To evaluate the use of standard apertures and range shifters for the treatment of brain metastasis in proton stereotactic radiosurgery.
Nine patients with localized brain metastasis who had been previously treated using our intracranial proton stereotactic radiosurgery procedure (ie, with a custom aperture and bolus) were randomly selected from our patient cohort. The custom aperture and bolus treatment plans were used as the standard of care in this case and comparative treatment plans using the standard aperture and range shifter concept were generated. Gantry/table angle and the number of treatment beams were optimized as part of this study to evaluate the ability of the standard aperture/range shifter system to deliver a comparable treatment to the patient. Conformity index, homogeneity index, isodose volumes, and integral dose were all evaluated to determine the degree of conformity of the plans created and for comparison to the custom aperture/bolus treatment modality.
The generated treatment plans demonstrated that the standard aperture and range shifter combination could be used to produce comparable conformity indices and isodose volumes to the custom aperture/bolus case in eight of the nine patients studied. A comparative conformity index was achieved in two patients by optimizing the angles of the 3 treatment beams, whereas, in the other cases, additional beams (1 or 2) were required. Additionally, this system exhibited efficiency gains of 60% to 90% over the custom aperture bolus system in reducing the time necessary for treatment planning, device manufacture, and quality assurance.
This work demonstrated that the largely spherical shape of brain metastases make this target well suited to an application of standard apertures while additionally providing efficiency gains in device manufacture and quality assurance for treatment.
Protons lend themselves well to radiosurgical applications through the way in which they deposit dose via the Bragg peak. Proton depth dose curves are characterized by a low entrance dose, followed by the high-dose Bragg peak, which can be placed at varying depths within the patient through modification of the initial proton energy . Beyond the Bragg peak, dose deposition falls off sharply, limiting dose to structures located beyond the target. Additionally, for energies typically used in intracranial radiosurgery, the penumbra of the proton beam is sharper than other external-beam modalities used for this treatment . These intrinsic properties allow clinicians to treat intracranial targets with fewer beams (typically 3 to 4) than used with photon-based radiosurgery modalities, drastically decreasing the integral dose experienced by the patient .
Intracranial radiosurgery has been practiced at the James M. Slater Proton Treatment and Research Center at Loma Linda University Medical Center (LLUMC; Loma Linda, California) since 1993 for arteriovenous malformations and brain metastases. The program was initiated by a team at LLUMC, including Richard Levy, Reinhard Schulte, Ken Frankel, Gary Steinberg, Michael Marks, Jerry Slater, and James Slater, after encouraging results from the Laurence Berkeley National Laboratory-Stanford University (Berkeley, California) helium radiosurgery program treating arteriovenous malformations . To date, 186 patients with arteriovenous malformation have been treated, with a similar number of patients with brain metastases also treated using proton radiosurgery at LLUMC. As imaging techniques continue to evolve and improve in the detection of small lesions, it is expected that the patient cohort of brain metastasis patients exhibiting smaller targets will increase, presenting a challenge to effectively and efficiently treat such lesions.
The LLUMC current standard for proton beam delivery to intracranial radiosurgery targets is passive scattering. In this modality of proton delivery, the proton beam is scattered by a one or two stage scattering system to create a beam of sufficient area to cover the target . A rotating, plastic wheel allows for the superposition of multiple Bragg peaks of varying depth and intensity to create a spread-out Bragg peak (SOBP) that is selected during the planning process to ensure uniform dose coverage of the target as a function of depth . The lateral margins of the beam are shaped using customized apertures manufactured from brass or Cerrobend, whereas the distal edge is shaped with a low-Z (wax or Lucite) material bolus or compensator . The custom beam-shaping devices (aperture and bolus) allow for the delivery of a very conformal dose to the target volume, yet their custom nature adds pressure during device manufacture, quality assurance (QA), and calibration, extending the time between initial consultation and the beginning of treatment. During the imaging and treatment process, the patient is immobilized with a vacuum-assisted bite-block fixation system . Alignment in the treatment room is achieved with a computed-tomography guided six degree of freedom patient positioner, with 2-dimensional, orthogonal x-ray imaging completed to verify positioning accuracy.
We are currently testing and developing the ability to use standard brass apertures of a particular diameter and uniform range shifters to treat intracranial radiosurgery targets with passively scattered protons. Such standard aperture systems are already used clinically in gamma knife and cyberknife systems to make the lateral boundaries of the radiation field conform to the target volume. These standard apertures and range shifters are precalibrated and linked to the treatment-planning system, allowing for selection at the planning level by the dosimetrist. Each treatment device is barcoded and, upon plan export, the associated treatment devices are linked to the DICOM RT Ion Plan to ensure that the correct device is used for each treatment field. Before treatment, output checks can be completed; however, those output checks would be significantly less time-intensive than those performed on custom aperture/bolus treatment fields, saving time and improving QA throughput. It is hypothesized that this system will be effective in the treatment of intracranial brain metastases, which generally have simpler spherical shapes, and also allow for efficiency gains in the treatment workflow. These efficiency gains are aimed toward achieving single-day treatments and potentially allowing for the use of rigid immobilization devices. By having available both custom and standard systems for use, the treatment team can determine which system is best suited to the needs of the patient and customize not only the dose distribution to the target but also the time to treatment.
To evaluate the efficacy of using standard apertures and range shifters in the treatment of localized brain metastases, we completed a comparative treatment planning study of 9 patients. The patients were selected at random from our intracranial proton radiosurgery patient cohort. These patients had been treated previously using a custom aperture and bolus configuration and that plan was used as the standard for comparison. Comparative plans were created using the standard aperture/range shifter configuration by first keeping all table/gantry angles the same and then by optimizing those for target coverage and normal tissue avoidance. A plan with 1 to 2 additional treatment beams was also completed to evaluate the effect those additional beams had on dose conformity to the target and integral dose. Such a study not only tests the efficacy of such a system for proton radiosurgery but also determines the potential time savings of such a system and provides treatment planning guidelines for clinical implementation.
Materials and Methods
Nine patients with localized brain metastases, who had previously been treated using our standard intracranial proton stereotactic radiosurgery procedure (ie, with a custom aperture and bolus) were randomly selected from our patient cohort. The custom aperture and bolus treatment plans were used as the standard of care, and all plans created with the standard aperture and range shifter system were compared back to that standard. The standard aperture system comprises of 5-, 8-, 10-, 12-, 15-, 18-, 20-, 25-, and 30-mm-diameter apertures and uniform standard range shifters with a water equivalent thickness range of 10 to 50 mm in 2-mm steps. Each of the devices (aperture and range shifter) were barcoded and linked to both the treatment plan and the treatment console (through export of the DICOM RT Ion Plan) ensuring that the dose could be delivered only when the appropriate devices were in place for treatment. The aperture and range shifter are housed in the radiosurgery cone (Figure 1) which is used for all stereotactic radiosurgery treatments with target diameters of < 4 cm. The radiosurgery cone allows for minimal air gaps between the beam-shaping devices and the patient, which leads to improved beam penumbra. Proton beam energies available for selection had a range in water of 9.7 or 15 cm with selectable beam modulations of 1.5, 2, 2.5, 3, 4, 5, and 6 cm.
Each tumor selected for testing had a physician-approved plan that was used in clinical patient treatment, which enlisted custom-built apertures and boluses that were not altered (referred to in this article as treated). These plans were identified as custom plans and served as controls. Four additional experimental plans were created for each tumor. The plans were created using the following guidelines:
Standard treatment (Std-Treat) contained the same number of beams, gantry, and table angles as the custom plan and replaced only the custom aperture and bolus for a standard aperture and range shifter
Standard optimized (Std-Optim) contained the same number of beams as the custom plan and adjusted the gantry and table angles and used the standard apertures and range shifters in all beams
Standard plus 1 beam (Std+1Beam) contained an additional proton beam and adjusted gantry and table angles and used the standard apertures and boluses in all beams.
Standard plus 2 beams (Std+2Beams) contained 2 additional proton beams and adjusted gantry and table angles and used standard apertures and boluses in all beams.
Each plan was optimized to ensure 80% isodose coverage of the target qualitatively with a dose to the prescription point of 25 Gy (ie, 100%) delivered in 1 or 2 treatment fractions. Additionally, to determine the quality of treatment for each tumor, the conformity index (CI) and homogeneity index (HI) were calculated based on the Radiation Therapy Oncology Group (Philadelphia, Pennsylvania) formalism. The CI was the most valuable quantitative metric in the comparison between the standard and custom plans because the HI is relatively stable in proton therapy due to the nature of the dose delivery. The CI is defined as the volume of reference isodose divided by the target volume [9, 10]; CI values between 1 and 2 are required for treatment, whereas values > 2 indicate that an unacceptable amount of healthy tissue has been radiated, and values < 1 indicate that the target volume is not adequately covered by the prescribed isodose . Treatment plans with the standard aperture/range shifter system needed to be within 0.2 of that achieved with the custom aperture and bolus while additionally meeting the criteria of 1 < CI < 2 to be considered acceptable. Additionally, the volume of tissue treated at 80%, 50%, and 10% isodose were calculated and used to assess the amount of healthy brain tissue radiated in addition to the tumor at a given isodose.
Figure 2 displays the transverse isodose distribution at the level of the prescription point for patient 4. The patient was treated with 3 beams and custom aperture/bolus, which provided a tight conformal isodose distribution with minimal integral dose to the surrounding brain tissue. When the custom aperture/bolus is replaced with a standard aperture/range shifter (Figure 2B), the dose distribution spreads as a direct consequence of the beams' distal edge not being shaped to the target volume, which can be compensated for by either optimizing the beam angles (Figure 2C) or by adding additional beams (Figure 2D and E), although additional treatment beams do affect the integral dose to the patient and may not guarantee improved conformity.
Table 1 displays the CI for each patient and each plan and gives an indication of which plan best reflects the treated case with a custom aperture and bolus. The replacement of the custom aperture and bolus with a standard system universally resulted in an unacceptable CI (the Radiation Therapy Oncology Group recommends that the CI lie between 1 and 2 for treatment ). The optimization of the gantry and table angles provided the best solution for patients 4 and 5 who had smaller targets (< 1 cc in volume) (Table 2), whereas additional beams were necessary for the larger targets in the other patients. Patient 3 was not able to achieve a CI with the standard aperture/range shifter combination to approach that delivered with the custom system. Because this was a shallow and relatively small target, a range shift and beam modulation combination was not available to enable sufficient optimization. Planning for patient 3 further highlighted that the standard aperture and range shift systems are designed, not as a replacement for the custom aperture and bolus, but rather as a supplement to those systems in a subset of patients.
The volume of tissue irradiated within the 80% isodose for each target/plan and the target volume is displayed in Table 2. These data were used to determine the CI shown in Table 1 and, as such, directly reflect the trends discussed previously. They were, however, included for completeness. The HI was also calculated for all plans generated as part of this study. Across all plans, the HI had an average value of 1.254 with an SD of 0.005. That very stable result demonstrates how the SOBP in proton therapy uniformly irradiates the tumor with depth, delivering homogeneous dose coverage of the target that is independent of the number of treatment beams used. The use of custom or standard beam-modifying devices, modification of gantry/table angles, and addition of treatment beams will affect the CI, as shown previously, yet the HI remains relatively unchanged and thus provides minimal information on proton plan optimization.
The volume of irradiated tissue within the 50% isodose and 10% isodose are shown in Tables 3 and 4, respectively. Those volumes are especially important to track during plan optimization because, although the CI can be reduced by adding beams or optimizing beam angles, doing so can lead to an increase in integral dose to the surrounding normal brain, which must be managed. The 12-Gy isodose volume is constrained in our treatment planning guidelines for a normal brain volume of < 10 cc to be deemed acceptable for a single-fraction treatment. In this case, the small target size and minimal beams required by proton radiosurgery produced an average 50% normal tissue isodose volume of 5.3 cc, which is well within limits. Nevertheless, these criteria become more challenging to meet as target size increases. An example is seen in patient 6, whose larger target of 4.36 cc resulted in a 50% normal tissue volume of 12.0 and 12.9 cc for the clinically acceptable plans (based on the CI) of Std+1beam and Std+2beams, respectively. On further evaluation, the normal brain isodose was below 10 cc (ie, when the skull and other normal tissues were removed from the volume), potentially allowing the target to be treated in a single fraction. Were the 12 Gy isodose volume > 10 cc, it would be possible to treat the target with 2 fractions, provided the 17-Gy normal brain isodose volume remained < 10 cc.
As shown in Table 1, patients 4 and 5 achieved a CI with the standard system that matched that of the custom system with the same number of beams yet with the angles of those beams optimized for the standard aperture/range shifter. With those optimized beam angles, the 50% isodose volume improved by 12% and 4%, respectively, whereas the 10% isodose volume for patient 4 was again superior to that of the custom aperture/bolus case, and, for patient 5, there was an increase of 4.33 cc. All other patients (with the exception of patient 3) required additional beams to achieve a CI that was comparable to that of the custom aperture and bolus case. The additional beams resulted in a negligible increase in the 50% isodose volume of 0.2 cc on average for those patients, whereas the 10% isodose volume increased by 41 cc, on average.
The standard circular aperture and range shifter system proved a useful tool for treating brain metastasis because of the largely spherical shape of the tumors. With a standard uniform range shifter, there is no shaping of the proton distal edge to conform to the target shape. As such, it was found that using treatment fields that are largely orthogonal to one another is key to a conformal dose distribution. In this arrangement, the circular aperture edges are used as the primary means for controlling dose conformity because of the largely spherical target volume. Selecting the appropriate range shifter and modulator wheel, although important to control the integral dose to the brain and shaping the proximal and distal edges of the SOBP, has less of an effect on making the prescription isodose line conform to the target. Such beam arrangements are made possible largely by a patient positioner with 6 degrees of freedom, which not only allows patient alignment in six-degrees but also makes available noncoplanar beams enabling perpendicular beam arrangements.
Although beam shaping along the beam direction (ie, with a range shifter or modulator wheel) is not as critical for conformity of prescription isodose to the target, it is an important consideration in determining which patients and targets are candidates for treatment with this system. Because of the gains in efficiency with this system, it is desirable to make the system applicable to a wide range of brain metastasis cases. This study demonstrated 3 potential areas of improvement in the system to allow treatment of a wider cohort of patients. The first is to modify the maximum available range shift from 50 mm (resulting in a minimum distal edge depth of 4.7 cm in water) to 60 or 70 mm. This allows smaller superficial lesions (such as those of patient 3), which have a maximum water equivalent depth of 2.7 to 3.7 cm, to be treated with the standard aperture and bolus system. The second potential improvement is to provide smaller aperture size increments of between a 2.0 and 3.0 cm diameter. The current specification of 5 mm limits the ability of the aperture to make the prescription isodose conform to the target volume, and steps of 2.5 mm are advantageous in treatment planning. Such a refinement would have been advantageous in planning for patients 6, 7, 8, and 9, with the potential to treat those patients' targets with an acceptable CI using fewer beams. The third potential improvement is to develop modulator wheels with a finer step size, especially at the smaller modulation widths. Currently, SOBP width can be varied in 5-mm increments from 15 to 30 mm and in 10-mm steps from 30 to 60 mm. Providing more options in this range and expanding the available beam modulations to those that are smaller than 15 mm will provide greater options in beam conformity, especially for smaller lesions, as demonstrated with patients 2 and 3. These modifications would have had a bearing on the system's ability to effectively treat patient 3 in this study.
The standard aperture/range shifter system is fully integrated and interchangeable with the custom aperture and bolus system. As such, in cases that lie outside the capabilities of the standard system (eg, patient 3), the custom system or a combination of both systems (ie standard aperture and custom bolus and vice versa) may be used, although that will affect the efficiency gains in treatment planning and QA. For example, although not part of this study, it is conceivable that the standard range shifter could be replaced with a patient-specific bolus in cases where additional distal edge shaping is necessary. As more patients are treated with the standard aperture/range shifter system, custom aperture/bolus system, and a combination of both, further refinements and reporting of outcomes is planned.
The real benefit of the standard aperture and range shifter system is that it provides a means for quickly planning a patient treatment and in completing QA, thus shortening the time necessary between patient simulation and treatment. The current standard of treatment for intracranial radiosurgery uses custom apertures and boluses, which is a relatively time-intensive process requiring design, manufacture, QA, and calibration of custom treatment devices. That process precludes the use of rigid immobilization and typically requires a number of days between the initial consultation and treatment. The standard aperture and range shift system can enable the treatment team to deliver a highly conformal dose to small spherical targets on the same day as imaging. This development presents the possibility of using rigid immobilization, and may have real benefit in preventing patient physical (ie, swelling from steroid use) and physiologic (ie, tumor growth) changes from affecting the treatment plan. In addition, the standard aperture and bolus system has real applications to functional radiosurgery, where highly conformal doses to simple targets are necessary with minimal setup uncertainty (indicating a need for rigid fixation).
The application of this system within our clinical practice is already underway for brain metastasis cases when the initiation of treatment is of some urgency from both a medical and patient well-being standpoint. Patients using the standard aperture/range shifter system are immobilized by our bite-block immobilization system, which provides a setup uncertainty of 1 mm . At the time of patient consultation, the physician identifies whether the patient should be treated with the standard aperture/range shifter system, and treatment planning is initiated. The physician reviews the case and determines whether it meets our planning criteria, including
80% isodose coverage of the target
1 < Conformity Index < 2
12 Gy normal brain isodose volume < 10 cc (single dose)
17 Gy normal brain isodose volume < 10 cc (2 fractions)
When the plan is approved by the physician, it is reviewed by our radiosurgery team before treatment. If the criteria are not met or if the physician (or radiosurgery team) feels that a more-optimal plan could be achieved, the treatment can be replanned with either a standard aperture and patient-specific bolus combination, custom aperture and standard range shifter combination, or with a patient-specific aperture/bolus combination. The ability to use both systems interchangeably, with external physician/physics review, allows the optimal plan to be delivered to the patient based on clinical need.
This work demonstrated that the largely spherical shape of brain metastases make this target well suited to an application of standard apertures and range shifters. In eight of the nine cases studied, this standard aperture/range shifter system generated treatment plans with CIs similar to the custom aperture/bolus and additionally provided efficiency gains in device manufacture and QA for treatment. This system has been clinically deployed at LLUMC for physician use in such cases. Further refinement of the system is planned that will improve its ability to be used on a wider range of target sizes and depths within the cranium. There are also plans to expand use of the systems to other targets within the cranium, including deployment for functional radiosurgery.
ADDITIONAL INFORMATION AND DECLARATIONS
Conflicts of Interest Disclosure: The authors have no conflicts of interest to disclose.
Acknowledgments: The authors would like to thank the treatment team in the Department of Radiation Medicine at Loma Linda University Medical Center who contributed to the intracranial radiosurgery program for more than 20 years.