, Im YS, Nam DH, et al: Preliminary report of multisession gamma knife radiosurgery for benign perioptic lesions: Visual outcome in 22 patients. J Korean Neurosurg Soc 44(2):67-71, 2008 Jo KI, Im YS, Kong DS, et al: Multisession Gamma Knife surgery for benign orbital tumors. J Neurosurg 117(suppl):102-107, 2012 Killory BD, Kresl JJ, Wait SD, et al: Hypofractionated CyberKnife radiosurgery for perichiasmatic pituitary adenomas: Early results. Neurosurgery 64(suppl 2):A19-A25, 2009 Liao HI, Wang CC, Wei KC, et al: Fractionated stereotactic radiosurgery using the Novalis system for the management of pituitary adenomas close to the optic apparatus. J Clin Neurosci 21(1):111-115, 2014 Iwata H, Tatewaki K, Inoue M, et al: Single and hypofractionated stereotactic radiotherapy with CyberKnife for craniopharyngioma. J Neurooncol 106(3):571-577, 2012 Iwata H, Sato K, Tatewaki K, et al: Hypofractionated stereotactic radiotherapy with CyberKnife for nonfunctioning pituitary adenoma: High local control with low toxicity. Neuro Oncol 13(8):916-922, 2011 Lyman JT: Complication probability as assessed from dose-volume histograms.. Radiat Res Suppl 8:S13-S19, 1985 Jackson A, Ten Haken RK, Robertson JM, et al: Analysis of clinical complication data for radiation hepatitis using a parallel architecture model. Int J Radiat Oncol Biol Phys 31(4):883-891, 1995 Levegrun S, Jackson A, Zelefsky MJ, et al: Fitting tumor control probability models to biopsy outcome after three-dimensional conformal radiation therapy of prostate cancer: Pitfalls in deducing radiobiologic parameters for tumors from clinical data. Int J Radiat Oncol Biol Phys 51 (4):1064-1080, 2001 Timmerman RD: An overview of hypofractionation and introduction to this issue of seminars in radiation oncology. Semin Radiat Oncol 18 (4):215-222, 2008 Mould RFS, RA, Bucholz RD, Gagnon GJ, et al: Robotic Radiosurgery. Sunnyvale, CA: CyberKnife Society Press, 2005 Adler JR, Murovic J, Liao YJ, et al: Extreme tolerance of the optic nerve to ionizing radiation: A case report revealing the role of the dose-volume effect. Cureus 6(7):e187, 2014. [DOI 110.7759/cureus.7187] Minniti G, Esposito V, Clarke E, et al: Fractionated stereotactic radiosurgery for patients with skull base metastases from systemic cancer involving the anterior visual pathway. Radiat Oncol 9:110, 2014 Deng X, Yang Z, Liu R, et al: The maximum tolerated dose of gamma radiation to the optic nerve during gamma knife radiosurgery in an animal study. Stereotact Funct Neurosurg 91(2):79-91, 2013 Leber KA, Bergloff J, Langmann G, et al: Radiation sensitivity of visual and oculomotor pathways. Stereotact Funct Neurosurg 64(suppl 1):233-238, 1995
*Department of Radiation Oncology, Stanford University, Stanford, CA. †Department of Radiation Oncology, Acibadem University, Istanbul, Turkey. ‡Department of Ophthalmology, Stanford University, Stanford, CA. §Department of Radiology, Stanford University, Stanford, CA. ║Department of Neurosurgery, Stanford University, Stanford, CA. Conflict of interest: none. Presented in part at the Radiosurgery Society SRS/SBRT Annual Meeting, May 2014. Address reprint requests to Scott G. Soltys, MD, Department of Radiation Oncology, Stanford Cancer Center, 875 Blake Wilbur Drive, Stanford, CA 94305-5847. E-mail: [email protected]
http://dx.doi.org/10.1016/j.semradonc.2015.11.008 1053-4296/& 2016 Elsevier Inc. All rights reserved.
97
S.M. Hiniker et al
98 25 Gy in 5 fractions from AAPM Task Group 101 all had less than 1% risk. In 262 patients with perioptic tumors treated with SRS, we found a risk of optic complications of less than 1%. These data support previously unvalidated estimates as safe guidelines, which may in fact underestimate the tolerance of the optic structures, particularly in patients without prior radiation. Further investigation would refine the estimated normal tissue complication probability for SRS near the optic apparatus. Semin Radiat Oncol 26:97-104 C 2016 Elsevier Inc. All rights reserved.
S
tereotactic radiosurgery (SRS) is a noninvasive, highly accurate form of radiation therapy that is increasingly used to treat benign and malignant intracranial tumors with high rates of local control. Recent large studies of patients treated with SRS for pituitary adenomas and meningiomas have shown long-term disease control of over 95% with reduced toxicity as compared to traditional external beam radiation therapy (EBRT).1,2 However, patients with tumors located near the optic nerves or optic chiasm (anterior optic apparatus) have been frequently excluded from SRS owing to concerns for visual toxicity. For tumors within 3 mm of the optic structures (“perioptic tumors”), single-fraction SRS risks vision loss either from radiation-induced optic neuropathy (RION) if the optic apparatus dose is too high, or from tumor progression if the tumor dose is too low to yield control.3,4 Historically, the SRS maximum dose (Dmax) limit to the optic pathway has been cited as 8 Gy in a single fraction. For single-fraction doses necessary to control benign tumors (13-16 Gy), a risk of blindness as high as 27% has been reported.5 Given these potential risks of singlefraction SRS (“SRS”), we have typically treated these perioptic tumors with hypofractionated SRS (“fSRS”) over 2-5 days, although evidence-based dose constraints for fSRS have not yet been reported. Here we summarize the literature on dose constraints for the optic structures and report our institutional experience with over 250 patients treated with singlefraction and hypofractionated SRS for perioptic tumors.
Optic Structure Tolerance to Radiation Optic apparatus dose constraints have been best defined for conventionally fractionated EBRT. The Emami data for EBRT dose causing a 5% risk of toxicity at 5 years (TD 5/5) for the chiasm and optic nerves is 50 Gy, and the 50% risk tolerance dose (TD 50/5) is 65 Gy.6 Of note, these estimates are for whole organ irradiation, and reports suggest they may be conservative, particularly in the setting of partial-organ dose. More recently, QUANTEC data estimates the risk of toxicity for optic nerve or chiasm maximum point dose (Dmax) of o55 Gy at o3%, for 55-60 Gy of 3%-7%, and for 460 Gy of 47%20%.7 Among 131 patients treated with EBRT at the University of Florida, no RION occurred at maximum doses at or below 59 Gy. The 15-year risk of RION was 11% with doses above 60 Gy with dose fractions of less than 1.9 Gy, and 47% with
doses above 60 Gy when fraction size was greater than or equal to 1.9 Gy.8 Other studies provide normal tissue complication probability (NTCP) calculations for optic structures, including a report of 39 patients treated for advanced paranasal sinus tumors, of which 13 patients were treated without optic nerve or chiasm sparing. Of these 13 patients, 3 patients experienced moderate or severe vision loss, and all 3 with vision loss received maximum dose to the optic apparatus of at least 64 Gy.9 Although hyperfractionation may reduce the risk of vision injury with EBRT, maximum dose to the optic apparatus best correlates with optic neuropathy.10
Single-Fraction SRS for Perioptic Tumors Early Studies Early studies of RION after SRS supported strict dose constraints for the optic structures. Among 62 patients treated with single-fraction SRS for cavernous sinus meningiomas, 17 who received an optic apparatus Dmax of 48 Gy developed visual complications; a single-fraction optic apparatus limit of 8 Gy was proposed.11 Another early report described 4 patients with RION after an SRS Dmax ranging from 7-14 Gy.12 Leber et al5 reported no cases of RION for a single-fraction dose of less than 10 Gy, but they reported a risk of 27% for 10-15 Gy and 78% for 415 Gy. Given these early reported complications, initial single-fraction SRS dose constraints for the optic apparatus were conservative at a Dmax of less than 8 Gy.
Dose Escalation Despite an early recommended optic apparatus Dmax of 8 Gy, multiple groups subsequently reported favorable vision outcomes with higher single-fraction doses. At the University of Maryland, 20 patients received an average Dmax of 9 Gy, with none experiencing RION with serial visual field testing.13 In a report from Norway of 100 patients treated for cavernous sinus meningiomas, 1 patient, with a visual pathway Dmax of 8.6 Gy, experienced RION, consistent with a 1% risk.14 Mayo Clinic reported on 88 patients treated with SRS for skull base meningiomas, with no RION at a median Dmax of 10 Gy (range: 1-16 Gy).15 These investigators later reported a rate of 1.1% for patients receiving up to 12 Gy to the optic apparatus.16 This series was later updated with a reported risk of RION of o1% at a maximum dose of o12 Gy.17 In this
Tolerance of the visual pathway to radiosurgery largest report to date, of 222 patients treated with SRS for perioptic tumors, 1 developed unilateral blindness after receiving a Dmax of 12.8 Gy. The above reports were primarily in patients treated with SRS for benign meningiomas or nonfunctioning pituitary adenomas. SRS doses are typically higher when treating functioning pituitary adenomas, yielding additional data for RION. A report from Yale described 31 patients treated with a marginal tumor dose of 35 Gy for functional pituitary adenomas with an optic apparatus Dmax of up to 13.8 Gy. One patient developed unilateral optic nerve pallor and temporal field defect 3 years after radiation, with a nerve Dmax of 7.4 Gy.18 The University of Virginia reported outcomes in 217 patients treated with GammaKnife radiosurgery (GKS) for recurrent pituitary adenoma, 131 of which were secretory, treated with median tumor dose of 23 Gy. In all, 3 patients (2%) had RION, although the specific optic apparatus Dmax for these patients was not provided. Although an increased number of isocenters was statistically related to RION, Dmax was not correlated.19 Other reports that suggest safety of single-fraction optic apparatus SRS doses greater than 12 Gy are limited by incomplete dosimetric reporting and patient heterogeneity. In a series of 100 patients treated with SRS for craniopharyngioma, Hasegawa et al reported 3 patients with RION. Of the 3 patients, 2 received 15 Gy to the optic apparatus, with the third treated with multiple courses of irradiation. The authors argue that their data potentially supports maximum doses of up to 14 Gy to a partial segment of the optic apparatus.20 Conversely, unilateral blindness was seen in a patient treated for chordoma following 14.8 Gy in a single fraction to the optic apparatus.21 The most detailed dose-volume analysis data for singlefraction SRS to date was reported in 2014 by Pollock et al, in which the risk of RION for single-fraction SRS for pituitary adenomas was calculated. Among 133 treated patients who had not received previous irradiation, with a total of 266 anterior visual sides analyzed, 29 anterior visual pathway sides (11%) received a maximum dose of greater than 12 Gy, and no RION occurred, with a 95% CI risk of optic neuropathy of 0%13.9% at 12 Gy.22 In a recent review of radiation dose-volume effects of the optic structures, QUANTEC threshold limits for increased toxicity are 60 Gy in 1.8 Gy/fraction, and 12 Gy for single-fraction SRS.4
Multiple Fraction SRS for Perioptic Tumors Dose constraints for hypofractionated SRS over 2-5 days for perioptic tumors have not been well described, though it is recognized that hypofractionation may reduce the risk of normal tissue toxicity. In our and other institutions, if the single-fraction optic apparatus Dmax is greater than the 1012 Gy reported above, fSRS over 2-5 days has been investigated. A prior series of 49 patients with perioptic tumors treated with fSRS at our institution showed excellent tumor control and visual field preservation at mean follow-up of 49 months, with 94% tumor control and 1 patient (2%)
99 experiencing RION.23 However, detailed optic apparatus dosimetry was not reported. Kim et al reported a Korean experience evaluating multisession GKS for perioptic tumors with tumor control in 21 of 22 patients. Vision deteriorated in the one patient with tumor progression, with no other complications.24 However, no detailed dosimetric analysis was reported. In a series of 23 patients treated with multisession GKS for perioptic tumors, Jo et al reported 12 with stable vision and 11 with improvement in visual fields or acuity or both. Patients were treated to a marginal dose of 22 Gy in 4 fractions, with optic nerve maximum dose ranging from 12.8-20.8 Gy.25 In another series of 20 patients treated with SRS for perichiasmatic pituitary adenomas who received 25 Gy in 5 fractions, vision remained intact in all patients and improved in 3 patients.26 Median maximum chiasm dose was 23.3 Gy (range: 18.3-25.1 Gy). Similarly, excellent results were reported in a series of patients treated for pituitary adenomas close to the optic apparatus to 21 Gy in 3 fractions, with dose to the optic apparatus reported as mean dose of 16.7 Gy to the nerve and 14.6 Gy to the chiasm.27 In a report from Japan of the treatment of craniopharyngioma with singlefraction or hypofractionated SRS, 43 patients received a median single-fraction marginal dose of 14.3 Gy, and median multiple fraction marginal dose of 21 Gy in 3 fractions or 25 Gy in 5 fractions. No patient experienced RION, though optic apparatus dosimetry was not reported.28 This group has also reported similar excellent outcomes using hypofractionated SRS for nonfunctioning pituitary adenomas with minimal vision toxicity, with median optic nerve maximum dose of 19.9 Gy in 5 fractions, and optic chiasm maximum dose of 20.3 Gy (range: 1.4-25.0 Gy).29 However, detailed dosimetric analysis has not yet been reported for hypofractionated SRS.
Dose-Response Model for Hypofractionated SRS As above, appropriate dose constraints for fSRS remain poorly defined. There have been no reports to date detailing the dosimetric parameters for hypofractionated SRS that yield long-term tumor control as well as acceptable toxicity to the optic pathway. In the present study, we report the tumor control and vision outcomes of 262 patients treated with single and hypofractionated SRS for perioptic tumors, specifically identifying dose-volume parameters for the optic structures in patients treated with single-fraction and hypofractionated SRS. We sought to construct a dose-response model for visual pathway tolerance to SRS delivered in 1-5 fractions.
Methods and Materials Patients We conducted a retrospective review of a prospectively maintained list of patients, identifying those with perioptic tumors within 3 mm of the optic nerves or chiasm, treated with radiosurgery at Stanford University from 2000-2013. This study was approved by the Stanford University Institutional
S.M. Hiniker et al
100 Review Board. Eligible patients included those having tumors within 3 mm of the optic apparatus treated with SRS inclusive of all histologies including benign and malignant disease. We identified 326 patients treated with SRS for perioptic tumors. We excluded patients for whom clinical follow-up was not available or for whom dose-volume histogram (DVH) data could not be retrieved because of software incompatibility between machine upgrades, yielding 262 evaluable patients. The median age was 53 years (range: 14-86 years). Among all, 62% of patients were women. Metastases or malignant disease comprised 26 cases, the remaining were benign tumors: meningiomas (132), secreting (25) and nonsecreting (43) pituitary adenomas, craniopharyngiomas (15), schwannomas (11); arteriovenous malformations, hemangioblastomas, hemangiopericytomas, chordomas (2 each); choroid plexus papilloma and hamartoma (1 each). A total of 154 patients (59%) had undergone at least 1 surgical resection prior to SRS, including craniotomy or trans-sphenoidal resection. A total of 34 patients (13%) had been treated previously with radiation therapy to the lesion in question, and of these, 27 patients were treated with EBRT, and 7 patients with SRS.
Treatment Details Radiosurgical planning was performed using thin-slice, highresolution computed tomography (CT) scans, with intravenous contrast unless there was a contraindication. Thinsection magnetic resonance imaging (MRI) scans were also performed in the majority of patients, and MRI-CT fusion was performed. The treating physician contoured both the anterior visual pathways and target tumor. As described previously,23 the standard inverse planning CyberKnife method was used for planning treatment. Dose selection was determined by the attending neurosurgeon and radiation oncologist and was based on tumor volume, histology, proximity to optic structures, and previous history of radiation. Radiosurgery was delivered over 1-5 fractions, separated by at least 12 hours. The median prescribed tumor dose was 18 Gy (range: 12-25 Gy) in 1 fraction (n ¼ 47), 24 Gy (range: 18-33 Gy) in 3 fractions (n ¼ 111), and 25 Gy (range: 18-40 Gy) in 5 fractions (n ¼ 66). The median Dmax to the optic nerve was 7.6 Gy (5-95 percentile range: 1.9-12.4 Gy) in 1 fraction, 13.4 Gy (2.7-23.3 Gy) in 3 fractions, and 19.6 Gy (3.8-29.4 Gy) in 5 fractions. Median values for other parameters in all fractionations are given in Tables e1-e2 of the Appendix A.
(CTCAE) grade 4 vision loss, corresponding to vision of 20/ 200 or worse in the affected eye. Response rates were classified as the presence or absence of local relapse. Local control was defined as decreased size or stability of the treated tumor. Local failure was classified as tumor growth or the need for subsequent treatment with surgery or repeat irradiation.
Analysis Time to event was measured from the date of SRS or fSRS. From the plan DVHs, a maximum-likelihood parameter fitting of the probit dose-response model was performed using DVH Evaluator software (DiversiLabs, LLC, Huntingdon Valley, PA). The 68% CIs were calculated using the profile likelihood method. The median number of fractions was 3; all doses were converted to 3-fraction equivalent doses with the linear quadratic model using α/β ¼ 2 Gy for the optic structures prior to the modeling. The following dose descriptors of the optic apparatus were analyzed: Dmin, Dmedian, Dmean, Dmax, V30 Gy, V25 Gy, V20 Gy, V15 Gy, V10 Gy, V5 Gy, D50%, D10%, D5%, D1%, D1 cc, D0.50 cc, D0.25 cc, D0.20 cc, D0.10 cc, D0.05 cc, D0.03 cc. Dose-response modeling was accomplished according to the flowchart in Figure 1. DVH data were loaded into the DVH Evaluator software and numerous dose descriptors Dx were extracted, where x is a specified volume and Dx is the DVH dosevolume point corresponding to the volume x. To ensure a relevant metric, it was required that 95% or more of the cases had a nonzero value for the selected Dx. For example, fewer than 95% of the cases had optic nerves exceeding a volume of 0.5 cc, so the D0.5 cc was not used. All Dx values were then converted to 3fraction equivalent doses using the linear quadratic model with α/ β ¼ 2 Gy for normal tissue tolerance. The probit dose-response model was used,30 with maximum-likelihood parameter fitting,31 with CIs via the profile likelihood method.32 Our model was compared to the previously unvalidated Dmax constraints of 12 Gy in 1 fraction from QUANTEC, 19.5 Gy in 3 fractions from Timmermann33, and 25 Gy in 5 fractions from AAPM Task Group 101, and all constraints had less than 1% risk.3,7,33
Results Tumor Control and Vision Outcomes Median follow-up was 36.8 months (range: 2-142 months). Among 262 patients, 19 experienced local progression of
Patient Follow-Up Patients were followed with clinical examination, brain imaging, and annual visual field testing. For benign tumors, followup is at 6 and 12 months post-SRS, then annually. For malignant tumors, follow-up is typically every 3 months for 1 year, then every 4 months. Evaluation of treatment response was assessed by clinical examination and radiographic studies. Evaluation of visual outcomes was assessed by formal visual field testing (performed by an ophthalmologist approximately yearly, when available) and clinical vision examination. RION was defined as common terminology criteria for adverse events
Figure 1 Flowchart of dose-response modeling process. (Color version of figure is available online.)
Tolerance of the visual pathway to radiosurgery
101 appeared stable, the patient underwent surgical decompression of the optic apparatus, without improvement, and the patient lost vision in that eye. In the second case of RION, the patient had 2 courses of irradiation previously: EBRT and SRS (20 Gy in 1 fraction), and surgery to a right parasellar grade I meningioma, measuring 14.33 cc. The patient was retreated to 25 Gy in 5 fractions, prescribed to the 78% isodose line; the maximum dose to the optic pathway was 27.7 Gy for the third course. Seven months after radiosurgery, the patient experienced gradual onset of vision loss in the right eye that became complete, attributed to optic neuropathy. For each patient, 3 critical structures could potentially cause the complication: the 2 optic nerves and the chiasm. Rather than combining all into a single optic pathway structure, an attempt was made to analyze the components separately, as the vision loss in both cases was unilateral. We used the structure receiving the highest dose as the likeliest to be responsible for vision loss, which was the optic nerve in the majority of patients. In the patient who experienced RION who had been previously treated with 2 courses of irradiation, the exact composite treatment plan for all courses was not available. The chiasm dose was the highest optic structure dose in the final treatment plan prior to the development of RION, so this was the dose used in the analysis. We noted that given the 2 courses of previous treatment, the dose used is likely to underestimate the total dose responsible for causing RION in this patient.
Normal Tissue Complication Probability
Figure 2 Optic nerve dose-response for small volumes: (A) D0.2 cc and (B) Dmax. AE, adverse event; MLE, maximum likelihood estimate; NfxED, N-fraction equivalent dose. (Color version of figure is available online.)
disease after treatment (7.3%). Of 26 patients with malignant disease, 6 (23%) experienced local failure, and 13 of 236 (5.5%) patients with benign tumors had local failure. Among the 86 patients with vision deficits prior to radiosurgery (33%), we found that 13% had improvement in vision after treatment. In all, 7 (2.7%) patients had worsening of vision following treatment; this was attributed to unequivocal tumor growth causing optic apparatus compression in 5 (1.9%) patients. A total of 2 patients (0.8%) experienced unilateral vision loss in the absence of progressive disease. The first patient was treated with 25 Gy in 5 fractions for a 7.84 cm3 sellar meningioma encasing the left optic nerve. The maximum dose to the optic nerve was 23.9 Gy. The patient experienced gradual onset of vision loss in the left eye 3 months after radiosurgery, and MRI revealed enhancement along the left optic nerve suspicious for RION. Though the tumor size
Graphs of the resulting NTCP curves for the optic nerve are shown in Figure 2, for D0.2 cc and Dmax, results for D50% and D10% are shown in the Appendix A, and model parameters are in Table 1. In the figures, the red squares represent the 2 cases with optic neuropathy, and the blue dots represent the doses for the non-complication cases. CIs are shown as the dashed green lines, which are reasonably tight in the clinical region where data exists, and diverge above these doses. Figure 2 (B) shows that for Dmax at 40 Gy in 3 fractions, the risk ranges from 5%-65%, whereas at 20 Gy, the CIs only range ⫾ 1% from the model estimate. This method of CI shows where the data are sparse and where the estimates mature. Dotted red lines intersect the NTCP curves at 1%, 2%, and 5% risk levels for visualization of the corresponding Dx values. The risk level of previously published dose tolerance limits was evaluated and is denoted in the second number in each cell of the DVH Risk Map (Fig. 3). Specifically, the QUANTEC single-fraction Dmax limit of 12 Gy has a risk level of 0.7% from this dataset, as shown in the upper-right cell of the table in Table 1 Optic Nerve Probit Model Parameters With 68% CI Dose Descriptor
TD50, Gy (68% CI)
m (68% CI)
D50% D10% D0.2cc D0.05cc Dmax
46.81 (29.56-242.84) 45.19 (32.98-93.99) 55.66 (34.78-434.24) 45.04 (33.25-91.00) 48.91 (37.12-94.44)
0.3387 (0.2950-0.3858) 0.2830 (0.2296-0.3450) 0.3361 (0.2907-0.3870) 0.2789 (0.2250-0.3416) 0.2459 (0.1878-0.3216)
S.M. Hiniker et al
102
Figure 3 DVH Risk Map showing published dose tolerance limits organized into high-risk (solid red line) and low-risk (dashed green line) categories. Estimates of risk for each selected limit are shown to the right of each cell in the table. Bold limits in the table indicate published limits, and italicized limits indicate new limits from the trendlines and from modeled estimates of risk. (Color version of figure is available online.)
model estimated a o1% risk of RION at 12 Gy in 1, 19.5 Gy in 3, and 25 Gy in 5 fractions.
Figure 3 and depicted graphically in the rightmost subplot. In the DVH Risk Map, each dose descriptor (D50%, D10%, D0.2cc, D0.05cc, and Dmax) is shown in a subplot, and all corresponding published dose tolerance limits are plotted as blue diamonds. For each dose descriptor in each fractionation, a reasonably high set of published limits were selected as high-risk limits and are labeled and plotted with a red trend line. At a lower dose, another set of low-risk limits were selected and labeled and were plotted with a dotted green trend line. From the model of the clinical dataset, the estimated risk levels for every dose tolerance limit along both high-risk and low-risk trend lines are displayed quantitatively in the tabular section of Figure 3. Estimated risk levels for optic nerve Dmax levels in 1-5 fractions are shown in Tables 2 and 3. To summarize, the
Discussion In this study of tolerance of the optic apparatus to singlefraction and hypofractionated SRS, we sought to review the literature and to better define the dose-volume parameters for the risk of RION. We report one of the largest studies to date examining vision outcomes after single-fraction and hypofractionated SRS for perioptic tumors. In this cohort of 262 perioptic tumors treated with SRS, there was a low rate of optic complications. Of the 262
Table 2 Estimated RION Risk Level of Common Dmax Levels for Optic Nerve One Fraction
Three Fractions
Five Fractions
Dmax (Gy)
Estimated Risk (%)
Dmax (Gy)
Estimated Risk (%)
Dmax (Gy)
Estimated Risk (%)
8 10 12 14 15
0.1 0.3 0.7 1.6 2.3
16 18 20 22 24
0.3 0.5 0.8 1.3 1.9
20 22.5 25 27.5 30
0.3 0.5 0.8 1.3 1.9
Tolerance of the visual pathway to radiosurgery
103
Table 3 Optic Nerve Dmax Values Corresponding to 1%, 2%, 3%, and 5% Risk of RION Number of Fractions
Dmax for Dmax for Dmax for Dmax for 1% Risk 2% Risk 3% Risk 5% Risk (Gy) (Gy) (Gy) (Gy)
1 2 3 4 5
12.7 17.5 20.9 23.7 26.1
14.6 20.2 24.2 27.5 30.3
15.9 21.9 26.3 29.9 32.9
17.5 24.2 29.1 33.1 36.6
analyzed patients, 2 (0.8%) experienced common terminology criteria for adverse events grade 4 vision loss in one eye, defined as vision of 20/200 or worse in the affected eye. Both cases were meningiomas treated with 25 Gy in 5 fractions, with optic pathway Dmax of 19.2 Gy and 22.2 Gy, respectively, in 3fraction equivalent doses. In one of these cases, the patient had received 2 previous courses of radiotherapy to this area. A model of these complications estimated the risk of vision injury for the previously unvalidated estimates of optic pathway tolerance, including a 0.2% risk for the estimated 3-fraction optic pathway constraint of Dmax r 15 Gy, and 0.7% risk for the estimated constraint of Dmax r 19.5 Gy.33,34 These data support the previously unvalidated estimates as safe guidelines, which may in fact underestimate the tolerance of the optic structures, particularly in patients without prior radiation. Patients were twice as likely to experience vision loss related to local progression of disease as compared to RION, and therefore it may be riskier to vision to adhere to dose constraints that may in fact be overly conservative rather than adopting more aggressive constraints to achieve tumor control.35 Our data suggest that perioptic tumors can be safely treated with SRS delivered over 1-5 fractions with a low risk of RION even with optic apparatus doses above the previously estimated constraints. We included patients with metastatic and malignant disease as long as they lived for at least 6 months, whereas most reports of RION for perioptic tumors focus on benign tumors. Minniti et al36 also report outcomes after fSRS for patients with skull base metastases involving the anterior visual pathway, and found a 2-year local control of 72%, with no RION and 51% of patients having improvement of preexisting cranial nerve deficits. In addition, we included patients with recurrent disease and 34 (13%) of our patients had received previous irradiation to the lesion in question, putting them at higher risk for RION. The predictors of optic neuropathy after radiation have not been well described. RION typically occurs within 3 years of treatment.4,12 The pretreatment condition of the optic nerve, even in those patients who have not received previous radiation, is likely to affect posttreatment vision outcomes. Animal studies revealed that a compressed optic nerve is more sensitive to SRS-induced RION than are noncompressed nerves. In cats, the minimum dose causing optic neuropathy is 12 Gy in noncompressed nerves, and 11 Gy in compressed nerves.37 Others have found that prior surgery or anterior visual pathway dysfunction may increase the risk of RION.38
One of our 2 patients with RION had surgery before, and this patient also had anterior visual pathway dysfunction prior to SRS. Other factors that may increase the risk of SRS toxicity include diabetes, hypertension, and collagen vascular disorders. Similar to others and not surprisingly, we found a higher rate of vision toxicity in patients undergoing retreatment with radiation (SRS or EBRT), with 1 of 34 retreated patients developing optic neuropathy, vs 1 of 228 previously untreated patients.16,19 Contouring the optic chiasm can be a challenge as it is a very small structure not well seen on CT scans. MRIs are beneficial in delineating the chiasm, but the distinction between chiasm and optic nerves may be difficult to determine and to some extent is arbitrary. Although absolute volumes are preferable in most situations, the small size of the optic structures and variability in contouring argues for a role of relative volume use in this setting. Similar to others, we found that the maximum dose to the optic apparatus was the most important predictor of toxicity, so volumetric constraints may be less important. The retrospective nature of our study is a notable limitation. Nevertheless, these results support the use of hypofractionated SRS for perioptic tumors, similar to other studies, and provide dose-volume constraints in a large cohort of patients. Further investigation is needed to refine the estimated NTCP for SRS near the optic apparatus.
Conclusions In summary, our data suggest that perioptic tumors can be safely treated with single-fraction and hypofractionated SRS with excellent local control and a low rate of vision complications. Our data provide support for the previously unvalidated estimates of hypofractionated dose constraints for the optic structures as safe guidelines, which may underestimate the actual tolerance of the optic apparatus. Further investigation and delineation of the dose-volume parameters of SRS for tumors in close proximity to the optic structures would be important, particularly in the setting of reirradiation. Our model suggests a less than 1% incidence of RION in our group of patients treated with an optic apparatus pathway maximum point dose of 12 Gy in 1, 19.5 Gy in 3, and 25 Gy in 5 fractions.
Appendix. Supplementary data Supplementary data associated with this article can be found in the online version at http://dx.doi.org/10.1016/j.semradonc. 2015.11.008.
References 1. Sheehan JP, Niranjan A, Sheehan JM, et al: Stereotactic radiosurgery for pituitary adenomas: An intermediate review of its safety, efficacy, and role in the neurosurgical treatment armamentarium. J Neurosurg 102 (4):678-691, 2005 2. Pollock BE, Stafford SL, Link MJ, et al: Single-fraction radiosurgery of benign intracranial meningiomas. Neurosurgery 71(3):604-612, 2012; [discussion 613]
S.M. Hiniker et al
104 3. Benedict SH, Yenice KM, Followill D, et al: Stereotactic body radiation therapy: The report of AAPM Task Group 101. Med Phys 37 (8):4078-4101, 2010 4. Mayo C, Martel MK, Marks LB, et al: Radiation dose-volume effects of optic nerves and chiasm. Int J Radiat Oncol Biol Phys 76(suppl 3): S28-S35, 2010 5. Leber KA, Bergloff J, Pendl G: Dose-response tolerance of the visual pathways and cranial nerves of the cavernous sinus to stereotactic radiosurgery. J Neurosurg 88(1):43-50, 1998 6. Emami B, Lyman J, Brown A, et al: Tolerance of normal tissue to therapeutic irradiation. Int J Radiat Oncol Biol Phys 21(1):109-122, 1991 7. Marks LB, Yorke ED, Jackson A, et al: Use of normal tissue complication probability models in the clinic. Int J Radiat Oncol Biol Phys 76(suppl 3): S10-S19, 2010 8. Parsons JT, Bova FJ, Fitzgerald CR, et al: Radiation optic neuropathy after megavoltage external-beam irradiation: Analysis of time-dose factors. Int J Radiat Oncol Biol Phys 30(4):755-763, 1994 9. Martel MK, Sandler HM, Cornblath WT, et al: Dose-volume complication analysis for visual pathway structures of patients with advanced paranasal sinus tumors. Int J Radiat Oncol Biol Phys 38(2):273-284, 1997 10. Bhandare N, Monroe AT, Morris CG, et al: Does altered fractionation influence the risk of radiation-induced optic neuropathy? Int J Radiat Oncol Biol Phys 62(4):1070-1077, 2005 11. Tishler RB, Loeffler JS, Lunsford LD, et al: Tolerance of cranial nerves of the cavernous sinus to radiosurgery. Int J Radiat Oncol Biol Phys 27 (2):215-221, 1993 12. Girkin CA, Comey CH, Lunsford LD, et al: Radiation optic neuropathy after stereotactic radiosurgery. Ophthalmology 104(10):1634-1643, 1997 13. Ove R, Kelman S, Amin PP, et al: Preservation of visual fields after perisellar gamma-knife radiosurgery. Int J Cancer 90(6):343-350, 2000 14. Skeie BS, Enger PO, Skeie GO, et al: Gamma knife surgery of meningiomas involving the cavernous sinus: Long-term follow-up of 100 patients. Neurosurgery 66(4):661-668, 2010; [discussion 668-669] 15. Morita A, Coffey RJ, Foote RL, et al: Risk of injury to cranial nerves after gamma knife radiosurgery for skull base meningiomas: Experience in 88 patients. J Neurosurg 90(1):42-49, 1999 16. Stafford SL, Pollock BE, Leavitt JA, et al: A study on the radiation tolerance of the optic nerves and chiasm after stereotactic radiosurgery. Int J Radiat Oncol Biol Phys 55(5):1177-1181, 2003 17. Leavitt JA, Stafford SL, Link MJ, et al: Long-term evaluation of radiationinduced optic neuropathy after single-fraction stereotactic radiosurgery. Int J Radiat Oncol Biol Phys 87(3):524-527, 2013 18. Grant RA, Whicker M, Lleva R, et al: Efficacy and safety of higher dose stereotactic radiosurgery for functional pituitary adenomas: A preliminary report. World Neurosurg 82(1-2):195-201, 2014 19. Cifarelli CP, Schlesinger DJ, Sheehan JP: Cranial nerve dysfunction following Gamma Knife surgery for pituitary adenomas: Long-term incidence and risk factors. J Neurosurg 116(6):1304-1310, 2012 20. Hasegawa T, Kobayashi T, Kida Y: Tolerance of the optic apparatus in single-fraction irradiation using stereotactic radiosurgery: Evaluation in 100 patients with craniopharyngioma. Neurosurgery 66(4):688-694, 2010; [discussion 694-685] 21. Hauptman JS, Barkhoudarian G, Safaee M, et al: Challenges in linear accelerator radiotherapy for chordomas and chondrosarcomas of the skull
22.
23.
24.
25. 26.
27.
28.
29.
30. 31.
32.
33.
34. 35.
36.
37.
38.
base: Focus on complications. Int J Radiat Oncol Biol Phys 83(2):542-551, 2012 Pollock BE, Link MJ, Leavitt JA, et al: Dose-volume analysis of radiationinduced optic neuropathy after single-fraction stereotactic radiosurgery. Neurosurgery 75(4):456-460, 2014; [discussion 460] Adler Jr JR, Gibbs IC, Puataweepong P, et al: Visual field preservation after multisession cyberknife radiosurgery for perioptic lesions. Neurosurgery 59(2):244-254, 2006; [discussion 244-254] Kim JW, Im YS, Nam DH, et al: Preliminary report of multisession gamma knife radiosurgery for benign perioptic lesions: Visual outcome in 22 patients. J Korean Neurosurg Soc 44(2):67-71, 2008 Jo KI, Im YS, Kong DS, et al: Multisession Gamma Knife surgery for benign orbital tumors. J Neurosurg 117(suppl):102-107, 2012 Killory BD, Kresl JJ, Wait SD, et al: Hypofractionated CyberKnife radiosurgery for perichiasmatic pituitary adenomas: Early results. Neurosurgery 64(suppl 2):A19-A25, 2009 Liao HI, Wang CC, Wei KC, et al: Fractionated stereotactic radiosurgery using the Novalis system for the management of pituitary adenomas close to the optic apparatus. J Clin Neurosci 21(1):111-115, 2014 Iwata H, Tatewaki K, Inoue M, et al: Single and hypofractionated stereotactic radiotherapy with CyberKnife for craniopharyngioma. J Neurooncol 106(3):571-577, 2012 Iwata H, Sato K, Tatewaki K, et al: Hypofractionated stereotactic radiotherapy with CyberKnife for nonfunctioning pituitary adenoma: High local control with low toxicity. Neuro Oncol 13(8):916-922, 2011 Lyman JT: Complication probability as assessed from dose-volume histograms.. Radiat Res Suppl 8:S13-S19, 1985 Jackson A, Ten Haken RK, Robertson JM, et al: Analysis of clinical complication data for radiation hepatitis using a parallel architecture model. Int J Radiat Oncol Biol Phys 31(4):883-891, 1995 Levegrun S, Jackson A, Zelefsky MJ, et al: Fitting tumor control probability models to biopsy outcome after three-dimensional conformal radiation therapy of prostate cancer: Pitfalls in deducing radiobiologic parameters for tumors from clinical data. Int J Radiat Oncol Biol Phys 51 (4):1064-1080, 2001 Timmerman RD: An overview of hypofractionation and introduction to this issue of seminars in radiation oncology. Semin Radiat Oncol 18 (4):215-222, 2008 Mould RFS, RA, Bucholz RD, Gagnon GJ, et al: Robotic Radiosurgery. Sunnyvale, CA: CyberKnife Society Press, 2005 Adler JR, Murovic J, Liao YJ, et al: Extreme tolerance of the optic nerve to ionizing radiation: A case report revealing the role of the dose-volume effect. Cureus 6(7):e187, 2014. [DOI 110.7759/cureus.7187] Minniti G, Esposito V, Clarke E, et al: Fractionated stereotactic radiosurgery for patients with skull base metastases from systemic cancer involving the anterior visual pathway. Radiat Oncol 9:110, 2014 Deng X, Yang Z, Liu R, et al: The maximum tolerated dose of gamma radiation to the optic nerve during gamma knife radiosurgery in an animal study. Stereotact Funct Neurosurg 91(2):79-91, 2013 Leber KA, Bergloff J, Langmann G, et al: Radiation sensitivity of visual and oculomotor pathways. Stereotact Funct Neurosurg 64(suppl 1):233-238, 1995