- Email: [email protected]
Initial Experience With Volumetric IMRT (RapidArc) for Intracranial Stereotactic Radiosurgery
Int. J. Radiation Oncology Biol. Phys., Vol. 78, No. 5, pp. 1457–1466, 2010 Copyright Ó 2010 Elsevier Inc. Printed in the USA. All rights reserved 0360-3016/$–see front matter
doi:10.1016/j.ijrobp.2009.10.005
CLINICAL INVESTIGATION
Central Nervous System
INITIAL EXPERIENCE WITH VOLUMETRIC IMRT (RAPIDARC) FOR INTRACRANIAL STEREOTACTIC RADIOSURGERY CHARLES S. MAYO, PH.D.,* LINDA DING, PH.D.,* ANTHONY ADDESA, M.D.,* SIDNEY KADISH, M.D.,* T. J. FITZGERALD, M.D.,* AND RICHARD MOSER, M.D.*y *Department of Radiation Oncology and yDivision of Neurosurgery, Department of Surgery, University of Massachusetts Medical School, Worcester, MA Purpose: Initial experience with delivering frameless stereotactic radiotherapy (SRT) using volumetric intensitymodulated radiation therapy (IMRT) delivered with RapidArc is presented. Methods and Materials: Treatment details for 12 patients (14 targets) with a mean clinical target volume (CTV) of 12.8 ± 4.0 cm3 were examined. Dosimetric indices for conformality, homogeneity, and dose gradient were calculated and compared with published results for other frameless, intracranial SRT techniques, including CyberKnife, TomoTherapy, and static-beam IMRT. Statistics on setup and treatment times and per patient dose validations were examined. Results: Dose indices compared favorably with other techniques. Mean conformality, gradient, and homogeneity index values were 1.10 ± 0.11, 64.9 ± 14.1, 1.083 ± 0.026, respectively. Median treatment times were 4.8 ± 1.7 min. Conclusion: SRT using volumetric IMRT is a viable alternative to other techniques and enables short treatment times. This is anticipated to have a positive impact on radiobiological effect and for facilitating wider use of SRT. Ó 2010 Elsevier Inc. RapidArc, Radiosurgery, Treatment time.
INTRODUCTION
tems, Palo Alto, CA) has been demonstrated by several investigators, indicating that it is a mature technology (17–22). Addition of volumetric arc IMRT to the armament of image-guided techniques for delivering intracranial stereotactic radiosurgery is a natural extension of the technology. The combination of modulated intensities with an arc-based approach positions the technique intermediate to arc-based techniques using cones and fixed-beam techniques using MLCs. This report details our initial experience in the use of volumetric IMRT with RapidArc.
Options for stereotactic radiosurgery with intracranial lesions have extended beyond frame-based with fixed multileaf collimator (MLC) or cone-based approaches on linear accelerators to included frameless image-guided approaches such as fixed-beam intensity-modulated radiation therapy (IMRT) (1, 2), TomoTherapy (3, 4), and CyberKnife (5–7). These allow for highly conformal treatment of lesions using technologies that are already used for routine treatment of nonstereotactic patients. In addition, the frameless approach facilitates the use of hypofractionated protocols. Recently, arc-based IMRT techniques have emerged as a promising progression from fixed-field techniques (8–16). Modulation of the intensity over the course of an arc can enable reduction in overall treatment time and reduced dose to normal tissue structures compared with IMRT. The accuracy of volumetric IMRT using RapidArc (Varian Medical Sys-
METHODS AND MATERIALS Twelve patients (Table 1) were treated between January 2009 and June 2009 with volumetric IMRT using RapidArc from Varian Medical Systems. Patients were immobilized in an alpha-cradle-based system for stereotactic radiosurgery (MayoMold, by CDR Systems www.cdrsys.ca, Alberta, Canada; Fig. 1) and then CT scanned in the treatment position using a helical scanner and slice spacing of 1.25 mm. Scans were fused with MRI. The MRI protocol adopted after Acknowledgment—We acknowledge presentations of Dr. Joseph Ting, in which he pointed out the potential role of reduced treatment time with RapidArc for achieving increased biological effect, and of Dr. Richard Popple illustrating use of RapidArc to treat multiple intracranial lesions. We thank Ms. Julie Trifone for demonstrating the immobilization system in Fig. 1. Received July 6, 2009, and in revised form Oct 6, 2009. Accepted for publication Oct 7, 2009.
Reprint requests to: Charles S. Mayo, Ph.D., Department of Radiation Oncology, HB200, University of Massachusetts Medical School, 55 Lake Avenue North, Worcester, MA 01655. Tel: (744) 442-5560; Fax: (774) 442-5006; E-mail: charles.mayo@ umassmemorial.org Conflicts of interest: The first author, has research grant support from Varian Medical Systems. 1457
1458
I. J. Radiation Oncology d Biology d Physics
Volume 78, Number 5, 2010
Table 1. Characteristics of treated patients Gender Male Female Age 56 11 years Histology Primary diagnosis Metastatic lung Melanoma Metastatic esophagus Metastatic breast Metastatic colon Metastatic renal cell Chemotherapy None Tarceva Previous cranial irradiation Whole brain Conformal Radiation Therapy None
8 4 Range, 49–81 years 7 1 1 1 1 1 10 2 7 1 4
March 2009 used T1-VIBE scans with a spacing of 1.25 mm. Target clinical target volumes (CTVs) were delineated on the fused scans and a planning target volume (PTV) was constructed with a 1- to 2-mm margin. Patients were treated on a Trilogy accelerator (Varian Medical Systems www.varian.com, Palo Alto, CA) equipped with a Millennium 120 multileaf collimator (MLC). The accelerator is equipped with a stereotactic beam mode with a beam energy of 6 MV delivered at 1,000 MU/min with a maximum field size of 15 15 cm. Patients were positioned using kV planar and cone-beam CT images. The setup and localization uncertainty of this system is <1 mm. This is confirmed on treatment days with a phantom designed for end-to-end testing from CT through delivery. The phantom is illustrated in Fig. 3. Treatment plans were created in the Eclipse (version 8.6, Varian Medical Systems, www.varian.com, Palo Alto, CA) treatment planning system with 2–3 arcs per isocenter. At least one of the arcs was noncoplanar. Figure 2 illustrates a typical approach of one 350 arc with no table rotation and a second vertex arc of 180 . In optimization, the minimum dose constraint on the PTV was set to the prescribed dose. A dose-limiting annulus (DLA) tuning structure (23) was created to facilitate creating a steep dose gradient beyond the PTV. The outer surface of the DLA was created with a 3-cm margin on the PTV; the inner surface was 1 mm from the PTV (Fig. 4a). The maximum dose constraint on the PTV was set to the prescribed dose. In addition, a normal tissue constraint was set that specified shape of the dose profile away from the PTV (Fig. 4b). Plans were normalized so that at least 96% of the PTV and 100% of the CTV would be covered by the prescribed dose. Dose distributions were compared with other published SRS options using dose-conformity and dose-gradient indices. Correlations among the indices were examined. Because SRS with static MLCs with small leaves is a common approach, a hypothetical machine was created in Eclipse using the same beam data but with a high-definition MLC (0.25 cm over central 10 cm). Static MLC plans were created for each of the targets, and dosimetric indices were compared. The conformity index of International Commission on Radiation Units and Measurements (ICRU) 62 is calculated as the ratio of the volume enclosed by the prescription isodose surface (VRx) to the volume of the PTV (VPTV): CIICRU ¼
VRx : VPTV
(1)
Fig. 1. (a) Patients are positioned in an alpha-cradle-based system that conforms to the posterior half of the patient from the crown of the head through the shoulders. An aquaplast mask indexes to both the cradle and the patient features when the patient is in the correct position. (b) The alpha cradle is indexed to a graphite board that cantilevers over the end of the treatment couch to facilitate used of posterior oblique angles.
Other authors have calculated the conformity index as the ratio of the volume encompassed by 95% of the prescription isodose surface (V95%Rx) (24). CI95%Rx ¼
V95%Rx VPTV
(2)
Paddick (25) noted that because the ICRU definition assumes but does not check for overlap of VRx and VPTV, false values of 1 were possible. The inverse of Paddick’s formulation is used by other authors. CIInvPaddick ¼
VRx VPTV ; 2 VRxXPTV
(3)
where VRxXPTV is the volume of the intersection of the PTV and the prescription isodose surface.
RapidArc SRT d C. S. MAYO et al.
1459
HISD ¼ 1 þ
DSD : DRx
(9)
The overall homogeneity index was calculated as the geometric average of the first two: pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi HIOverall ¼ HIMax HISTD : (10)
Fig. 2. Patients were treated with two to three arcs. The figure illustrates a three-dimensional view of two arcs, one in the transverse plane and one in the sagittal plane. SRT = stereotactic radiotherapy.
The ability of a plan to spare normal tissues outside of the PTV depends on steepness gradient of decreasing dose away from the PTV. Wagner (26) introduced the gradient score of the Conformality Gradient Index (CGIg) for stereotactic radiosurgery: CGIg ¼ 100 {100 REff ;50%Rx REff ;Rx 0:3 } (4) The effective radius of the volume encompassed by the prescription isodose surface is rffiffiffiffiffiffiffiffiffi 3 3VRx REff ;Rx ¼ : (5) 4p The effective radius of the volume encompassed by 50% of the prescription isodose surface (V50%Rx) is: rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 3 3V50%Rx REff ;50%Rx ¼ : (6) 4p
RESULTS
In addition to CGIg, the gradient was calculated as GrEff ¼
50% : REff ;50%Rx REff ;Rx
(7)
The gradient was also measured (GrMeas) as the average gradient in the right-left, anterior-posterior, and cranial-caudal planes through the isocenter. Uniformity of the dose within the PTV volume was compared using three homogeneity indices (HI). The first, which is used by many authors, is the ratio of the maximum dose within the PTV (Dmax) to the prescribed dose (DRx): HIMax ¼
DMax : DRx
Dose distributions were validated with two measurement methods. In the first method, RapidArc plans were projected onto a phantom consisting of the Matrixx ionization chamber array (Scandiatronix-Wellhoffer www.scanditronix-wellhofer.com, Bartlett, TN) with 8 cm of plastic water above and below the array. The phantom was positioned on the accelerator as for the patient and treated with the same RapidArc beams. Calculated and measured doses were compared in Matrixx to the dose distribution calculated in Eclipse. This can be problematic for small targets, because cylindrical ionization chambers in the Matrixx array (4.5 mm diameter, 5.0 mm length) are spaced on a 7.62-mm grid. The nearest chambers to the center of the grid array are at 0.38 cm in the x and y directions. For the second method, measurements were repeated using Extended Dose Range (EDR) film (Eastman Kodak, www.kodak.com, Rochester, NY) in plastic water at the same plane as the Matrixx array. Monitor units were scaled to dose levels that would not saturate the EDR film. A set of dose calibration films was taken in the same session. Films were scanned with a Vidar scanner. Predicted and measured distributions with the scaled monitor units were evaluated using the RIT113 (Radiological Imaging Technology, www.radimage.com, Colorado Springs, CO) software package. An independent monitor unit (MU) check was also performed. The doses delivered at isocenter were calculated in RadCalc software (Life Line Software, www.lifelinesoftware.com, Tyler, TX) and compared with predicted values from Eclipse. Because the Matrixx and RadCalc measures are independent, the average agreement of the two methods was also calculated. Correlations among values of various indices in the sample were examined by calculation of correlation coefficients and of 95% confidence intervals (95% CI) for the coefficients using Fisher’s transforms. Values were judged to be uncorrelated in the sample if the 95% CI contained the value 0. Significance of differences between means was assessed using a two-tailed t test with level of significance set at 0.025.
(8)
To gain a perspective on the statistical range of doses within the PTV, a second homogeneity index was calculated using the standard deviation of the dose within the PTV (DSD):
Treatment plan characteristics for each of the patient targets are detailed in Table 2. Couch angles and angular range of arcs for the targets treated are presented in Table 2. The majority (11 of 14) were treated with two arcs, and the remainder were treated with three arcs. A transverse arc (couch 0) was used for all patients. The cumulative arc length used to treat each target ranged from 500 to 720 with a mean of 563 72.4. No correlation was observed between total arc angle and CIICRU (r = –0.20; 95% CI, –0.66 to 0.37) or CGIg (r = 0.44; 95% CI, –0.12 to 0.79). Transverse isodose distributions and dose profiles for targets with CTV volumes of 0.1, 1.3, and 12.6 cm3 are illustrated in Fig. 5. Table 3 shows median CTV and PTV volumes were 1.20 3.69 and 2.35 6.01. Mean values for CIICRU, CI95%Rx, and CIInv-Paddick were 1.10 0.11,
1460
I. J. Radiation Oncology d Biology d Physics
Volume 78, Number 5, 2010
Fig. 3. (a) A phantom for end-to-end testing consisting of a pattern of cerrobend and air-filled holes on a scribed grid allows checking congruence of light field, MV, kV, cone-beam CT, and bite block isocenters. (b) Cerrobend markers are easily visualized on MV and kV plane images. Here a misalignment of <1 mm in the coronal plane is visualized for cone-beam CT using the air markers. (c) After alignment, a film placed at isocenter is irradiated with the RapidArc beams. The central cerrobend dot produces an artifact on demonstrating agreement of radiation and phantom isocenters. Phantom will be discussed in a separate publication.
1.54 0.34, and 1.32 0.20, respectively. Becausse CIInvPaddick is a ratio of products of two volumes, whereas the other indices are ratios of volumes, it was of interest what would result from a more comparable function. Calculating the square root of CIInv-Paddick for each target, the mean was 1.15 0.08. The conformity index CIICRU had a weak positive correlation with CI95%Rx (r = 0.73; 95% CI, 0.32–0.91) and a weak negative correlation with CIInv-Paddick (r = –0.54; 95% CI, –0.01 to –0.83). No correlation was found between CI95%Rx and CIInv-Paddick or to volume of the CTV or PTV. Thus, in this sample, none of the conformity indices was an accurate surrogate for another. Gradient indices and homogeneity indices are examined in Table 4. Measured gradients GrMeas were slightly steeper than the effective gradients GrEff with mean values of 83.7% 15.8%/cm and 80.0% 16.6%/cm, respectively. Median value for CGIg was 64.9 14.1. A linear regression on CGIg found an increased with GrMeas with a slope of 0.834 0.095 and intercept of –4.9 8.0. No correlation was
found between gradient indices (GrMeas,GrEff, CGIg) and conformity indices (CIICRU, CI95%Rx, CIInv-Paddick). However, CTV volume correlated negatively with CGIg (r = –0.71; 95% CI, –0.9 to –0.28) and GrMeas (r = –0.58; 95% CI, –0.85 to –0.07). Figure 6 shows the measured gradient for CTV and PTV volumes in this study decreasing from values >100 %/cm for small volumes toward a plateau of approximately 60%/cm for CTV volumes >10 cm3. Dose distributions within the PTV volume were homogeneous. Mean values for HIMax, HISD, and HIOverall were 1.083 0.026, 1.017 0.005 and 1.049 0.014, respectively. HIMax did not correlate with either CIICRU or CGIg. However, HISD did correlate with CIICRU (r = –0.85, 95% CI, –0.37 to –0.93), CGIg (r = 0.78, 95% CI, 0.43–0.92) and GrMeas(r = 0.81, 95% CI, 0.50–0.94). This indicates that as target volumes decreased allowing steeper gradients, dose homogeneity within the target also decreased. Most dosimetric indices for plans with static 0.25-cm leaves were not significantly different from RapidArc with
RapidArc SRT d C. S. MAYO et al.
1461
Fig. 4. (a) A dose-limiting annulus tuning structure is created to help control dose conformality and gradients near the surface of the planning target volume (PTV). (b) This is used in conjunction with normal tissue constraint in the optimizer that specifies the dose profile up to the PTV. CTV = clinical target volume.
0.5-cm leaves. Mean conformality values of CIICRU and CI95%Rx were 1.1 0.11 and 1.56 0.19, respectively. Gradient indices GrMeas and CGIg were 76.2 13.3 and 61.3 9.0, respectively. Homogeneity indices HIMax and HISD were 1.099 0.013 and 1.024 0.004, respectively. The small increase in HISD was significant (p = 0.004). Direct measurements of dose distributions in a phantom are illustrated in Fig. 7. Mean agreement of central axis dose was –1.4 3.6% and 1.6 2.4% for Matrixx and film dosimetry, respectively. Independent monitor unit calculations with RadCalc calculations of central axis dose showed good agreement with a mean of –1.8% 5.0%. Table 2. Couch angles and angular ranges used to treat patients Arc 1
Arc 2
Mean setup for the 14 targets treated with hypofractionation were 19 9 min. Setup time for the single fraction SRS was 20 min. The time required to acquire (image and save to server) a pair of orthogonal kV images or a conebeam CT are 3 and 4 min, respectively. Thus the additional setup time is due to waiting for the necessary staff to assemble and determine adjustments. Mean treatment time for the hypofractionated treatments was 4.8 1.7 min. Treatment time for the single fraction treatment (15 Gy) was 7 min. Median follow-up for the patients was 3 months (range, 1–6). There were no cases of progression or adverse events. One patient progressed at other sites in the central nervous system. None of the patients receiving 3 7 Gy required the addition or increase of steroids secondary to idiosyncratic reactions.
Arc 3
DISCUSSION Couch Arc Couch Arc Couch Arc Patient Target angle range angle range angle range 1 2 3 4 5* 6 7 8 9 10 11 12
1 1 1 1 1 1 2 1 1 1 1 1 1 2
0 0 0 0 0 0 0 0 0 0 0 0 0 0
350 350 350 350 350 350 350 350 350 358 358 350 350 350
300 300 270 290 300 90 90 267 270 90 270 270 270 270
175 145 185 175 175 175 175 175 175 179 180 200 150 175
60 60
175 210
45
175
* 15 Gy in one fraction. All others are 7 Gy in three fractions.
In this study, results for dosimetric indices with RapidArc SRT on an accelerator equipped with 0.5-cm MLC leaves compared favorably to other techniques. In their introduction of the gradient score index (CGIg), Wagner et al. (26) indicated that scores $90 were typical for small targets treated with cones and simple geometries, whereas scores in the range of 60–80 are attainable for larger or more complex targets. Bolsi et al. (24) reported on a planning comparison of intracranial targets for 12 patients using several treatment modalities. Median target volume was 2.37 cm3 (range, 0.49–14.32 cm3). Mean values of CI95%Rx for stereotactic arc therapy (four–five noncoplanar arcs with cones), threedimensional conformal therapy (4 coplanar fields), IMRT (4 coplanar fields), spot scanning protons (1–3 fields), and
I. J. Radiation Oncology d Biology d Physics
1462
Volume 78, Number 5, 2010
Fig. 5. Transverse plane isodose distributions are shown for the range of clinical target volume (CTV) volumes in this study: (a) 0.1 cm3; (b) 1.1 cm3; and (c) 12.6 cm3.
passive scattering protons were 3.8 1.6, 2.8 0.6, 4.4 1.7, 3.2 1.0, and 2.5 0.7, respectively. In a subsequent study, Cozzi et al. (11) reported on a planning comparison on intracranial targets for 12 patients using various treatment modalities. Median target volume was 2.37 cm3 (range, 0.49–14.32 cm3). Mean values of CI95%Rx for stereotactic arcs (3–5 noncoplanar arcs with cones), IMRT (3–4 coplanar fields, 5-mm MLC), dynamic IMRT arcs (Ergo++, 3–5 noncoplanar arcs, 3-mm MLC), Helical TomoTherapy, and Cyberknife were 2.9 1.2, 3.0 0.9, 3.9 1.5, 1.8 0.6, and 1.8 0.3, respectively. Han et al. (4) reported on a planning comparison study of 16 patients with mean target volumes of 14.9 12.41 cm3 (range, 1.45–40.78 cm3). Mean value of CIInv-Paddick for coplanar IMRT (6–8 fields, 0.5-cm MCL), noncoplanar IMRT (9–12 fields, 0.5-cm MLC), and Helical TomoTherapy were 1.53 0.38, 1.35 0.15, and 1.26 0.15, respectively.
Mean homogeneity indices (HIDMax) were 1.15 0.05, 1.13 0.04, and 1.18 0.06 for coplanar IMRT, noncoplanar IMRT, and Helical TomoTherapy, respectively. Mean gradient score indices (CGIg) were 13.7 19.08, 22.32 19.20, and 43.28 13.78 for coplanar IMRT, noncoplanar IMRT, and Helical TomoTherapy, respectively. Colombo et al. (5) reviewed results for 199 patients treated for benign meningiomas with CyberKnife (5). Median target volume was 6.9 cm3 (range, 0.1–64 cm3. The mean CIICRU was 1.18 (range, 1.01–1.48). They calculated a modified conformity index corresponding to CIInv-Paddick with a mean value of 1.29 (range, 1.14–1.52). Mean homogeneity index (HIMax) was 1.35 (range, 1.18–2.01) Collins et al. (6) reported on patients treated with CyberKnife. Median values for CIInv-Paddick and HIMax were 1.64 (range, 1.04–3.11) and 1.19 (range, 1.11–1.54) respectively for similar patients (Group II). No gradient indices were reported.
Table 3. Conformity indices for each of the targets displayed in the context of CTV and PTV volumes Patient 1 2 3 4 5 6 7 8 9 10 11 12 Mean SD
Target 1 1 1 1 1 1 2 1 1 1 1 1 1 2
CTV (cm3)
PTV (cm3)
CIICRU
CI95%Rx
CIInv-Paddick
0.3 0.1 3.3 1.1 1.3 0.1 0.8 0.1 0.9 3.3 9.1 12.6 3.2 3.1 2.8 4.0
0.7 0.6 8.7 2 2.3 0.7 2.2 0.4 2.4 7 16.8 19.3 4.8 4.7 5.2 6.01
1.14 1.17 1.09 0.90 0.87 1.14 1.05 1.25 1.17 1.17 1.21 1.12 1.13 1.06 1.10 0.11
1.86 1.67 1.29 1.25 1.09 1.57 1.36 2.50 1.63 1.49 1.54 1.34 1.56 1.47 1.54 0.34
1.56 1.17 1.20 1.84 1.42 1.56 1.27 1.25 1.27 1.21 1.21 1.17 1.22 1.21 1.32 0.2
Abbreviations: CI = conformity index; CTV = clinical target volume; PTV = planning target volume. CIICRU < 1.0 and CI95%Rx > 1.0 corresponds the PTV not fully covered by the prescribed dose but receiving at least 95% of the prescribed dose. Figure 8 illustrates the sensitivity of CIICRU to plan normalization.
RapidArc SRT d C. S. MAYO et al.
1463
Table 4. Gradient indices compared with measurements of gradients and homogeneity indices Patient 1 2 3 4 5 6 7 8 9 10 11 12
Target
CGIg
GrEff (%/cm)
GrMeas (%/cm)
1 1 1 1 1 1 2 1 1 1 1 1 1 2
68.4 82.7 67.4 77.4 80.1 79.4 76.5 65.3 62.1 64.5 46.3 45.4 53.4 39.3 64.9 14.1
81.1 105.6 79.8 95.0 100.2 98.9 93.4 77.2 73.6 76.3 59.7 59.1 65.3 55.1 80.0 16.6
85.2 103.4 94.3 96.2 96.8 107.9 96.8 82.4 71.4 76.9 69.8 65.8 62.8 61.7 83.7 15.8
Mean SD
HIMax
HISD
HIOverall
1.083 1.108 1.120 1.067 1.107 1.125 1.099 1.054 1.039 1.087 1.075 1.082 1.064 1.060 1.083 0.026
1.018 1.020 1.020 1.023 1.026 1.019 1.023 1.011 1.011 1.014 1.010 1.014 1.011 1.013 1.017 0.005
1.050 1.063 1.069 1.045 1.066 1.071 1.060 1.032 1.025 1.050 1.042 1.048 1.037 1.036 1.049 0.014
Abbreviations: GR = gradient index, HI = Homogeneity Index, CGIg = gradient score of Conformality Gradient Index.
No significant difference was found between the conformality indices in this study and those for the other frameless SRT modalities. Mean CGIg values were in agreement with the range found by Wagner. Compared with the TomoTherapy results of Han, the gradient index was significantly (p = 0.0002) larger for our study. Values for the homogeneity index, HIMax, were significantly smaller than Han (p < 0.0001) and for other modalities. The low value of HISD = 1.016, demonstrated uniformity of the dose distribution throughout the PTV volume. Thus, dose distributions obtained using RapidArc with an MLC having a 0.5-cm leaf width are at minimum comparable to other, more widely used techniques for frameless SRT. Sharper dose gradients and better homogeneity are possible. It is worth noting when making comparisons that dependence of conformality index values can depend significantly on normalization chosen in a plan. For small volumes, the difference between normalizing to deliver full dose to 95% of the PTV volume or to the corresponding larger volume is significant. For a hypothetical spherical PTV and perfectly
Measured Gradient (%/cm)
110 CTV
100
PTV
90 80 70 60 50
0
5
10
15
20
25
Volume (cm3)
Fig. 6. Measured dose gradients decrease as the volume of the clinical target volume (CTV) increases. A threshold in the range of 2 cm3 is evident, with gradients larger than 100%/cm for volumes <0.5 cm3. PTV = planning target volume.
conformal plan, CIICRU would be equal to 1 if normalized to cover exactly the radius of the PTV. The distance between the 100% and 95% isodose lines is on the order of 0.7 mm. Figure 8 illustrates the strong impact on CIICRU if the plan is instead normalized to cover inside vs. outside the PTV volume by 0.7 mm. RapidArc treatment times in the range of 4–7 min can be significantly smaller compared with other modalities. Han (4) reported the treatment time for Helical TomoTherapy patients to be 42 16 min. Colombo (5) did not report on treatment times for CyberKnife. In our experience, treatment times for frame-based SRS using a mini-multileaf collimator are in the range of 45–60 min for a single lesion, not including setup time. The short treatment time required for delivery may translate into larger biological effect. Paganetti (27) calculated dose rate effects for nine cell lines using the lethal/potentially lethal model. Recasting data from that study to plot HX34 cell line survival fraction vs. dose rate as an example, Fig. 9 illustrates survival decreased by a factor of 1.06 for 2 Gy per fraction delivered at 0.2 vs. 2.0 Gy/min. However, at a larger dose fraction, 7 Gy, survival was reduced by a factor of 2.75. In addition, reduction in treatment time may also reduce the magnitude of intrafraction motions. Recently, in a study of 104 patients, Rades et al. (28) demonstrated equivalence in control for whole brain (WB) radiotherapy plus stereotactic radiosurgery (SRS) vs. surgery plus WB radiotherapy for 1–3 brain metastases. A multi-institutional study by Sneed et al. (29) of 569 patients comparing outcomes of WB to WB + SRS WB radiation therapy to found equivalent survival rates for both groups. Reporting on a study of 236 patients with metastatic lesions, Pirzkal et al. (30) demonstrated improved survival for WB + SRS vs. SRS alone. Giubilei et al. (31) demonstrated safe use (with a different technology Ergo++ and 3DLine) of a frameless hypofractionated stereotactic radiation therapy (SRT) with concomitant whole brain irradiation in a group of 30 patients.
I. J. Radiation Oncology d Biology d Physics
1464
Volume 78, Number 5, 2010
Fig. 7. Typical isodose and profile measurements from plan quality assurance with (a) Matrixx and (b) EDR film.
These and other studies demonstrate the value of enabling clinics to bring SRS or SRT to patients as a treatment option. In our study, the ability to image and treat intracra-
nial lesions in time slots of 20 min on a linear accelerator routinely used to treat other sites with larger volumes (e.g., breast, lung, prostate) is potentially a significant factor for Effect of Dose Rate on Survival (HX34) 1
1.8
Effect of dose normalization point on CIICRU for spherical PTV
1.6
2 Gy/fx
Survival N/N N0
Hypothetical CIICRU
1.4 12 . 1 0.8
0.1
7 Gy/fx 10 Gy/fx
0.01
0.6 0.4
PTV radius - 0.7 mm
0.2
0.001 0.01
0 0
1
2
3
4
5
6
7
8
9
10
11
12
13
7Gy/4.5 min
7Gy/45 min
PTV radius + 0.7 mm
14
PTV Volume (cm3)
Fig. 8. Values for CIICRU vs. volume are calculated for a hypothetical, spherical planning target volume (PTV) when the plan is normalized to deliver the prescribed dose 0.7 mm from the PTV surface.
0.1
1
10
100
Dose Rate (Gy/min)
Fig. 9. Data on survival fraction vs. single dose fraction for the HX34 cell line in Paganetti (27, Fig. 3) are recast as survival fraction vs. dose rate to highlight the potential of smaller treatment times for producing enhanced biological effect for single and hypofractionated doses.
RapidArc SRT d C. S. MAYO et al.
1465
Fig. 10. Multiple lesions may be treated with a single pair of RapidArcs. The figure illustrates a hypothetical case of eight lesions distributed throughout the cranium.
enabling more routine use of SRS and SRT to treat metastatic lesions. The use of a hypofractionated SRS regimen (SRT) permits delivery of stereotactic ablative doses to lesions within eloquent areas of the brain. Preliminary results of studies using hypofractionated SRS are comparable to both surgery and SRS data for brain metastases in terms of local control and overall survival with acceptable morbidity (32, 33). In addition, it is possible to use volumetric IMRT to treat multiple metastatic lesions using a single isocenter and 2–3 arcs. Figure 10 illustrates a hypothetical patient with eight lesions distributed throughout the cranium treated to 15 Gy in one fraction. Special equipment to adjust the pitch, roll, and yaw angles would be necessary to ensure that small rotations do not translate into shifts beyond the PTV margins. We an-
ticipate that the ability to treat multiple lesions simultaneously will have a significant impact on care patterns for patients with metastatic lesions.
CONCLUSION Volumetric IMRT with RapidArc for frameless intracranial stereotactic radiosurgery is well positioned alongside other technologies, producing dose distributions with comparable conformality and dose gradients. In addition to all the benefits of a frameless SRS system, such as patient comfort and the ability to use hypofractionated regimens, RapidArc frameless SRS provides for a more efficient use of both personnel and equipment resources.
REFERENCES 1. Chen JCT, Rahimian J, Girvigian MR, Miller MJ. Contemporary methods of radiosurgery treatment with the Novalis Linear Accelerator System. Neurosurg Focus 2007;23. E4: 1–9. 2. Benedict SH, Cardinale RM, Wu Q, et al. Intensity-modulated stereotactic radiosurgery using dynamic micro-multileaf collimation. Int J Radiat Oncol Biol Phys 2001;50:751–758. 3. Baisden JM, Benedict SH, Sheng K, et al. Helical TomoTherapy in the treatment of central nervous system metastasis. Neurosurg Focus 2007;22:E8. 4. Han C, Liu A, Schultheiss TE, et al. Dosimetric comparisons of Helical TomoTherapy treatment plans and step-and-shoot inten-
sity-modulated radiosurgery treatment plans in intracranial stereotactic radiosurgery. Int J Radiat Oncol Biol Phys 2006;65: 608–616. 5. Colombo F, Casentini L, Cavedon C, et al. CyberKnife radiosurgery for benign meningiomas: Short-term results in 199 patients. Neurosurgery 2009;64(Suppl):A7–A13. 6. Collins SP, Coppa ND, Zhang Y, et al. CyberKnife radiosurgery in the treatment of complex skull base tumors: Analysis of treatment planning parameters. Radiat Oncol 2006;1:46. 7. Yu C, Jozef G, Apuzzo MLJ, Petrovich Z. Dosimetric comparison of CyberKnife with other radiosurgical modalities for an ellipsoidal target. Neurosurgery 2003;53:1155–1162.
1466
I. J. Radiation Oncology d Biology d Physics
8. Verbakel WF, Cuijpers JP, Hoffmans D, et al. Volumetric intensity-modulated arc therapy vs. conventional IMRT in head-andneck cancer: A comparative planning and dosimetric study. Int J Radiat Oncol Biol Phys 2009;74:252–259. 9. Fogliata A, Clivio A, Nicolini G, et al. Intensity modulation with photons for benign intracranial tumours: A planning comparison of volumetric single arc, helical arc and fixed gantry techniques. Radiother Oncol 2008;89:254–262. 10. Clivio A, Fogliata A, Franzetti-Pellanda A, et al. Volumetricmodulated arc radiotherapy for carcinomas of the anal canal: A treatment planning comparison with fixed field IMRT. Radiother Oncol 2009;92:118–124. 11. Cozzi L, Clivio A, Gauman G, et al. Comparison of advanced irradiation techniques with photons for benign intracranial tumors. Med Phys 2006;80:268–273. 12. Vanetti E, Clivio A, Nicolini G, et al. Volumetric modulated arc radiotherapy for carcinomas of the oro-pharynx, hypo-pharynx and larynx: A treatment planning comparison with fixed field IMRT. Radiother Oncol 2009;92:111–117. 13. Palma D, Vollans E, James K, et al. Volumetric modulated arc therapy for delivery of prostate radiotherapy: Comparison with intensity-modulated radiotherapy and three-dimensional conformal radiotherapy. Int J Radiat Oncol Biol Phys 2008;72: 996–1001. 14. Cozzi L, Dinshaw KA, Shrivastava SK, et al. A treatment planning study comparing volumetric arc modulation with RapidArc and fixed field IMRT for cervix uteri radiotherapy. Radiother Oncol 2008;89:180–191. 15. Kjær-Kristoffersen F, Ohlhues L, Medin J, Korreman S. RapidArc volumetric modulated therapy planning for prostate cancer patients. Acta Oncol 2008;48:227–232. 16. Fogliata A, Yartsev S, Nicolini G, et al. On the performance of intensity modulated Protons, RapidArc, and Helical TomoTherapy for selected paediatric cases. Radiat Oncol 2009;4:2. 17. Ling CC, Zhang P, Archambault Y, et al. Commissioning and quality assurance of RapidArc radiotherapy delivery. Int J Radiat Oncol Biol Phys 2008;72:575–581. 18. Nicolini G, Vanetti E, Clivio A, et al. The GLAaS algorithm for portal dosimetry and quality assurance of RapidArc an intensity modulated rotational therapy. Radiother Oncol 2008;3:24. 19. Korreman S, Medin J, Kjær-Kristoffersen F. Dosimetric verification of RapidArc treatment delivery. Acta Oncol 2008;1–7. 20. Vanetti E, Nicolini G, Clivio A, et al. The impact of treatment couch modelling on RapidArc. Phys Med Biol 2009;54: N157–N166.
Volume 78, Number 5, 2010
21. Gagne IM, Ansbacher W, Zavgorodni S, et al. A Monte Carlo evaluation of RapidArc dose calculations for oropharynx radiotherapy. Phys Med Biol 2008;53:7167–7185. 22. Bush K, Townson R, Zavgorodni S. Monte Carlo simulation of RapidArc radiotherapy delivery. Phys Med Biol 2008;53: N359–370. 23. Mayo C, Urie M. Eclipse IMRT—A practical treatment planning guide. Franklin, MA: KnowNexus; 2004. 24. Bolsi A, Fogliata A, Cozzi L. Radiotherapy of small intracranial tumours with different advanced techniques using photon and proton beams: A treatment planning study. Radiother Oncol 2003;68:1–14. 25. Paddick I. A simple scoring ratio to index the conformity of radiosurgical treatment plans. Technical note. J Neurosurg 2000; 93(Suppl 3):219–222. 26. Wagner TH, Bova FJ, Friedman WA, et al. A simple and reliable index for scoring rival stereotactic radiosurgery plans. Int J Radiat Oncol Biol Phys 2003;57:1141–1149. 27. Paganetti H. Changes in tumor cell response due to prolonged dose delivery times in fractionated radiation therapy. Int J Radiat Oncol Biol Phys 2005;63:892–900. 28. Rades D, Kueter JD, Veninga T, et al. Whole brain radiotherapy plus stereotactic radiosurgery (WBRT + SRS) versus surgery plus whole brain radiotherapy (OP + WBRT) for 1–3 brain metastases: Results of a matched pair analysis. Eur J Cancer 2009; 45:400–404. 29. Sneed PK, Suh JH, Goetsch SJ, et al. A multi-institutional review of radiosurgery alone vs. radiosurgery with whole brain radiotherapy as the initial management of brain metastases. Int J Radiat Oncol Biol Phys 2002;53:519–526. 30. Pirzkall A, Bebus J, Lohr F, et al. Radiosurgery alone or in combination with whole-brain radiotherapy for brain metastases. J Clin Oncol 1998;16:3563–3569. 31. Giubilei C, Ingrosso G, D’Andrea M, et al. Hypofractionated stereotactic radiotherapy in combination with whole brain radiotherapy for brain metastases. J Neurooncol 2009;91:207– 212. 32. Narayana A, Chang J, Yenice K, et al. Hypofractionated stereotactic radiotherapy using intensity-modulated radiotherapy in patients with one or two brain metastases. Stereotact Funct Neurosurg 2007;85:82–87. 33. Molenaar R, Wiggenraad R, Verbeek-de Kanter A, et al. Relationship between volume, dose and local control in stereotactic radiosurgery of brain metastasis. Br J Neurosurg 2009;23:170– 178.
doi:10.1016/j.ijrobp.2009.10.005
CLINICAL INVESTIGATION
Central Nervous System
INITIAL EXPERIENCE WITH VOLUMETRIC IMRT (RAPIDARC) FOR INTRACRANIAL STEREOTACTIC RADIOSURGERY CHARLES S. MAYO, PH.D.,* LINDA DING, PH.D.,* ANTHONY ADDESA, M.D.,* SIDNEY KADISH, M.D.,* T. J. FITZGERALD, M.D.,* AND RICHARD MOSER, M.D.*y *Department of Radiation Oncology and yDivision of Neurosurgery, Department of Surgery, University of Massachusetts Medical School, Worcester, MA Purpose: Initial experience with delivering frameless stereotactic radiotherapy (SRT) using volumetric intensitymodulated radiation therapy (IMRT) delivered with RapidArc is presented. Methods and Materials: Treatment details for 12 patients (14 targets) with a mean clinical target volume (CTV) of 12.8 ± 4.0 cm3 were examined. Dosimetric indices for conformality, homogeneity, and dose gradient were calculated and compared with published results for other frameless, intracranial SRT techniques, including CyberKnife, TomoTherapy, and static-beam IMRT. Statistics on setup and treatment times and per patient dose validations were examined. Results: Dose indices compared favorably with other techniques. Mean conformality, gradient, and homogeneity index values were 1.10 ± 0.11, 64.9 ± 14.1, 1.083 ± 0.026, respectively. Median treatment times were 4.8 ± 1.7 min. Conclusion: SRT using volumetric IMRT is a viable alternative to other techniques and enables short treatment times. This is anticipated to have a positive impact on radiobiological effect and for facilitating wider use of SRT. Ó 2010 Elsevier Inc. RapidArc, Radiosurgery, Treatment time.
INTRODUCTION
tems, Palo Alto, CA) has been demonstrated by several investigators, indicating that it is a mature technology (17–22). Addition of volumetric arc IMRT to the armament of image-guided techniques for delivering intracranial stereotactic radiosurgery is a natural extension of the technology. The combination of modulated intensities with an arc-based approach positions the technique intermediate to arc-based techniques using cones and fixed-beam techniques using MLCs. This report details our initial experience in the use of volumetric IMRT with RapidArc.
Options for stereotactic radiosurgery with intracranial lesions have extended beyond frame-based with fixed multileaf collimator (MLC) or cone-based approaches on linear accelerators to included frameless image-guided approaches such as fixed-beam intensity-modulated radiation therapy (IMRT) (1, 2), TomoTherapy (3, 4), and CyberKnife (5–7). These allow for highly conformal treatment of lesions using technologies that are already used for routine treatment of nonstereotactic patients. In addition, the frameless approach facilitates the use of hypofractionated protocols. Recently, arc-based IMRT techniques have emerged as a promising progression from fixed-field techniques (8–16). Modulation of the intensity over the course of an arc can enable reduction in overall treatment time and reduced dose to normal tissue structures compared with IMRT. The accuracy of volumetric IMRT using RapidArc (Varian Medical Sys-
METHODS AND MATERIALS Twelve patients (Table 1) were treated between January 2009 and June 2009 with volumetric IMRT using RapidArc from Varian Medical Systems. Patients were immobilized in an alpha-cradle-based system for stereotactic radiosurgery (MayoMold, by CDR Systems www.cdrsys.ca, Alberta, Canada; Fig. 1) and then CT scanned in the treatment position using a helical scanner and slice spacing of 1.25 mm. Scans were fused with MRI. The MRI protocol adopted after Acknowledgment—We acknowledge presentations of Dr. Joseph Ting, in which he pointed out the potential role of reduced treatment time with RapidArc for achieving increased biological effect, and of Dr. Richard Popple illustrating use of RapidArc to treat multiple intracranial lesions. We thank Ms. Julie Trifone for demonstrating the immobilization system in Fig. 1. Received July 6, 2009, and in revised form Oct 6, 2009. Accepted for publication Oct 7, 2009.
Reprint requests to: Charles S. Mayo, Ph.D., Department of Radiation Oncology, HB200, University of Massachusetts Medical School, 55 Lake Avenue North, Worcester, MA 01655. Tel: (744) 442-5560; Fax: (774) 442-5006; E-mail: charles.mayo@ umassmemorial.org Conflicts of interest: The first author, has research grant support from Varian Medical Systems. 1457
1458
I. J. Radiation Oncology d Biology d Physics
Volume 78, Number 5, 2010
Table 1. Characteristics of treated patients Gender Male Female Age 56 11 years Histology Primary diagnosis Metastatic lung Melanoma Metastatic esophagus Metastatic breast Metastatic colon Metastatic renal cell Chemotherapy None Tarceva Previous cranial irradiation Whole brain Conformal Radiation Therapy None
8 4 Range, 49–81 years 7 1 1 1 1 1 10 2 7 1 4
March 2009 used T1-VIBE scans with a spacing of 1.25 mm. Target clinical target volumes (CTVs) were delineated on the fused scans and a planning target volume (PTV) was constructed with a 1- to 2-mm margin. Patients were treated on a Trilogy accelerator (Varian Medical Systems www.varian.com, Palo Alto, CA) equipped with a Millennium 120 multileaf collimator (MLC). The accelerator is equipped with a stereotactic beam mode with a beam energy of 6 MV delivered at 1,000 MU/min with a maximum field size of 15 15 cm. Patients were positioned using kV planar and cone-beam CT images. The setup and localization uncertainty of this system is <1 mm. This is confirmed on treatment days with a phantom designed for end-to-end testing from CT through delivery. The phantom is illustrated in Fig. 3. Treatment plans were created in the Eclipse (version 8.6, Varian Medical Systems, www.varian.com, Palo Alto, CA) treatment planning system with 2–3 arcs per isocenter. At least one of the arcs was noncoplanar. Figure 2 illustrates a typical approach of one 350 arc with no table rotation and a second vertex arc of 180 . In optimization, the minimum dose constraint on the PTV was set to the prescribed dose. A dose-limiting annulus (DLA) tuning structure (23) was created to facilitate creating a steep dose gradient beyond the PTV. The outer surface of the DLA was created with a 3-cm margin on the PTV; the inner surface was 1 mm from the PTV (Fig. 4a). The maximum dose constraint on the PTV was set to the prescribed dose. In addition, a normal tissue constraint was set that specified shape of the dose profile away from the PTV (Fig. 4b). Plans were normalized so that at least 96% of the PTV and 100% of the CTV would be covered by the prescribed dose. Dose distributions were compared with other published SRS options using dose-conformity and dose-gradient indices. Correlations among the indices were examined. Because SRS with static MLCs with small leaves is a common approach, a hypothetical machine was created in Eclipse using the same beam data but with a high-definition MLC (0.25 cm over central 10 cm). Static MLC plans were created for each of the targets, and dosimetric indices were compared. The conformity index of International Commission on Radiation Units and Measurements (ICRU) 62 is calculated as the ratio of the volume enclosed by the prescription isodose surface (VRx) to the volume of the PTV (VPTV): CIICRU ¼
VRx : VPTV
(1)
Fig. 1. (a) Patients are positioned in an alpha-cradle-based system that conforms to the posterior half of the patient from the crown of the head through the shoulders. An aquaplast mask indexes to both the cradle and the patient features when the patient is in the correct position. (b) The alpha cradle is indexed to a graphite board that cantilevers over the end of the treatment couch to facilitate used of posterior oblique angles.
Other authors have calculated the conformity index as the ratio of the volume encompassed by 95% of the prescription isodose surface (V95%Rx) (24). CI95%Rx ¼
V95%Rx VPTV
(2)
Paddick (25) noted that because the ICRU definition assumes but does not check for overlap of VRx and VPTV, false values of 1 were possible. The inverse of Paddick’s formulation is used by other authors. CIInvPaddick ¼
VRx VPTV ; 2 VRxXPTV
(3)
where VRxXPTV is the volume of the intersection of the PTV and the prescription isodose surface.
RapidArc SRT d C. S. MAYO et al.
1459
HISD ¼ 1 þ
DSD : DRx
(9)
The overall homogeneity index was calculated as the geometric average of the first two: pffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi HIOverall ¼ HIMax HISTD : (10)
Fig. 2. Patients were treated with two to three arcs. The figure illustrates a three-dimensional view of two arcs, one in the transverse plane and one in the sagittal plane. SRT = stereotactic radiotherapy.
The ability of a plan to spare normal tissues outside of the PTV depends on steepness gradient of decreasing dose away from the PTV. Wagner (26) introduced the gradient score of the Conformality Gradient Index (CGIg) for stereotactic radiosurgery: CGIg ¼ 100 {100 REff ;50%Rx REff ;Rx 0:3 } (4) The effective radius of the volume encompassed by the prescription isodose surface is rffiffiffiffiffiffiffiffiffi 3 3VRx REff ;Rx ¼ : (5) 4p The effective radius of the volume encompassed by 50% of the prescription isodose surface (V50%Rx) is: rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 3 3V50%Rx REff ;50%Rx ¼ : (6) 4p
RESULTS
In addition to CGIg, the gradient was calculated as GrEff ¼
50% : REff ;50%Rx REff ;Rx
(7)
The gradient was also measured (GrMeas) as the average gradient in the right-left, anterior-posterior, and cranial-caudal planes through the isocenter. Uniformity of the dose within the PTV volume was compared using three homogeneity indices (HI). The first, which is used by many authors, is the ratio of the maximum dose within the PTV (Dmax) to the prescribed dose (DRx): HIMax ¼
DMax : DRx
Dose distributions were validated with two measurement methods. In the first method, RapidArc plans were projected onto a phantom consisting of the Matrixx ionization chamber array (Scandiatronix-Wellhoffer www.scanditronix-wellhofer.com, Bartlett, TN) with 8 cm of plastic water above and below the array. The phantom was positioned on the accelerator as for the patient and treated with the same RapidArc beams. Calculated and measured doses were compared in Matrixx to the dose distribution calculated in Eclipse. This can be problematic for small targets, because cylindrical ionization chambers in the Matrixx array (4.5 mm diameter, 5.0 mm length) are spaced on a 7.62-mm grid. The nearest chambers to the center of the grid array are at 0.38 cm in the x and y directions. For the second method, measurements were repeated using Extended Dose Range (EDR) film (Eastman Kodak, www.kodak.com, Rochester, NY) in plastic water at the same plane as the Matrixx array. Monitor units were scaled to dose levels that would not saturate the EDR film. A set of dose calibration films was taken in the same session. Films were scanned with a Vidar scanner. Predicted and measured distributions with the scaled monitor units were evaluated using the RIT113 (Radiological Imaging Technology, www.radimage.com, Colorado Springs, CO) software package. An independent monitor unit (MU) check was also performed. The doses delivered at isocenter were calculated in RadCalc software (Life Line Software, www.lifelinesoftware.com, Tyler, TX) and compared with predicted values from Eclipse. Because the Matrixx and RadCalc measures are independent, the average agreement of the two methods was also calculated. Correlations among values of various indices in the sample were examined by calculation of correlation coefficients and of 95% confidence intervals (95% CI) for the coefficients using Fisher’s transforms. Values were judged to be uncorrelated in the sample if the 95% CI contained the value 0. Significance of differences between means was assessed using a two-tailed t test with level of significance set at 0.025.
(8)
To gain a perspective on the statistical range of doses within the PTV, a second homogeneity index was calculated using the standard deviation of the dose within the PTV (DSD):
Treatment plan characteristics for each of the patient targets are detailed in Table 2. Couch angles and angular range of arcs for the targets treated are presented in Table 2. The majority (11 of 14) were treated with two arcs, and the remainder were treated with three arcs. A transverse arc (couch 0) was used for all patients. The cumulative arc length used to treat each target ranged from 500 to 720 with a mean of 563 72.4. No correlation was observed between total arc angle and CIICRU (r = –0.20; 95% CI, –0.66 to 0.37) or CGIg (r = 0.44; 95% CI, –0.12 to 0.79). Transverse isodose distributions and dose profiles for targets with CTV volumes of 0.1, 1.3, and 12.6 cm3 are illustrated in Fig. 5. Table 3 shows median CTV and PTV volumes were 1.20 3.69 and 2.35 6.01. Mean values for CIICRU, CI95%Rx, and CIInv-Paddick were 1.10 0.11,
1460
I. J. Radiation Oncology d Biology d Physics
Volume 78, Number 5, 2010
Fig. 3. (a) A phantom for end-to-end testing consisting of a pattern of cerrobend and air-filled holes on a scribed grid allows checking congruence of light field, MV, kV, cone-beam CT, and bite block isocenters. (b) Cerrobend markers are easily visualized on MV and kV plane images. Here a misalignment of <1 mm in the coronal plane is visualized for cone-beam CT using the air markers. (c) After alignment, a film placed at isocenter is irradiated with the RapidArc beams. The central cerrobend dot produces an artifact on demonstrating agreement of radiation and phantom isocenters. Phantom will be discussed in a separate publication.
1.54 0.34, and 1.32 0.20, respectively. Becausse CIInvPaddick is a ratio of products of two volumes, whereas the other indices are ratios of volumes, it was of interest what would result from a more comparable function. Calculating the square root of CIInv-Paddick for each target, the mean was 1.15 0.08. The conformity index CIICRU had a weak positive correlation with CI95%Rx (r = 0.73; 95% CI, 0.32–0.91) and a weak negative correlation with CIInv-Paddick (r = –0.54; 95% CI, –0.01 to –0.83). No correlation was found between CI95%Rx and CIInv-Paddick or to volume of the CTV or PTV. Thus, in this sample, none of the conformity indices was an accurate surrogate for another. Gradient indices and homogeneity indices are examined in Table 4. Measured gradients GrMeas were slightly steeper than the effective gradients GrEff with mean values of 83.7% 15.8%/cm and 80.0% 16.6%/cm, respectively. Median value for CGIg was 64.9 14.1. A linear regression on CGIg found an increased with GrMeas with a slope of 0.834 0.095 and intercept of –4.9 8.0. No correlation was
found between gradient indices (GrMeas,GrEff, CGIg) and conformity indices (CIICRU, CI95%Rx, CIInv-Paddick). However, CTV volume correlated negatively with CGIg (r = –0.71; 95% CI, –0.9 to –0.28) and GrMeas (r = –0.58; 95% CI, –0.85 to –0.07). Figure 6 shows the measured gradient for CTV and PTV volumes in this study decreasing from values >100 %/cm for small volumes toward a plateau of approximately 60%/cm for CTV volumes >10 cm3. Dose distributions within the PTV volume were homogeneous. Mean values for HIMax, HISD, and HIOverall were 1.083 0.026, 1.017 0.005 and 1.049 0.014, respectively. HIMax did not correlate with either CIICRU or CGIg. However, HISD did correlate with CIICRU (r = –0.85, 95% CI, –0.37 to –0.93), CGIg (r = 0.78, 95% CI, 0.43–0.92) and GrMeas(r = 0.81, 95% CI, 0.50–0.94). This indicates that as target volumes decreased allowing steeper gradients, dose homogeneity within the target also decreased. Most dosimetric indices for plans with static 0.25-cm leaves were not significantly different from RapidArc with
RapidArc SRT d C. S. MAYO et al.
1461
Fig. 4. (a) A dose-limiting annulus tuning structure is created to help control dose conformality and gradients near the surface of the planning target volume (PTV). (b) This is used in conjunction with normal tissue constraint in the optimizer that specifies the dose profile up to the PTV. CTV = clinical target volume.
0.5-cm leaves. Mean conformality values of CIICRU and CI95%Rx were 1.1 0.11 and 1.56 0.19, respectively. Gradient indices GrMeas and CGIg were 76.2 13.3 and 61.3 9.0, respectively. Homogeneity indices HIMax and HISD were 1.099 0.013 and 1.024 0.004, respectively. The small increase in HISD was significant (p = 0.004). Direct measurements of dose distributions in a phantom are illustrated in Fig. 7. Mean agreement of central axis dose was –1.4 3.6% and 1.6 2.4% for Matrixx and film dosimetry, respectively. Independent monitor unit calculations with RadCalc calculations of central axis dose showed good agreement with a mean of –1.8% 5.0%. Table 2. Couch angles and angular ranges used to treat patients Arc 1
Arc 2
Mean setup for the 14 targets treated with hypofractionation were 19 9 min. Setup time for the single fraction SRS was 20 min. The time required to acquire (image and save to server) a pair of orthogonal kV images or a conebeam CT are 3 and 4 min, respectively. Thus the additional setup time is due to waiting for the necessary staff to assemble and determine adjustments. Mean treatment time for the hypofractionated treatments was 4.8 1.7 min. Treatment time for the single fraction treatment (15 Gy) was 7 min. Median follow-up for the patients was 3 months (range, 1–6). There were no cases of progression or adverse events. One patient progressed at other sites in the central nervous system. None of the patients receiving 3 7 Gy required the addition or increase of steroids secondary to idiosyncratic reactions.
Arc 3
DISCUSSION Couch Arc Couch Arc Couch Arc Patient Target angle range angle range angle range 1 2 3 4 5* 6 7 8 9 10 11 12
1 1 1 1 1 1 2 1 1 1 1 1 1 2
0 0 0 0 0 0 0 0 0 0 0 0 0 0
350 350 350 350 350 350 350 350 350 358 358 350 350 350
300 300 270 290 300 90 90 267 270 90 270 270 270 270
175 145 185 175 175 175 175 175 175 179 180 200 150 175
60 60
175 210
45
175
* 15 Gy in one fraction. All others are 7 Gy in three fractions.
In this study, results for dosimetric indices with RapidArc SRT on an accelerator equipped with 0.5-cm MLC leaves compared favorably to other techniques. In their introduction of the gradient score index (CGIg), Wagner et al. (26) indicated that scores $90 were typical for small targets treated with cones and simple geometries, whereas scores in the range of 60–80 are attainable for larger or more complex targets. Bolsi et al. (24) reported on a planning comparison of intracranial targets for 12 patients using several treatment modalities. Median target volume was 2.37 cm3 (range, 0.49–14.32 cm3). Mean values of CI95%Rx for stereotactic arc therapy (four–five noncoplanar arcs with cones), threedimensional conformal therapy (4 coplanar fields), IMRT (4 coplanar fields), spot scanning protons (1–3 fields), and
I. J. Radiation Oncology d Biology d Physics
1462
Volume 78, Number 5, 2010
Fig. 5. Transverse plane isodose distributions are shown for the range of clinical target volume (CTV) volumes in this study: (a) 0.1 cm3; (b) 1.1 cm3; and (c) 12.6 cm3.
passive scattering protons were 3.8 1.6, 2.8 0.6, 4.4 1.7, 3.2 1.0, and 2.5 0.7, respectively. In a subsequent study, Cozzi et al. (11) reported on a planning comparison on intracranial targets for 12 patients using various treatment modalities. Median target volume was 2.37 cm3 (range, 0.49–14.32 cm3). Mean values of CI95%Rx for stereotactic arcs (3–5 noncoplanar arcs with cones), IMRT (3–4 coplanar fields, 5-mm MLC), dynamic IMRT arcs (Ergo++, 3–5 noncoplanar arcs, 3-mm MLC), Helical TomoTherapy, and Cyberknife were 2.9 1.2, 3.0 0.9, 3.9 1.5, 1.8 0.6, and 1.8 0.3, respectively. Han et al. (4) reported on a planning comparison study of 16 patients with mean target volumes of 14.9 12.41 cm3 (range, 1.45–40.78 cm3). Mean value of CIInv-Paddick for coplanar IMRT (6–8 fields, 0.5-cm MCL), noncoplanar IMRT (9–12 fields, 0.5-cm MLC), and Helical TomoTherapy were 1.53 0.38, 1.35 0.15, and 1.26 0.15, respectively.
Mean homogeneity indices (HIDMax) were 1.15 0.05, 1.13 0.04, and 1.18 0.06 for coplanar IMRT, noncoplanar IMRT, and Helical TomoTherapy, respectively. Mean gradient score indices (CGIg) were 13.7 19.08, 22.32 19.20, and 43.28 13.78 for coplanar IMRT, noncoplanar IMRT, and Helical TomoTherapy, respectively. Colombo et al. (5) reviewed results for 199 patients treated for benign meningiomas with CyberKnife (5). Median target volume was 6.9 cm3 (range, 0.1–64 cm3. The mean CIICRU was 1.18 (range, 1.01–1.48). They calculated a modified conformity index corresponding to CIInv-Paddick with a mean value of 1.29 (range, 1.14–1.52). Mean homogeneity index (HIMax) was 1.35 (range, 1.18–2.01) Collins et al. (6) reported on patients treated with CyberKnife. Median values for CIInv-Paddick and HIMax were 1.64 (range, 1.04–3.11) and 1.19 (range, 1.11–1.54) respectively for similar patients (Group II). No gradient indices were reported.
Table 3. Conformity indices for each of the targets displayed in the context of CTV and PTV volumes Patient 1 2 3 4 5 6 7 8 9 10 11 12 Mean SD
Target 1 1 1 1 1 1 2 1 1 1 1 1 1 2
CTV (cm3)
PTV (cm3)
CIICRU
CI95%Rx
CIInv-Paddick
0.3 0.1 3.3 1.1 1.3 0.1 0.8 0.1 0.9 3.3 9.1 12.6 3.2 3.1 2.8 4.0
0.7 0.6 8.7 2 2.3 0.7 2.2 0.4 2.4 7 16.8 19.3 4.8 4.7 5.2 6.01
1.14 1.17 1.09 0.90 0.87 1.14 1.05 1.25 1.17 1.17 1.21 1.12 1.13 1.06 1.10 0.11
1.86 1.67 1.29 1.25 1.09 1.57 1.36 2.50 1.63 1.49 1.54 1.34 1.56 1.47 1.54 0.34
1.56 1.17 1.20 1.84 1.42 1.56 1.27 1.25 1.27 1.21 1.21 1.17 1.22 1.21 1.32 0.2
Abbreviations: CI = conformity index; CTV = clinical target volume; PTV = planning target volume. CIICRU < 1.0 and CI95%Rx > 1.0 corresponds the PTV not fully covered by the prescribed dose but receiving at least 95% of the prescribed dose. Figure 8 illustrates the sensitivity of CIICRU to plan normalization.
RapidArc SRT d C. S. MAYO et al.
1463
Table 4. Gradient indices compared with measurements of gradients and homogeneity indices Patient 1 2 3 4 5 6 7 8 9 10 11 12
Target
CGIg
GrEff (%/cm)
GrMeas (%/cm)
1 1 1 1 1 1 2 1 1 1 1 1 1 2
68.4 82.7 67.4 77.4 80.1 79.4 76.5 65.3 62.1 64.5 46.3 45.4 53.4 39.3 64.9 14.1
81.1 105.6 79.8 95.0 100.2 98.9 93.4 77.2 73.6 76.3 59.7 59.1 65.3 55.1 80.0 16.6
85.2 103.4 94.3 96.2 96.8 107.9 96.8 82.4 71.4 76.9 69.8 65.8 62.8 61.7 83.7 15.8
Mean SD
HIMax
HISD
HIOverall
1.083 1.108 1.120 1.067 1.107 1.125 1.099 1.054 1.039 1.087 1.075 1.082 1.064 1.060 1.083 0.026
1.018 1.020 1.020 1.023 1.026 1.019 1.023 1.011 1.011 1.014 1.010 1.014 1.011 1.013 1.017 0.005
1.050 1.063 1.069 1.045 1.066 1.071 1.060 1.032 1.025 1.050 1.042 1.048 1.037 1.036 1.049 0.014
Abbreviations: GR = gradient index, HI = Homogeneity Index, CGIg = gradient score of Conformality Gradient Index.
No significant difference was found between the conformality indices in this study and those for the other frameless SRT modalities. Mean CGIg values were in agreement with the range found by Wagner. Compared with the TomoTherapy results of Han, the gradient index was significantly (p = 0.0002) larger for our study. Values for the homogeneity index, HIMax, were significantly smaller than Han (p < 0.0001) and for other modalities. The low value of HISD = 1.016, demonstrated uniformity of the dose distribution throughout the PTV volume. Thus, dose distributions obtained using RapidArc with an MLC having a 0.5-cm leaf width are at minimum comparable to other, more widely used techniques for frameless SRT. Sharper dose gradients and better homogeneity are possible. It is worth noting when making comparisons that dependence of conformality index values can depend significantly on normalization chosen in a plan. For small volumes, the difference between normalizing to deliver full dose to 95% of the PTV volume or to the corresponding larger volume is significant. For a hypothetical spherical PTV and perfectly
Measured Gradient (%/cm)
110 CTV
100
PTV
90 80 70 60 50
0
5
10
15
20
25
Volume (cm3)
Fig. 6. Measured dose gradients decrease as the volume of the clinical target volume (CTV) increases. A threshold in the range of 2 cm3 is evident, with gradients larger than 100%/cm for volumes <0.5 cm3. PTV = planning target volume.
conformal plan, CIICRU would be equal to 1 if normalized to cover exactly the radius of the PTV. The distance between the 100% and 95% isodose lines is on the order of 0.7 mm. Figure 8 illustrates the strong impact on CIICRU if the plan is instead normalized to cover inside vs. outside the PTV volume by 0.7 mm. RapidArc treatment times in the range of 4–7 min can be significantly smaller compared with other modalities. Han (4) reported the treatment time for Helical TomoTherapy patients to be 42 16 min. Colombo (5) did not report on treatment times for CyberKnife. In our experience, treatment times for frame-based SRS using a mini-multileaf collimator are in the range of 45–60 min for a single lesion, not including setup time. The short treatment time required for delivery may translate into larger biological effect. Paganetti (27) calculated dose rate effects for nine cell lines using the lethal/potentially lethal model. Recasting data from that study to plot HX34 cell line survival fraction vs. dose rate as an example, Fig. 9 illustrates survival decreased by a factor of 1.06 for 2 Gy per fraction delivered at 0.2 vs. 2.0 Gy/min. However, at a larger dose fraction, 7 Gy, survival was reduced by a factor of 2.75. In addition, reduction in treatment time may also reduce the magnitude of intrafraction motions. Recently, in a study of 104 patients, Rades et al. (28) demonstrated equivalence in control for whole brain (WB) radiotherapy plus stereotactic radiosurgery (SRS) vs. surgery plus WB radiotherapy for 1–3 brain metastases. A multi-institutional study by Sneed et al. (29) of 569 patients comparing outcomes of WB to WB + SRS WB radiation therapy to found equivalent survival rates for both groups. Reporting on a study of 236 patients with metastatic lesions, Pirzkal et al. (30) demonstrated improved survival for WB + SRS vs. SRS alone. Giubilei et al. (31) demonstrated safe use (with a different technology Ergo++ and 3DLine) of a frameless hypofractionated stereotactic radiation therapy (SRT) with concomitant whole brain irradiation in a group of 30 patients.
I. J. Radiation Oncology d Biology d Physics
1464
Volume 78, Number 5, 2010
Fig. 7. Typical isodose and profile measurements from plan quality assurance with (a) Matrixx and (b) EDR film.
These and other studies demonstrate the value of enabling clinics to bring SRS or SRT to patients as a treatment option. In our study, the ability to image and treat intracra-
nial lesions in time slots of 20 min on a linear accelerator routinely used to treat other sites with larger volumes (e.g., breast, lung, prostate) is potentially a significant factor for Effect of Dose Rate on Survival (HX34) 1
1.8
Effect of dose normalization point on CIICRU for spherical PTV
1.6
2 Gy/fx
Survival N/N N0
Hypothetical CIICRU
1.4 12 . 1 0.8
0.1
7 Gy/fx 10 Gy/fx
0.01
0.6 0.4
PTV radius - 0.7 mm
0.2
0.001 0.01
0 0
1
2
3
4
5
6
7
8
9
10
11
12
13
7Gy/4.5 min
7Gy/45 min
PTV radius + 0.7 mm
14
PTV Volume (cm3)
Fig. 8. Values for CIICRU vs. volume are calculated for a hypothetical, spherical planning target volume (PTV) when the plan is normalized to deliver the prescribed dose 0.7 mm from the PTV surface.
0.1
1
10
100
Dose Rate (Gy/min)
Fig. 9. Data on survival fraction vs. single dose fraction for the HX34 cell line in Paganetti (27, Fig. 3) are recast as survival fraction vs. dose rate to highlight the potential of smaller treatment times for producing enhanced biological effect for single and hypofractionated doses.
RapidArc SRT d C. S. MAYO et al.
1465
Fig. 10. Multiple lesions may be treated with a single pair of RapidArcs. The figure illustrates a hypothetical case of eight lesions distributed throughout the cranium.
enabling more routine use of SRS and SRT to treat metastatic lesions. The use of a hypofractionated SRS regimen (SRT) permits delivery of stereotactic ablative doses to lesions within eloquent areas of the brain. Preliminary results of studies using hypofractionated SRS are comparable to both surgery and SRS data for brain metastases in terms of local control and overall survival with acceptable morbidity (32, 33). In addition, it is possible to use volumetric IMRT to treat multiple metastatic lesions using a single isocenter and 2–3 arcs. Figure 10 illustrates a hypothetical patient with eight lesions distributed throughout the cranium treated to 15 Gy in one fraction. Special equipment to adjust the pitch, roll, and yaw angles would be necessary to ensure that small rotations do not translate into shifts beyond the PTV margins. We an-
ticipate that the ability to treat multiple lesions simultaneously will have a significant impact on care patterns for patients with metastatic lesions.
CONCLUSION Volumetric IMRT with RapidArc for frameless intracranial stereotactic radiosurgery is well positioned alongside other technologies, producing dose distributions with comparable conformality and dose gradients. In addition to all the benefits of a frameless SRS system, such as patient comfort and the ability to use hypofractionated regimens, RapidArc frameless SRS provides for a more efficient use of both personnel and equipment resources.
REFERENCES 1. Chen JCT, Rahimian J, Girvigian MR, Miller MJ. Contemporary methods of radiosurgery treatment with the Novalis Linear Accelerator System. Neurosurg Focus 2007;23. E4: 1–9. 2. Benedict SH, Cardinale RM, Wu Q, et al. Intensity-modulated stereotactic radiosurgery using dynamic micro-multileaf collimation. Int J Radiat Oncol Biol Phys 2001;50:751–758. 3. Baisden JM, Benedict SH, Sheng K, et al. Helical TomoTherapy in the treatment of central nervous system metastasis. Neurosurg Focus 2007;22:E8. 4. Han C, Liu A, Schultheiss TE, et al. Dosimetric comparisons of Helical TomoTherapy treatment plans and step-and-shoot inten-
sity-modulated radiosurgery treatment plans in intracranial stereotactic radiosurgery. Int J Radiat Oncol Biol Phys 2006;65: 608–616. 5. Colombo F, Casentini L, Cavedon C, et al. CyberKnife radiosurgery for benign meningiomas: Short-term results in 199 patients. Neurosurgery 2009;64(Suppl):A7–A13. 6. Collins SP, Coppa ND, Zhang Y, et al. CyberKnife radiosurgery in the treatment of complex skull base tumors: Analysis of treatment planning parameters. Radiat Oncol 2006;1:46. 7. Yu C, Jozef G, Apuzzo MLJ, Petrovich Z. Dosimetric comparison of CyberKnife with other radiosurgical modalities for an ellipsoidal target. Neurosurgery 2003;53:1155–1162.
1466
I. J. Radiation Oncology d Biology d Physics
8. Verbakel WF, Cuijpers JP, Hoffmans D, et al. Volumetric intensity-modulated arc therapy vs. conventional IMRT in head-andneck cancer: A comparative planning and dosimetric study. Int J Radiat Oncol Biol Phys 2009;74:252–259. 9. Fogliata A, Clivio A, Nicolini G, et al. Intensity modulation with photons for benign intracranial tumours: A planning comparison of volumetric single arc, helical arc and fixed gantry techniques. Radiother Oncol 2008;89:254–262. 10. Clivio A, Fogliata A, Franzetti-Pellanda A, et al. Volumetricmodulated arc radiotherapy for carcinomas of the anal canal: A treatment planning comparison with fixed field IMRT. Radiother Oncol 2009;92:118–124. 11. Cozzi L, Clivio A, Gauman G, et al. Comparison of advanced irradiation techniques with photons for benign intracranial tumors. Med Phys 2006;80:268–273. 12. Vanetti E, Clivio A, Nicolini G, et al. Volumetric modulated arc radiotherapy for carcinomas of the oro-pharynx, hypo-pharynx and larynx: A treatment planning comparison with fixed field IMRT. Radiother Oncol 2009;92:111–117. 13. Palma D, Vollans E, James K, et al. Volumetric modulated arc therapy for delivery of prostate radiotherapy: Comparison with intensity-modulated radiotherapy and three-dimensional conformal radiotherapy. Int J Radiat Oncol Biol Phys 2008;72: 996–1001. 14. Cozzi L, Dinshaw KA, Shrivastava SK, et al. A treatment planning study comparing volumetric arc modulation with RapidArc and fixed field IMRT for cervix uteri radiotherapy. Radiother Oncol 2008;89:180–191. 15. Kjær-Kristoffersen F, Ohlhues L, Medin J, Korreman S. RapidArc volumetric modulated therapy planning for prostate cancer patients. Acta Oncol 2008;48:227–232. 16. Fogliata A, Yartsev S, Nicolini G, et al. On the performance of intensity modulated Protons, RapidArc, and Helical TomoTherapy for selected paediatric cases. Radiat Oncol 2009;4:2. 17. Ling CC, Zhang P, Archambault Y, et al. Commissioning and quality assurance of RapidArc radiotherapy delivery. Int J Radiat Oncol Biol Phys 2008;72:575–581. 18. Nicolini G, Vanetti E, Clivio A, et al. The GLAaS algorithm for portal dosimetry and quality assurance of RapidArc an intensity modulated rotational therapy. Radiother Oncol 2008;3:24. 19. Korreman S, Medin J, Kjær-Kristoffersen F. Dosimetric verification of RapidArc treatment delivery. Acta Oncol 2008;1–7. 20. Vanetti E, Nicolini G, Clivio A, et al. The impact of treatment couch modelling on RapidArc. Phys Med Biol 2009;54: N157–N166.
Volume 78, Number 5, 2010
21. Gagne IM, Ansbacher W, Zavgorodni S, et al. A Monte Carlo evaluation of RapidArc dose calculations for oropharynx radiotherapy. Phys Med Biol 2008;53:7167–7185. 22. Bush K, Townson R, Zavgorodni S. Monte Carlo simulation of RapidArc radiotherapy delivery. Phys Med Biol 2008;53: N359–370. 23. Mayo C, Urie M. Eclipse IMRT—A practical treatment planning guide. Franklin, MA: KnowNexus; 2004. 24. Bolsi A, Fogliata A, Cozzi L. Radiotherapy of small intracranial tumours with different advanced techniques using photon and proton beams: A treatment planning study. Radiother Oncol 2003;68:1–14. 25. Paddick I. A simple scoring ratio to index the conformity of radiosurgical treatment plans. Technical note. J Neurosurg 2000; 93(Suppl 3):219–222. 26. Wagner TH, Bova FJ, Friedman WA, et al. A simple and reliable index for scoring rival stereotactic radiosurgery plans. Int J Radiat Oncol Biol Phys 2003;57:1141–1149. 27. Paganetti H. Changes in tumor cell response due to prolonged dose delivery times in fractionated radiation therapy. Int J Radiat Oncol Biol Phys 2005;63:892–900. 28. Rades D, Kueter JD, Veninga T, et al. Whole brain radiotherapy plus stereotactic radiosurgery (WBRT + SRS) versus surgery plus whole brain radiotherapy (OP + WBRT) for 1–3 brain metastases: Results of a matched pair analysis. Eur J Cancer 2009; 45:400–404. 29. Sneed PK, Suh JH, Goetsch SJ, et al. A multi-institutional review of radiosurgery alone vs. radiosurgery with whole brain radiotherapy as the initial management of brain metastases. Int J Radiat Oncol Biol Phys 2002;53:519–526. 30. Pirzkall A, Bebus J, Lohr F, et al. Radiosurgery alone or in combination with whole-brain radiotherapy for brain metastases. J Clin Oncol 1998;16:3563–3569. 31. Giubilei C, Ingrosso G, D’Andrea M, et al. Hypofractionated stereotactic radiotherapy in combination with whole brain radiotherapy for brain metastases. J Neurooncol 2009;91:207– 212. 32. Narayana A, Chang J, Yenice K, et al. Hypofractionated stereotactic radiotherapy using intensity-modulated radiotherapy in patients with one or two brain metastases. Stereotact Funct Neurosurg 2007;85:82–87. 33. Molenaar R, Wiggenraad R, Verbeek-de Kanter A, et al. Relationship between volume, dose and local control in stereotactic radiosurgery of brain metastasis. Br J Neurosurg 2009;23:170– 178.