Required target margins for image-guided lung SBRT: Assessment of target position intrafraction and correction residuals

Transcription

1 Practical Radiation Oncology (2013) 3, Original Report Required target margins for image-guided lung SBRT: Assessment of target position intrafraction and correction residuals Chirag Shah MD a, Larry L. Kestin MD a, Andrew J. Hope MD b, Jean-Pierre Bissonnette PhD b, Matthias Guckenberger MD c, Ying Xiao PhD d, Jan-Jakob Sonke PhD e, Jose Belderbos MD, PhD e, Di Yan DSc a, Inga S. Grills MD a, a Department of Radiation Oncology, William Beaumont Hospital, Royal Oak, Michigan b Princess Margaret Hospital, Toronto, Ontario, Canada c University of Würzburg, Würzburg, Germany d Thomas Jefferson University and Hospitals, Philadelphia, Pennsylvania e Netherlands Cancer Institute, Amsterdam, Netherlands Received 25 January 2012; revised 8 March 2012; accepted 9 March 2012 Abstract Purpose: With increased use of stereotactic body radiotherapy (SBRT) for early-stage lung cancer, quantification of intrafraction variation (IFV) is required to develop adequate target margins. Methods and Materials: A total of 409 patients with 427 tumors underwent 1593 fractions of lung SBRT between 2005 and Translational target position correction of the mean target position (MTP) was performed via onboard cone-beam computed tomography (CBCT). IFV was measured as the difference in MTP between the post-correction CBCT and the post-treatment CBCT and was calculated on 1337 fractions. Results: Mean IFV-MTP was 0.0 ± 1.7 mm, 0.6 ± 2.2 mm, and 1.0 ± 2.0 mm in the mediolateral (ML), anteroposterior (AP), and craniocaudal (CC) dimensions, and the vector was 3.1 ± 2.0 mm; 67.8% of fractions had an IFV vector greater than 2 mm, and 14.3% greater than 5 mm. Weight, excursion, forced expiratory volume in the first second of expiration, diffusing capacity of the lung for carbon monoxide, and treatment time were found to be significant predictors of IFV-MTP greater than 2 mm and 5 mm. Significant differences in IFV-MTP were seen between immobilization devices with a mean IFV of 2.3 ± 1.4 mm, 2.7 ± 1.6 mm, 3.0 ± 1.7 mm, 3.0 ± 2.5 mm, 3.3 ± 1.7 mm, and 3.3 ± 2.2 mm for the body frame, hybrid device, alpha cradle, body fix, wing board, and no immobilization, respectively (P b.001). Estimated required target margins for the entire cohort were 4.3, 6.1, and 6.0 mm in the ML, AP, and CC dimensions, with differences in margins based on immobilization. Conclusions: IFV is dependent on several factors: immobilization device, treatment time, pulmonary function, and bodyweight. These factors are responsible for a significant portion of Conflicts of interest: None. Corresponding author. Department of Radiation Oncology, William Beaumont Hospital, 3601 West Thirteen Mile Rd, Royal Oak, MI address: igrills@beaumont.edu (I.S. Grills) /$ see front matter 2013 American Society for Radiation Oncology. Published by Elsevier Inc. All rights reserved.

2 68 C. Shah et al Practical Radiation Oncology: January-March 2013 target margins with a mean IFV vector of 3 mm. Target margins of 6 mm or greater are required to encompass IFV in all dimensions when using four-dimensional CT with CBCT without respiratory gating or compression American Society for Radiation Oncology. Published by Elsevier Inc. All rights reserved. Introduction Increasingly, stereotactic body radiotherapy (SBRT) is being utilized in the management of early-stage lung cancer. The basis for this shift in paradigm is the increasing number of publications demonstrating the safety and efficacy of SBRT. Recent studies have published local control rates approaching or exceeding 90% and the overall survival rates of potential surgical patients treated with SBRT in nonrandomized studies have approached that of some surgical series. 1-3 Furthermore, while surgical management has been traditionally considered the standard of care for early-stage lung cancer, at least 20% of patients are considered inoperable, and a significant fraction of operable patients refuse surgical intervention due to the associated risks. 4 While increasing data regarding lung SBRT outcomes are becoming available, there are limited published data regarding intrafraction variation (IFV) during lung SBRT. Currently, the reported dose per fraction with lung SBRT range from 4 to 34 Gy, which is significantly higher than traditional dose fractions (1.8 to 2.0 Gy/fraction), and therefore longer treatment times are required. Data from William Beaumont Hospital have previously demonstrated that IFV increases proportionately to treatment time during hypofractionated prostate treatment and therefore one might expect similar findings with lung SBRT, which further magnifies the significance of IFV. 5 The significance of these findings is that if IFV is found to represent a significant component of required target margins, planning target volumes may need to be expanded in order to ensure adequate target coverage. The largest previous series to evaluate IFV in patients undergoing lung SBRT found IFV to be 0.2 ± 1.8 mm, 0.1 ± 1.9 mm, and 0.01 ± 1.5 mm in the craniocaudal (CC), anteroposterior (AP), and mediolateral (ML) dimensions, and an IFV-mean target position (MTP) vector of 2.3 ± 2.1 mm. 6 Further, this study identified factors associated with increased IFV including treatment time, respiratory excursion, patient weight, and immobilization device. 6 These data have been confirmed by smaller series from other institutions as well. 7-9 However, these prior reports suffer from relatively small sample sizes and a limited number of fractions to evaluate factors associated with IFV. Therefore, the purpose of our analysis was to quantify IFV in a large cohort of patients undergoing lung SBRT, with different immobilization devices with online image guidance at multiple institutions, and to determine target margin requirements based on the IFV and correction residuals. Methods and materials From 2005 to 2010, 409 patients with stage I-II nonsmall cell lung cancer underwent image-guided lung SBRT to 427 lung tumors at 5 institutions that were all members of a Lung Research Group: William Beaumont Hospital, Princess Margaret Hospital, Thomas Jefferson University, Netherlands Cancer Institute, and the University of Wurzburg. Institutional Review Board approval was granted for this study at each institution. A total of 1593 fractions were delivered with cone-beam computed tomography (CBCT). Of these fractions all fractions had pre-correction CBCT data, 1498 (94%) post-correction CBCT data, and 1337 (84%) the post-treatment CBCT data necessary for IFV calculation. Respiratory excursion data were available for 1543 fractions. Inclusion criteria varied by institution. However, all institutions had a GTV maximum of 8 cm while only one institution limited treatment to tumors greater than 2 cm from the proximal bronchial tree. IFV could only be calculated from data from William Beaumont Hospital, Princess Margaret Hospital, and Netherlands Cancer Institute secondary to post-treatment CBCT being performed. At simulation, patients underwent a treatment planning CT scan that included both lungs in entirety. After 2005, all patients underwent a respiratory-correlated (4-dimensional [4D]) CT scan, while prior patients underwent a 3D CT scan with fluoroscopy used to assess tumor motion. Regarding 4D CT, 2 institutions utilized 10 respiratory phases, 2 institutions utilized maximum inhalation and exhalation (2 phases), and 1 used a mean time-weighted tumor position. With the exception of the Netherlands Cancer Institute, all patients were immobilized with 1 of 5 immobilization devices: stereotactic body frame (Elekta, Stockholm, Sweden), custom Alpha Cradle (KGF Enterprises, Chesterfield, MI), BodyFIX (Elekta), a hybrid customized device combining an Alpha Cradle + BodyFIX vacuum suction ( hybrid ), or a wing board. Variations did exist in treatment planning; 3 of the 5 institutions added no additional margin when expanding the gross tumor volume (GTV) to the clinical tumor volume while 2 added 4-7 mm to the GTV to create the clinical tumor volume. All institutions used patientspecific planning target volume (PTV) margins that were at least 5 mm. Treatment was delivered using 5-16 coplanar or noncoplanar beams that were either manually or inversely optimized. Three of the 5 institutions utilized heterogeneity corrections with all prescriptions being volumetric to the PTV edge. Fractionation schedules varied from single fraction treatment delivering 26 Gy to

3 Practical Radiation Oncology: January-March fractions of 4.2 Gy. Dosimetric constraints varied across institutions according to institutional protocol and 10%-40% heterogeneity was allowed within the target volume. MTP was defined as the mean position of the tumor during the respiratory cycle. This was calculated as the average position of the center of mass of the target volume on the CT scan. Patients underwent daily online CBCTbased volumetric (3D) image-guided radiation therapy (IGRT) using grayscale soft-tissue-based image registration in X-ray volumetric imaging on Elekta Synergy or Elekta Axesse linear accelerators (Elekta). Volumetric modulated arc therapy was not utilized on this cohort of patients. Prior to 2005, offline CT target verification just before beam-on was used at a single institution. MTP localization was performed by soft tissue registration of the pre-correction CBCT with the composite ( average ) planning 4DCT prior to delivery of each fraction. 10 Most centers used 2-3 CBCT scans per fraction, including precorrection, post-correction, and post-treatment scans. Treatment time was recorded as the time from initial precorrection CBCT to post-treatment CBCT, as well as from post-correction CBCT to post-treatment CBCT. Respiratory gating was not used on patients and abdominal compression was used on a small fraction of patients. IFV-MTP was quantified by measuring the difference in the MTP between the post-correction CBCT and the post-treatment CBCT. The mean and the standard deviation of IFV-MTP were calculated in the CC, AP, and ML dimensions, respectively, for all fractions, with the mean representing an average of the true values, not absolute values. A composite IFV-MTP value was created using a 3D vector of the 3 dimensions as previously utilized. 11 Post-correction residual error was quantified by measuring the difference in the MTP between the precorrection CBCT and the post-correction CBCT. Respiratory excursion was defined using the simulation 4D CT scan based on the maximal displacement of the center of mass over the respiratory phases or for a small subset of patients based on fluoroscopy. Interfraction variation was calculated based on the shift in MTP from baseline to pretreatment CBCT. Final position vector was calculated based on the shift in MTP based on the post-treatment CBCT. Multiple variables were analyzed for their association with IFV including age, performance status, weight, forced expiratory volume in the first second of expiration (FEV 1 ), diffusing capacity of the lung for carbon monoxide (DLCO), treatment time, respiratory excursion, and GTV size. Differences between immobilization groups were analyzed, including age, weight, treatment time, GTV size, and excursion. Target margins (ie, internal target volume to PTV expansions) were calculated using the 2-parameter (geometric margin) and 4-parameter model (dosimetric margin) with a 90% confidence level, as previously discussed in Yan et al. 12 The 2-parameter model utilized the mean and standard deviation of all displacements while the 4-parameter model utilized mean and standard deviation of the individual patient means as well as the root mean-square and standard deviation of the individual standard deviations. 12 Margins incorporated IFV and residual error (remaining setup error). Dose penumbra was based on measurements performed on the Elekta Axesse. Statistical associations between variables were analyzed using the Fisher exact test (2-tailed), logistic regression, and linear regression. The unpaired Student t test was used to compare differences in mean values between groups. A P value of.05 was considered statistically significant. Statistical analyses were performed with SPSS version 17.0 (SPSS Inc, Chicago, IL). Results A total of 409 patients with 427 tumors were evaluated. The mean age was 73.6 years old (range, years old), with a mean Eastern Cooperative Oncology Group performance status of 0.97 (range, 0-3) and weight of 71.0 kg (range, kg). The mean GTV maximum dimension for the cohort was 2.7 cm (range, cm). The median follow-up was 1.2 years for the entire group, with a range of years. The mean IFV-MTP was 0.0 ± 1.7 mm, 0.6 ± 2.2 mm, and 1.0 ± 2.0 mm in the ML, AP, and CC dimensions, respectively. The mean 3D-vector IFV-MTP was 3.1 ± 2.0 mm. Figure 1 plots IFV-MTP for the entire cohort. The mean respiratory excursion was 1.5 ± 1.6 mm, 2.4 ± 2.3 mm, and 5.0 ± 5.6 mm in the ML, AP, and CC dimensions, respectively, and the vector was 6.2 ± 5.8 mm. Figure 2 plots excursion for the 1543 fractions evaluated. With regard to treatment time, the mean time from post-correction CBCT to post-treatment CBCT was 24 minutes 28 seconds (range, 6:36-77:47) and the mean time from pre-correction CBCT to post-treatment CBCT was 32 minutes 47 seconds (range, 12: ). The mean interfraction variation was 2.8 ± 5.5 mm, 3.0 ± 6.2 mm, and Figure 1 Lung SBRT margins 69 Frequency of intrafraction variation.

4 70 C. Shah et al Practical Radiation Oncology: January-March 2013 Figure 2 Excursion vector frequency ± 6.3 mm, in the ML, AP, and CC dimensions, respectively. The mean correction residual error following correction was 0.04 ± 1.3 mm, 0.4 ± 1.8 mm, and 0.6 ± 1.7 mm in the ML, AP, and CC dimensions. A total of 67.8% of all fractions evaluated had an IFV- MTP 3D vector greater than 2 mm, with 19.7%, 30.4%, and 32.5% of fractions having an IFV-MTP greater than 2 mm in the ML, AP, and CC dimensions, respectively. Further, 56.7% of all fractions had at least 1 dimension greater than 2 mm. An analysis comparing fractions with an IFV-MTP vector b2 mm with fractions with an IFV- MTP vector 2 mm was performed and presented in Table 1. Age (74.1 years vs 73.5 years, P =.48), Eastern Cooperative Oncology Group performance status (0.94 vs 1.01, P =.29), GTV maximum dimension (2.8 vs 2.8 cm, P =.49), absolute DLCO (10.93 vs 11.53, P =.15), gender (P =.15), and local recurrence (P =.68) were not found to be significant predictors of IFV-MTP greater than 2 mm. However, patient weight, excursion, absolute FEV 1,DLCO % predicted, treatment time, and immobilization device were found to be significant predictors of IFV-MTP greater than 2 mm compared with fractions with IFV-MTP less than 2 mm. A total of 14.3% of all fractions evaluated had an IFV- MTP vector greater than 5 mm with 1.6%, 4.6%, and 4.6% of fractions having an IFV-MTP greater than 5 mm in the ML, AP, and CC dimensions, respectively. Further, 10.1% of all fractions had at least 1 dimension greater than 5 mm. Analysis of fractions with IFV-MTP greater than 5 mm was similar to the 2 mm analysis and is presented in Table 2. Weight (P =.009), respiratory excursion (P b.001), absolute FEV 1 (P b.001), DLCO % predicted (P b.001), and treatment time (P b.001) were continuous variables associated with IFV-MTP greater than 5 mm, while immobilization device was associated as well. Of note, 19% of fractions had an IFV-MTP greater than 5 mm with no immobilization compared with only 4% with the body frame. Treatment time was also analyzed as a categoric variable with a cutoff of 24 minutes and 40 seconds Table 1 Analysis of fractions with intrafraction variation greater than or less than 2 mm Variables IFV b 2mm IFV 2mm t test P value Continuous variables (mean values) Age (y) ECOG performance status Weight (kg) GTV max dimension (cm) Excursion vector (cm) b.001 FEV 1 (L) b.001 FEV 1 % predicted 65.1% 65.8% b.001 DLCO (ml CO/min/mm Hg) DLCO % predicted 51.0% 55.1% b.001 Pre-correction to 7:51 8:43 b.001 post-correction time (min) 23:06 25:12 b.001 post-treatment time (min) Total time (min) 31:09 33:35 b.001 Categoric variables (percentage of fractions) Male 33% 68%.15 Female 35% 65% No local recurrence 34% 66%.68 Local recurrence 26% 74% Immobilization device.02 α-cradle 28% 72% Body frame 51% 49% BodyFIX 40% 60% None 33% 67% Hybrid 36% 65% Wing board 27% 73% DLCO, diffusing capacity of the lung for carbon monoxide; ECOG, Eastern Cooperative Oncology Group; FEV 1, forced expiratory volume in the first second of expiration; GTV, gross tumor volume; IFV, intrafraction variation. based on the median treatment time from postcorrection CBCT to post-treatment CBCT. A significant increase in fractions with IFV-MTP greater than 5 mm was noted when treatment time exceeded this threshold of 24:40 (P b.001). Multivariate analysis was performed for all fractions to evaluate associations with IFV-MTP. On multivariate analysis, treatment time (P =.01), DLCO % predicted (P =.008), and excursion vector (P =.03) were independently significantly associated with IFV-MTP. Of note, immobilization device, FEV 1, weight, and institution were not significant. In order to evaluate the impact of immobilization, patient characteristics were compared by device and are presented in Table 3. Significant differences between groups did exist with regard to weight, GTV max dimension, excursion, FEV 1, DLCO % predicted, and treatment time. IFV-MTP, interfraction variation, correction residual error, and target margins by immobilization device are presented in Table 4. Differences were noted in IFV-MTP vector with the smallest IFV-MTP

5 Practical Radiation Oncology: January-March 2013 Table 2 Analysis of fractions with intrafraction variation greater than or less than 5 mm Variables IFV b 5mm IFV 5mm P value Continuous variables (mean values) Age (y) ECOG performance status Weight (kg) GTV max dimension (cm) Excursion vector (cm) b.001 FEV 1 (L) b.001 FEV 1 % predicted 64.6% 71.7% b.001 DLCO (ml CO/min/mm Hg) DLCO % predicted 53.1% 58.2% b.001 Pre-correction to 8:10 10:03 b.001 post-correction time (min) 24:02 27:18 b.001 post-treatment time (min) Total time (min) 32:06 36:57 b.001 Categoric variables (percentage of fractions) Male 84% 16%.15 Female 88% 12% No local recurrence 86% 14%.68 Local recurrence 93% 7% Immobilization device:.02 α-cradle 90% 10% Body frame 96% 4% BodyFIX 85% 15% None 81% 19% Hybrid 92% 8% Wing board 84% 16% Treatment time b % 9% post-treatment time b24:40 post-treatment time 24:40 81% 19% DLCO, diffusing capacity of the lung for carbon monoxide; ECOG, Eastern Cooperative Oncology Group; FEV 1, forced expiratory volume in the first second of expiration; IFV, intrafraction variation; GTV, gross tumor volume. associated with the stereotactic body frame (2.3 ± 1.4 mm) and the largest associated with both the wing board (3.3 ± 1.7 mm) and no immobilization device (3.3 ± 2.2 mm). The smallest correction residual error was noted with the hybrid device (1.8 ± 0.9 mm). Target margins for the entire cohort were 3.7, 6.2, and 5.7 mm in the ML, AP, and CC dimensions using the 4-parameter model and 4.3, 6.1, and 6.0 mm in the ML, AP, and CC dimensions using the 2-parameter model. The smallest target margins by immobilization were noted using the stereotactic body frame. Target margins for the stereotactic body frame were 2.3, 4.0, and 3.7 mm in the ML, AP, and CC dimensions using the 4-parameter model and 2.9, 4.1, and 4.1 mm in the ML, AP, and CC dimensions using the 2-parameter model. Discussion Lung SBRT margins 71 With over 400 patients and 1300 fractions evaluated, our study found the mean IFV-MTP during 427 courses of lung SBRT were 0.0 ± 1.7 mm, 0.6 ± 2.2 mm, and 1.0 ± 2.0 mm in the ML, AP, and CC dimensions, respectively, with a mean 3D vector of 3.1 ± 2.0 mm. Further, when analyzing factors that are associated with IFV-MTP, our analysis determined that multiple patient, tumor, and treatment characteristics, including weight, respiratory excursion, FEV 1,DLCO, and treatment time were significantly associated with fractions having increased IFV. Our results are consistent with data from previous smaller series that evaluated IFV during lung SBRT. A previous analysis from William Beaumont Hospital found that IFV was 0.2 ± 1.8 mm, 0.1 ± 1.9 mm, and 0.01 ± 1.5 mm in the CC, AP, and ML dimensions, respectively, with a mean 3D vector of 2.3 ± 2.1 mm. Further, in this series, 41% of fractions in that series had an IFV-MTP vector greater than 2 mm and 7% had an IFV-MTP vector greater than 5 mm, which is similar but lower than the rates of 68% and 14% from this analysis. 6 The current results are also comparable with a smaller series from University of Wurzburg that found that 39% of lung SBRT fractions had IFV greater than 2 mm and 16% greater than 3 mm and data from the Netherlands Cancer Institute which found the mean IFV to be 1.2, 1.8, and 1.2 mm in the CC, AP, and ML dimensions. 8,9 This series, which used a comparable CBCT-based IGRT regimen, found a 3D vector of IFV to be 2.8 ± 1.6 mm. 8 IFV results from Princess Margaret Hospital reported lower rates of IFV with 18% of fractions having an IFV greater than 2 mm and 1% greater than 5 mm, which may be due to small numbers of patients or differences in imaging protocols. 11 Data from the same institution demonstrated that the majority of IFV values were less than 1 mm in all dimensions while utilizing 4DCT for planning and online CBCT correction. 7 While it is important to quantify IFV in order to create appropriate target margins, it is just as important to identify factors that drive IFV. With these factors identified, patient-specific target margins that incorporate the patient, tumor, and treatment characteristics associated with IFV can be generated. This can allow for smaller target margins potentially reduce normal tissue toxicity, improve dosimetric coverage, and reduce the potential for target miss. Our current analysis identified multiple factors including weight, respiratory excursion, FEV 1,DLCO, and treatment time. While little can be done to change the first 4 factors in patients with increased weight and excursion, increased target margins may be required to ensure adequate dosimetric coverage. While treatment time has not consistently been found to be associated with increased IFV, recent series from Princess Margaret Hospital and William Beaumont Hospital have identified an association

6 72 C. Shah et al Practical Radiation Oncology: January-March 2013 Table 3 Mean patient, tumor, and treatment characteristics by immobilization device Type Institutions Age (y) Weight (kg) GTV max dimension (cm) Excursion vector (cm) FEV 1 (L) FEV 1 % predicted DLCO (ml CO/ min/mm Hg) DLCO % predicted Treatment time (min) α-cradle PMH, WBH % % 30:03 Body PMH, WBH, TJU, % % 28:19 frame UW BodyFIX PMH, WBH, TJU, % % 28:29 UW None NKI % 57.1% 38:13 Hybrid WBH % % 28:04 Wing PMH, TJU % % 31:00 board P value b.001 b.001 b.001 b b.001 b.001 DLCO, diffusing capacity of the lung for carbon monoxide; FEV 1, forced expiratory volume in the first second of expiration; GTV, gross tumor volume; PMH, Princess Margaret Hospital; WBH, William Beaumont Hospital; TJU, Thomas Jefferson University; UW, University of Wurzburg; NKI, Netherlands Cancer Insititute. while data from the Netherlands Cancer Institute failed to confirm these findings, likely secondary to a narrow treatment time window in that study. 6,9,11 Another potential explanation for these confounding results is that patients treated at the Netherlands Cancer Institute underwent intrafraction CBCT, which may eliminate the association between treatment time and IFV. This hypothesis seems to be confirmed by data from Li et al 13 in which 133 patients underwent lung SBRT with intrafraction correction and no correlation between IFV and treatment time was noted. Another potential solution for treatment time may be to utilize high output linear accelerators or arc therapies such as volumetric modulated arc therapy that have been shown to reduce treatment time. 14,15 Further, planning techniques may be optimized to reduce the number of segments per gantry angle in an effort to reduce small segments that deliver small numbers of monitor units but increase treatment time. Finally, as explained above, the relationship between IFV and treatment time has been shown to be eliminated with intrafraction correction, which may be a solution in plans requiring prolonged treatment time. In an effort to reduce target margins, the impact of immobilization devices has been assessed. In the series from William Beaumont Hospital reduced target margins were noted for patients immobilized with the stereotactic frame and hybrid device. 6 Our analysis confirmed these findings. However, because of significant differences in patient characteristics between the various immobilization devices (Table 3), definitive conclusions cannot be made at this time. However, target margins did vary between immobilization devices with the smallest margins required for the stereotactic body frame with all margins being less than 4.1 mm. Also, use of the wing board was associated with the largest IFV vector along with no immobilization but with the wing board target margins were similar to other immobilization devices. There are limitations to our analysis. While data were collected prospectively, this represents an unplanned analysis and therefore suffers from the limitations of a retrospective analysis. The patient cohort analyzed was treated at 5 separate institutions and there was heterogeneity with regard to treatment planning (immobilization, treatment time), delivery, and imaging protocols, which may preclude definitive conclusions. However, all patients did undergo online cone-beam CT-based image guidance with correction and with a variety of techniques utilized; the data may be more meaningful as it reflects multiple techniques rather than a single institutional paradigm. Another limitation with regard to our immobilization Table 4 Type Intrafraction, interfraction, correction residuals, and target margins by immobilization device Interfraction variation vector (mm) Correction residuals vector (mm) IFV-vector (mm) IFV N 2mm IFV N 5mm Final position variation vector (mm) 2-parameter margins (mm) 4-parameter margins (mm) ML AP CC ML AP CC α-cradle 8.3 ± ± ± % 10.0% 2.9 ± Body frame 6.9 ± ± ± % 3.9% 2.2 ± BodyFIX 10.7 ± ± ± % 15.1% 3.2 ± None 7.8 ± ± ± % 19.2% 3.3 ± Hybrid 12.6 ± ± ± % 8.4% 2.7 ± Wing board 7.4 ± ± ± % 15.9% 3.2 ± AP, anteroposterior; CC, craniocaudal; IFV, intrafraction variation; ML, mediolateral.

7 Practical Radiation Oncology: January-March 2013 device analysis is that significant differences existed between immobilization device cohorts, potentially limiting a statement on the ideal immobilization device to limit IFV. However, to date, this series represents the largest dataset on which IFV and target margins have been evaluated for lung SBRT. Conclusions In the largest series evaluating IFV and correction residual error during lung SBRT, IFV was found to substantially impact target margins with target margins of 4.3, 6.1, and 6.0 mm in the ML, AP, and CC dimensions with post-igrt residuals consistently exceeding 2 mm when using 4D-CT for planning with online CBCT correction and without respiratory gating or compression. Patient and treatment factors including treatment time, patient weight, and immobilization device are associated with IFV and may be optimized to reduce their impact of IFV and therefore, on planning target margins during lung SBRT. Alternative strategies to accommodate for IFV include increasing target margins, correcting for IFV during treatment, and reducing treatment time with high output accelerators. References 1. Fakiris AJ, McGarry RC, Yiannoutsos CT, et al. Stereotactic body radiation therapy for early-stage non-small-cell lung carcinoma: four-year results of a prospective phase II study. Int J Radiat Oncol Biol Phys. 2009;75: Onishi H, Shirato H, Nagata Y, et al. Hypofractionated stereotactic radiotherapy (HypoFXSRT) for stage I non-small cell lung cancer: updated results of 257 patients in a Japanese multi-institutional study. J Thorac Oncol. 2007;2(7 Suppl 3):S94-S Timmerman R, Paulus R, Galvin J, et al. Stereotactic body radiation therapy for inoperable early stage lung cancer. JAMA. 2010;303: Lung SBRT margins Raz DJ, Zell JA, Ou SH, Gandara DR, Anton-Culver H, Jablons DM. Natural history of stage I non-small cell lung cancer: implications for early detection. Chest. 2007;132: Adamson J, Wu Q. Prostate intrafraction motion assessed by simultaneous kv fluoroscopy at MV delivery I: clinical observations and pattern analysis. Int J Radiat Oncol Biol Phys. 2010;78: Shah C, Grills IS, Kestin LL, et al. Intrafraction variation of mean tumor position during image-guided hypofractionated stereotactic body radiotherapy for lung cancer. Int J Radiat Oncol Biol Phys. 2012;82: Bissonnette JP, Franks KN, Purdie TG, et al. Quantifying interfraction and intrafraction tumor motion in lung stereotactic body radiotherapy using respiration-correlated cone beam computed tomography. Int J Radiat Oncol Biol Phys. 2009;75: Guckenberger M, Meyer J, Wilbert J, et al. Intra-fractional uncertainties in cone-beam CT based image-guided radiotherapy (IGRT) of pulmonary tumors. Radiother Oncol. 2007;83: Sonke JJ, Rossi M, Wolthaus J, van Herk M, Damen E, Belderbos J. Frameless stereotactic body radiotherapy for lung cancer using fourdimensional cone beam CT guidance. Int J Radiat Oncol Biol Phys. 2009;74: Grills IS, Hugo G, Kestin LL, et al. Image-guided radiotherapy via daily online cone-beam CT substantially reduces margin requirements for stereotactic lung radiotherapy. Int J Radiat Oncol Biol Phys. 2008;70: Purdie TG, Bissonnette JP, Franks K, et al. Cone-beam computed tomography for on-line image guidance of lung stereotactic radiotherapy: localization, verification, and intrafraction tumor position. Int J Radiat Oncol Biol Phys. 2007;68: Yan D, Lockman D, Martinez A, et al. Computed tomography guided management of interfractional patient variation. Semin Radiat Oncol. 2005;15: Li W, Purdie TG, Taremi M, et al. Effect of immobilization and performance status on intrafraction moation for stereotactic lung radiotherapy: analysis of 133 patients. Int J Radiat Oncol Biol Phys. 2010;81: Brock J, Bedford J, Partridge M, et al. Optimising stereotactic body radiotherapy for non-small cell lung cancer with volumetric intensity-modulated arc therapy a planning study. Clin Oncol Clin Oncol (R Coll Radiol). 2012;24: McGrath SD, Matuszak MM, Yan D, Kestin LL, Martinez AA, Grills IS. Volumetric modulated arc therapy for delivery of hypofractionated stereotactic lung radiotherapy: A dosimetric and treatment efficiency analysis. Radiother Oncol. 2010;95: