- Email: [email protected]
Comparative clinical dosimetry with X-knife and gamma knife
Physica Medica (2012) 28, 269e272
available at www.sciencedirect.com
journal homepage: http://intl.elsevierhealth.com/journals/ejmp
TECHNICAL NOTE
Comparative clinical dosimetry with X-knife and gamma knife M.K. Semwal a,*, Sukhvir Singh a, A. Sarin a, S. Bhatnagar a, H.C. Pathak b a b
Department of Radiotherapy, Army Hospital Research & Referral, New Delhi 110010, India Department of Neurosurgery, Army Hospital Research & Referral, New Delhi 110010, India
Received 24 December 2010; received in revised form 6 July 2011; accepted 10 July 2011 Available online 30 July 2011
KEYWORDS X-knife; Gamma knife; Dosimetric comparison
Abstract X-knife and gamma knife techniques are well-established for cranial stereotactic radiosurgery (SRS). Due to differences in their radiation delivery methods, some of the dosimetric parameters of these two techniques differ which may have clinical significance. There are many dosimetric studies comparing linear accelerator based techniques such as X-knife with gamma knife but generally from different institutions. We carried out a retrospective comparative study of the dosimetric parameters of the SRS treatments performed at our centre with X-knife (circular cones) and gamma knife. Our results indicate that the dose conformity and dose fall-off in the vicinity of the target volumes were better for patients treated with gamma knife as compared to X-knife. However, the dose fall-off pattern shows a reversal at a larger distance from the target. It was better for the X-knife as compared to gamma knife in the low dose region. ª 2011 Associazione Italiana di Fisica Medica. Published by Elsevier Ltd. All rights reserved.
Introduction Stereotactic radiosurgery (SRS) is a term used to describe a sterotactically guided high precision conformal irradiation of a target volume in a single session as against the fractional dose delivery in conventional radiotherapy. If the same 3-D localization technique is used for delivering multiple fractionations then it is called as stereotactic
* Corresponding author. Tel.: þ91 11 25691181. E-mail address: [email protected] (M.K. Semwal).
radiotherapy (SRT). The main attributes of a SRS/SRT technique are high geometric accuracy, high conformality, and sharp dose fall-off beyond the target volume. The tumours that are suitable for SRS should be small and welldefined with clearly demarcated edges on imaging modalities such as CT, MRI and PET. These conditions are often consistent with the treatment of cerebral lesions. Many technologies are currently in use for SRS/SRT with their advantages and disadvantages. The X-knife technique using mega voltage (MV) x-rays from a linear accelerator (linac), and gamma knife based on gamma-rays from cobalt-60 sources are two such well-established techniques for SRS/SRT. The mechanical accuracy associated with
1120-1797/$ - see front matter ª 2011 Associazione Italiana di Fisica Medica. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.ejmp.2011.07.003
270 treatment delivery is known to be better for the gamma knife as compared to the linac based systems such as the Xknife. The linac isocentric and frame application accuracy are the main contributors for the difference in accuracy of the two systems [1e4]. Our hospital has both a gamma knife and an X-knife facility operated and used by a common team of medical physicists, neurosurgeons, and radiation oncologists. The Xknife facility was the first to be commissioned which now mostly is used for SRT after the commissioning of the gamma knife facility. We carried out a retrospective comparative study on dosimetric parameters of the treatments delivered with X-knife and gamma knife techniques at our centre. Having a common team of experts at one centre for both the techniques will bring out similarities and differences that are intrinsic to the two techniques. The other extrinsic factors such as human bias, and variations in treatment planning protocols among different centres are eliminated in this case. The X-knife technology with circular cones and the gamma knife with circular external collimator helmets, though not the latest SRS technologies in their respective categories, have been used and continue to be used for treating a considerable number of patients. It is believed that a dosimetric comparison presented in this paper has relevance to the comparative clinical outcomes from these two modalities of SRS.
Material and methods
M.K. Semwal et al. head ring, in head neutral position was obtained as the BRW frame was not compatible with MRI. A 1 Tesla MRI machine model Magnetom Harmony (Siemens, Germany) was used for the purpose. In most of the cases, T1 weighted (plain and contrast enhanced), and T2 weighted spin echo imaging sequences (2D/3D) with 1 mm thick contiguous slices (pixel size 1 mm 1 mm) were obtained. The images were transferred to the X-knife treatment planning system (version XKnife RT) through a DICOM network. The CT and MRI axial scans were fused with a manual land mark based algorithm in the treatment planning system and then contouring of the tumour and the critical structures was carried out. Cone size was decided based on the tumour size and shape. Non-coplanar beam arcs were placed ensuring that there was minimum overlap between the arcs out-side the tumour volume. Dose volume histograms (DVH), isodose distribution on images (axial, coronal and sagittal), and surface dose summary were used to arrive at the optimal plan for delivery. Dose prescription was usually at 85e90% isodose surface. Isocenter coordinates along with the beam plan were then printed out and implemented on the linac. All quality assurance checks on the linac and accessories were performed before treatment delivery. All the treatment plans were generated with a single isocenter to avoid complexities in planning and delivery of treatment as this was the beginning of the SRS programme at our centre. The average number of arcs used was 5 (range 4e7) and average arc angle 469.6 (range 340e628 ).
Patients Gamma knife As per our policy, patients with benign tumours/metastases with volumes less than about 13 cc were considered for SRS after being found inoperable and/or if patients’ were unwilling for other modalities of treatment. In a rare condition, a primary cranial malignancy was also chosen for SRS. All the 14 cases of SRS treated with X-knife were included in the study. They comprised of 7 acoustic neuromas, 03 single metastases, 01 meningioma, 01 pituitary adenoma, and 02 other benign tumours. For comparison, the first 21 cases treated on gamma knife were selected. The diagnosis-wise break up was as follows: 07 acoustic neuromas, 04 pituitary adenomas, 04 meningiomas, 02 arteriovenus malformations, and 03 other benign tumours. The average tumour size was 4.64 cc (range 0.62e11.8 cc), and 2.94 cc (range 0.41e5.8 cc) for X-knife and gamma knife respectively.
X-knife X-knife system (Radionics, USA) with circular collimators (cones) of aperture size from 12.5 to 40 mm (at isocenter) in steps of 2.5 mm was used for the delivery of SRS. The Xknife accessories were mated with a linac model Primus (Siemens, Germany) and the irradiation was carried out with a 6 MV photon beam with a fixed jaw opening of 6 cm 6 cm. After fixing the BRW (BrowneRobertseWells) head ring on a patient, CT images were obtained with the CT localizer frame on a CT machine model HiSpeed (GE Healthcare, USA). Prior to head ring fixation, MRI of the patient, without
The Leksell Gamma Knife (LGK) model 4C (Elekta, Sweden) used in this study has a distinctive addition over the earlier LGK models, called the automatic positioning system (APS). It has 4 collimator helmet sizes namely 4, 8, 14 and 18 mm. In the machine, 201 cobalt-60 sources (each with an initial nominal activity of about 30 Ci) are distributed on a hemispherical surface in such a way that when used with an external helmet (collimator), all the 201 radiation beams converge at one point called the focus (isocenter). The source to focal point distance is 40 cm. In LGK planning, usually multiple isocenters (shots) and helmet sizes are used to achieve the desired conformality of dose distribution. The APS moves the patient head from one shot to another within one helmet size (run) without the need for the operator to enter the treatment room for manual change of shot co-ordinates, as was the practice with the earlier trunnion based systems. The main advantages of the APS include speeding up of the treatment delivery, less chances of human error in setting up the co-ordinates manually, and a possible reduction in man-power requirement. However, another less appreciated advantage is the fact that a treatment planner can better afford to use a larger number of shots to improve on conformality of a treatment plan, if needed. The Leksell gamma plan (LGP) 4C release 5.34 (Elekta, Sweden) was used for treatment planning. After fixation of the stereotactic frame (Leksell coordinate frame G) on the patient, which is MRI compatible, all patients underwent MRI scans either on a 1.0 Tesla Magnetom Harmony
Comparative clinical dosimetry with X-knife and gamma knife (Siemens, Germany) or a 1.5 Tesla Signa HDx (GE Healthcare, USA) MRI machines with the MRI localizer fixed on the frame. In the case of arteriovenus malformation (AVM), an additional imaging namely planar digital subtraction angiography (DSA) was carried out on an Axiom Artis BA machine (Siemens, Germany) for better delineation of the AVM nidus. The images were always checked for distortions at the imaging consoles using fiducial markers before transferring them to the LGP console. Skull measurements with a perspex spherical box called bubble (Elekta, Sweden) were carried out to reconstruct the patient head geometry. The LGP algorithm assumes homogeneous head density for dose calculations. After tumour delineation on MRI and/or DSA, an adequate number of isocenters was placed to achieve an acceptable treatment plan. The average number of isocenters used was 5.7 (range 1e28). The dose prescription was usually at 50% (of the maximum dose within the target) isodose surface that covered more than 95% of the target volume. Dose volume histograms (DVHs) and isodose distributions on imaging slices were used for plan analysis, prescription and reporting.
Dosimetric parameters Tumour coverage, Paddick conformity index (PCI), and gradient indices were estimated from DVH parameters. The PCI was calculated as, PCI Z (VTP/VP) (VTP/VT), where VTP is target volume covered by prescription isodose surface, VP is tissue volume covered by prescription isodose surface, and VT is target volume [5]. As is evident, better conformity and coverage will yield a PCI value closer to unity The other commonly used conformity index (CI) was also calculated as CI Z (VP/VTP), to be able to compare it with some published literature. The dose fall-off was assessed from the gradient index (g value) as per the equation: V=V0 ZðD=D0 Þg ; or jgjZðlog V=V0 Þ=log ðD=D0 Þ;
where V is volume covered by isodose surface of dose D, V0 is the volume covered by the isodose of the reference dose D0 [6]. In our case, D0 is the prescription isodose normalized to 100%. The decrease in absolute value of g indicates sharper dose fall-off in the dose range considered. As stated by Ma et al. [6], the advantage of g index is that its value is not constrained by a fixed dose range unlike Paddick Gradient Index.
Table 1
LGK 4C X-knife p Value
271
Results and discussion Table 1 presents the mean values of the evaluated dosimetric parameters for the study. The p-values shown in the table were calculated using Student’s t-test for independent groups. It can be seen that the conformity achieved with LGK is significantly better than X-knife (p < 0.001). This is mainly due to the use of multiple isocenters in the LGK as compared to a single isocenter in the case of Xknife. An additional factor in our case could be the nonusage of linac collimator jaws with the circular collimator for shaping of the fields for achieving better conformity. The CI (mean) calculated for the LGK was 1.38 which is in agreement with published literature [7e9]. Fig. 1 shows the scatter plot of PCI and gamma values for X-knife and LGK respectively. It is evident from the figure that in the dose range 100e80% there is wide variation in the gamma values for both the X-knife and LGK. In the case of X-knife, size of the circular cone used, and less than optimum separation between non-coplanar arcs can impact the dose fall-off substantially in the close vicinity of the target. Also, nonusage of jaws for field shaping with the circular collimator would add to this variation depending upon target shape. For LGK, it could be the number of shots used and their overlap that can cause large change in dose fall-off pattern in the close vicinity of the target. The homogeneity index (HI) defined as the ratio of maximum to minimum dose within a target volume, varied from 1.8 to 2.2 for LGK as compared to the range of 1.1e1.2 for X-knife in our case. Typically, as reported in literature, plans generated with the LGK have homogeneity index (HI) in the range of 2.0e3.0 while the linac plans have HI in the range of 1.0e1.2 [10]. Multiple isocenters in the case of LGK as compared to a single isocentre in the case of X-knife is the main reason for this vast difference in HI of the two modalities. The difference in the g values of X-knife and LGK in the dose ranges 100e80% and 100e50% is large and statistically significant (p < 0.02). Smaller absolute g values for LGK means sharper dose fall-off as compared to X-knife in both the dose ranges. Interestingly, the absolute gamma value in the dose range 100e20% is higher for LGK as compared to Xknife, though the difference is small and statistically insignificant (p Z 0.582). The sharper dose fall-off in LGK close to the high dose volume could partly be due to the lower beam penumbral width for LGK as compared to a 6 MV linac based SRS system among other reasons [11,12]. This implies that LGK will spare normal tissues better in the higher dose range of 10e12 Gy as compared to X-knife. This agrees with the observation of some authors that overall
Conformity and gradient indices for LGK 4C and X-knife. PCI (Mean std dev)
g1 (Mean std dev)
g2 (Mean std dev)
g3 (Mean std dev)
0.664 0.048 0.501 0.240 <0.001
1.703 0.247 2.251 0.450 <0.001
1.559 0.099 1.800 0.327 0.018
1.536 0.058 1.515 0.131 0.580
Paddick Conformity Index (PCI), and absolute g values for Leksell Gamma Knife (LGK) 4C and X-knife. The three dose ranges in terms of percentage of prescription dose for g value calculations are: 100-80% (g1); 100-50%( g2); 100e20% (g3).
272
M.K. Semwal et al. indication of smaller low dose volume in X-knife as compared to LGK can have favourable implications on the chances of secondary malignancies for the former. The present study has the limitation of a smaller data set, and hence it is felt that a similar study on a larger patient data base for each tumour site/diagnosis would be desirable for validation of our results and better correlation with clinical outcomes.
Conflict of interest None.
References
Figure 1 Estimated Paddick conformity index (PCI), and (g) index values in the dose ranges 100e80% (g1), 100e50% (g2), and 100e20% (g3) for Leksell Gamma Knife (LGK) 4C and X-knife.
dose fall-off may be improved if the prescription isodose line is lowered as is the case with LGK as compared with Xknife [6]. However, in the low dose region away from the target (dose range 100e20%), the trend seems to be reversing. In this region, X-knife seems to have the potential of sparing the normal tissues (brain) better than LGK. Ma L et al. [6] reported a g value of 1.57 for dose range 100e50% for LGK 4C as compared to our value of 1.557. The other gamma values estimated by us are also close to their results.
Conclusions Dosimetric parameters namely CI, HI, and g values for patients treated at one centre on X-knife and LGK systems were compared. The differences observed in our case generally followed the published reports. As for the clinical significance of the differences observed, it can be said that the vastly superior dose homogeneity in X-knife as compared to LGK may provide potential advantage to the former in the treatment of tumours where an organ at risk (OAR) traverses the target volume. In such a case injury to the OAR could be more for LGK as compared to X-knife as there is a much higher dose region within the target in LGK, and the fact that location of the higher dose region within the target is unpredictable in it. However, in cases where higher dose at tumour core is desirable, LGK offers advantage. The sharper dose fall-off in the vicinity of the target and the better dose conformity in LGK can help reduce peripheral high dose volume and thus decrease the probability of radiation necrosis incidences. The tentative
[1] Winston KR, Lutz W. Linear accelerator as a neurosurgical tool for stereotactic radiosurgery. Neurosurgery 1988;22:454e64. [2] Maciunas RJ, Galloway RL, Latimer JW. The application accuracy of stereotactic frames. Neurosurgery 1994;35: 682e94. [3] Wu A. Physics and dosimetry of the gamma knife. Neurosurg Clin N Am 1992;3:35e50. [4] Schell MC, Bova FJ, Larson DA, Leavitt DD, Luz WR, Podgorsak EB. Stereotactic radiosurgery. American Association of Physicists in Medicine report 54. Report of task group 42. New York: Radiation Therapy Committee. American Institute of Physics; 1995. [5] Paddick I. A simple scoring ratio to index the conformity of radiosurgical treatment plans. Technical note. J Neurosurg 2000;93(Suppl. 3):219e22. [6] Ma L, Verhey L, Chuang C, Descovich M, Smith V, Huang K, et al. Effect of composite sector collimation on average dose fall-off for Gamma Knife Perfexion. J Neurosurg 2008;109: 15e20. [7] Nakamura JL, Verhey LJ, Smith V, Petti PL, Lamborn KR, Larson DA, et al. Dose conformity of gamma knife radiosurgery and risk factors for complications. Int J Radiat Oncol Biol Phys 2001;51:1313e9. [8] Perks JR, George EJS, Hamri KE, Blackburn P, Plowman PN. Stereotactic radiosurgery XVI: Isodosimetric comparison of photon stereotactic radiosurgery techniques (gamma knife versus micromultileaf collimator linear accelerator) for acoustic neuroma e and potential clinical significance. Int J Radiat Oncol Biol Phys 2003;57:1450e9. [9] Plowman PN, Doughty D. Stereotactic radiosurgery, X: clinical Isodosimetry of gamma knife versus linear accelerator X-Knife for pituitary and acoustic tumours. Clin Oncol 1999;11:321e9. [10] Shaw E, Scott C, Souhami L, Dinapoli R, Kline R, Loeffler J, et al. Single dose radiosurgical treatment of recurrent previously irradiated primary brain tumors and brain metastases: final report of RTOG protocol 90-05. Int J Radiat Oncol Biol Phys 2000;47:291e8. [11] Gerbi BJ, Higgins PD, Cho KH, Hall WA. LINAC-based stereotactic radiosurgery for treatment of trigeminal neuralgia. J Appl Clin Med Phys 2004;5(3):80e90, http://www.jacmp.org. Retrieved on 15 Sep 2010. [12] Lindquist C, Paddick I. The Leksell Gamma Knife Perfection and comparisons with its predecessors. Neurosurgery 2007;61: 130e1.
available at www.sciencedirect.com
journal homepage: http://intl.elsevierhealth.com/journals/ejmp
TECHNICAL NOTE
Comparative clinical dosimetry with X-knife and gamma knife M.K. Semwal a,*, Sukhvir Singh a, A. Sarin a, S. Bhatnagar a, H.C. Pathak b a b
Department of Radiotherapy, Army Hospital Research & Referral, New Delhi 110010, India Department of Neurosurgery, Army Hospital Research & Referral, New Delhi 110010, India
Received 24 December 2010; received in revised form 6 July 2011; accepted 10 July 2011 Available online 30 July 2011
KEYWORDS X-knife; Gamma knife; Dosimetric comparison
Abstract X-knife and gamma knife techniques are well-established for cranial stereotactic radiosurgery (SRS). Due to differences in their radiation delivery methods, some of the dosimetric parameters of these two techniques differ which may have clinical significance. There are many dosimetric studies comparing linear accelerator based techniques such as X-knife with gamma knife but generally from different institutions. We carried out a retrospective comparative study of the dosimetric parameters of the SRS treatments performed at our centre with X-knife (circular cones) and gamma knife. Our results indicate that the dose conformity and dose fall-off in the vicinity of the target volumes were better for patients treated with gamma knife as compared to X-knife. However, the dose fall-off pattern shows a reversal at a larger distance from the target. It was better for the X-knife as compared to gamma knife in the low dose region. ª 2011 Associazione Italiana di Fisica Medica. Published by Elsevier Ltd. All rights reserved.
Introduction Stereotactic radiosurgery (SRS) is a term used to describe a sterotactically guided high precision conformal irradiation of a target volume in a single session as against the fractional dose delivery in conventional radiotherapy. If the same 3-D localization technique is used for delivering multiple fractionations then it is called as stereotactic
* Corresponding author. Tel.: þ91 11 25691181. E-mail address: [email protected] (M.K. Semwal).
radiotherapy (SRT). The main attributes of a SRS/SRT technique are high geometric accuracy, high conformality, and sharp dose fall-off beyond the target volume. The tumours that are suitable for SRS should be small and welldefined with clearly demarcated edges on imaging modalities such as CT, MRI and PET. These conditions are often consistent with the treatment of cerebral lesions. Many technologies are currently in use for SRS/SRT with their advantages and disadvantages. The X-knife technique using mega voltage (MV) x-rays from a linear accelerator (linac), and gamma knife based on gamma-rays from cobalt-60 sources are two such well-established techniques for SRS/SRT. The mechanical accuracy associated with
1120-1797/$ - see front matter ª 2011 Associazione Italiana di Fisica Medica. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.ejmp.2011.07.003
270 treatment delivery is known to be better for the gamma knife as compared to the linac based systems such as the Xknife. The linac isocentric and frame application accuracy are the main contributors for the difference in accuracy of the two systems [1e4]. Our hospital has both a gamma knife and an X-knife facility operated and used by a common team of medical physicists, neurosurgeons, and radiation oncologists. The Xknife facility was the first to be commissioned which now mostly is used for SRT after the commissioning of the gamma knife facility. We carried out a retrospective comparative study on dosimetric parameters of the treatments delivered with X-knife and gamma knife techniques at our centre. Having a common team of experts at one centre for both the techniques will bring out similarities and differences that are intrinsic to the two techniques. The other extrinsic factors such as human bias, and variations in treatment planning protocols among different centres are eliminated in this case. The X-knife technology with circular cones and the gamma knife with circular external collimator helmets, though not the latest SRS technologies in their respective categories, have been used and continue to be used for treating a considerable number of patients. It is believed that a dosimetric comparison presented in this paper has relevance to the comparative clinical outcomes from these two modalities of SRS.
Material and methods
M.K. Semwal et al. head ring, in head neutral position was obtained as the BRW frame was not compatible with MRI. A 1 Tesla MRI machine model Magnetom Harmony (Siemens, Germany) was used for the purpose. In most of the cases, T1 weighted (plain and contrast enhanced), and T2 weighted spin echo imaging sequences (2D/3D) with 1 mm thick contiguous slices (pixel size 1 mm 1 mm) were obtained. The images were transferred to the X-knife treatment planning system (version XKnife RT) through a DICOM network. The CT and MRI axial scans were fused with a manual land mark based algorithm in the treatment planning system and then contouring of the tumour and the critical structures was carried out. Cone size was decided based on the tumour size and shape. Non-coplanar beam arcs were placed ensuring that there was minimum overlap between the arcs out-side the tumour volume. Dose volume histograms (DVH), isodose distribution on images (axial, coronal and sagittal), and surface dose summary were used to arrive at the optimal plan for delivery. Dose prescription was usually at 85e90% isodose surface. Isocenter coordinates along with the beam plan were then printed out and implemented on the linac. All quality assurance checks on the linac and accessories were performed before treatment delivery. All the treatment plans were generated with a single isocenter to avoid complexities in planning and delivery of treatment as this was the beginning of the SRS programme at our centre. The average number of arcs used was 5 (range 4e7) and average arc angle 469.6 (range 340e628 ).
Patients Gamma knife As per our policy, patients with benign tumours/metastases with volumes less than about 13 cc were considered for SRS after being found inoperable and/or if patients’ were unwilling for other modalities of treatment. In a rare condition, a primary cranial malignancy was also chosen for SRS. All the 14 cases of SRS treated with X-knife were included in the study. They comprised of 7 acoustic neuromas, 03 single metastases, 01 meningioma, 01 pituitary adenoma, and 02 other benign tumours. For comparison, the first 21 cases treated on gamma knife were selected. The diagnosis-wise break up was as follows: 07 acoustic neuromas, 04 pituitary adenomas, 04 meningiomas, 02 arteriovenus malformations, and 03 other benign tumours. The average tumour size was 4.64 cc (range 0.62e11.8 cc), and 2.94 cc (range 0.41e5.8 cc) for X-knife and gamma knife respectively.
X-knife X-knife system (Radionics, USA) with circular collimators (cones) of aperture size from 12.5 to 40 mm (at isocenter) in steps of 2.5 mm was used for the delivery of SRS. The Xknife accessories were mated with a linac model Primus (Siemens, Germany) and the irradiation was carried out with a 6 MV photon beam with a fixed jaw opening of 6 cm 6 cm. After fixing the BRW (BrowneRobertseWells) head ring on a patient, CT images were obtained with the CT localizer frame on a CT machine model HiSpeed (GE Healthcare, USA). Prior to head ring fixation, MRI of the patient, without
The Leksell Gamma Knife (LGK) model 4C (Elekta, Sweden) used in this study has a distinctive addition over the earlier LGK models, called the automatic positioning system (APS). It has 4 collimator helmet sizes namely 4, 8, 14 and 18 mm. In the machine, 201 cobalt-60 sources (each with an initial nominal activity of about 30 Ci) are distributed on a hemispherical surface in such a way that when used with an external helmet (collimator), all the 201 radiation beams converge at one point called the focus (isocenter). The source to focal point distance is 40 cm. In LGK planning, usually multiple isocenters (shots) and helmet sizes are used to achieve the desired conformality of dose distribution. The APS moves the patient head from one shot to another within one helmet size (run) without the need for the operator to enter the treatment room for manual change of shot co-ordinates, as was the practice with the earlier trunnion based systems. The main advantages of the APS include speeding up of the treatment delivery, less chances of human error in setting up the co-ordinates manually, and a possible reduction in man-power requirement. However, another less appreciated advantage is the fact that a treatment planner can better afford to use a larger number of shots to improve on conformality of a treatment plan, if needed. The Leksell gamma plan (LGP) 4C release 5.34 (Elekta, Sweden) was used for treatment planning. After fixation of the stereotactic frame (Leksell coordinate frame G) on the patient, which is MRI compatible, all patients underwent MRI scans either on a 1.0 Tesla Magnetom Harmony
Comparative clinical dosimetry with X-knife and gamma knife (Siemens, Germany) or a 1.5 Tesla Signa HDx (GE Healthcare, USA) MRI machines with the MRI localizer fixed on the frame. In the case of arteriovenus malformation (AVM), an additional imaging namely planar digital subtraction angiography (DSA) was carried out on an Axiom Artis BA machine (Siemens, Germany) for better delineation of the AVM nidus. The images were always checked for distortions at the imaging consoles using fiducial markers before transferring them to the LGP console. Skull measurements with a perspex spherical box called bubble (Elekta, Sweden) were carried out to reconstruct the patient head geometry. The LGP algorithm assumes homogeneous head density for dose calculations. After tumour delineation on MRI and/or DSA, an adequate number of isocenters was placed to achieve an acceptable treatment plan. The average number of isocenters used was 5.7 (range 1e28). The dose prescription was usually at 50% (of the maximum dose within the target) isodose surface that covered more than 95% of the target volume. Dose volume histograms (DVHs) and isodose distributions on imaging slices were used for plan analysis, prescription and reporting.
Dosimetric parameters Tumour coverage, Paddick conformity index (PCI), and gradient indices were estimated from DVH parameters. The PCI was calculated as, PCI Z (VTP/VP) (VTP/VT), where VTP is target volume covered by prescription isodose surface, VP is tissue volume covered by prescription isodose surface, and VT is target volume [5]. As is evident, better conformity and coverage will yield a PCI value closer to unity The other commonly used conformity index (CI) was also calculated as CI Z (VP/VTP), to be able to compare it with some published literature. The dose fall-off was assessed from the gradient index (g value) as per the equation: V=V0 ZðD=D0 Þg ; or jgjZðlog V=V0 Þ=log ðD=D0 Þ;
where V is volume covered by isodose surface of dose D, V0 is the volume covered by the isodose of the reference dose D0 [6]. In our case, D0 is the prescription isodose normalized to 100%. The decrease in absolute value of g indicates sharper dose fall-off in the dose range considered. As stated by Ma et al. [6], the advantage of g index is that its value is not constrained by a fixed dose range unlike Paddick Gradient Index.
Table 1
LGK 4C X-knife p Value
271
Results and discussion Table 1 presents the mean values of the evaluated dosimetric parameters for the study. The p-values shown in the table were calculated using Student’s t-test for independent groups. It can be seen that the conformity achieved with LGK is significantly better than X-knife (p < 0.001). This is mainly due to the use of multiple isocenters in the LGK as compared to a single isocenter in the case of Xknife. An additional factor in our case could be the nonusage of linac collimator jaws with the circular collimator for shaping of the fields for achieving better conformity. The CI (mean) calculated for the LGK was 1.38 which is in agreement with published literature [7e9]. Fig. 1 shows the scatter plot of PCI and gamma values for X-knife and LGK respectively. It is evident from the figure that in the dose range 100e80% there is wide variation in the gamma values for both the X-knife and LGK. In the case of X-knife, size of the circular cone used, and less than optimum separation between non-coplanar arcs can impact the dose fall-off substantially in the close vicinity of the target. Also, nonusage of jaws for field shaping with the circular collimator would add to this variation depending upon target shape. For LGK, it could be the number of shots used and their overlap that can cause large change in dose fall-off pattern in the close vicinity of the target. The homogeneity index (HI) defined as the ratio of maximum to minimum dose within a target volume, varied from 1.8 to 2.2 for LGK as compared to the range of 1.1e1.2 for X-knife in our case. Typically, as reported in literature, plans generated with the LGK have homogeneity index (HI) in the range of 2.0e3.0 while the linac plans have HI in the range of 1.0e1.2 [10]. Multiple isocenters in the case of LGK as compared to a single isocentre in the case of X-knife is the main reason for this vast difference in HI of the two modalities. The difference in the g values of X-knife and LGK in the dose ranges 100e80% and 100e50% is large and statistically significant (p < 0.02). Smaller absolute g values for LGK means sharper dose fall-off as compared to X-knife in both the dose ranges. Interestingly, the absolute gamma value in the dose range 100e20% is higher for LGK as compared to Xknife, though the difference is small and statistically insignificant (p Z 0.582). The sharper dose fall-off in LGK close to the high dose volume could partly be due to the lower beam penumbral width for LGK as compared to a 6 MV linac based SRS system among other reasons [11,12]. This implies that LGK will spare normal tissues better in the higher dose range of 10e12 Gy as compared to X-knife. This agrees with the observation of some authors that overall
Conformity and gradient indices for LGK 4C and X-knife. PCI (Mean std dev)
g1 (Mean std dev)
g2 (Mean std dev)
g3 (Mean std dev)
0.664 0.048 0.501 0.240 <0.001
1.703 0.247 2.251 0.450 <0.001
1.559 0.099 1.800 0.327 0.018
1.536 0.058 1.515 0.131 0.580
Paddick Conformity Index (PCI), and absolute g values for Leksell Gamma Knife (LGK) 4C and X-knife. The three dose ranges in terms of percentage of prescription dose for g value calculations are: 100-80% (g1); 100-50%( g2); 100e20% (g3).
272
M.K. Semwal et al. indication of smaller low dose volume in X-knife as compared to LGK can have favourable implications on the chances of secondary malignancies for the former. The present study has the limitation of a smaller data set, and hence it is felt that a similar study on a larger patient data base for each tumour site/diagnosis would be desirable for validation of our results and better correlation with clinical outcomes.
Conflict of interest None.
References
Figure 1 Estimated Paddick conformity index (PCI), and (g) index values in the dose ranges 100e80% (g1), 100e50% (g2), and 100e20% (g3) for Leksell Gamma Knife (LGK) 4C and X-knife.
dose fall-off may be improved if the prescription isodose line is lowered as is the case with LGK as compared with Xknife [6]. However, in the low dose region away from the target (dose range 100e20%), the trend seems to be reversing. In this region, X-knife seems to have the potential of sparing the normal tissues (brain) better than LGK. Ma L et al. [6] reported a g value of 1.57 for dose range 100e50% for LGK 4C as compared to our value of 1.557. The other gamma values estimated by us are also close to their results.
Conclusions Dosimetric parameters namely CI, HI, and g values for patients treated at one centre on X-knife and LGK systems were compared. The differences observed in our case generally followed the published reports. As for the clinical significance of the differences observed, it can be said that the vastly superior dose homogeneity in X-knife as compared to LGK may provide potential advantage to the former in the treatment of tumours where an organ at risk (OAR) traverses the target volume. In such a case injury to the OAR could be more for LGK as compared to X-knife as there is a much higher dose region within the target in LGK, and the fact that location of the higher dose region within the target is unpredictable in it. However, in cases where higher dose at tumour core is desirable, LGK offers advantage. The sharper dose fall-off in the vicinity of the target and the better dose conformity in LGK can help reduce peripheral high dose volume and thus decrease the probability of radiation necrosis incidences. The tentative
[1] Winston KR, Lutz W. Linear accelerator as a neurosurgical tool for stereotactic radiosurgery. Neurosurgery 1988;22:454e64. [2] Maciunas RJ, Galloway RL, Latimer JW. The application accuracy of stereotactic frames. Neurosurgery 1994;35: 682e94. [3] Wu A. Physics and dosimetry of the gamma knife. Neurosurg Clin N Am 1992;3:35e50. [4] Schell MC, Bova FJ, Larson DA, Leavitt DD, Luz WR, Podgorsak EB. Stereotactic radiosurgery. American Association of Physicists in Medicine report 54. Report of task group 42. New York: Radiation Therapy Committee. American Institute of Physics; 1995. [5] Paddick I. A simple scoring ratio to index the conformity of radiosurgical treatment plans. Technical note. J Neurosurg 2000;93(Suppl. 3):219e22. [6] Ma L, Verhey L, Chuang C, Descovich M, Smith V, Huang K, et al. Effect of composite sector collimation on average dose fall-off for Gamma Knife Perfexion. J Neurosurg 2008;109: 15e20. [7] Nakamura JL, Verhey LJ, Smith V, Petti PL, Lamborn KR, Larson DA, et al. Dose conformity of gamma knife radiosurgery and risk factors for complications. Int J Radiat Oncol Biol Phys 2001;51:1313e9. [8] Perks JR, George EJS, Hamri KE, Blackburn P, Plowman PN. Stereotactic radiosurgery XVI: Isodosimetric comparison of photon stereotactic radiosurgery techniques (gamma knife versus micromultileaf collimator linear accelerator) for acoustic neuroma e and potential clinical significance. Int J Radiat Oncol Biol Phys 2003;57:1450e9. [9] Plowman PN, Doughty D. Stereotactic radiosurgery, X: clinical Isodosimetry of gamma knife versus linear accelerator X-Knife for pituitary and acoustic tumours. Clin Oncol 1999;11:321e9. [10] Shaw E, Scott C, Souhami L, Dinapoli R, Kline R, Loeffler J, et al. Single dose radiosurgical treatment of recurrent previously irradiated primary brain tumors and brain metastases: final report of RTOG protocol 90-05. Int J Radiat Oncol Biol Phys 2000;47:291e8. [11] Gerbi BJ, Higgins PD, Cho KH, Hall WA. LINAC-based stereotactic radiosurgery for treatment of trigeminal neuralgia. J Appl Clin Med Phys 2004;5(3):80e90, http://www.jacmp.org. Retrieved on 15 Sep 2010. [12] Lindquist C, Paddick I. The Leksell Gamma Knife Perfection and comparisons with its predecessors. Neurosurgery 2007;61: 130e1.