- Email: [email protected]
A comparison of clinical and radiologic outcomes between frame-based and frameless stereotactic radiosurgery for brain metastases
A Comparison of Clinical and Radiologic Outcomes Between Frame-based and Frameless Stereotactic Radiosurgery for Brain Metastases Nathan R. Bennion MD, Timothy Malouff BS, MD, Vivek Verma, Kyle Denniston MD, Abhijeet Bhirud, Weining Zhen MD, Andrew Wahl MD, Chi Lin MD, PhD PII: DOI: Reference:
S1879-8500(16)30046-7 doi: 10.1016/j.prro.2016.05.001 PRRO 622
To appear in:
Practical Radiation Oncology
Received date: Revised date: Accepted date:
1 September 2015 26 April 2016 5 May 2016
Please cite this article as: Bennion Nathan R., Malouff Timothy, Verma Vivek, Denniston Kyle, Bhirud Abhijeet, Zhen Weining, Wahl Andrew, Lin Chi, A Comparison of Clinical and Radiologic Outcomes Between Frame-based and Frameless Stereotactic Radiosurgery for Brain Metastases, Practical Radiation Oncology (2016), doi: 10.1016/j.prro.2016.05.001
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
T
A Comparison of Clinical and Radiologic Outcomes Between Frame-based and Frameless Stereotactic Radiosurgery for Brain Metastases Nathan R. Bennion, MD, Timothy Malouff, BS, MD, Vivek Verma, Kyle Denniston, MD, Abhijeet Bhirud, Weining Zhen, MD, Andrew Wahl, MD, Chi Lin, MD PhD 1
RI P
Department of Radiation Oncology, Nebraska Medicine, University of Nebraska Medical Center Campus, Omaha, NE USA
SC
Key Words: stereotactic radiosurgery, frame, frameless, brain metastases
NU
Shortened Title: Frame versus Frameless stereotactic radiosurgery for brain metastases
Corresponding Author:
PT
ED
Nathan Bennion, MD Department of Radiation Oncology University of Nebraska Medical Center Omaha, NE
MA
Acknowledgments: None
AC
CE
Phone: (402)-552-3844 Fax: (402)-552-3844 Email: [email protected]
ACCEPTED MANUSCRIPT Abstract
RI P
T
Purpose/Objective(s): Modern experiences in stereotactic radiosurgery (SRS) report non-invasive frameless techniques as an effective alternative to frame-based SRS. Frameless techniques potentially increase positional uncertainty and planning target volume (PTV) margins are frequently employed. Here we compare rates of local control and radiation necrosis in frameless versus frame-based SRS.
SC
Materials/Methods: Ninety-eight patients (170 lesions) with radiologic and clinical follow-up were analyzed. Group 1 contained 34 patients (61 lesions) immobilized with an invasive stereotactic frame. Group 2 had 64 patients (109 lesions) immobilized with a frameless SRS mask. Patient, tumor, and treatment characteristics were recorded, as were intervals to local recurrence and radiation necrosis (asymptomatic and symptomatic).
ED
MA
NU
Results: Median patient age was 59 years (range 25 -89), and KPS was 80 (range 50-100). Median radiologic and clinical follow-up was 6.5 months (range 0.7-44.3) and 7 months (range 0.7-45.7). A median of 2 tumors were treated per course (range 1-5) with a median dose of 18 Gy (range 13-24 Gy). The median time to local failure was not reached, and Kaplan-Meier estimates of local failure were not statistically significant between groups (p=0.303). Actuarial 6-month local failure rates were 7.2% in Group 1 and 12.6% in Group 2 (p=0.295), with 12-month local failure rates of 14.5% and 26.8% (p=0.185), respectively. There was no statistically significant difference in symptomatic (p=0.391) or asymptomatic (p=0.149) radiation necrosis. Six-month radiation necrosis was 0% in Group 1 and 1.6% in Group 2 (p=0.311) with 12-month rates of 20.2% and 3.8%, respectively (p=0.059). Median time to necrosis was not reached in Group 1, but was 44 months in Group 2.
CE
INTRODUCTION
PT
Conclusion: Frameless SRS demonstrates clinical outcomes comparable to frame-based techniques with respect to local failure and radiation necrosis.
AC
Metastases account for more than one-half of brain tumors. With more sensitive imaging and improved therapy for extracranial disease, the incidence of metastatic brain disease is likely to rise.1,2 Treatment for these patients has historically been limited to whole brain radiation (WBRT) and corticosteroids; however, in our age of evolving therapy, treatment options have become more sophisticated. Advances in technology and prospective randomized trials have established stereotactic radiosurgery (SRS) as a safe means of treating intracranial tumors while minimizing radiation to normal brain parenchyma.3,4 Leksell first described this technique including the concept of “center of arc” targeting using an invasive head frame for rigid immobilization.5 This early work culminated with the introduction of the first Leksell Gamma Knife system in 1968. Clinicians adapted similar frame-based methods for linear accelerator-based radiosurgery.6 Though frame-based methods reliably immobilize patients, they also present several obstacles. Invasive fixation to the skull may be painful or provoke patient anxiety. It requires the patient to remain in the department for extended periods and places time constraints on the plan design and quality assurance in order to complete the simulation, planning, and delivery in the same day. Now, modern frameless techniques provide a non-invasive alternative for patient immobilization and repositioning when used in conjunction with proper image guidance and precision delivery.7 Frameless
ACCEPTED MANUSCRIPT
SC
RI P
T
techniques also provide more time for planning which can increase safety and flexibility. However, frameless SRS is suspected to increase set up uncertainty when compared with frame-based SRS, and may require expanded margins around the gross tumor volume (GTV).8 While expanded target volumes may compensate for set up error, larger treatment volumes may also correlate with higher rates of radiation necrosis.9,10 Rigidly interpreting these two obstacles in frameless SRS would lead to the assumption that increasing margins would result in excess rates of radiation necrosis, or that dismissing the need for additional margins would compromise local control. However, there is little clinical data comparing local control and radionecrosis rates for frame-based versus frameless techniques. We aim to compare the rates of local control and radiation necrosis in frameless versus frame-based SRS. PATIENTS/METHODS
NU
Frame-based SRS Procedure
CE
PT
ED
MA
Each patient was simulated supine with the head frame locked into the simulation table and thin-slice computed tomography (CT) was performed. Thin-slice T1 post-contrast (T1c) magnetic resonance imaging (MRI) and T2/T2 Fluid-attenuated inversion recovery sequences were reconstructed in the coronal, sagittal and axial planes as necessary and fused to the simulation CT. Organs at risk (OAR) were constructed manually. GTVs were contoured using T1c MRI sequences. PTV expansions were created at the discretion of the treating physician. Dose was prescribed to the PTV. Plans were evaluated for optimum coverage and conformity. Patients were positioned on the treatment table with rigid headframe fixation. Positioning was confirmed with orthogonal X-rays using the ExacTrac system (BrainLAB, Feldkirchen, Germany). After verifying patient positioning, treatment was delivered, and the head frame was removed. Frameless SRS Procedure
AC
Each patient was simulated in the supine position. Rigid head rests and radiosurgery specific thermoplastic masks were used for fixation with stereotactic box localization. A thin slice CT was performed in the treatment position, and patients were sent home. MRI fusions, target and OAR construction and treatment planning were as per outlined in the frame-based methods above, with the noted exception that simulation and treatment delivery were not conducted on the same day. On the day of treatment, patients were positioned on the treatment table with the thermoplastic stereotactic mask. Setup accuracy was confirmed for each patient prior to treatment with ExacTrac or kV cone beam CT image guidance. After verifying patient positioning, treatment was delivered and the stereotactic mask removed. Follow-up There was not a strict follow-up program for the entire cohort. However, it is the general practice in our institution to see patients 4-6 weeks after their initial treatment and every 3 months after that. Initial post-treatment MRIs were obtained between 2-4 months after SRS, then with subsequent follow-ups. Patients with signs or symptoms suggestive of disease progression or radiation necrosis were seen in clinic earlier with imaging when appropriate. Tumor control was determined by MRI in all patients. Our
ACCEPTED MANUSCRIPT
T
institutions MRI protocol included T2-weighted, T2/FLAIR, thin cut T1-weighted with and without contrast, diffusion and gradient echo sequences. Magnetic resonance perfusion and spectroscopy were ordered in cases where radiation necrosis is indistinguishable from local progression by standard imaging.
RI P
Data Collection
NU
SC
After obtaining IRB approval, records were reviewed for all patients who underwent SRS at our institution between 2008 and 2014. Patient, disease, and treatment characteristics are as stated in Table 1. We also recorded target volumes, PTV margins, and frame-based (Group 1) and frameless (Group 2) delivery. Last dates of radiologic and clinical follow-up were documented as were dates of local failure and radiation necrosis. Tumor Size
MA
Gross tumor (GTV) and planning target volumes (PTV) were calculated using the iPlan treatment planning system. Tumor Control
Radiation Necrosis
CE
PT
ED
Local failures were assessed from the date of treatment to the first radiologic evidence of progression. Complete response (CR) required radiographic disappearance of brain metastases. Greater than 50% reduction in the size of each lesion radiographically, using perpendicular diameters, was considered a partial response (PR). Up to 50% decrease and up to 25% increase in the size of each lesion radiographically, using perpendicular diameters was deemed stable disease (SD). Lastly, disease progression (DP) was deemed to be an increase of > 25% in the size of a lesion.11
AC
Radiologic radionecrosis was defined by characteristic changes on MRI in the absence of attributable symptoms. In cases where there was ambiguity in determining necrosis versus local tumor progression, serial scans aided in retrospectively categorizing the event. Biopsies were not required for event categorization. Radiographic, asymptomatic radiation necrosis was assessed separately from symptomatic radionecrosis. Time to radiation necrosis was defined as the interval from the date of treatment to the first radiologic evidence of radiation necrosis. Statistical Methods Tumor sizes, measured by GTV and PTV, were compared between each group with the Wilcoxon ranksum test. Tumor control, symptomatic radiation necrosis and asymptomatic radiation necrosis were all analyzed with the Kaplan-Meier method and log-rank test. Six- and 12-month rates of local failure and necrosis were compared using the Chi-square test. Univariate and multivariate analyses were conducted using the Cox proportional hazards model. All reported p-values are two-sided and are accompanied by their associated 95% confidence intervals (95% CI).
ACCEPTED MANUSCRIPT
RESULTS
MA
NU
SC
RI P
T
Ninety-eight patients with available patient, tumor and treatment information underwent SRS at our institution between 2008 and 2014, for a total of 170 treated lesions (Table 1). Median patient age was 59 years (range 25-89), and KPS was 80 (range, 50-100). A median of 2 tumors were treated per treatment course (range 1-5) with a median dose of 18 Gy (range 13-24). Thirty-four patients with a total of 61 lesions were immobilized with a stereotactic frame (Group 1) with a median radiologic follow-up of 7 months (range 6.1-29.8) and clinical follow-up of 8.7 months (range, 6.1-45.7). Sixty-four patients with 109 lesions immobilized with a non-invasive SRS-specific mask (Group 2) had a median radiologic follow-up of 5.6 months (range, 0.7-44.3) and clinical follow-up of 6.1 months (range, 0.7-44). Group 1 had a higher percentage of lesions treated early in the study period, with 28 of 61 (46%) of lesions treated between 2008 and 2010 and the remainder treated between 2011 and 2014. In Group 2, 19 of 109 (17%) of lesions were treated between 2008 and 2010, and 90 of 109 (83%) of lesions were treated between 2011 and 2014.
CE
PT
ED
The Wilcoxon rank-sum analysis showed no differences between the gross tumor volumes in each group. Group 1 had a median GTV of 0.54 cc (range 0.01-15.24) compared to 0.89 cc (range 0.01-13.41) in Group 2 (p=0.118). Thirty-one (51%) of 61 lesions in Group 1 had a PTV margin with a median radial expansion of 1 mm (range 0-3). In Group 2, 87% of the lesions had radial expansions with a median of 1 mm (range 0-3). PTVs were significantly larger in Group 2 with a median volume of 1 cc in Group 1 and 1.89 cc in Group 2 (p=0.016). The volume added by PTV expansion also significantly differed between Groups 1 and 2 with median increases of 0.05 cc and 0.66 cc, respectively (p=<0.001). The percent increase was also statistically greater in Group 2 compared to Group 1 (69% versus 23%) with a p-value of 0.008.
AC
Local failures were identified in 6 targets (10%) from Group 1 and 15 (14%) in Group 2. Median time to local failure was not reached in either group, and Kaplan-Meier estimates of time to local failure in Group 1 (95% CI, 30.7-38 mo) and Group 2 (95% CI, 29.1-39.2 mo) were not statistically different (p=0.303; Figure 1). Six-month estimated risk of local failures were 7.2% (95% CI, 0.4-14%) in Group 1 and 12.6% (95% CI, 5.1-20.1%) in Group 2 (p=0.295), with 12-month estimates of 14.5% (95% CI, 2.926.1%) and 26.8% (95% CI, 12.9-40.7%), respectively (p=0.185). The rates of complete response, partial response, stable disease and disease progression are presented in Table 2. Univariate analysis identified male gender (HR 2.4, 95% CI, 1.025-5.668, p=0.044) as a predictor of local control. It was not significant on multivariate analysis, however. Treatment group was not predictive of local control by either univariate or multivariate analysis with p-values of 0.153 and 0.691, respectively (Table 3). Radiographic, asymptomatic radiation necrosis was present in 11 (18%) of Group 1 lesions and 14 (13%) in Group 2. Kaplan-Meier analysis showed no significant difference between the two groups (p=0.391; Figure 2). Median time to asymptomatic necrosis was not reached in Group 1 and was 44.3 months in Group 2. The risk of developing asymptomatic radiation necrosis was 1.7% (95% CI, 0-5%) in Group 1 and
ACCEPTED MANUSCRIPT 3.1% (95% CI, 0-7.3%) in Group 2 at six months (p=0.609) and 37.5% (95% CI, 18.5-55.6%) in Group 1 and 16.5 (95% CI, 4.1-28.9%) at 12 months (p=0.073).
SC
RI P
T
Symptomatic necrosis was confirmed by MRI in 5 lesions (8%) from Group 1 and 6 lesions (6%) from Group 2. Kaplan-Meier analysis showed no significant difference between the two groups with a logrank p=0.149 (Figure 3). Once again, median time to symptomatic necrosis was not reached in Group 1 and was 44.3 months in Group 2. The risk of developing symptomatic necrosis was 0% (95% CI, 0-0%) in Group 1 and 1.6% (95% CI, 0-4.7%) in Group 2 at six months (p=0.312) and 20.2% (95% CI, 3.9-36.5%) in Group 1 and 3.8% (95% CI, 0-8.9%) in Group 2 at 12 months (p=0.059).
ED
MA
NU
Univariate analysis identified SRS dose >17 Gy (HR 0.212, 95% CI, 0.082-0.549, p=0.001), GTV >2.4 cc (HR 2.522, 95% CI, 1.139-5.581, p=0.022), tumor diameter of over 12 mm (HR 2.257, 95% CI, 1.024-4.973, p=0.043), and presence of neurologic symptoms prior to treatment (HR 0.3, 95% CI, 0.123-0.732, p=0.008) as predictors of any (asymptomatic or symptomatic) radiation necrosis (Table 4). On multivariate analysis, only SRS dose of more than 17 Gy was associated with lower rates of necrosis (HR 0.3, 95% CI, 0.1-0.9, p=0.030). Treatment group was not predictive of radiation necrosis on either univariate (p=0.396) or multivariate analysis (p=0.304).
DISCUSSION
AC
CE
PT
Our investigation reflects the results of a modern cohort with access to current imaging and treatment technologies. To our knowledge, it is the largest clinical comparison between frame and frameless stereotactic radiosurgery. We verified that the treatment groups were numerically and statistically similar in GTV size, but showed the expected variations in PTV volumes due to larger and more frequent expansions made in the frameless group. Treatment groups were otherwise similar in histology, previous radiation treatment, SRS dose, and radiologic follow-up. This investigation demonstrates similar local control and radionecrosis rates in patients with brain metastases treated with SRS via both frame and frameless techniques. Prior studies report local control rates ranging from 61% to 100% with SRS. Better control rates were associated with radiation dose, smaller tumors, and SRS in conjunction with WBRT.12-16 Three recent cohorts also described acceptable local control using frameless immobilization. Breneman et al. reported 90% local control at 6 months and 80% local control at 12 months after frameless SRS. Their analysis also agreed with prior studies showing that tumor dose of at least 18 Gy and tumors 0.2 cm3 or smaller were associated with better control.17 Minniti et al. demonstrated 91% local control at 1 year and 82% at 2 years.18 Comparison is difficult because of differences in reporting statistics. Nonetheless, our local control rates are consistent with previous reports for both frame and frameless. Our actuarial local control at 6- and 12-months was 92.8% and 85.5% in Group 1 compared to 87.4% and 73.3% in Group 2. The numerical differences were not statistically significant. Because approximately half of our patients received prior WBRT, and thus have better control rates, we expected our local control rates to be well within the historically reported range of 60-90%.
ACCEPTED MANUSCRIPT
MA
NU
SC
RI P
T
Reporting of radiation necrosis lacks standardization and can be biased by clinical judgment in the absence of tissue diagnosis. Previous studies are not uniform in reporting radiographic asymptomatic radiation necrosis or symptomatic radiation necrosis, with the prior having much higher incidences. Our asymptomatic necrosis rates compare favorably to Minniti et al., who reported asymptomatic radiation necrosis rates of approximately 14%.9 Our crude rates of symptomatic radiation necrosis in the framebased (8%) and frameless groups (6%) were also comparable to those reported by Minniti et al. (10%) and Voges et al. (12.5%).9,19 Taggar et al. also reported 8% radiation necrosis rates, but they were not further specified.20 However, when comparing our frame-based cohort’s crude rate of 8%, the actuarial 12-month rate (20%) is surprisingly high. Vogue et al. reported a similar crude rate of 12.5% but with an actuarial rate of only 14% at 2 years. One would also expect our actuarial rate only to increase minimally. This difference could be explained by the Kaplan-Meier method accounting for lesions that were censored due to either death or loss of follow-up. The actuarial number reports the number of events over the number of lesions at risk which is decreased with every censored event. Especially in a disease with such a short life-expectancy, the Kaplan-Meier method may be overly sensitive to censored data. This would cause an overestimation of radiation necrosis or local recurrence. The magnitude of these statistical effects may also be exaggerated in the frame-based group due to the smaller sample size when compared to greater numbers in the frameless group. One could suggest that patients who were symptomatic from radiation necrosis would be more likely to return for follow-up than to be censored, rendering the crude rate as a reasonable estimate.
AC
CE
PT
ED
In addition, frame-based SRS was more common earlier in our experience. Lesions treated earlier in this cohort may have been subject to less refined planning constraints, and resulted in less conformal therapy. Though we do not have the conformality index data for each lesion, we do see that the framebased group had a higher proportion of patients (46%) treated in the earlier era of 2008-2011 compared to the frameless cohort (17%). This could also explain the numerically higher rates of radionecrosis in Group 1. Furthermore, we did not see a significantly different proportion of patients with previous whole brain radiation in the frame-based group, nor were the target volumes larger. In fact, the PTVs were shown to be statistically larger in the frameless group. Another unexpected finding in our study was that patients who had >17 Gy were less likely to have radiation necrosis. Because we consider the size of the lesion when selecting a prescription dose, per the past and current protocols,11,21 higher doses are likely used with smaller target volumes, which may lead to lower doses to the normal brain tissue. While these data suggest that frameless SRS is effective and safe when compared to frame-based immobilization, acceptable results are heavily dependent upon accurate imaging, treatment planning, dose calculations and delivery. Determining when to use PTV margins and choosing the appropriate margin size seem to be important treatment factors. Margins may be increasingly important in the setting of frameless SRS. One study observed up to 2.9 mm shifts in isocenter displacement when patients were immobilized without an invasive head ring.18 Another study used mathematical formulas to estimate the margin needed for the CTV to receive the prescribed dose in 95% of patients and recommended approximately 3 mm.22 Larger margins, however, have been challenged in a recent prospective trial by Kirkpatrick et al. Forty-nine patients with 80 brain metastases were treated with SRS using a thermoplastic mask and randomized to either 1 mm or 3 mm margins. There was no difference in local control, but 3 mm margins were associated with a higher rate of biopsy-proven radiation necrosis and volume receiving ≥12 Gy.
ACCEPTED MANUSCRIPT
T
This study is subject to the biases inherent in a retrospective single-institution analysis. As a regional referral center, there were a number of patients who traveled long distances for SRS at our institution, which may have limited their clinical and radiographic follow-up. Though it is potentially underpowered, the current analysis is the best data available to evaluate frame-based versus frameless radiosurgery, and the outcomes in both of the treatment groups are in keeping with contemporary series.
NU
SC
RI P
In conclusion, this retrospective comparison suggests that the rates of local control and radiation necrosis are similar in patients treated with SRS for brain metastases whether they are immobilized by frame-based or frameless techniques. Optimal outcomes, however, rely on evidence-based decisionmaking and sound technique.
BIBLIOGRAPHY
MA
1. Chia SK, Speers CH, D'yachkova Y, et al. The impact of new chemotherapeutic and hormone agents on survival in a population-based cohort of women with metastatic breast cancer. Cancer. 2007;110(5):973-
ED
979.
PT
2. Weber M, Zoubaa S, Schlieter M, et al. Diagnostic performance of spectroscopic and perfusion MRI for
CE
distinction of brain tumors. Cancer Imaging. 2006;6(Spec No A):S32.
3. Andrews DW, Scott CB, Sperduto PW, et al. Whole brain radiation therapy with or without
AC
stereotactic radiosurgery boost for patients with one to three brain metastases: Phase III results of the RTOG 9508 randomised trial. Lancet. 2004;363(9422):1665-1672.
4. Aoyama H, Shirato H, Tago M, et al. Stereotactic radiosurgery plus whole-brain radiation therapy vs stereotactic radiosurgery alone for treatment of brain metastases: A randomized controlled trial. JAMA. 2006;295(21):2483-2491.
5. LEKSELL L. The stereotaxic method and radiosurgery of the brain. Acta Chir Scand. 1951;102(4):316319.
ACCEPTED MANUSCRIPT 6. Lutz W, Winston KR, Maleki N. A system for stereotactic radiosurgery with a linear accelerator. Int J Radiat Oncol Biol Phys. 1988;14(2):373-381.
RI P
T
7. Kamath R, Ryken TC, Meeks SL, Pennington EC, Ritchie J, Buatti JM. Initial clinical experience with frameless radiosurgery for patients with intracranial metastases. Int J Radiat Oncol Biol Phys.
SC
2005;61(5):1467-1472.
NU
8. Zhang Q, Chan MF, Burman C, Song Y, Zhang M. Three independent one-dimensional margins for single-fraction frameless stereotactic radiosurgery brain cases using CBCT. Med Phys.
MA
2013;40(12):121715.
ED
9. Minniti G, Clarke E, Lanzetta G, et al. Stereotactic radiosurgery for brain metastases: Analysis of
PT
outcome and risk of brain radionecrosis. Radiat Oncol. 2011;6:48-717X-6-48.
10. Blonigen BJ, Steinmetz RD, Levin L, Lamba MA, Warnick RE, Breneman JC. Irradiated volume as a
CE
predictor of brain radionecrosis after linear accelerator stereotactic radiosurgery. Int J Radiat Oncol Biol
AC
Phys. 2010;77(4):996-1001.
11. Brown P, Roberge D. N107C: Phase III trial of post-surgical stereotactic radiosurgery (SRS) compared with whole brain radiotherapy (WBRT) for resected metastatic brain disease. . 2013:11.11.
12. Bindal AK, Bindal RK, Hess KR, et al. Surgery versus radiosurgery in the treatment of brain metastasis. J Neurosurg. 1996;84(5):748-754.
13. Flickinger JC, Kondziolka D, Lunsford LD, et al. A multi-institutional experience with stereotactic radiosurgery for solitary brain metastasis. Int J Radiat Oncol Biol Phys. 1994;28(4):797-802.
ACCEPTED MANUSCRIPT 14. Kondziolka D, Patel A, Lunsford LD, Kassam A, Flickinger JC. Stereotactic radiosurgery plus whole brain radiotherapy versus radiotherapy alone for patients with multiple brain metastases. Int J Radiat
RI P
T
Oncol Biol Phys. 1999;45(2):427-434.
15. O'Neill BP, Iturria NJ, Link MJ, Pollock BE, Ballman KV, O'Fallon JR. A comparison of surgical resection
SC
and stereotactic radiosurgery in the treatment of solitary brain metastases. Int J Radiat Oncol Biol Phys.
NU
2003;55(5):1169-1176.
16. Shiau CY, Sneed PK, Shu HK, et al. Radiosurgery for brain metastases: Relationship of dose and
MA
pattern of enhancement to local control. Int J Radiat Oncol Biol Phys. 1997;37(2):375-383.
ED
17. Breneman JC, Steinmetz R, Smith A, Lamba M, Warnick RE. Frameless image-guided intracranial stereotactic radiosurgery: Clinical outcomes for brain metastases. Int J Radiat Oncol Biol Phys.
PT
2009;74(3):702-706.
CE
18. Minniti G, Scaringi C, Clarke E, Valeriani M, Osti M, Enrici RM. Frameless linac-based stereotactic radiosurgery (SRS) for brain metastases: Analysis of patient repositioning using a mask fixation system
AC
and clinical outcomes. Radiat Oncol. 2011;6:158-717X-6-158.
19. Voges J, Treuer H, Sturm V, et al. Risk analysis of linear accelerator radiosurgery. Int J Radiat Oncol Biol Phys. 1996;36(5):1055-1063.
20. Taggar A, Lau H, Lim G, et al. Outcomes from first 3 years of frameless stereotactic radiosurgery in treating brain metastases. International Journal of Radiation Oncology • Biology • Physics. 2015;90(5):S68.
ACCEPTED MANUSCRIPT 21. Shaw E, Scott C, Souhami L, et al. Single dose radiosurgical treatment of recurrent previously irradiated primary brain tumors and brain metastases: Final report of RTOG protocol 90-05. Int J Radiat
RI P
T
Oncol Biol Phys. 2000;47(2):291-298.
22. Zhang Q, Chan M, Song Y, Burman C. Three dimensional expansion of margins for single-fraction
AC
CE
PT
ED
MA
NU
Engineering and Radiation Oncology. 2012;1(02):15.
SC
treatments: Stereotactic radiosurgery brain cases. International Journal of Medical Physics, Clinical
ACCEPTED MANUSCRIPT Table 1. Lesion and treatment characteristics by treatment group. P-values were obtained using the Chi-
T
square test.
RI P
Figure 1. Probability of local failure in frame-based SRS (Group 1: blue line) and frameless SRS (Group 2:
SC
red line) with their associated confidence intervals.
NU
Table 2. Outcomes by treatment group. P-values were obtained using the Chi-square test.
MA
Table 3. Analysis of predictive factors for local control. P-values were obtained using the Chi-square test.
ED
Figure 2. Probability of radiographic asymptomatic radiation necrosis in frame-based SRS (Group 1: blue
PT
line) and frameless SRS (Group 2: red line) and their associated confidence intervals.
CE
Figure 3. Probability of symptomatic radiation necrosis in frame-based SRS (Group 1: blue line) and
AC
frameless SRS (Group 2: red line) and their associated confidence intervals.
Table 4. Analysis of predictive factors for radiation necrosis. P-values were obtained using the Chi-square test.
ACCEPTED MANUSCRIPT Table 1 Lesion and Treatment Characteristics Frameless (109)
61 (31-75)
57 (24-88)
30 (49%) 31 (51%)
61 (56%) 48 (44%)
1 (2%) 19 (31%) 6 (10%) 13 (21%) 22 (33%) 0.54 (0.01-15.23)
6 (6%) 42 (39%) 11 (10%) 25 (23%) 25 (23%) 0.89 (0.01-13.41)
0.224 0.281 0.957 0.807 0.043 0.117
0.89 (0.01-13.41) 1.89 (0.13-20.79) 95 (87%) 1.21 0.66 cc (0-9.194)
0.118 0.016 <0.001 <0.001 <0.001
69% 18 (14-24) 5 (5%) 55 (50%) 63 (58%) 33 (30%) 2 (2%) 37 (35%) 37 (34%) 57 (52%) 61 (56%)
0.008 0.127 0.694 0.327 0.957 0.576 0.045 0.049 0.003 0.315 0.311
PT
CE
AC
RI P
SC
NU
MA
0.54 ( 0.01-15.24) 1 (0.05-18.6) 31 (51%) 0.67 0.05 cc (0-3.37) 23% 18(13-22) 3 (5%) 26 (43%) 35 57%) 9 (15%) 5 (8%) 12 (20%) 8 (13%) 27 (44%) 39 (64%)
T
Frame-based (61)
ED
Characteristics Patient Factors Age (median, range) Gender Male Female Disease Factors Histology Squamous cell carcinoma Adenocarcinoma Renal cell Melanoma Other Tumor Volume (median, range) Treatment Factors GTV volume in cc (median, range) PTV volume in cc (median, range) Number with Margin added (%) Average Margin size in mm Volume added from expansion (median, range) % volume increase to PTV (median) Dose (median, range) Previous resection (%) Age: <58 years SRS Dose: >17 Gy PTV: >3.4 cc GTV: >2.4 cc PTV-GTV: >1 cc Margin: >1 mm Previous WBRT Neurologic symptoms
p 0.502 0.395
ACCEPTED MANUSCRIPT
Frameless (109) 6.1 (0.7-44)
7.0 (6.1-39.8)
5.6 (0.7-44.3)
13 (21%) 12 (20%) 30 (49%) 6 (10%) 25 (41%) 11 (18%)
21 (19%) 14 (13%) 59 (54%) 15 (14%) 42 (39%) 14 (13%)
0.749 0.235 0.666 0.456 0.754 0.360
6 (6%)
0.083
RI P
NU
5 (8%)
ED PT CE AC
T
Frame-based (61) 8.7 (0.7-45.7)
MA
Outcomes Clinical follow-up in months (median, range) Radiologic follow-up in months (median, range) Local Failure CR PR (>50 % decrease) SR (up to 50% decrease) DP (>25% increase) Elsewhere brain failure (%) Radiologic Asymptomatic Radiation Necrosis (%) Symptomatic Radiation Necrosis (%)
SC
Table 2 Follow Up and Lesion Outcomes p 0.025 0.095
ACCEPTED MANUSCRIPT
Table 3 Variable Analysis for Local Control Multivariate
T
Univariate N/N
HR (95% CI)
Frame: Frame/Mask
61/109
0.509 (0.201-1.285)
Gender: Male/Female
91/79
2.410 (1.025-5.668)
Age: <58/≥58
81/89
1.045 (0.468-2.335)
SRS Dose: >17/≤17 Gy
98/72
1.042 (0.455-2.384)
0.9226
PTV: >3.4/≤3.4 cc
49/121
2.204 (0.976-4.978)
0.0572
GTV: >2.4/≤2.4 cc
51/119
1.925 (0.852-4.347)
0.1151
PTV-GTV: >1/≤1cc
48/122
1.787 (0.793-4.026)
0.1613
Margin: >1/≤1 mm
44/126
1.456 (0.636-3.333)
0.3735
Tumor diameter: >12/≤12 mm
63/107
1.688 (0.753-3.782)
0.2036
84/86
0.570 (0.236-1.380)
0.2128
64/106
0.513 (0.203-1.298)
0.1591
71/99
0.73(7 0.329-1.648)
0.4570
AC
Neurologic sign: Yes/No
p
HR (95% CI)
p
0.1527
0.691 (0.243-1.962)
0.691
0.0438
1.407 (0.510-3.879)
1.407
0.924 (0.343-2.493)
0.924
SC
NU
MA
ED
CE
MRI: diagnostic/planning
PT
Previous WBRT: yes/No
RI P
Predicting Factors
0.9139
ACCEPTED MANUSCRIPT Table 4 Variable Analysis for Radiation Necrosis Univariate
Multivariate
N/N
ED PT CE AC
p 0.3044
0.303 (0.103-0.890)
0.0299
0.660 (0.278-1.567)
0.3466
0.497 (0.191-1.291)
0.1510
T
HR (95% CI) 1.596 (0.654-3.897)
RI P
p 0.3957 0.1488 0.2676 0.0014 0.4479 0.0225 0.4327 0.0434 0.2128 0.4798 0.0082
SC
61/109 91/79 81/89 98/72 49/121 51/119 44/126 63/107 84/86 64/106 71/99
MA
Frame: Frame/Mask Gender: Male/Female Age: <58/≥58 years SRS Dose: >17/≤17 Gy PTV: >3.4/≤3.4 cc GTV: >2.4/≤2.4 cc Margin: >1/≤1 mm Tumor diameter: >12/≤12 mm Previous WBRT: yes/No MRI: diagnostic/planning Neurologic symptoms: Yes/No
HR (95% CI) 1.451 (0.614-3.429) 1.964 (0.786-4.910) 0.605 (0.249-1.471) 0.212 (0.082-0.549) 1.403 (0.585-3.367) 2.522 (1.139-5.581) 1.440 (0.579-3.576) 2.257 (1.024-4.973) 0.933 (0.386-2.256) 1.369 (0.573-3.268) 0.300 (0.123-0.732)
NU
Predicting Factors
PT
ED
MA
NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
Figure 1
PT
ED
MA
NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
Figure 2
PT
ED
MA
NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
Figure 3
S1879-8500(16)30046-7 doi: 10.1016/j.prro.2016.05.001 PRRO 622
To appear in:
Practical Radiation Oncology
Received date: Revised date: Accepted date:
1 September 2015 26 April 2016 5 May 2016
Please cite this article as: Bennion Nathan R., Malouff Timothy, Verma Vivek, Denniston Kyle, Bhirud Abhijeet, Zhen Weining, Wahl Andrew, Lin Chi, A Comparison of Clinical and Radiologic Outcomes Between Frame-based and Frameless Stereotactic Radiosurgery for Brain Metastases, Practical Radiation Oncology (2016), doi: 10.1016/j.prro.2016.05.001
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
T
A Comparison of Clinical and Radiologic Outcomes Between Frame-based and Frameless Stereotactic Radiosurgery for Brain Metastases Nathan R. Bennion, MD, Timothy Malouff, BS, MD, Vivek Verma, Kyle Denniston, MD, Abhijeet Bhirud, Weining Zhen, MD, Andrew Wahl, MD, Chi Lin, MD PhD 1
RI P
Department of Radiation Oncology, Nebraska Medicine, University of Nebraska Medical Center Campus, Omaha, NE USA
SC
Key Words: stereotactic radiosurgery, frame, frameless, brain metastases
NU
Shortened Title: Frame versus Frameless stereotactic radiosurgery for brain metastases
Corresponding Author:
PT
ED
Nathan Bennion, MD Department of Radiation Oncology University of Nebraska Medical Center Omaha, NE
MA
Acknowledgments: None
AC
CE
Phone: (402)-552-3844 Fax: (402)-552-3844 Email: [email protected]
ACCEPTED MANUSCRIPT Abstract
RI P
T
Purpose/Objective(s): Modern experiences in stereotactic radiosurgery (SRS) report non-invasive frameless techniques as an effective alternative to frame-based SRS. Frameless techniques potentially increase positional uncertainty and planning target volume (PTV) margins are frequently employed. Here we compare rates of local control and radiation necrosis in frameless versus frame-based SRS.
SC
Materials/Methods: Ninety-eight patients (170 lesions) with radiologic and clinical follow-up were analyzed. Group 1 contained 34 patients (61 lesions) immobilized with an invasive stereotactic frame. Group 2 had 64 patients (109 lesions) immobilized with a frameless SRS mask. Patient, tumor, and treatment characteristics were recorded, as were intervals to local recurrence and radiation necrosis (asymptomatic and symptomatic).
ED
MA
NU
Results: Median patient age was 59 years (range 25 -89), and KPS was 80 (range 50-100). Median radiologic and clinical follow-up was 6.5 months (range 0.7-44.3) and 7 months (range 0.7-45.7). A median of 2 tumors were treated per course (range 1-5) with a median dose of 18 Gy (range 13-24 Gy). The median time to local failure was not reached, and Kaplan-Meier estimates of local failure were not statistically significant between groups (p=0.303). Actuarial 6-month local failure rates were 7.2% in Group 1 and 12.6% in Group 2 (p=0.295), with 12-month local failure rates of 14.5% and 26.8% (p=0.185), respectively. There was no statistically significant difference in symptomatic (p=0.391) or asymptomatic (p=0.149) radiation necrosis. Six-month radiation necrosis was 0% in Group 1 and 1.6% in Group 2 (p=0.311) with 12-month rates of 20.2% and 3.8%, respectively (p=0.059). Median time to necrosis was not reached in Group 1, but was 44 months in Group 2.
CE
INTRODUCTION
PT
Conclusion: Frameless SRS demonstrates clinical outcomes comparable to frame-based techniques with respect to local failure and radiation necrosis.
AC
Metastases account for more than one-half of brain tumors. With more sensitive imaging and improved therapy for extracranial disease, the incidence of metastatic brain disease is likely to rise.1,2 Treatment for these patients has historically been limited to whole brain radiation (WBRT) and corticosteroids; however, in our age of evolving therapy, treatment options have become more sophisticated. Advances in technology and prospective randomized trials have established stereotactic radiosurgery (SRS) as a safe means of treating intracranial tumors while minimizing radiation to normal brain parenchyma.3,4 Leksell first described this technique including the concept of “center of arc” targeting using an invasive head frame for rigid immobilization.5 This early work culminated with the introduction of the first Leksell Gamma Knife system in 1968. Clinicians adapted similar frame-based methods for linear accelerator-based radiosurgery.6 Though frame-based methods reliably immobilize patients, they also present several obstacles. Invasive fixation to the skull may be painful or provoke patient anxiety. It requires the patient to remain in the department for extended periods and places time constraints on the plan design and quality assurance in order to complete the simulation, planning, and delivery in the same day. Now, modern frameless techniques provide a non-invasive alternative for patient immobilization and repositioning when used in conjunction with proper image guidance and precision delivery.7 Frameless
ACCEPTED MANUSCRIPT
SC
RI P
T
techniques also provide more time for planning which can increase safety and flexibility. However, frameless SRS is suspected to increase set up uncertainty when compared with frame-based SRS, and may require expanded margins around the gross tumor volume (GTV).8 While expanded target volumes may compensate for set up error, larger treatment volumes may also correlate with higher rates of radiation necrosis.9,10 Rigidly interpreting these two obstacles in frameless SRS would lead to the assumption that increasing margins would result in excess rates of radiation necrosis, or that dismissing the need for additional margins would compromise local control. However, there is little clinical data comparing local control and radionecrosis rates for frame-based versus frameless techniques. We aim to compare the rates of local control and radiation necrosis in frameless versus frame-based SRS. PATIENTS/METHODS
NU
Frame-based SRS Procedure
CE
PT
ED
MA
Each patient was simulated supine with the head frame locked into the simulation table and thin-slice computed tomography (CT) was performed. Thin-slice T1 post-contrast (T1c) magnetic resonance imaging (MRI) and T2/T2 Fluid-attenuated inversion recovery sequences were reconstructed in the coronal, sagittal and axial planes as necessary and fused to the simulation CT. Organs at risk (OAR) were constructed manually. GTVs were contoured using T1c MRI sequences. PTV expansions were created at the discretion of the treating physician. Dose was prescribed to the PTV. Plans were evaluated for optimum coverage and conformity. Patients were positioned on the treatment table with rigid headframe fixation. Positioning was confirmed with orthogonal X-rays using the ExacTrac system (BrainLAB, Feldkirchen, Germany). After verifying patient positioning, treatment was delivered, and the head frame was removed. Frameless SRS Procedure
AC
Each patient was simulated in the supine position. Rigid head rests and radiosurgery specific thermoplastic masks were used for fixation with stereotactic box localization. A thin slice CT was performed in the treatment position, and patients were sent home. MRI fusions, target and OAR construction and treatment planning were as per outlined in the frame-based methods above, with the noted exception that simulation and treatment delivery were not conducted on the same day. On the day of treatment, patients were positioned on the treatment table with the thermoplastic stereotactic mask. Setup accuracy was confirmed for each patient prior to treatment with ExacTrac or kV cone beam CT image guidance. After verifying patient positioning, treatment was delivered and the stereotactic mask removed. Follow-up There was not a strict follow-up program for the entire cohort. However, it is the general practice in our institution to see patients 4-6 weeks after their initial treatment and every 3 months after that. Initial post-treatment MRIs were obtained between 2-4 months after SRS, then with subsequent follow-ups. Patients with signs or symptoms suggestive of disease progression or radiation necrosis were seen in clinic earlier with imaging when appropriate. Tumor control was determined by MRI in all patients. Our
ACCEPTED MANUSCRIPT
T
institutions MRI protocol included T2-weighted, T2/FLAIR, thin cut T1-weighted with and without contrast, diffusion and gradient echo sequences. Magnetic resonance perfusion and spectroscopy were ordered in cases where radiation necrosis is indistinguishable from local progression by standard imaging.
RI P
Data Collection
NU
SC
After obtaining IRB approval, records were reviewed for all patients who underwent SRS at our institution between 2008 and 2014. Patient, disease, and treatment characteristics are as stated in Table 1. We also recorded target volumes, PTV margins, and frame-based (Group 1) and frameless (Group 2) delivery. Last dates of radiologic and clinical follow-up were documented as were dates of local failure and radiation necrosis. Tumor Size
MA
Gross tumor (GTV) and planning target volumes (PTV) were calculated using the iPlan treatment planning system. Tumor Control
Radiation Necrosis
CE
PT
ED
Local failures were assessed from the date of treatment to the first radiologic evidence of progression. Complete response (CR) required radiographic disappearance of brain metastases. Greater than 50% reduction in the size of each lesion radiographically, using perpendicular diameters, was considered a partial response (PR). Up to 50% decrease and up to 25% increase in the size of each lesion radiographically, using perpendicular diameters was deemed stable disease (SD). Lastly, disease progression (DP) was deemed to be an increase of > 25% in the size of a lesion.11
AC
Radiologic radionecrosis was defined by characteristic changes on MRI in the absence of attributable symptoms. In cases where there was ambiguity in determining necrosis versus local tumor progression, serial scans aided in retrospectively categorizing the event. Biopsies were not required for event categorization. Radiographic, asymptomatic radiation necrosis was assessed separately from symptomatic radionecrosis. Time to radiation necrosis was defined as the interval from the date of treatment to the first radiologic evidence of radiation necrosis. Statistical Methods Tumor sizes, measured by GTV and PTV, were compared between each group with the Wilcoxon ranksum test. Tumor control, symptomatic radiation necrosis and asymptomatic radiation necrosis were all analyzed with the Kaplan-Meier method and log-rank test. Six- and 12-month rates of local failure and necrosis were compared using the Chi-square test. Univariate and multivariate analyses were conducted using the Cox proportional hazards model. All reported p-values are two-sided and are accompanied by their associated 95% confidence intervals (95% CI).
ACCEPTED MANUSCRIPT
RESULTS
MA
NU
SC
RI P
T
Ninety-eight patients with available patient, tumor and treatment information underwent SRS at our institution between 2008 and 2014, for a total of 170 treated lesions (Table 1). Median patient age was 59 years (range 25-89), and KPS was 80 (range, 50-100). A median of 2 tumors were treated per treatment course (range 1-5) with a median dose of 18 Gy (range 13-24). Thirty-four patients with a total of 61 lesions were immobilized with a stereotactic frame (Group 1) with a median radiologic follow-up of 7 months (range 6.1-29.8) and clinical follow-up of 8.7 months (range, 6.1-45.7). Sixty-four patients with 109 lesions immobilized with a non-invasive SRS-specific mask (Group 2) had a median radiologic follow-up of 5.6 months (range, 0.7-44.3) and clinical follow-up of 6.1 months (range, 0.7-44). Group 1 had a higher percentage of lesions treated early in the study period, with 28 of 61 (46%) of lesions treated between 2008 and 2010 and the remainder treated between 2011 and 2014. In Group 2, 19 of 109 (17%) of lesions were treated between 2008 and 2010, and 90 of 109 (83%) of lesions were treated between 2011 and 2014.
CE
PT
ED
The Wilcoxon rank-sum analysis showed no differences between the gross tumor volumes in each group. Group 1 had a median GTV of 0.54 cc (range 0.01-15.24) compared to 0.89 cc (range 0.01-13.41) in Group 2 (p=0.118). Thirty-one (51%) of 61 lesions in Group 1 had a PTV margin with a median radial expansion of 1 mm (range 0-3). In Group 2, 87% of the lesions had radial expansions with a median of 1 mm (range 0-3). PTVs were significantly larger in Group 2 with a median volume of 1 cc in Group 1 and 1.89 cc in Group 2 (p=0.016). The volume added by PTV expansion also significantly differed between Groups 1 and 2 with median increases of 0.05 cc and 0.66 cc, respectively (p=<0.001). The percent increase was also statistically greater in Group 2 compared to Group 1 (69% versus 23%) with a p-value of 0.008.
AC
Local failures were identified in 6 targets (10%) from Group 1 and 15 (14%) in Group 2. Median time to local failure was not reached in either group, and Kaplan-Meier estimates of time to local failure in Group 1 (95% CI, 30.7-38 mo) and Group 2 (95% CI, 29.1-39.2 mo) were not statistically different (p=0.303; Figure 1). Six-month estimated risk of local failures were 7.2% (95% CI, 0.4-14%) in Group 1 and 12.6% (95% CI, 5.1-20.1%) in Group 2 (p=0.295), with 12-month estimates of 14.5% (95% CI, 2.926.1%) and 26.8% (95% CI, 12.9-40.7%), respectively (p=0.185). The rates of complete response, partial response, stable disease and disease progression are presented in Table 2. Univariate analysis identified male gender (HR 2.4, 95% CI, 1.025-5.668, p=0.044) as a predictor of local control. It was not significant on multivariate analysis, however. Treatment group was not predictive of local control by either univariate or multivariate analysis with p-values of 0.153 and 0.691, respectively (Table 3). Radiographic, asymptomatic radiation necrosis was present in 11 (18%) of Group 1 lesions and 14 (13%) in Group 2. Kaplan-Meier analysis showed no significant difference between the two groups (p=0.391; Figure 2). Median time to asymptomatic necrosis was not reached in Group 1 and was 44.3 months in Group 2. The risk of developing asymptomatic radiation necrosis was 1.7% (95% CI, 0-5%) in Group 1 and
ACCEPTED MANUSCRIPT 3.1% (95% CI, 0-7.3%) in Group 2 at six months (p=0.609) and 37.5% (95% CI, 18.5-55.6%) in Group 1 and 16.5 (95% CI, 4.1-28.9%) at 12 months (p=0.073).
SC
RI P
T
Symptomatic necrosis was confirmed by MRI in 5 lesions (8%) from Group 1 and 6 lesions (6%) from Group 2. Kaplan-Meier analysis showed no significant difference between the two groups with a logrank p=0.149 (Figure 3). Once again, median time to symptomatic necrosis was not reached in Group 1 and was 44.3 months in Group 2. The risk of developing symptomatic necrosis was 0% (95% CI, 0-0%) in Group 1 and 1.6% (95% CI, 0-4.7%) in Group 2 at six months (p=0.312) and 20.2% (95% CI, 3.9-36.5%) in Group 1 and 3.8% (95% CI, 0-8.9%) in Group 2 at 12 months (p=0.059).
ED
MA
NU
Univariate analysis identified SRS dose >17 Gy (HR 0.212, 95% CI, 0.082-0.549, p=0.001), GTV >2.4 cc (HR 2.522, 95% CI, 1.139-5.581, p=0.022), tumor diameter of over 12 mm (HR 2.257, 95% CI, 1.024-4.973, p=0.043), and presence of neurologic symptoms prior to treatment (HR 0.3, 95% CI, 0.123-0.732, p=0.008) as predictors of any (asymptomatic or symptomatic) radiation necrosis (Table 4). On multivariate analysis, only SRS dose of more than 17 Gy was associated with lower rates of necrosis (HR 0.3, 95% CI, 0.1-0.9, p=0.030). Treatment group was not predictive of radiation necrosis on either univariate (p=0.396) or multivariate analysis (p=0.304).
DISCUSSION
AC
CE
PT
Our investigation reflects the results of a modern cohort with access to current imaging and treatment technologies. To our knowledge, it is the largest clinical comparison between frame and frameless stereotactic radiosurgery. We verified that the treatment groups were numerically and statistically similar in GTV size, but showed the expected variations in PTV volumes due to larger and more frequent expansions made in the frameless group. Treatment groups were otherwise similar in histology, previous radiation treatment, SRS dose, and radiologic follow-up. This investigation demonstrates similar local control and radionecrosis rates in patients with brain metastases treated with SRS via both frame and frameless techniques. Prior studies report local control rates ranging from 61% to 100% with SRS. Better control rates were associated with radiation dose, smaller tumors, and SRS in conjunction with WBRT.12-16 Three recent cohorts also described acceptable local control using frameless immobilization. Breneman et al. reported 90% local control at 6 months and 80% local control at 12 months after frameless SRS. Their analysis also agreed with prior studies showing that tumor dose of at least 18 Gy and tumors 0.2 cm3 or smaller were associated with better control.17 Minniti et al. demonstrated 91% local control at 1 year and 82% at 2 years.18 Comparison is difficult because of differences in reporting statistics. Nonetheless, our local control rates are consistent with previous reports for both frame and frameless. Our actuarial local control at 6- and 12-months was 92.8% and 85.5% in Group 1 compared to 87.4% and 73.3% in Group 2. The numerical differences were not statistically significant. Because approximately half of our patients received prior WBRT, and thus have better control rates, we expected our local control rates to be well within the historically reported range of 60-90%.
ACCEPTED MANUSCRIPT
MA
NU
SC
RI P
T
Reporting of radiation necrosis lacks standardization and can be biased by clinical judgment in the absence of tissue diagnosis. Previous studies are not uniform in reporting radiographic asymptomatic radiation necrosis or symptomatic radiation necrosis, with the prior having much higher incidences. Our asymptomatic necrosis rates compare favorably to Minniti et al., who reported asymptomatic radiation necrosis rates of approximately 14%.9 Our crude rates of symptomatic radiation necrosis in the framebased (8%) and frameless groups (6%) were also comparable to those reported by Minniti et al. (10%) and Voges et al. (12.5%).9,19 Taggar et al. also reported 8% radiation necrosis rates, but they were not further specified.20 However, when comparing our frame-based cohort’s crude rate of 8%, the actuarial 12-month rate (20%) is surprisingly high. Vogue et al. reported a similar crude rate of 12.5% but with an actuarial rate of only 14% at 2 years. One would also expect our actuarial rate only to increase minimally. This difference could be explained by the Kaplan-Meier method accounting for lesions that were censored due to either death or loss of follow-up. The actuarial number reports the number of events over the number of lesions at risk which is decreased with every censored event. Especially in a disease with such a short life-expectancy, the Kaplan-Meier method may be overly sensitive to censored data. This would cause an overestimation of radiation necrosis or local recurrence. The magnitude of these statistical effects may also be exaggerated in the frame-based group due to the smaller sample size when compared to greater numbers in the frameless group. One could suggest that patients who were symptomatic from radiation necrosis would be more likely to return for follow-up than to be censored, rendering the crude rate as a reasonable estimate.
AC
CE
PT
ED
In addition, frame-based SRS was more common earlier in our experience. Lesions treated earlier in this cohort may have been subject to less refined planning constraints, and resulted in less conformal therapy. Though we do not have the conformality index data for each lesion, we do see that the framebased group had a higher proportion of patients (46%) treated in the earlier era of 2008-2011 compared to the frameless cohort (17%). This could also explain the numerically higher rates of radionecrosis in Group 1. Furthermore, we did not see a significantly different proportion of patients with previous whole brain radiation in the frame-based group, nor were the target volumes larger. In fact, the PTVs were shown to be statistically larger in the frameless group. Another unexpected finding in our study was that patients who had >17 Gy were less likely to have radiation necrosis. Because we consider the size of the lesion when selecting a prescription dose, per the past and current protocols,11,21 higher doses are likely used with smaller target volumes, which may lead to lower doses to the normal brain tissue. While these data suggest that frameless SRS is effective and safe when compared to frame-based immobilization, acceptable results are heavily dependent upon accurate imaging, treatment planning, dose calculations and delivery. Determining when to use PTV margins and choosing the appropriate margin size seem to be important treatment factors. Margins may be increasingly important in the setting of frameless SRS. One study observed up to 2.9 mm shifts in isocenter displacement when patients were immobilized without an invasive head ring.18 Another study used mathematical formulas to estimate the margin needed for the CTV to receive the prescribed dose in 95% of patients and recommended approximately 3 mm.22 Larger margins, however, have been challenged in a recent prospective trial by Kirkpatrick et al. Forty-nine patients with 80 brain metastases were treated with SRS using a thermoplastic mask and randomized to either 1 mm or 3 mm margins. There was no difference in local control, but 3 mm margins were associated with a higher rate of biopsy-proven radiation necrosis and volume receiving ≥12 Gy.
ACCEPTED MANUSCRIPT
T
This study is subject to the biases inherent in a retrospective single-institution analysis. As a regional referral center, there were a number of patients who traveled long distances for SRS at our institution, which may have limited their clinical and radiographic follow-up. Though it is potentially underpowered, the current analysis is the best data available to evaluate frame-based versus frameless radiosurgery, and the outcomes in both of the treatment groups are in keeping with contemporary series.
NU
SC
RI P
In conclusion, this retrospective comparison suggests that the rates of local control and radiation necrosis are similar in patients treated with SRS for brain metastases whether they are immobilized by frame-based or frameless techniques. Optimal outcomes, however, rely on evidence-based decisionmaking and sound technique.
BIBLIOGRAPHY
MA
1. Chia SK, Speers CH, D'yachkova Y, et al. The impact of new chemotherapeutic and hormone agents on survival in a population-based cohort of women with metastatic breast cancer. Cancer. 2007;110(5):973-
ED
979.
PT
2. Weber M, Zoubaa S, Schlieter M, et al. Diagnostic performance of spectroscopic and perfusion MRI for
CE
distinction of brain tumors. Cancer Imaging. 2006;6(Spec No A):S32.
3. Andrews DW, Scott CB, Sperduto PW, et al. Whole brain radiation therapy with or without
AC
stereotactic radiosurgery boost for patients with one to three brain metastases: Phase III results of the RTOG 9508 randomised trial. Lancet. 2004;363(9422):1665-1672.
4. Aoyama H, Shirato H, Tago M, et al. Stereotactic radiosurgery plus whole-brain radiation therapy vs stereotactic radiosurgery alone for treatment of brain metastases: A randomized controlled trial. JAMA. 2006;295(21):2483-2491.
5. LEKSELL L. The stereotaxic method and radiosurgery of the brain. Acta Chir Scand. 1951;102(4):316319.
ACCEPTED MANUSCRIPT 6. Lutz W, Winston KR, Maleki N. A system for stereotactic radiosurgery with a linear accelerator. Int J Radiat Oncol Biol Phys. 1988;14(2):373-381.
RI P
T
7. Kamath R, Ryken TC, Meeks SL, Pennington EC, Ritchie J, Buatti JM. Initial clinical experience with frameless radiosurgery for patients with intracranial metastases. Int J Radiat Oncol Biol Phys.
SC
2005;61(5):1467-1472.
NU
8. Zhang Q, Chan MF, Burman C, Song Y, Zhang M. Three independent one-dimensional margins for single-fraction frameless stereotactic radiosurgery brain cases using CBCT. Med Phys.
MA
2013;40(12):121715.
ED
9. Minniti G, Clarke E, Lanzetta G, et al. Stereotactic radiosurgery for brain metastases: Analysis of
PT
outcome and risk of brain radionecrosis. Radiat Oncol. 2011;6:48-717X-6-48.
10. Blonigen BJ, Steinmetz RD, Levin L, Lamba MA, Warnick RE, Breneman JC. Irradiated volume as a
CE
predictor of brain radionecrosis after linear accelerator stereotactic radiosurgery. Int J Radiat Oncol Biol
AC
Phys. 2010;77(4):996-1001.
11. Brown P, Roberge D. N107C: Phase III trial of post-surgical stereotactic radiosurgery (SRS) compared with whole brain radiotherapy (WBRT) for resected metastatic brain disease. . 2013:11.11.
12. Bindal AK, Bindal RK, Hess KR, et al. Surgery versus radiosurgery in the treatment of brain metastasis. J Neurosurg. 1996;84(5):748-754.
13. Flickinger JC, Kondziolka D, Lunsford LD, et al. A multi-institutional experience with stereotactic radiosurgery for solitary brain metastasis. Int J Radiat Oncol Biol Phys. 1994;28(4):797-802.
ACCEPTED MANUSCRIPT 14. Kondziolka D, Patel A, Lunsford LD, Kassam A, Flickinger JC. Stereotactic radiosurgery plus whole brain radiotherapy versus radiotherapy alone for patients with multiple brain metastases. Int J Radiat
RI P
T
Oncol Biol Phys. 1999;45(2):427-434.
15. O'Neill BP, Iturria NJ, Link MJ, Pollock BE, Ballman KV, O'Fallon JR. A comparison of surgical resection
SC
and stereotactic radiosurgery in the treatment of solitary brain metastases. Int J Radiat Oncol Biol Phys.
NU
2003;55(5):1169-1176.
16. Shiau CY, Sneed PK, Shu HK, et al. Radiosurgery for brain metastases: Relationship of dose and
MA
pattern of enhancement to local control. Int J Radiat Oncol Biol Phys. 1997;37(2):375-383.
ED
17. Breneman JC, Steinmetz R, Smith A, Lamba M, Warnick RE. Frameless image-guided intracranial stereotactic radiosurgery: Clinical outcomes for brain metastases. Int J Radiat Oncol Biol Phys.
PT
2009;74(3):702-706.
CE
18. Minniti G, Scaringi C, Clarke E, Valeriani M, Osti M, Enrici RM. Frameless linac-based stereotactic radiosurgery (SRS) for brain metastases: Analysis of patient repositioning using a mask fixation system
AC
and clinical outcomes. Radiat Oncol. 2011;6:158-717X-6-158.
19. Voges J, Treuer H, Sturm V, et al. Risk analysis of linear accelerator radiosurgery. Int J Radiat Oncol Biol Phys. 1996;36(5):1055-1063.
20. Taggar A, Lau H, Lim G, et al. Outcomes from first 3 years of frameless stereotactic radiosurgery in treating brain metastases. International Journal of Radiation Oncology • Biology • Physics. 2015;90(5):S68.
ACCEPTED MANUSCRIPT 21. Shaw E, Scott C, Souhami L, et al. Single dose radiosurgical treatment of recurrent previously irradiated primary brain tumors and brain metastases: Final report of RTOG protocol 90-05. Int J Radiat
RI P
T
Oncol Biol Phys. 2000;47(2):291-298.
22. Zhang Q, Chan M, Song Y, Burman C. Three dimensional expansion of margins for single-fraction
AC
CE
PT
ED
MA
NU
Engineering and Radiation Oncology. 2012;1(02):15.
SC
treatments: Stereotactic radiosurgery brain cases. International Journal of Medical Physics, Clinical
ACCEPTED MANUSCRIPT Table 1. Lesion and treatment characteristics by treatment group. P-values were obtained using the Chi-
T
square test.
RI P
Figure 1. Probability of local failure in frame-based SRS (Group 1: blue line) and frameless SRS (Group 2:
SC
red line) with their associated confidence intervals.
NU
Table 2. Outcomes by treatment group. P-values were obtained using the Chi-square test.
MA
Table 3. Analysis of predictive factors for local control. P-values were obtained using the Chi-square test.
ED
Figure 2. Probability of radiographic asymptomatic radiation necrosis in frame-based SRS (Group 1: blue
PT
line) and frameless SRS (Group 2: red line) and their associated confidence intervals.
CE
Figure 3. Probability of symptomatic radiation necrosis in frame-based SRS (Group 1: blue line) and
AC
frameless SRS (Group 2: red line) and their associated confidence intervals.
Table 4. Analysis of predictive factors for radiation necrosis. P-values were obtained using the Chi-square test.
ACCEPTED MANUSCRIPT Table 1 Lesion and Treatment Characteristics Frameless (109)
61 (31-75)
57 (24-88)
30 (49%) 31 (51%)
61 (56%) 48 (44%)
1 (2%) 19 (31%) 6 (10%) 13 (21%) 22 (33%) 0.54 (0.01-15.23)
6 (6%) 42 (39%) 11 (10%) 25 (23%) 25 (23%) 0.89 (0.01-13.41)
0.224 0.281 0.957 0.807 0.043 0.117
0.89 (0.01-13.41) 1.89 (0.13-20.79) 95 (87%) 1.21 0.66 cc (0-9.194)
0.118 0.016 <0.001 <0.001 <0.001
69% 18 (14-24) 5 (5%) 55 (50%) 63 (58%) 33 (30%) 2 (2%) 37 (35%) 37 (34%) 57 (52%) 61 (56%)
0.008 0.127 0.694 0.327 0.957 0.576 0.045 0.049 0.003 0.315 0.311
PT
CE
AC
RI P
SC
NU
MA
0.54 ( 0.01-15.24) 1 (0.05-18.6) 31 (51%) 0.67 0.05 cc (0-3.37) 23% 18(13-22) 3 (5%) 26 (43%) 35 57%) 9 (15%) 5 (8%) 12 (20%) 8 (13%) 27 (44%) 39 (64%)
T
Frame-based (61)
ED
Characteristics Patient Factors Age (median, range) Gender Male Female Disease Factors Histology Squamous cell carcinoma Adenocarcinoma Renal cell Melanoma Other Tumor Volume (median, range) Treatment Factors GTV volume in cc (median, range) PTV volume in cc (median, range) Number with Margin added (%) Average Margin size in mm Volume added from expansion (median, range) % volume increase to PTV (median) Dose (median, range) Previous resection (%) Age: <58 years SRS Dose: >17 Gy PTV: >3.4 cc GTV: >2.4 cc PTV-GTV: >1 cc Margin: >1 mm Previous WBRT Neurologic symptoms
p 0.502 0.395
ACCEPTED MANUSCRIPT
Frameless (109) 6.1 (0.7-44)
7.0 (6.1-39.8)
5.6 (0.7-44.3)
13 (21%) 12 (20%) 30 (49%) 6 (10%) 25 (41%) 11 (18%)
21 (19%) 14 (13%) 59 (54%) 15 (14%) 42 (39%) 14 (13%)
0.749 0.235 0.666 0.456 0.754 0.360
6 (6%)
0.083
RI P
NU
5 (8%)
ED PT CE AC
T
Frame-based (61) 8.7 (0.7-45.7)
MA
Outcomes Clinical follow-up in months (median, range) Radiologic follow-up in months (median, range) Local Failure CR PR (>50 % decrease) SR (up to 50% decrease) DP (>25% increase) Elsewhere brain failure (%) Radiologic Asymptomatic Radiation Necrosis (%) Symptomatic Radiation Necrosis (%)
SC
Table 2 Follow Up and Lesion Outcomes p 0.025 0.095
ACCEPTED MANUSCRIPT
Table 3 Variable Analysis for Local Control Multivariate
T
Univariate N/N
HR (95% CI)
Frame: Frame/Mask
61/109
0.509 (0.201-1.285)
Gender: Male/Female
91/79
2.410 (1.025-5.668)
Age: <58/≥58
81/89
1.045 (0.468-2.335)
SRS Dose: >17/≤17 Gy
98/72
1.042 (0.455-2.384)
0.9226
PTV: >3.4/≤3.4 cc
49/121
2.204 (0.976-4.978)
0.0572
GTV: >2.4/≤2.4 cc
51/119
1.925 (0.852-4.347)
0.1151
PTV-GTV: >1/≤1cc
48/122
1.787 (0.793-4.026)
0.1613
Margin: >1/≤1 mm
44/126
1.456 (0.636-3.333)
0.3735
Tumor diameter: >12/≤12 mm
63/107
1.688 (0.753-3.782)
0.2036
84/86
0.570 (0.236-1.380)
0.2128
64/106
0.513 (0.203-1.298)
0.1591
71/99
0.73(7 0.329-1.648)
0.4570
AC
Neurologic sign: Yes/No
p
HR (95% CI)
p
0.1527
0.691 (0.243-1.962)
0.691
0.0438
1.407 (0.510-3.879)
1.407
0.924 (0.343-2.493)
0.924
SC
NU
MA
ED
CE
MRI: diagnostic/planning
PT
Previous WBRT: yes/No
RI P
Predicting Factors
0.9139
ACCEPTED MANUSCRIPT Table 4 Variable Analysis for Radiation Necrosis Univariate
Multivariate
N/N
ED PT CE AC
p 0.3044
0.303 (0.103-0.890)
0.0299
0.660 (0.278-1.567)
0.3466
0.497 (0.191-1.291)
0.1510
T
HR (95% CI) 1.596 (0.654-3.897)
RI P
p 0.3957 0.1488 0.2676 0.0014 0.4479 0.0225 0.4327 0.0434 0.2128 0.4798 0.0082
SC
61/109 91/79 81/89 98/72 49/121 51/119 44/126 63/107 84/86 64/106 71/99
MA
Frame: Frame/Mask Gender: Male/Female Age: <58/≥58 years SRS Dose: >17/≤17 Gy PTV: >3.4/≤3.4 cc GTV: >2.4/≤2.4 cc Margin: >1/≤1 mm Tumor diameter: >12/≤12 mm Previous WBRT: yes/No MRI: diagnostic/planning Neurologic symptoms: Yes/No
HR (95% CI) 1.451 (0.614-3.429) 1.964 (0.786-4.910) 0.605 (0.249-1.471) 0.212 (0.082-0.549) 1.403 (0.585-3.367) 2.522 (1.139-5.581) 1.440 (0.579-3.576) 2.257 (1.024-4.973) 0.933 (0.386-2.256) 1.369 (0.573-3.268) 0.300 (0.123-0.732)
NU
Predicting Factors
PT
ED
MA
NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
Figure 1
PT
ED
MA
NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
Figure 2
PT
ED
MA
NU
SC
RI P
T
ACCEPTED MANUSCRIPT
AC
CE
Figure 3