Dosimetric performance of the new high‐definition multileaf collimator for intracranial stereotactic radiosurgery

The objective was to evaluate the performance of a high‐definition multileaf collimator (MLC) of 2.5 mm leaf width (MLC2.5) and compare to standard 5 mm leaf width MLC (MLC5) for the treatment of intracranial lesions using dynamic conformal arcs (DCA) technique with a dedicated radiosurgery linear accelerator. Simulated cases of spherical targets were created to study solely the effect of target volume size on the performance of the two MLC systems independent of target shape complexity. In addition, 43 patients previously treated for intracranial lesions in our institution were retrospectively planned using DCA technique with MLC2.5 and MLC5 systems. The gross tumor volume ranged from 0.07 to 40.57 cm3 with an average volume of 5.9 cm3. All treatment parameters were kept the same for both MLC‐based plans. The plan evaluation was performed using figures of merits (FOM) for a rapid and objective assessment on the quality of the two treatment plans for MLC2.5 and MLC5. The prescription isodose surface was selected as the greatest isodose surface covering ≥95% of the target volume and delivering 95% of the prescription dose to 99% of target volume. A Conformity Index (CI) and conformity distance index (CDI) were used to quantifying the dose conformity to a target volume. To assess normal tissue sparing, a normal tissue difference (NTD) was defined as the difference between the volume of normal tissue receiving a certain dose utilizing MLC5 and the volume receiving the same dose using MLC2.5. The CI and normal tissue sparing for the simulated spherical targets were better with the MLC2.5 as compared to MLC5. For the clinical patients, the CI and CDI results indicated that the MLC2.5 provides better treatment conformity than MLC5 even at large target volumes. The CI's range was 1.15 to 2.44 with a median of 1.59 for MLC2.5 compared to 1.60–2.85 with a median of 1.71 for MLC5. Improved normal tissue sparing was also observed for MLC2.5 over MLC5, with the NTD always positive, indicating improvement, and ranging from 0.1 to 8.3 for normal tissue receiving 50% (NTV50), 70% (NTV70) and 90% (NTV90) of the prescription dose. The MLC2.5 has a dosimetric advantage over the MLC5 in Linac‐based radiosurgery using DCA method for intracranial lesions, both in treatment conformity and normal tissue sparing when target shape complexity increases. PACS number: 87.56J‐, 87.56 jk

The objective was to evaluate the performance of a high-definition multileaf collimator (MLC) of 2.5 mm leaf width (MLC 2.5 ) and compare to standard 5 mm leaf width MLC (MLC 5 ) for the treatment of intracranial lesions using dynamic conformal arcs (DCA) technique with a dedicated radiosurgery linear accelerator. Simulated cases of spherical targets were created to study solely the effect of target volume size on the performance of the two MLC systems independent of target shape complexity. In addition, 43 patients previously treated for intracranial lesions in our institution were retrospectively planned using DCA technique with MLC 2.5 and MLC 5 systems. The gross tumor volume ranged from 0.07 to 40.57 cm 3 with an average volume of 5.9 cm 3 . All treatment parameters were kept the same for both MLC-based plans. The plan evaluation was performed using figures of merits (FOM) for a rapid and objective assessment on the quality of the two treatment plans for MLC 2.5 and MLC 5 . The prescription isodose surface was selected as the greatest isodose surface covering ≥ 95% of the target volume and delivering 95% of the prescription dose to 99% of target volume. A Conformity Index (CI) and conformity distance index (CDI) were used to quantifying the dose conformity to a target volume. To assess normal tissue sparing, a normal tissue difference (NTD) was defined as the difference between the volume of normal tissue receiving a certain dose utilizing MLC 5 and the volume receiving the same dose using MLC 2.5 . The CI and normal tissue sparing for the simulated spherical targets were better with the MLC 2.5 as compared to MLC 5 . For the clinical patients, the CI and CDI results indicated that the MLC 2.5 provides better treatment conformity than MLC 5 even at large target volumes. The CI's range was 1.15 to 2.44 with a median of 1.59 for MLC 2.5 compared to 1.60-2.85 with a median of 1.71 for MLC 5 . Improved normal tissue sparing was also observed for MLC 2.5 over MLC 5 , with the NTD always positive, indicating improvement, and ranging from 0.1 to 8.3 for normal tissue receiving 50% (NTV 50 ), 70% (NTV 70 ) and 90% (NTV 90 ) of the prescription dose. The MLC 2.5 has a dosimetric advantage over the MLC 5 in Linac-based radiosurgery using DCA method for intracranial lesions, both in treatment conformity and normal tissue sparing when target shape complexity increases.

I. INtRODUCtION
Stereotactic radiosurgery (SRS) is the process of delivering a high dose of external beam radiation to a small intracranial target in a single fraction using 60 Co sources, medical linear accelerators, or charged particle beams guided by an external frame system. Collimated radiation beams are precisely positioned and focused onto a target within the brain to deliver a high dose of localized radiation. An advantage of SRS is the prevention of damage to surrounding healthy tissue because of the steep dose gradient around the target volume. As suggested by normal tissue complication probability modeling for radiosurgery, (1) a high degree of conformity of the prescription dose to target volume should be achieved to allow safe treatment of the target. Conformity Index (CI) is the ratio of the prescription volume to the target volume, as defined in the Radiation Therapy Oncology Group (RTOG) radiosurgery guidelines in 1993. (2)(3)(4) CI is useful for evaluating competing plans for the same patient or comparing different modalities. Dose volume histograms (DVH) summarize the dose distribution information for a region of interest or anatomical structure and identify characteristics such as dose uniformity and hot or cold spots. DVHs may be a preliminary step in evaluating statistics such as tumor control and normal tissue complication probabilities. (5,6) However, when comparing a large number of plans, DVHs contain large amounts of data and make the comparison difficult and cumbersome.
Radiosurgery has evolved over the past decade with the development of new treatment delivery technologies such as dedicated radiosurgery linear accelerators. In the past, it has been reported that linear accelerator-based SRS is less conformal than gamma knife SRS. (4,7,8) However, linear accelerator-based SRS has become highly sophisticated, evolving from circular arc and multiple isocenters per target to dynamic conformal arcs (DCA) based on a single isocenter. The dynamic conformal arc is a method of linear accelerator-based SRS that uses multiple arcs rotating about a single isocenter. (9,10) This method uses the MLC to conform to the target volume every 10 degrees of arc. Recently, it has been reported that improvement in linear accelerator-based SRS techniques have allowed comparable conformity to that of a gamma knife. (11) The impacts of linear accelerator MLC leaf width on stereotactic radiosurgery and radiotherapy plans have been investigated and previously reported. (12)(13)(14)(15)(16)(17)(18) Kubo et al. (12) compared the conformity of 3D conformal plans using 1.7 mm, 3 mm and 10 mm leaf width MLCs and found that the smaller leafs produced more normal tissue sparing. IMRT plans for cranial cases were compared using 5 mm and 10 mm MLC leaf widths, and noticeably better sparing of optic structure was observed using 5 mm MLC. (13) Monk et al. (14) compared the Varian Millennium 120-MLC (minimal 5 mm leaf) with the BrainLAB (BrainLAB, Munich, Germany) micro-MLC (minimal 3 mm leaf width) with plans using fixed non-coplanar beams. They reported a small but statistically significant improvement in dose conformity and organ at risk (OAR) sparing with the 3 mm MLC compared with the 5 mm MLCs. Jin et al. (17) used dynamic conformal arcs and IMRS/IMRT techniques to compare the 3 mm micro-MLC and the 5 mm and 10 mm MLC, and found significant dosimetric differences in the conformity indices between the three MLCs -with the 3 mm leaf width scoring better. This study also reported that the difference in the conformity index decreases with the target volume and that, as the MLC margin increases in increments of 1 mm, the difference in the conformity indices decreases. Chern et al. (18) also compared the 3 mm BrainLAB micro-MLC (minimal leaf width is 3 mm) and Varian Millennium 120-MLC (minimal leaf width is 5 mm) using the DCA technique. This paper reported improved dosimetric results using 3 mm as compared to 5-MLC and for small target volume (< 1 cm 3 ), they reported as high as 10% improvement, on average, in CI.
Our current study compares the dose distributions between a high-definition 2.5 mm MLC leaf width (MLC 2.5 ) (High Definition (HD) MLC, Varian Medical Systems, Palo Alto, CA) and a standard 5 mm MLC leaf width (MLC 5 ) (Millennium MLC, Varian Medical Systems, Palo Alto, CA) in the treatment of intracranial lesion using DCA treatment techniques. Clinical patient cases were planned with both MLCs using the same DCA beam paths. Figures of merit (FOM) for plan comparison such as CI, target volume dose coverage, and normal tissue sparing were used to evaluate the superiority of the resulting treatment plans. The effect of target volume on the performance of the two MLCs also was analyzed using simulated and actual patient case data.

A. treatment delivery MLCs
A Novalis Tx (Varian Medical Systems, Palo Alto, CA) radiosurgical linear accelerator with a high definition (HD) MLC is one machine used by this study. This linear accelerator has 6 MV photon beam energy with a 1000 MU/min maximum dose rate. The Novalis Tx has a maximum possible MLC field size of 22 × 40 cm 2 at isocenter. The 22 cm is formed by 32 leaf pairs of 2.5 mm leaf width in the central part and 28 leaf pairs of 5 mm leaf width in the outer part of the MLC. This HD MLC is referred to as the MLC 2.5 in this paper. A Trilogy (Varian Medical Systems, Palo Alto, CA) linear accelerator using a standard 120-leaf MLC is used as the comparison system in this study. The Trilogy has a 6 MV and 18 MV photon beam energy with a maximum dose rate of 1000 MU/min. The 120-leaf MLC has two leaf widths with the inner 40 leaf pairs having a 5 mm leaf width and the outer 20 leaf pairs having 10 mm leaf width. The maximum field size for the standard120-leaf MLC is 40 × 40 cm 2 . This standard Trilogy MLC is referred to as the MLC 5 in this paper. All target volumes in this study were small and utilized the inner set of leaves to enable a direct comparison of 2.5 mm and 5 mm leaf widths.

B. Simulated target volume
A simulated patient case was created to study solely the effect of target volume size and shape on the performance of the two MLC systems. For an example patient case, a sphere was created as the target volume with volumes ranging from 0.11 to 39 cm 3 . The simulated sphere was centrally located within the patient head.

C. Study patients
A total of 43 stereotactic radiosurgery patients were selected for this study. These patients were selected from 278 patients treated since January 2007 in our institution. All of these (except one patient treated with IMRS) were treated with DCA method for SRS. Patient selection was based on including a wide variety of shapes and volumes of radiosurgery targets. The selected cases included 11 patients with acoustic neuromas, 3 with meningiomas, 4 with arteriovenous malformations, 19 with metastatic lesions, 3 with glomus tumors, 1 with an astrocytoma, 1 with a pineocytoma and 1 with a pituitary adenoma. The patients' ages ranged from 26 to 81 yrs old, with an average age of 55 yrs. The volume of targets ranged from 0.07 to 40.57 cm 3 and the average volume was 5.9 cm 3 . All these patients were planned using DCA technique for both MLC 2.5 and MLC 5 .

D. treatment plan creation
The DCA technique was used for all the plans in this study. In this technique, the MLC automatically conforms to the target volume outline via software methods. Figure 1 shows the difference via beam's eye view (BEV) between the MLC 2.5 and MLC 5 conforming to a target volume. The treatment plans were created using a dedicated radiosurgical planning system (iPlan RT Dose 3.0, BrainLAB, Germany). A planning CT image study was obtained for each patient. Patient images were acquired with a stereotactic localization frame (BrainLAB, Germany) attached to the patient's head. The CT imaging slice thickness was 0.625 mm. Magnetic resonance (MR) images were carried out for each patient at a similar slice thickness. The CT and MR were registered together to allow for MR-based target delineation. Target volumes and critical structures were outlined by our radiation oncologist and neurosurgeon during planning process. A DCA plan was created using 4 or 5 non-coplanar dynamic conformal arcs averaging 100° per arc. The treatment planning system using the MLC automatically creates a field shape that conforms to the target outlines as shown in Fig. 1. The collimator angles were set to 90° and, in few cases, it was changed to improve MLC conformity with the target volume. For non-metastatic lesions, no margin was added to create a planning target volume (PTV). For all metastatic lesions, a margin of 1 mm was added to create a PTV. To compare the MLC 2.5 and MLC 5 plans, all treatment parameters were identical for planning purposes.

E. treatment plan evaluation
The evaluation of competing MLC-based plans was performed using isodose displays with medical images, DVHs, and figures of merits (FOM). The use of isodose displays and DVHs are standard plan evaluation tools used in the clinical environment. Dose volume histograms (DVHs) were generated for each lesion and for the surrounding normal brain tissue. In this analysis, the use of FOM allows for a rapid and objective assessment on the quality of the two treatment plans. This section describes the FOMs used for evaluating the plan differences in this study. . These two views show MLC's leaf position for every ten degrees surrounding the target volume (in blue). The leaf width between the two views for the 2.5 mm and 5 mm leaf width is shown; the yellow outline demonstrates the difference in conformality of the two MLC systems.

E.1 Target coverage
Target coverage (TC) is defined as the percent volume of the tumor volume receiving the prescription dose. Typically, the coverage index should be at least 95%. In this study, the prescription isodose surface was selected as the greatest isodose surface covering ≥ 95% of the target volume and delivering 95% of the prescription dose to 99% of target volume. (11,18) After the prescription isodose surface was determined for the MLC 2.5 , the equivalent isodose surface was used for the MLC 5 . Table 1 shows data for an example patient case with a target volume of 4.93 cm 3 . Using this example data in Table 1, the 88% isodose surface covers 95.5% of the target volume, which satisfies the first condition. The second condition is also satisfied since 95% of this prescription isodose value (88% × 95% = 83.6%) covers 99% of the target volume. In this analysis, if the first condition was fulfilled and the second condition was not fulfilled, then a lower isodose surface was selected until both conditions are satisfied.

E.2 Normal tissue sparing
To evaluate the normal tissue sparing associated with the MLC systems, an anatomical structure consisting of an adjacent tissue shell was created to surround the target volume by adding a 1 cm margin. This is similar to the method described by Chern et al. (18) The DVH for this normal tissue structure was computed to assess the dose volume values for selected points on the DVH curve. The normal tissue volume 50 (NTV 50 ) was calculated which is the normal tissue receiving 50% of the prescription isodose (PI). The NTV 70 and NTV 90 receiving 70% and 90% of prescription isodose, respectively, were also computed from the DVH.
To assess normal tissue sparing, a normal tissue difference (NTD) was calculated. NTD is the difference between the volume of normal tissue receiving a certain dose utilizing MLC 5 and the volume receiving the same dose using MLC 2.5 . This is illustrated in Fig. 2 with a schematic diagram showing the striped region as the area calculated for the NTD. For example, the NTD for tissue receiving 50% of the prescription isodose (NTD 50 ) is calculated as follows: (1) where and are the volume of normal tissue receiving 50% of PI with the use of MLC 5 and MLC 2.5 , respectively.

E.3 Target conformity
Conformity indices are used to compare competing plans, evaluate treatment technique, and assess clinical complications by quantifying the dose conformity to a target volume. Several different indices have been reported to evaluate the conformity of prescription isodose surface to the target volume. In 1993, Radiation Therapy Oncology Group (RTOG) proposed routine evaluation of stereotactic radiotherapy treatment plans based on reference isodose value and reference isodose volume and target volume. (2)(3)(4)7,19) The RTOG proposed the conformity index as PITV (4,7,8,11,19,20) which is defined as: (2) where TV is the target volume and PIV is the prescription isodose volume.
The RTOG Conformity index, or PITV ratio, is the most frequently used conformity index to compare the conformity of treatment plans from different radiosurgery delivery systems. (4,7,8,(21)(22)(23)(24) According to the PITV ratio, a PITV value equal to 1 corresponds to ideal conformation. A PITV greater than 1 indicates that the irradiated volume is greater than the target volume, which means it includes healthy tissues. If the PITV is less than 1, it indicates that the target volume is partially covered. According to RTOG guidelines, a PITV (19) between 1 and 2 is considered a treatment plan of acceptable dose conformity. The treatment plan receives a minor violation judged by RTOG guidelines for a PITV between 2 and 2.5 or 0.9 and 1. The treatment plan is rated a major violation by RTOG standards if the PITV is less than 0.9 or more than 2.5.
The drawback of the PITV ratio is that it does not take into account the spatial location and the shape of the prescription isodose volume relative to the TV. If the volume of tissue receiving the prescribed dose is equal to the TV, the PITV ratio will be 1 and the treatment plan will receive the same perfect score of 1 regardless whether the prescribed isodose perfectly enclosed the TV or completely missed the TV (i.e. 0% of the TV received the prescribed dose). Therefore, PITV ratio can be improved by accounting for coverage of the target volume.
Paddick (20) accounted for target volume coverage by proposing a new conformity index as follows: where TV PIV is the target volume within the prescribed isodose volume PIV. This becomes the inverse of the PTIV when the prescription isodose fully covers the target volume.
Paddick's Conformity CI was modified by Nakamura et al. (7) and expressed as follows: where is the target volume, PIV is the prescription isodose volume, PVTV is the TV included in the prescription isodose surface. The PIV equals to the PVTV plus the normal tissue NT encompassed by the prescription isodose surface. The numerator of the above equation measures the excess volume of normal tissue within the prescription isodose surface and the denominator measures the target coverage. In this study, we used a similar conformity index method, which is as follows: where PIV is the total volume encompassed by the prescription isodose surface, PVTV is the volume of target TV encompassed by the prescription isodose surface, NT V is the volume of the normal tissue encompassed by the prescription isodose surface. It has been demonstrated that prescription isodose surface associated with the minimal CI does not necessarily produce provide adequate coverage. This is due to the fact that, as the coverage of the TV increases, a large amount of NT is included in the PIV and the plan will have higher CI. Therefore, the prescription isodose surface should be chosen that balances conformity and target coverage. In another report, Paddick et al. (25) also proposed a dose gradient index (GI), which is the ratio of the volume of 50% of the prescription isodose to the volume of the prescription isodose to compare plans of equal conformity indices. The GI shows which plan gives the steepest dose falloff outside the target. Other investigators introduced a conformity distance index (CDI). (26) The CDI is defined as the average distance between the prescription isodose and the target contour. This parameter accounts for the influence of target size and shape complexity on the conformity of the plan. The CDI is expressed as follows: (6) where S PIV and S TV are the surfaces of TV PIV and PIV, respectively. The CDI values were accurately calculated using custom-developed software by the authors.

A. Simulated target study
In our study, simulated spherical target volumes (ranging from 0.11 to 39 cm 3 ) were analyzed. These targets were created to reduce target shape effects on the CI and evaluate its dependence on the target volume only, for both MLCs. The plotted results show that, while CIs for both MLCs decrease with the target volume, the CI for MLC 2.5 for each TV volume is clearly lower than the corresponding CI of MLC 5 across the entire range of volumes (Fig. 3). Figure 4 demonstrates that NTV 70 and NTV 90 also increase with the volume of a spherical target using both MLC and that the MLC 2.5 provides better normal tissue sparing.  Table 2 presents the MLC 2.5 conformity indices which range from 1.15 to 2.44 with a median of 1.59. The CI range for MLC 5 was 1.60 to 2.85 with a median of 1.71. Similar to our results with the simulated target study, these results indicate that the MLC 2.5 provides better treatment conformity than MLC 5 as judged by the CI values in a range of actual patient tumors. The graph of percentage difference of the conformity indices between MLC 5 and MLC 2.5 versus target volume is shown (Fig. 5). While differences in CIs between the two MLCs decrease with target volume, even at large volumes the MLC 2.5 provides better conformity. Figure 6 shows dose volume histograms of the 1 cm normal tissue shell surrounding representative target volumes. In this figure, the normal tissue volume receiving doses between 20-80% of the prescription dose is reduced when MLC 2.5 is used. Comparison of the normal tissue sparing capabilities of the MLC 2.5 and MLC 5 is shown in Fig. 7 and Table 3. As illustrated by this data, NTV 50 , NTV 70 and NTV 90 for MLC 2.5 plans were reduced significantly, resulting in positive NTD values in each case. Table 3 also shows the values of CDI for all cases using both MLCs. The values of this parameter ranged were 0.9 to 4.8 mm for MLC 2.5 and from 1.37 to 6.15 mm for MLC 5 . The values of CDI indicate that the MLC 2.5 is more suitable than MLC 5 for highly complex shaped targets.

IV. DISCUSSION
Our study shows that the dose conformity of the MLC 2.5 is significantly better than MLC 5 . For target volumes ranging from 0.07 to 45.6 cm 3 , the MLC 2.5 CI improved by an average of 7.3%.
For small targets such as acoustic nueroma (< 1.1 cm 3 ), an average CI improvement was 11% for MLC 2.5 plans. For our cohort of patients studied (see Table 2), the CI was always better when MLC 2.5 was used, with the greatest CI difference of 18% seen for a lesion of 0.35 cm 3 .
Our study showed that the overall average difference in the CI between the two MLCs increases from 4.6% for the simulated spherical cases to 7.3% for the patients in this study. Since the shape of a spherical target is simple and targets of actual patients are more complex, these results suggest that MLC 2.5 yields better conformity for complex tumor shape than MLC 5 .
Our results showed that the CDI values for all tumor sizes and shapes studied were less when MLC 2.5 was used compared to those with MLC 5 . The CDI values were higher for highly complex targets (such as in case 16 which is an AVM) and low for uniform shapes (such as in case 27) as shown in Table 3.
The MLC 2.5 is a better choice to treat SRS target volume with DCA treatment technique particularly for small lesions and geometrically complex tumors. Other investigators have reported modest but statistically significant improvements using small leaf MLCs. (9,11,13,18) In particular, investigators from other institutions compared 3 mm leaf width MLC (MLC 3mm ) with MLC 5 and reported better target conformity and tissue sparing with the MLC 3mm . (14,17,18) Hazard et al. (11) compared linear accelerator-based SRS using DCA with a Gamma Knife system and reported that accelerator-based SRS provided comparable treatment conformity. Thus, MLC 2.5 linear accelerator-based SRS system should offer at least equivalent conformity when compared with Gamma Knife systems. With both simulated targets and actual patient tumors, the difference in the conformity indices between MLC 5 and MLC 2.5 decreased with target volume (Fig. 3 and  5). Two separate studies reported in the literature (17,18) found similar trends, whereas Monk et al. (14) did not report any dependence of CI on target volume. Figures 4, 6 and 7 show that using MLC 2.5 improves normal tissue sparing. Figure 7 shows that NTD increased with target volume and its values were always positive, suggesting better normal tissue sparing with the use of MLC 2.5 compared to MLC 5 . Based on our analysis, MLC 2.5 provides better normal tissue sparing than MLC 5 for intracranial lesions even where the CI difference is modest. This is significant for a patient where the tumor is located close to the brainstem or the optic chiasm. In this scenario, the MLC 2.5 could offer a conformal therapeutic dose to the target volume while providing a lower dose for the adjacent critical structure. For such a patient, the dose to critical structure could prohibit the treatment using the MLC 5 .

V. CONCLUSIONS
The MLC 2.5 has a dosimetric advantage over the MLC 5 both in the treatment conformity (CI and CDI) and normal tissue sparing when target shape complexity increases. The MLC 2.5 has an advantage particularly when the critical structures are adjacent to the target volume. The use of MLC 2.5 improved the plan conformity for small lesions (< 1.1 cm 3 ) on average by 11%, as compared to the MLC 5 . In summary, the MLC 2.5 has the potential to provide improved dose conformity to the target volume and lower doses to critical structures compared with standard MLCs such as MLC 5 . The 2.5 mm leaf width for an MLC represents a new standard for linear accelerator-based radiosurgery.