Gamma Knife® Stereotactic Radiosurgery and Hypo-Fractionated Stereotactic Radiotherapy

Abstract

Gamma Knife radiosurgery is credited with the establishment of radiosurgery as a mainstay in the management of both benign and malignant tumors. Evolution in Gamma Knife technology has opened up new possibilities for the choices of dose and fractionation for these patients. This section discusses the technology, practical aspects of dose planning, and selection of dose for various indications and the results thereof.

Learning Objectives

• Understand the technology and technique of Gamma Knife radiosurgery.

• Develop a framework for planning a radiosurgery case.

• Understand dose prescription guidelines and the key Gamma Knife literature supporting them.

Description and Evolution of Modality

Swedish neurosurgeon, Lars Leksell, proposed the concept of stereotactic radiosurgery [1] as the use of stereotactically directed ionizing beams to ablate intracranial targets in 1951. Sixteen years later, together with physicist Börje Larsson [2], he completed the design of the Leksell Gamma Knife (LGK) as a dedicated tool to perform brain radiosurgery. The LGK in all its various models consists of a number of independent Co⁶⁰ sources that emit gamma radiation in the 1.1 MeV range that are focused through a series of collimators to one focal point (isocenter). The diameter of the isovolume created by the cross-firing of approximately 200 beams can be varied from 4 to 16 mm (18 mm in the early models). Treatment plans are generated by superimposing multiple such dose clouds creating a multi-isocenter dose plan. The target is then stereotactically aligned with the focal point of the unit, and treatment delivery is one isocenter at a time.

The key features of the LGK and the evolution of the technology over the various models are summarized in Fig. 45.1.

Fig. 45.1

Key technological highlights for various models of the Leksell Gamma Knife arranged chronologically from left to right

Immobilization Techniques and Image Guidance

Precise delivery of the treatment plan is dependent on the ability of the system to localize the target in stereotactic coordinate space. This requires immobilization of the patient. There are two immobilization techniques that can be used with the LGK: the Leksell stereotactic frame and a thermoplastic mask.

Leksell Stereotactic Frame

Originally designed for stereotactic neurosurgery, this device is made of high-grade anodized aluminum and uses a Cartesian coordinate system to localize targets in stereotactic space (Fig. 45.2). The coordinates are expressed as a triplet of x, y, and z. The origin of the coordinate system (0,0,0) is on the right—superior—posterior aspect of the skull. The x-axis runs from the right to the left; the y-axis runs from posterior to anterior and the z-axis from superior to inferior. The center of the coordinate system has a value of 100,100,100.

Fig. 45.2

Leksell coordinate space and the Leksell stereotactic frame model G. [Courtesy of Elekta]

Using the frame provides a very high level of accuracy allowing the device to perform at its calibrated specification which is always better than 0.3 mm. Inaccuracy in imaging and frame displacement as a result of improper application are the major sources of error with this setup.

Thermoplastic Mask

With the LGK Icon®, it is possible to immobilize the patient using thermoplastic mask and a deformable at cushion (Fig. 45.3). This fixation system permits single as well as hypo-fractionated stereotactic treatments with the Gamma Knife. Stereotactic coordinates are obtained by performing a cone beam computerized tomography (CBCT) using the onboard CBCT system.

Fig. 45.3

Thermoplastic mask , moldable cushion, and cradle for hypo-fractionated treatments. [Courtesy of Elekta]

Cone Beam CT

The CBCT in the ICON is integrated into the patient positioning and source unit as one rigid entity. This makes it operate in true Leksell coordinate space—and every voxel in the reconstructed image has known Leksell coordinates, requiring a transformation along only the z-axis, while determining true x and y coordinates. This distinguishes it from all other image guidance systems in use. The unit operates at two different computerized tomography dose index (CTDI) settings 2.5 mGy and 6.3 mGy. It is customary to use the higher CTDI setting for the localizing scan when the co-registration is being performed to a pre-planning MRI, allowing more detail for mutual information matching. The lower CTDI setting is used for the daily delivery CBCT which is co-registered to the reference CT.

Accuracy of the CBCT system has been verified experimentally (Dalhalwi) and shows excellent concordance with frame-based coordinates in phantom studies. Insert error values here.

Intra-Fraction Motion Management

In order to ensure “intra-fraction patient position” when a thermoplastic mask is in use, a high-definition motion management system (HDMM) is coupled with the thermoplastic mask (Fig. 45.4). The HDMM uses an infrared reflective marker placed on the nose and tracked by an infrared camera relative to static reflectors located on the patient head cradle. A motion trace for the marker is then displayed on the operator console allowing the operator to set a tolerance level for the maximum deviation of the patient from the initial position. The system automatically suspends (gates) delivery of radiation if the patient exceeds the programmed tolerance. Should the patient return within tolerance in a predefined time interval, treatment delivery will continue; however, if the patient remains out of position for a single or repeated periods exceeding 20 s, then the treatment is interrupted and the patient is ejected from the machine. The operator then decides whether to override the deviation or to perform a new cone beam CT and realign the patient for continued delivery. In practice with a cooperative patient, it is not unusual to have deviations of the nasal marker less than 0.5 mm. Most patients can be delivered to treatment with nose marker deviations under 1.5 mm. It should be pointed out that the marker on the nose is a surrogate for target position, and the relative target deviation depends on target location in the brain. Based on studies conducted on a test system, the corresponding deviation at the target was on average half of that displayed as the nose marker deviation.

Fig. 45.4

High-definition motion management for mask-based delivery with the LGK Icon®. [Courtesy of Elekta]

Treatment Planning

Treatment planning for the LGK is performed on a dedicated planning system—Leksell Gamma Plan®—and can be performed manually, semiautomatically as a forward plan with optimization assistance or as a fully automated inverse plan.

At the outset, it is important to understand that the treatment planning with LGK is more akin to brachytherapy rather than a conventional external beam plan. This is due to the fact that the plan is often multi-isocenter and comprised of multiple superimposed dose clouds, each with its own isocenter called a “shot” in LGK parlance. The goal is to create a confluent dose cloud that conformally encloses the target. This fundamental aspect of an LGK plan makes it inherently more conformal but also more prone to heterogeneity. Another key difference from traditional IMRT and SRS plans on a LINAC-based system is the fact that dose is prescribed (or normalized) to isodose lines varying from 90% to 30% (with the mode and median prescription IDL in plans being the 50%). This derives from the fact that the LGK dose profile offers the steepest dose gradient between the 40% and 55% isodose lines based on its physical design characteristics.

Typical Dose Distribution

At initial loading, LGK houses upward of 6000 curies of radioactivity and a resultant dose rate of 3.3–3.6 Gy/min. Based on the collimator output factors, this dose rate is modified by a factor of 0.8–1.0 for the three collimators 4, 8, and 16 mm. Their numeric designation refers to the diameter in mm of the 80% isovolume of a single shot of a given size. Dose profiles are shown in Figs. 45.5 and 45.6, and typical dose distributions of the three collimators as depicted in Gamma Plan are shown in Fig. 45.7.

Fig. 45.5

Dose profiles along x-axis for the three collimators 4, 8, and 16 mm for LGK Perfexion and ICON

Fig. 45.6

Penumbra widths for all sources combined for Leksell Gamma Knife Perfexion or ICON

Fig. 45.7

Typical dose distribution of single 4, 8, and 16 mm collimators . The isodose line in yellow is the 50%. Also represented are 10, 20, 30, 40, 60, 70, 80, and 90% isodose lines in green, axial plan is represented in the left-most column, coronal in the middle column, and saggital in the right-most column

In practice the user has to develop a sense for the isovolume generated by each collimator in three dimensions; since visual inspection is dependent on the magnification of the images being viewed, it is best to start by placing a test shot of a given size on the images in question.

When more than one isocenter/shot is present, the superimposed distribution depends on the size of collimators used and the separation between the isocenters. It is common practice to place adjacent shots such that they are overlapping. As shown in Fig. 45.8, the inter-isocenter distance has an effect on both the prescription isodose (yellow) and the appearance of cold and hot spots.

Fig. 45.8

Two 8 mm collimator shots placed adjacent to each other will yield a resultant dose distribution with a cold spot if they are non-overlapping. As they are moved closer (top to bottom), the cold spot diminishes in size, and a hot spot appears, axial plan is represented in the left-most column, coronal in the middle column, and saggital in the right-most column

For the Perfexion and Icon models , sources are mounted in groups of 24 on movable conical sections called sectors. Each of the 8 sectors for a given isocenter can be configured to be blocked or open and collimated to the 4, 8, or 16 setting. Since the sectors are arranged along the z-axis, the most intuitive effect of blocking a single sector is in the axial (x–y) plane. The influence on the other planes is not intuitive and is illustrated in Fig. 45.9 since in those planes the skull geometry affects the dose rate from different sectors differently. In addition to choosing the size of the collimation , sector blocks and composite shots (with different collimator settings for different sectors) are other ways to creating shaped dose distributions.

Fig. 45.9

Influence of sector blocking on the dose distribution from a 4 mm collimator, illustrating the difference between sectors

Prescriptions for the LGK were historically normalized to the 50% IDL, but in fact as the more automated models of LGK became available and the number of isocenters used increased, this is no longer true. Once multiple isocenter penumbras are combined, the steepest gradient can actually fall anywhere in the 40–55% IDL range. As discussed later this effect will reflect itself in steeper dose gradients and lower values of gradient index with normalization to less than 50% IDL. Care must be exercised in recognizing that the associated peak dose prescription and mean energy delivered by the plan will increase (Fig. 45.10). This can have consequences on the target and its response to the treatment.

Fig. 45.10

Dose profile along the long axis of a tumor demonstrating the impact of renormalization of the dose plan from the 50% IDL to the 40% IDL: steeper gradient in normal tissue and elevated hot spot in target. [Courtesy of Ian Paddick]

Measures of Plan Quality

Before discussing planning and prescribing techniques with the LGK, it is important to discuss and define the parameters used to assess the quality of a dose plan.

Coverage : is defined as the proportion of the target volume (TV) that is covered by the prescription isodose volume (PIV), that is, Volume (PIV∩TV)/Volume (TV).

Selectivity : is defined as the proportion of the prescription isodose volume (PIV) that is inside the target volume (TV), that is, Volume (PIV∩TV)/Volume (PIV).

Gradient Index : is defined [3] as the quotient between the half-prescription isodose volume size and the prescription isodose volume size, that is, Volume (PIV25%)/Volume (PIV50%) if the planning isodose is 50%. Gradient index is commonly used to quantify the steepness of the dose falloff.

In addition, there are several measures of plan conformality:

RTOG PITV ratio : Defined by Shaw et al. [4], this is simply the ratio given by PIV/TV. It has the advantages of being easy to calculate. A value of >1 suggests a treatment volume that exceeds the target and implies irradiation of non-target surrounding tissue.

A value of <1 suggests a that treatment volume is smaller than the target and therefore indicates under-coverage of target. However, this ratio fails to reflect the actual overlap of the two volumes, leaving that determination to the planner.

Paddick conformity index : Defined by Paddick [5] this is calculated as the product of the coverage and selectivity. It is therefore ((PIV∩TV) ²)/(PIV × TV). It has a theoretical maximum value of 1. This index has the advantage of accounting for the concordance between dose and target as well as being a number that ranges from 0 to 1.

When designing a dose plan, the planner strives to achieve coverage as close to 1 as possible, although in most clinical situations, any value greater than 0.95 is found to be acceptable. Likewise, selectivity should be maximized, and values in 0.75 are easy to achieve and a reasonable target for the plan. Coverage and selectivity are often inversely related to each other, particularly in irreglar targets. Clinical judgment should be exercised to decide the balance netween the two, as long as a minimum of 0.95 in coverage has been achieved.

The achievable gradient index (GI) has a theoretical limit based on the physical characteristics of the LGK around a value of 2.5. The factors that determine the GI include number and size of collimators used as well as the IDL to which the dose is prescribed. For example, in the curve shown in Fig. 45.11 based on one plan shows that the lowest GI would correspond to the 40% IDL. In addition to the global dose gradient, local gradients close to critical structures are also important for treatment plan quality assessment.

Fig. 45.11

Dependence of the gradient index value on the prescription isodose line chosen

Forward Planning

The process of forward planning begins by delineating a target volume and designating it as such in the planning system. The user then sets a dose grid (called target or matrix) centered on the desired target. There is a separate grid on each target in the dose plan such as in multiple metastases.

The dose plan is then constructed by placing individual shots of varying collimator size and or composite shots in the target. The dose filling strategy varies greatly by operator and both “center out” and “periphery in” filling paradigms are used. In general, while the largest collimator provides the most coverage, it can also provide the sloppiest dose gradient in normal tissue. Thus, an efficient dose plan includes large collimators used away from critical structures and smaller collimators closer to critical structures. Fig. 45.12 demonstrates the construction of a dose plan for a cavernous sinus meningioma.

Fig. 45.12

Development of a forward dose plan for a right cavernous sinus meningioma. The plan begins (a) with the a composite 8 and 16 mm shot placed in the center of the tumor (shown are the axial, coronal, and shot configuration representations of the plan), followed by the addition of other shots (b and c) and the final dose plan in axial and coronal views (d) and sagittal and 3D representations (e). Note that doses are depicted in Gy. The isodose lines represented in green from outside in are 8, 10, 16, 18, and 20 Gy. The prescription line is 14 Gy shown in yellow. Target volume – red and optic chiasm and pathway (pink)

Optimizer-Assisted Forward Planning

The optimizer provided with LGP uses a cost function for optimization . The values of this cost function lie between 0 and 1, and the higher the value, the better the plan quality. The equation for the cost function is:

$F=\frac{c^{\min \left(2\alpha, 1\right)}{s}^{\min \left(\left(2-2\alpha \right),1\right)}+\beta G+\gamma T}{1+\beta +\gamma }$

c is the target coverage, and s is the selectivity (defined previously).

α, β, and γ are weights between 0 and 1, set by the user.

G and T are functions whose values lie between 0 and 1 that describe how “good” the gradient index, g, and treatment time, t, are. (G = 1 if g < 2.6, G = 0 if g > 6 & T = 1 if t < 0.25 T ₀, T = 0 if T > 1.5 T _0, where T ₀ is the beam-on time at the start of the optimization).

α, β, and γ are set by the user with interactive sliders in the inverse planning settings dialog box (Fig. 45.13). The coverage and selectivity sliders are interconnected since they visually reflect the effect of the α parameter which drives the c and s variables in the cost function. The gradient index slider is the β parameter and drives the G parameter. Beam-on time represents the γ parameter and drives the T function.

Fig. 45.13

Options in the inverse planning dialog for Leksell Gamma Plan

Default settings in LGP are α = 0.5, β = 0.25, and γ = 0. Therefore, if no changes are made, then the cost function ignores the time of plan delivery and concentrates on a good balance of selectivity and coverage with a gradient index which is as low as possible. The algorithm for optimization can change the already placed shots by the user and can even delete shots that are deemed redundant, using simulated annealing to maximize the cost function. The user can restrict the system from changing the manual plan in various ways. It is recommended that optimization be performed on a plan copy. If the lock positions and lock collimator settings boxes are checked (Fig. 45.13), then the only possible changes to the plan that the optimizer can make are shot weights. It is unlikely that this strategy will yield a big improvement in the plan. When only “lock collimator settings” is checked, the planning system will optimize the position and weights of the shots and will weight shots that are not necessary to zero. This is perhaps the most useful setting for small to mid-size targets where the planner is seeking to optimally place a few shots. It is possible to use this function repeatedly at various stages of the plan as one places for instance different size collimators. When no boxes are checked, the optimizer will introduce composite shots, and while the global dose gradient may improve, the user should evaluate the result with the lower isodose lines displayed so as to prevent locally sloppy gradients in high-risk areas since the optimizer does not take into account avoidance structures. At this stage, no sectors (other than those that were manually blocked by the user) will be blocked by the optimizer. Allowing sector blocking enables that functionality. It is important to reiterate that sector blocking usually negatively impacts gradient – as does normalization to a higher isodose line.

Choosing optimization parameters is often a personal matter and depends on the manner in which the shots are placed by the user and the individualized planning goals for the case in question. Table 45.1 enumerates typical settings for various targets.

Table 45.1 Optimizer settings for various targets to obtain the best results with inverse planning with Gamma Plan

Inverse Planning

Inverse planning requires no a priori placement of shots and uses the fill function of the program before using the optimization techniques described above. The fill dialog (Fig. 45.14) has a few options. The user can choose to use composite collimators or simple collimators. By unchecking the composite box, the dialog changes to display the 3 collimators 16, 8, and 4 and allows the user to decide which collimators are likely to best suit the plan intent.

Fig. 45.14

The fill dialog

If composite collimators are allowed, then the software uses the size slider as a guide to choosing composite sectors. If set to small, then there are unlikely to be size 16 sectors in the placed shots. The program uses a bevy of preconfigured shot templates to fill, starting from the periphery and then populating the center. It initially places the shots without overlap of their 50% isovolume. The final setting is the gamma angle that can be chosen for patient docking based on lesion location and frame placement to avoid collisions (only usable with the stereotactic frame). Once shot filling is completed, the optimizer is used to optimize the dose plan as described in the previous section. Finally, manual placement of shots and or manipulation of the existing shots may be required to complete the process.

Dose Specification

Traditionally dose prescriptions with the LGK are defined as the minimum dose or dose to the margin of the target along with the percent IDL to which it is normalized. The corresponding maximum and mean doses while recorded are not used in publications when discussing efficacy and side effects. The exception to this pattern are functional targets where it is customary to prescribe the maximum dose to a point (given dose). Care must be taken not to confuse the two approaches to prescription.

Single Session: SRS

The doses used in single session radiosurgery are well worked out in LGK literature, and representative range is presented in Table 45.2. The outcomes of GKRS for vestibular schwannoma (Table 45.3), meningioma (Table 45.4), and metastatic tumors (Table 45.5) are provided for reference.

Table 45.2 Representative single session dose ranges by pathology

Table 45.3 Outcomes for vestibular schwannomas treated with single-fraction GKRS

Table 45.4 Tumor control rates for radiosurgery for meningiomas

Table 45.5 Outcomes for metastatic tumors treated with single session SRS with GK

Multisession SRS (Hypo-Fractionated SRT)

With the introduction of the LGK Icon, it has become more practical to perform multisession SRS (hypo-fractionated SRT), although frame-based hypo-fractionation has been performed both with the traditional frame and the Extend® frame .

Using the extend system, McTyre et al. [107] reported 34 cases where they used hypo-fractionation for benign tumors >10 cc in volume or abutting the optic pathway, vestibular schwannoma with the intent of hearing preservation, or a tumor previously irradiated with single-fraction GKRS.

The most challenging aspect of hypo-fractionation is developing an understanding of iso-effective doses with different fractionation schemes in the face of limited literature. One approach is to calculate the equivalent dose in 2 Gy fractions used in standard fractionation the EQD2. Martens et al. [108] reported a significant difference in median LC 14.9 months for EQD2 > 35 Gy and 3.4 months for EQD2 ≤ 35 Gy(p < 0.004). In order to allow the reader a quick tool to decide the number and size of fractional doses that will exceed this threshold , Fig. 45.15 is useful.

Fig. 45.15

Threshold EQD2 based on Martens et al. [108] for local control plotted against different fractionation schemes

In single session, the threshold EQD2 dose is exceeded by a dose of 17 Gy, but it takes 2 fractions of 10 Gy, 3 of 8 Gy, and so on to achieve the same EQD2. In more general terms, the most common doses used in treating metastatic tumors can be plotted against the EQD2 for various dose fractionation schemes as shown in Fig. 45.16.

Fig. 45.16

Single doses commonly used for treating brain metastasis (assume alpha beta ratio of 10) and fractionated schemes using EQD2. This chart is provided as a quick tool and should be used in conjunction with clinical judgment with the understanding that radiobiologic modeling is an imperfect science

Hypo-fractionated treatment of brain metastasis has been reported by many authors using a wide variety of fractionation schemes and varied success and complication rates. These are summarized in Table 45.6.

Table 45.6 Outcomes of hypo-fractionated treatment of brain metastases and/or surgical cavities

Treatment Delivery

Treatment delivery can follow one of several workflow patterns depending on the fixation used and the fractionation or single session model. It is also dependent on the model of LGK being used. For brevity, we will discuss workflows with the Perfexion and Icon units.

Frame-Based SRS

The first step in the delivery of frame-based SRS on the day of treatment is the application of the Leksell stereotactic frame to the patients’ head. This is performed as a clean or sterile procedure depending on the preference of the surgeon, under anxiolytics, or mild sedation in adults and general anesthesia in children. Local anesthetic is used at the four points on the scalp where the fixation screws penetrate the skin. Fixation is achieved with titanium screws (aluminum if only CT imaging will be used). Following frame application, a collision check cap is placed on the head to ensure that there is no risk of collision with the interior of the LGK in any desired target position.

In a linear workflow , this is followed by stereotactic MRI, CT, and catheter angiography as needed (Fig. 45.17). Since the frame limits the size of imaging coils that can be used with the MRI and the sequences that can be performed, sometimes it is preferred to perform the MRI imaging without a frame and co-registration with a frame-based CT (Perfexion), or onboard CBCT (Icon) is used to align the MRI into stereotactic space (Fig. 45.18).

Fig. 45.17

Treatment delivery in a frame-based linear workflow . Red ring indicates image that serves as source of stereotactic coordinates, and green ring is indicative of predelivery verification

Fig. 45.18

Treatment delivery in a pre-imaging/pre-planning frame-based workflow . Orange ring indicates images that are used for getting the anatomical data for dose planning, red ring indicates image that serves as source of stereotactic coordinates, and green ring is indicative of predelivery verification

Mask-Based SRS or Hypo-Fractionated SRT

Planning for mask-based treatments is performed on MRI or diagnostic CT imaging which is obtained with at least one sequence covering the whole head with a slice including air at the top and is designated as the pre-plan reference. More specialized sequences that delineate anatomy relevant to the plan can be used by co-registration with the pre-plan reference as long as the imaging covers at least a 50 mm thick slab of brain. Thermoplastic mask immobilization can be accomplished on the same day as treatment or prior to treatment. The stereotactic coordinates are obtained from the CBCT obtained at any point after the immobilization has been designed. These images are registered to the pre-plan allowing real Leksell coordinates to be acquired for the treatment plan. Prior to delivery of the actual treatment, several steps are required: (1) the HDMM (IFMM) nose marker has to be placed on the patient, (2) the HDMM camera has to be deployed, (3) the delivery fraction has to brought up on the console, and (4) one Gamma Plan station has to be in treatment mode. At this point a delivery CBCT is obtained and transferred to the Gamma Plan station in treatment mode. A co-registration window opens automatically and permits co-registration and verification of shifts. On accepting the shifts, a treatment evaluation window appears allowing the user to evaluate the influence of the repositioning on the dose plan. This can be viewed both for the current delivery as well as the cumulative delivery of all fractions delivered to that point in time. Since the shifts are applied, the comparison is made between the dynamic re-planned dose and the original plan. This workflow is illustrated in Fig. 45.19.

Fig. 45.19

Treatment delivery in a pre-imaging/pre-planning mask-based workflow . Orange ring indicates images that are used for getting the anatomical data for dose planning, red ring indicates image that serves as source of stereotactic coordinates, and green ring is indicative of predelivery verification

Quality Assurance and AAPM Task Group

Purpose

This section describes a procedure for investigating and verifying the precision of the dose delivery. Various factors may affect the dose distribution, such as the strength of each radiation source, the exact alignment to the collimator system, and the tolerances to which the collimators are manufactured.

Method

In Leksell Gamma Knife®, with very steep dose gradients and complex geometry, it is recommended to use film dosimetry because of good spatial resolution and low energy dependence. Due to the designs with a large number of sources (201 sources for Leksell Gamma Knife® B, C, 4, and 4C and 192 sources for Leksell Gamma Knife® Perfexion™ and Leksell Gamma Knife® Icon™), it is not possible to measure and investigate the beams from every single source. For Leksell Gamma Knife® B, C, 4, and 4C, the transmission through the collimator helmet would be too high, and more than 50 beams are required to have an excessive transmission of less than 1%. For Leksell Gamma Knife® Perfexion™ and Leksell Gamma Knife® Icon™, it is not possible to use only 1 beam at all, because they are designed with sectors of 24 sources each and individual sources cannot be blocked.

To investigate the precision in dose delivery, it is recommended to test the dose distributions from all beams in a standard geometry at the Leksell coordinate x,y,z = 100 mm for the various collimator sizes available on the treatment unit. The standard geometry is a sphere of 8 cm diameter. It is recommended to use the spherical phantom or the Elekta Dosimetry Phantom.

For each collimator size to be investigated:

• (a) Prepare two films to the appropriate size for phantom (and collimator size).

• (b) In the selected phantom type, mount the film in the center plane of the phantom in the XY plane.

• (c) Prepare a test plan with the coordinates X, Y, Z = 100 mm, and select an appropriate dose for the film type used (e.g., 5 Gy for Gafchromic EBT type film).

• (d) Expose the film to the selected dose.

• (e) Repeat steps 2–4 for the XZ plane.

• (f) Prepare eight films to the appropriate size for phantom (and collimator size).

• (g) Create a dose-intensity calibration curve.

Case Study

A 59-year-old female with known metastatic melanoma presents with short onset ataxia and mild headache. MRI of the brain reveals a hemorrhagic metastatic deposit in the R middle cerebellar peduncle. The lesion has a 9 mm solid tumor (0.3 cc) component and a 32 mm hemorrhage (7.9 cc).

Location precluded surgical removal. Given the radio-resistant nature of the primary radiosurgery would be a superior method for controlling disease. Since this was a single metastasis, whole brain radiotherapy would not be appropriate. However the volume of the hemorrhage would make single session dose to this area to be restricted to reduce side effects, compromising efficacy.

Leveraging the ability to hypo-fractionate with the ICON, the entire hemorrhagic cavity was treated to15 Gy in three fractions and the solid tumor received an additional boost dose of 5 Gy. Figure 45.20 (a) shows the original plan with the hemorrhage covered by the yellow (15 Gy) line and the solid tumor portion in turquoise that received additional 5 Gy. Follow-up imaging at 2 months (b), 4 months (c), and 1 year (d) reveals resolution of hematoma and shrinkage of tumor.

Fig. 45.20

(a–d) Metastatic melanoma with hemorrhage in the cerebellar peduncle treated with hypo-fractionated SRS showing progressive resolution of tumor and hemorrhage

Summary

Gamma Knife radiosurgery introduced the concept of radiosurgery in the CNS and has become an important tool in the management of CNS tumors in conjunction with or in lieu of microsurgery, fractionated radiotherapy, and chemotherapy. It provides the ability to deliver high doses of radiation with high precision and steep gradients for falloff in normal structures. Long-term results reflect high efficacy and low toxicity rates for the procedure. With the introduction of the ICON®, there are new indications and possibilities for the patients.

Self-Assessment Questions

• 1. Tolerance of the optic chiasm to single session SRS is commonly accepted to be:

  • A. 12–14 Gy

  • B. 2–4 Gy

  • C. 5 Gy

  • D. 8–10 Gy

• 2. Tolerance dose to the brain stem in single session SRS is

  • A. 50 Gy

  • B. 25 Gy

  • C. 12 Gy

  • D. 10 Gy

• 3. Gradient index refers to

  • A. Homogeneity of dose inside the target

  • B. Rapidity of falloff of dose in normal structures

  • C. Extent of target covered in adequate dose

  • D. Volume of target included in the prescription dose

• 4. Frame-based radiosurgery with Gamma Knife is capable of achieving a precision of

  • A. Less than 2.0 mm

  • B. Less than 1.0 mm

  • C. Less than 0.5 mm

  • D. Less than 0.1 mm

• 5. Dose normalization with the Gamma Knife is typically in the range of 40–70% because:

  • A. It provides more choices for treatment.

  • B. Minimizes treatment time.

  • C. Allows more coverage.

  • D. Permits optimal gradients in normal tissue.

Answers

• 1. D

• 2. C

• 3. B

• 4. C

• 5. D