WO2016191622A1 - Unconstrained radiosurgery with greatly improved dosage fall-off - Google Patents

Abstract

Off-target radiation delivery is a significant issue in radiosurgery. Provided herein are novel radiosurgury methods and hardware that greatly increase the number of different potential firing positions/angles from which beams can be delivered to a target. These inventions enable sharper dosage profiles which spare surrounding healthy tissues from supra-optimal radiation. In one aspect, the invention encompasses methods and devices which move the patient's head so that various beam angles are enabled. In another aspect, the invention encompasses radiosurgery systems which can emit higher numbers of beams and/or wherein the independent movement of individual beams emitters is enabled. In yet another aspect, the invention encompasses a system for image-guided delivery of beams to a patient's head, wherein the patient moves their head in a predetermined and simple movement pattern to enable delivery higher numbers of potential beams.

Description

Title: Unconstrained Radiosurgery with Greatly Improved Dosage Fall- Off

CROSS-REFERENCE TO RELATED APPLICATIONS: This application claims the benefit of priority to United States Provisional Application Serial Number 62/167,464 entitled "Unconstrained Radiosurgery with Greatly Improved Dosage Fall-Off," filed May 28, 2015, the contents which are hereby incorporated by reference.

STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT: Not applicable

BACKGROUND OF THE INVENTION

[0001] Various stereotactic radiosurgery (SRS) methods are known in the art, including gamma knife, proton beam, and linear accelerator treatment methodologies. These systems can deliver radiation to a target, for example a tumor located in the brain, which is not readily accessible to normal surgical methods. In a given treatment session, multiple beams, typically in the range of dozens to hundreds of beams, are delivered to the target from multiple angles, resulting in an isocenter at the target which receives a high dosage of radiation, capable of killing the tumor cells, while delivering less energy to healthy tissue surrounding the target.

[0002] While current methods of stereotactic radiosurgery SRS can effectively deliver sufficient dosages to kill cells at the target isocenter, the fall-off of the dosage is often suboptimal, meaning that higher than optimal radiation dosages are experienced by healthy tissues surrounding the target. Radiation necrosis is a serious side effect of SRS in which healthy tissues surrounding a treated tumor are inadvertently damaged or killed as a result of insufficient drop- off of delivered radiation. The effects of radiation necrosis include bleeding, pain, impaired cognitive abilities, and, rarely, can be fatal. The incidence of radiation necrosis varies among patient pools based on the type of tumor treated, tumor volume and location, and treatment modality. For example, in a study of patients treated for brain metastases, radiation necrosis was observed in 24% of treated lesions, with 10% of treated patients experience symptomatic necrosis (Minnitti et al., 2011, Stereotactic radiosurgery for brain metastases: analysis of outcome and risk of brain radionecrosis, Radiation Oncology 6:48). In another study of patients treated with radiosurgery for glioma, 5% of treated patients had radionecrosis (Ruben, et al., 2006, Cerebral radiation necrosis: incidence, outcomes, and risk factors with emphasis on radiation parameters and chemotherapy. Int J Radiat Oncol Biol Phys 65:499-508).

[0003] Accordingly, there remains a strong need in the art for systems that improve dosage fall-off in order to spare healthy tissue surrounding radiosurgery targets.

Summary of the Invention

[0004] Various methods and associated hardware devices for improving the degree of dosage fall-off in radiosurgery are provided herein. The basic objective of the several inventions described herein is to remove the constraints in beam delivery that limit the number of potential firing positions for beams. These constraints, inherent in the design of current radiosurgery systems, result in treatment solutions that deliver too much off-target radiation. By greatly increasing the number of different potential firing positions and angles that can be delivered to a target, sharper dosage profiles can be created and surrounding healthy tissues can be spared supra-optimal dosages of radiation. Accordingly, the invention comprises a set of solutions to the various constraints of the prior art. These solutions may be utilized individually or in combination to create an expanded set of potential beams. Thereafter, mathematical solutions are applied to the set of potential beams to create a treatment plan with a fall-off dosage profile that is superior to that attained with prior art methodologies.

[0005] In one aspect, the invention encompasses methods and devices which move the patient's head so that various beam angles are enabled. In another aspect, the methods of the invention encompass radiosurgery treatment methods with higher numbers of beams delivered than in currently known treatment techniques. In another aspect, the invention comprises improvements to currently used radiosurgery equipment which remove constraints on the potential number of beams that can be delivered at different angles and/or intensities. In yet another aspect, the methods and devices of the invention encompass a system for image-guided delivery of beams to a patient's head, including methods wherein the patient moves their head in a predetermined and simple movement pattern to enable high numbers of potential beams. Description of the Figures

Fig. 1. Fig. 1 depicts the 3-axis system referred to in the description.

Fig. 2. Fig. 2 depicts an exemplary work flow for a head touring treatment method.

Fig. 3A, 3B, and 3C. Fig. 3A, 3B, and 3C depict the 3-dimensional tumor mapping utilized to generate targets in a head touring treatment plan. A series of points defining the tumor in a plane are mapped in 3A. A 3-dimensional map of the tumor generated from the planar maps is depicted in Fig. 3B. An optimized head touring and treatment plan which will hit 250 separate spots defining the tumor is depicted in Fig. 3C.

Detailed Description of the Invention

[0006] The various embodiments of the description are described below. As used herein, a "treatment system" will refer to any system for the delivery of radiation to a target within a patient, such as brain tumor. A given treatment system will comprise one or more radiation beam-emitting elements. For example, a standard prior art gamma knife treatment system may comprise 192 individual beam emitting elements arranged in a hemispherical array. The beam emitting elements are typically present in a single unit, such as a collimator helmet in a gamma knife system or a gantry in a LIN AC system.

[0007] The treatment systems are used in the performance of a treatment session. In a treatment session, beams are emitted at one or more targets. The target, as used herein, is an isocenter through which multiple beams pass. The beams may pass through the isocenter simultaneously, such that a high dosage of ionizing radiation is delivered in one instance. Alternatively, the beams in a treatment session may be delivered

sequentially, such that a cumulative high dosage of radiation is delivered to the target over the duration of the treatment session. In keeping with the principles of radiosurgery, each individual beam will have a sufficiently low deposited energy that such individual beam will cause little or no damage to cells encountered by it and cell killing occurs only at the convergence of many beams. [0008] Radiosurgury is typically directed at tumors and other neoplasms of the central nervous system. The target may be any tissue within a living creature, for example a tumor or a portion of a tumor in a human patient. Exemplary tumors include those of the brain, spine, and other organs. Exemplary brain tumors include acoustic neuroma, astrocytoma, gliomas, meduUoblastoma, metastatic brain tumors, schwannomas, and other types of brain tumors. Large tumors or tumors having complex shapes are typically divided into multiple targets. In a given treatment session, one or more targets covering the tumor volume may be irradiated. In some cases, especially in brain metastases, multiple tumors are treated simultaneously.

[0009] For each beam-emitting element of a given treatment system, one or more potential beam-target angles is possible. The potential beam-target angle refers to an angle between the beam emitter and a selected target that is enabled and available for use in a treatment session. In some systems, for example, in a traditional gamma knife system, a single beam-target angle is typically enabled for each emitter. In a system such as a standard LINAC system, wherein a moving gantry positions the beam emitter along an arc over the patient, multiple co-planar beam-target angles are available around the arc.

[0010] A given beam emitter system may also comprise a series of potential beam intensities. The beam intensities can be modulated by changing the beam diameter and/or the on-time of the emitted beam. In the traditional gamma knife system, only a single beam intensity was enabled. In newer treatment systems such as the LINAC system, modulation of beam intensities is made possible by moving partitions ("leaves") which can partially occlude collimator channels.

[0011] For ease of reference, the description provided herein will make reference to a 3-axis 3-dimensional coordinate system with respect to a human patient, as depicted in Fig. 1. The long axis of the patient's body, oriented from the top of the head down to the patient's feet, will be defined as the Z axis. The X axis will be defined as a line drawn transversally along the longitudinal axis of the patient, oriented from left to right. The Y axis is defined as a line drawn from the anterior to the posterior of the patient, for example from the tip of the nose to the back of the head. [0012] Patient Positioning Devices that Expand the Number of Potential Beam-Target Angles. A major constraint in radiosurgery dosage planning is the number of potential angles at which beams can be delivered to a target. For example, in the case of a standard gamma knife collimator, the beam sources, for example 192 beam emitters, are in a fixed configuration with respect to one another. Some prior art systems allow for changes in the head tilt angle of the patient, for example the Leksell PERFEXION(TM) system allows for three preset tilt-angles of 70, 90, and 110 degrees. In a standard LINAC system, the gantry moves in a single arc around the Z axis of the patient, and the patient may be rotated somewhat around his or her Y axis in order to effect a greater number of potential shot angles.

[0013] The patient positioning systems of the present invention greatly expand on the number of angles from which the beams may potentially be delivered to a target. The patient movement systems of the present invention may comprise two elements. In a first embodiment, the patient movement system comprises a headrest that gently moves the patient's head in different directions. The human neck is a remarkable joint, and an ambulatory person of normal flexibility can typically rotate their head over 100 degrees around their Z axis, at least 60 degrees around their X axis, and at least 60 degrees around their Y axis. In one embodiment, the moving headrest of the invention can rotate the patient's head to some tolerable amount around their Z axis, for example plus or minus 1 to 50 degrees around the Z axis. In one embodiment, the moving headrest of the invention can rotate the patient's head to some tolerable amount around their Y axis, for example plus or minus 1 to 30 degrees around the Y axis. In another embodiment, the moving headrest of the invention can rotate the patient' s head to some tolerable amount around their X axis, for example plus or minus between 1 and 30 degrees around the X axis. Movement along two axes may be accomplished as well. Such systems may be used to create any number of tilt angles, greatly expanding the number of potential angles which each emitter can hit a target. The patient' s head can be secured in the headrest by any number of structures, including straps, mouthpieces, etc. and movement of the head can be effected by any number of actuators including padded structures that gently push, lift, or otherwise tilt the patient' s head, such actuators being under the control of a computer control system. [0014] The patient's head may also be moved by the use of a moving "couch" or "bed," comprising an assembly on which the patient may lie or recline and on which the patient may be comfortably immobilized. The bed is motorized such that the entire patient (or a portion of the patient, such as the upper body) is moved by the use of motors and other actuators. In such a system the patient is held securely, for example with a headrest, leg rests, arm rests, body straps, and other devices to hold the patient comfortably but firmly to the bed so that he or she can be raised, lowered, rotated, tilted, rolled and otherwise positioned in three dimensions with respect to the radiation sources. Such a moving bed can optionally be used in combination with a moving headrest, as described above, to increase the degrees of freedom of movement of the patient.

[0015] The patient positioning systems described herein are enabled by the use of precision motorized systems and actuators that allow fine tuning of patient position to a very low margin of error. Such systems are also enabled by the use of computerized control systems, as known in the art. Patient movement can be paused at specific orientations for administration of beams, or the movement may be continuous with beam delivery occurring throughout the motion. Head position can optionally be monitored and confirmed during/after movement of the headrest or moving couch by the use of 3- dimensional mapping means (i.e. camera plus processor plus software) which detect landmarks such as facial features, anatomical landmarks, or ink markings drawn on the patient. Target position is then derived based on its known position relative to the landmarks, to confirm that the target is correctly positioned before and/or during the time each beam or cluster of beams is fired.

[0016] Modified Gamma Knife System. The gamma knife system, pioneered by Lars Leksell has been in use for decades. The basic system comprises a dome of tungsten or like metal, having about 200 channels which act as collimators. Radioactive sources, typically cobalt-60, are present on upper side of the dome and can be moved into position above the channels, which collimate gamma rays emitted by the cobalt, resulting in beams that converge on the target below the dome. The patient's head is positioned at the isocenter of the beams by a special helmet which is fitted with the dome for precise targeting of the isocenter on the tumor or other target. [0017] The newest incarnation of the gamma knife system is the Leksell

PERFEXION(TM) system. In this system, radioactive sources can be moved from a blocked position (wherein no radiation reaches the patient) to a position over a collimator channel, directing gamma rays through the channel at the target. By moving the sources to different positions, various channels of different diameters can be engaged to control beam diameter, giving more potential beam intensities for improved planning. Eight

independent sources, each serving about 24 different channels can be operated

independently of each other, providing additional flexibility in planning. The

PERFEXION(TM) system represents an increase in the number of potential beam-target angles and intensities compared to previous gamma knife systems.

[0018] In one aspect, the scope of the invention comprises an improvement to the current gamma knife systems by increasing the number of independently moveable sources. The number independently moveable sources may, for example, be 10 or greater, 20 or greater, 40 or greater, 60 or greater, 100 or greater, 200 or greater, etc. Each independently moveable source can, by mechanical actuators (e.g. motors under control of a computerized firing system) be moved from a blocked position (wherein no radiation is emitted from the collimator dome) to an "on" emitting position over or near one or more channels. Channel diameters may be, for example, 1-20 mm in diameter. The increase in the number of independently moveable sources over current systems greatly increases the flexibility of the system by allowing different combinations of beams from different positions of the collimator dome to be pulsed independently of each other. This combinatorial increase in the number of potential is centers that can be generated greatly increases the ability to plan treatments with high dosage drop-off around the target.

[0019] Multi-Beam LINAC. Linear particle accelerator treatment systems emit high energy photons or electrons which damage target cells. In photon based systems, microwaves are used to bombard a heavy metal such as tungsten, resulting in the guided emission of X-rays (high energy photons) from the emitter. In electron based systems, high energy RF waves are used to generate guided electron beams from the emitter.

[0020] LINAC systems advantageously can utilized moveable shutters or "leaves" to change the width of the emitted beam, in what is called intensity modulated radiation therapy (IMRT) regimes. [0021] In current LIN AC methodologies, a single beam can be delivered at each time point from an emitter called the gantry. The gantry moves in an arc over the patient. The couch on which the patient is positioned under the gantry can also be moved to enable a greater number of potential beam-target angles. In one aspect, the invention

encompasses novel radiosurgery hardware devices wherein the number of LINAC beams sources is increased over that found in the prior art.

[0022] For example, in one embodiment, the invention comprises a LINAC system having greater than one beam source, for example, having 2-50 beam emitters, for example, 5, 10,15, 20, 40 or more beam emitters. Because of the bulk and weight of the hardware elements of each beam source, in such multi-beam configurations, the beam emitters are arranged in a static device that surrounds the patient, for example in a hemispherical or arc arrangement. In this embodiment, rather than the single beam source affixed to a moving gantry found in current LINAC systems, multiple LINAC beam sources are present in a static configuration (or a configuration having limited some ability to move) and the patient is moved, for example as by the patient positioning systems described above.

[0023] Image Guided Radiosurgery. In another aspect, the scope of the invention encompasses novel methods of radiosurgery wherein the patient's body or target region of the body (e.g.head) is not immobilized, and instead the patient's movements are utilized to effect multiple potential beam-target angles that enable high treatment sessions with steep drop off of radiation dosage. This novel treatment method will be referred to as a "head touring" treatment, although it will be understood that movement and treatment of other portions of the body are within the scope of the invention.

[0024] The head touring radiosurgery methods of the invention encompass several elements. A first element is a planning element wherein the size, shape, and location of the tumor are assessed, as in standard radiosurgery planning. A computerized system then maps a number of targets or is centers to completely cover the volume and exterior of the tumor. For example, the tumor can be mapped as a three dimensional object comprising one or more voxels or spatial units, for example 0.1 mm units. [0025] A planning step is then executed using computer programs which take into account (1) the tumor position; (2) the potential beam angles that can be effected by the selected treatment system; and (3) potential motion patterns of the patient's body, for example, the patient's head. Also taken into account, as in standard planning steps, is the required dosage for the effective treatment of the particular tumor or other target.

Exemplary motion patterns include motions such as the patient tilting their head up, tilting their head down, turning to the left, turning to the right, etc. Taking into account these three variables, an optimized motion pattern or head touring sequence, and an associated series of shots (multiple simultaneous beams) or single beams is planned. Any appropriate software methods can be used to determine the head touring pattern and associated series of shots, including travelling salesman algorithms and like solutions such as greedy search algorithms, MINOS software, and other optimization tools known in the art.

[0026] In a next step, the patient is trained to perform the head touring motion generated by the planning system. For example, the patient may be trained to turn their head 10-45 degrees to the left, then 10-45 degrees to the right, then back to a centered position, followed by tilting the head up 10-45 degrees and then down 10-45 degrees. The patient is trained to follow the touring pattern at a speed which will enable the shots to be efficiently and accurately delivered, for example moving the head at a speed of 1 to 45 degrees of rotation per second.

[0027] In a next step the patient is placed in position with a treatment system. For example, in one embodiment, the patient is seated or reclining with their head under the dome of a gamma knife system. In an alternative embodiment, the patient is seated or reclining on a couch beneath a LINAC gantry or multiple LINAC beam emitter. An element of the system is a real time mapping system, comprising elements which can accurately map the patient's head position, and thus, the position of the one or more targets to be treated. The patient then moves their head in the predetermined motion pattern and real time computational tools assess the position of the target or targets in the patient's head and beams are delivered to targets as enabled by the transit of the target through potential beam-target angles. The motion pattern is repeated until all beams required to fulfill the treatment plan have been delivered. [0028] The goal of the system is to enable comfortable and rapid treatment wherein an enormous number of potential beam-target angles are enabled by the motion of the patient's head, exposing targets to attack from beam emitters at numerous positions, enabling optimized treatment sessions with sharp dosage fall off at the target periphery.

[0029] Short firing on times, in the range of 10- 100 milliseconds will generally be necessary to keep beams accurately focused on the target, which is in motion. Such on- times can be enabled by fast-action actuation mechanisms for radioactive sources in gamma knife systems. Current gamma knife radioactive sources are housed on a movable plate that can shift via a linear encoder to align with pre-drilled apertures on a piece of heavy tungsten. Employing a faster encoder, faster motor, or shortening the distance that a source travels would significantly shorten the firing on time by several folds. For LIN AC system, the firing on time is pulsed and digitally controlled to be sufficiently fast for the target motion speed enable by the skull movements.

[0030] An exemplary workflow for the head touring treatment methods that produces the best dosage that a user could freely select in real-time of the invention is depicted in Fig. 2.

[0031] In current systems, mapping of patient or target position is used as a safety feature. For example, during a specialized treatment in the Gamma Knife Icon system where the patient's skull is secured via a head mask, wherein detection of the patient's head being out of position results in halting of treatment until desired position is reestablished. The novel system of the current invention takes advantage of advances in computational power and speed which now enable accurate targeting of beams to a moving target, contrary to previous systems where head movement was undesirable because systems could not adjust in real time to movement of the target.

[0032] Exemplary systems for detecting current target position may encompass any system for mapping the 3D position of a target object. An exemplary system would be an imaging means (camera) and image processing means (processor and software) that detects facial position by detection and tracking of facial landmarks, followed by a mapping step which indicates the current location of the target, based on its known position relative to facial landmarks. Software which predicts the location of a moving 3- dimensional object in advance may be employed to keep beams firing on target as the patient moves his or her head in the motion pattern. Exemplary systems for facial tracking and 3-dimensional mapping of objects includes those systems and methods described in United States Patent Publication Number 20150268058, by Samarasekera et al., entitled "Real Time System for Multimodal 3-D Geospatial Mapping, Object Recognition, Scene Annotation, and analytics; United States Patent Number 9,268,993, by Wus et al., entitled "Real Time Face Detection Using Combinations of Global and Local Features,"

ALIGNRT(TM) surface guided radiation systems (by VisionRT), and other systems known in the art.

[0033] Treatment Planning. As described above, the invention encompasses various means of expanding the number of potential beams beyond that enabled by the prior art. For a given system, wherein one or more improvements of the invention has been applied to expand the number of potential beams, the set of potential beams is input to a planning algorithm or software package in order to generate an optimal treatment solution. The optimal treatment solution will be that which creates the sharpest dosage fall-off between the target and the surrounding healthy tissue, within a maximum treatment duration limit. The dosage fall-off is generally quantified by the gradient index (GI) which is defined as ratio of the peripheral isodose volume such as 50% of the prescription isodose volume and 100% of the prescription isodose volume. Typical GI value for the current radiosurgical delivery is estimated to be approximately 3.0 and an optimized solutions of the invention will improve this index by 15% or more representing additional sparing of the entire normal tissue by tens to hundreds cc in volume. The maximum treatment duration limit is the maximum time period deemed acceptable for the treatment, for example 15 to 20 minutes may be set as the maximum treatment time. The treatment plan will be generated taking into account the speed at which the beam emitter unit and/or patient can move to the various positions dictated by the treatment plan and the on-time at which the emitter unit is fired at a given position.

[0034] The methods and novel hardware devices described herein, by creating a sharp delineation between treated and normal tissue, are especially amenable to targeting the exterior surface of a growing tumor, which is potentially the area of the tumor experiencing the highest rates of growth. The methods described herein can be described as "painting" the tumor by delivering a series of is centers that cover the periphery of the target structure. The sharp delineations also allow for precise targeting of blood supply sites to the tumor.

[0035] Any treatment plan algorithm, software, or mathematical solution known in the art may be utilized, so long as it is able to take into account the relevant variables of a given treatment system. Exemplary treatment planning methods include the use of constrained optimization algorithms, Monte Carlo simulated annealing algorithms, genetic algorithms and swarm particle algorithms, as known in the art. Current commercial treatment planning systems employ various types of these algorithms for planning Gamma Knife or LINAC -based treatments (cf J Neurosurg. 2010 Dec; 113 Suppl: 199-206.;

Pubmed PMID: 26676060, Med Phys. 2015 Mar;42(3): 1367-77. doi: 10, 1 1.18/1.4908224. etc).

[0036] In one embodiment, the resulting treatment plan may be a plan that encompasses the use of many more emitted beams than utilized in standard treatment plans. For example, standard treatment plans typically encompass 1-15 for LINAC based treatments or 100-200 beams per target or treatment center for Gamma knife based treatments. Utilizing the methods of the invention, treatment plans comprising 500- 50,000 emitted beams may be implemented. Exemplary treatment plans include plans having greater than 500 beams, greater than 1,000 beams, greater than 2,000 beams, greater than 5,000 beams, or greater than 20,000 beams may be used.

[0037] The methods and novel hardware devices of the invention may be applied using any radiosurgery modality, including X-rays, charged particles, protons, and other energy sources known in the art. The methods and novel hardware devices of the invention may be applied in any radiosurgery context. While the foregoing description has placed special emphasis on systems for delivery of beams to the head, it will be understood that the various embodiments of the invention may be applied in the treatment of any organs or tissues of the body, including the head, breast, prostate or others. The methods and novel hardware devices of the invention may be applied in the treatment of any condition, including cancers and other neoplastic conditions and the treatment of arteriovenous malformations. The methods and novel hardware devices of the invention may be utilized in the treatment of human patients as well as in the treatment of other animal species in veterinary medicine or research contexts. [0038] Examples. Included herein are Examples which describe various implementations of the invention that greatly increase the number of potential beams, allowing for the generation of treatment solutions that have improved fall-off relative to standard treatment methodologies. All methods and devices described in the following examples demonstrate superior dosage fall-off compared to prior art treatment systems.

[0039] EXAMPLE 1. Creating a large number of focused beams with variable patient head tilt angle to improve dose fall-off near a target and reduce damage to peripheral normal tissue for brain radiosurgery

[0040] In this retrospective study, our aim was to investigate a novel treatment delivery and planning strategy by increasing the number of beams to minimize dose to brain tissue surrounding a target, while maximizing dose coverage to the target. By multiplying the number of beams by a certain factor, and delivering these beams at variable patient-head tilt angles we hoped to improve dose fall-off. In analyzing 12 different treatment plans via Leksell Gamma Knife, prior treatment plans were simulated to increase the number of beams from the original treatment plan by varying head-tilt angles, while maintaining original isocenter and beam positions in the x-, y-, and z-axes, collimator size and beam blocking. Optimized treatment plans were compared

dosimetrically with original treatment plans. When comparing total normal tissue isodose volumes between original and optimized plans, the low-level percentage isodose volumes decreased in all plans, with an average of 5.03 ± 3.22% decrease. Despite the addition of multiple beams up to a factor of 25, beam-on times for 1 tilt angle versus 3 or more tilt angles were comparable (<1 min.). The addition of more tilt angles correlated to a greater decrease in normal brain irradiated volume, and shows improvement for dose fall-off for brain surgery. The study demonstrates methods to decrease target dose fall-off and increase brain tissue sparing. Variations in volume decrease may be related to shape or location of tumor, but further investigation is needed in this area.

[0041] It was hypothesized that dose fall-off near a target may be enhanced via delivering beams at variable patient-head tilt angles. By increasing the number of beams in the treatment plan, the aim is to add potential beams through variable angles in order to minimize dosage of each individual beam while keeping total radiation dosage the same. Twelve randomly chosen patient treatment plans from plans within the UCSF medical center database were examined and modified to incorporate the aforementioned changes. Improvement was measured by the decrease and minimization of peripheral isodose volume fall-off. In the current treatment paradigm, the total number of beams is physically limited by the maximum number of apertures that can be drilled into a metal collimator.

[0042] Increasing the number of beams while also increasing the tilt-angles of delivery was found to influence peripheral isodose volume significantly in some cases, with up to an 11.4% decrease in prescription isodose volume. On average, total average volume percentage decrease in volume was 5.03 ± 3.22%.

[0043] Beam-on treatment time was not significantly changed with decreasing peripheral isodose volume in each of the twelve plans. Therefore, creating a large number of focused beams with variable patient head-tilt angle shows promise in improving dose fall-off near a target and reducing damage to peripheral normal tissue for brain

radiosurgery.

[0044] Twelve random patient treatment plans via the Leksell PERFEXION™ system (PFX) were chosen, and for each, plans were generated as well as with 4C, the previous version of Leksell Gamma Knife planning system. For standardization, only cases with a single tumor treatment were chosen. Nine cases had diagnoses of acoustic schwannoma, and three were meningiomas.

[0045] The reason for using two different treatment-planning systems is due to certain limitations in each. PFX allows for an unlimited number of beams to be added by the user, but only allows for three pre-set tilt-angles of 70, 90, and 110 degrees due to the design of the couch. Leksell 4C, on the other hand, allows the user to input a user-defined patient tilt-angle, but has a maximum beam number input of 50.

[0046] In all cases, the original treatment plan was compared with two "optimized" plans created in PFX and 4C respectively. The term optimized is used to designate plans in which the number of beams are increased in treatment plans by varying tilt angles of the patient head, while maintaining original beam positions in the x-, y- and z-axes, collimator size, and beam blocking. Newly created treatment plans were compared dosimetrically with the original treatment plan, and beam-on times for original and optimized plans were recorded.

[0047] The method for optimizing these plans was to increase the number of beams by a factor of 3, thereby increasing the total number of beams to 39. The x-, y-, z- coordinates, weight, collimator size, and beam blocking positions for each of the first 13 beams are noted, and added to the first 13 beams two times. The only change between the first, second, and third set of 13 beams are the delivery tilt-angle degree. In 4C, the first set of 13 beams are delivered at 30 degrees, the second set of 13 beams are delivered at 90 degrees, and the third set of 13 beams are delivered at 150 degrees. Because 4C has a limitation of 50 beams per treatment plan, a fourth set was not added. In PFX, the three sets of beams are delivered at the three preset angles of 70, 90, and 110.

[0048] Data from original and optimized plans were collected from the dose volume histogram files that can be saved from within the Leksell Gamma Knife program. Each extracted dose- volume-histogram file provides information about the target bin center dose in gy, the percentage volume of the target receiving the dose, as well as the volume (mm ). To determine the average decrease in volume percentage from the original treatment plan beam volume, a percentage change was first calculated from the optimized plan volume and the original treatment plan volume. Percent changes from 0% to 50% prescription isodose were averaged. The majority of cases were prescribed at 12.5gy at 50% isodose volume.

[0049] Comparing total normal tissue isodose volumes between original and optimized plans in the twelve selected treatment plans, the low-level percentage isodose volumes decreased, on average, in 100% of the plans when adding beams at varied patient head tilt angles in 4C and in PFX. The total normal tissue volume getting any radiation at all is decreased in >1 tilt angle, compared to 1 tilt angle. In 4C, the average percent volume decrease ranged from -0.35% to -11.41%, with a total average percentage volume change of -5.03 ± 3.22%, and in PFX, the average percent volume decrease ranged from - 0.31% to -2.66%, with a total average percentage volume change of -1.40 ± 0.58%.

[0050] The higher decrease in average isodose volume in 4C compared to PFX was due to the limitation in adding additional patient head tilt angles in PFX, at angles 70, 90, and 100 deg. The system in 4C does not have such a limitation, and allows for adding beam coordinates at angles ranging between 30 and 150 deg. Because of this, 4C allows for beam coordinates to be added in factors up to 25, while PFX only allowed for beams to be added in factors of 2 or 3.

[0051] Decreasing isodose volume percentage in the optimized treatment plans was observed. The data indicated that when more beams are added to treatment plans, isodose volume is decreased. In one plan, as dosage fall-off level drops from 50% to 10%, the isodose volume is decreased in increasing amounts as the number of beams is increased from 9 to 45. In another, it was observed that increasing the number of beams from 5 to any range between 10 and 35 beams will actually cause the prescription isodose volumes between 10 to 30% to increase. However, when the beam number is increased from 5 to 45, at all isodose levels at the fall-off, the volume decreases. Similarly, another plan showed that increasing the number of beams from 12 to 48 does not improve dose fall-off as much as increasing the number of beams from 12 to 36. This may be an indication that while increasing the number of beams often improves dose fall-off, there may be an increase threshold that when reached, further addition of beams does not improve dose-fall off, and decreases the effectiveness of avoiding peripheral tissue radiation. Several other factors that influence this optimum likely involved.

[0052] In 92% of the plans (11 out of 12), the addition of more shots at multiple angles led to a decrease in volume. In cases where the original beam number is increased by the same factor in 4C and PFX, the difference lies in the variability in patient head-tilt angles.

[0053] The beam on time differences (min.) between one tilt angle versus three or more tilt angles are minimal. In 4C, the treatment time for optimized plans decreased on average by 0.34 + 2.00 min., and in PFX, the treatment time for optimized plans decreased on average by 0.07 ± 0.76 min. Results indicate that this method is feasible without altering treatment beam-on times by a significant amount.

[0054] The results demonstrate that adding an increased number of additional focused beams with variable patient head tilt yields improvement for dose fall-off for brain radiosurgery. The study demonstrates technical feasibility of adding beams to decrease target volume and increase brain tissue sparing without increasing treatment time significantly. Although the PFX system was developed after 4C, it shows limitations in varying patient head-tilt angle to user-set values. Incorporating the option of varying angles more than just 70, 90, 110 in the PFX couch system will allow for 4C levels of isodose volume decrease in PFX. Variations in volume decrease may be related to the shape or location of the tumor.

[0055] EXAMPLE 2. Minimizing Spillage Dose Via a Broad-Range Optimization Approach of Hundreds of Beams for Treating Multiple Brain Metastases

[0056] Large variable normal brain dose and inter-target dose interplay effects have been reported for volumetric modulated arc therapy (VMAT) of multiple brain metastases. In order to maximize normal brain sparing for such a treatment, a Broad- Range Optimization of Modulated Beam Approach (BROOMBA) has been developed for this study. In this approach, hundreds of intensity-modulated beams surrounding the central axis of the skull were applied. BROOMBA first selects and then optimizes orientations and intensity levels of these beams while simultaneously optimizing dose to all the targets within the brain. To demonstrate its feasibility and dosimetric strength, treatment planning via BROOMBA for a multi-institutional benchmark case was carried out and its results were compared with clinical VMAT treatment plans. As a result, BROOMBA was found to outperform multi-arc VMAT plans in terms of significant reduction in the low-level background dose as well as the inter-target dose interplay effects. For example, when planning for 12 brain metastases, BROOMBA lowered the ambient background dose by approximately 110% and the inter-target dose interplay effects were reduced to be negligible across the 8-Gy to 12-Gy isodose levels. In summary, a multi-target BROOMBA has been developed and demonstrated to be a potentially powerful approach for multiple brain metastases treatments via high-output linac -based deliveries.

[0057] Recent advancements in early diagnosis and systematic therapy have made multiple brain metastases a leading indication currently managed by stereotactic radiosurgery and radiotherapy. The role of focal radiation therapy in managing limited number (N < 3) of brain metastases as well as a relatively large number of brain metastasis (N>3) is rapidly evolving. Recent meta-analysis of three randomized clinical trial has even pointed to quality-of-life and survival benefits of focal radiation therapy for younger and high performing patients. In addition, technical advancements such as the introduction of volumetric modulated arc therapy (VMAT) delivery via a high output, flattening-filter- free digital linear accelerators have also enabled rapid treatments of N>3 brain metastases within a conventional treatment delivery window of 15-20 minutes.

[0058] However, several recent studies have noted a significantly increased low- level of normal tissue spillage dose associated with the reported VMAT delivery techniques for treating multiple brain metastases lesions. Further investigation of such effects via a multi-institutional benchmark study has pointed to the dose interplay effects (i.e. dose delivered to one target affecting the dose to another target and vice versa) as a major contributive factor for an increased low-level normal brain dose. By definition, such dose interplay effect only occurs for multi-target treatments and the effect is found in the study to increase rapidly with increasing number of targets. [0059] For this reason, conventional treatment planning approaches such as using pre-arranged beam configurations via equally spaced axial/parasagittal arcs or a limited number (N<10) of fixed beams are unlikely to produce an optimal solution for individual patient cases since multiple brain metastases can be randomly located throughout the brain parenchyma. Therefore, the goal of this study is to develop a new approach to overcome such a challenge. Specifically, we aimed to develop an effective approach that significantly reduces the normal brain spillage dose and the dose-interplay effects for focal radiation treatments of multiple brain metastases.

[0060] For the study, we first developed a Broad Range Optimization Of Modulated Beam Approach (BROOMBA), where two basic principles are applied in planning multiple targets: (1) expand the total number of beams by orders of magnitude surrounding the central skull axis (2) simultaneous optimizing individual beam orientations and intensity levels for all the targets under consideration. As the first proof- of-concept study, the BROOMBA has been implemented on a standalone workstation and its planning results compared with those of clinical VMAT treatment plans for a benchmark case that has previously been published in a multi-institutional study.

[0061] The data set for the benchmark case was created from an actual patient case previously treated with stereotactic radiosurgery. The data set consisted of CT and MR images, and DicomRT contours of 12 brain tumors distributed inside the brain as described in the previous study. In short, the largest tumor measured approximately 1.0 cm in diameter and smallest 0.3 cm in diameter with a mean target volume 0.45+0.34 mL. Similar to the benchmark study, different target combinations of 3 to 12 targets were adopted for BROOMBA and VMAT treatment planning comparisons.

[0062] For BROOMBA treatment planning, one thousand one hundred sixty-two non-coplanar beams were first placed surrounding the central axis of the skull with 6 degrees of separation between two adjacent beams. All the beams were isocentric with the isocenter placed at the center of mass of all the targets under consideration. BROOMBA first eliminated all the beams that are physically inaccessible or constrained by the hardware such as those aiming superiorly close to the central axis of the skull. The remaining beams were subdivided into 2 mm x 2 mm beamlets, and the dose distribution matrices of each beamlet were calculated using a previously published collapsed-cone convolution algorithm with 6-MV x-ray polyenergetic kernels. The dose calculation was matched to 6-MV machine commissioning data. The dose calculation resolution was 2 mm as previously published(Dong et ah , 2013). [0063] The BROOMBA optimization routine is given as follow: we here denote D i_j as the dose delivered to a voxel j from beamlet i t N bin beam b e B. F(z) is the objective function for which the optimization problem is formulated as follows:

minimize F(z) subject to Equation 1 :

Equation 1 : dose for voxel j: z_j =∑_hzB∑_lZBh D_hljx_hl

where Xbi is the beamlet intensity that needs to be optimized.

[0064] A greedy algorithm was used to determine the beam orientation while explicitly taking into account the treatment plan quality. The optimization started from an empty solution set, and for each iteration, a new beam from the remainder of the candidate beam pool was added to the selected beam set for solving the free modulation

optimization (FMO) problem, i .e., the problem of determining the optimal beamlet intensity levels for the fixed beam angles. The iterative process continued until the desired number of beams was reached or the objective function plateaued. To select the new beam, solving the FMO problem with all potential beam candidates and choosing one beam that had the lowest objective function value possible, the computation time would have been clinically impractical. Instead, the benefit of adding a beam was predicted rather than explicitly computed. The first-order information also known as the shadow price in constrained optimization was used to select the new beam. The shadow price is the instantaneous change, per unit of the constraint, in the objective value of the optimal solution of an optimization problem obtained by relaxing the constraint. In our problem, the constraint is Equation 1. Each new beam will add values to those constraints. The beam with largest shadow price was selected.

[0065] We used an objective function F(z) that is based on a linear approximation of an equivalent uniform dose:

Equation 2: 6 (z) = y_smeaft(z ) + (1— x_s)m x(z_/) for OARs

Equation 3 : G_r(z) = Y_rmean(z^) + (1 — y_s)min(z_/) for P'TVr

Equation 4: F(z) =∑_mg.s. _f. a_mG_m{z) where a_m < 0 for PTV and > 0 for OARs. The weights among multi-objectives cc_m were fine-tuned to reach individual planning objectives. The assignment of a voxel that that lay within multiple OARs was given to the OAR with greatest optimization priority, which was manually determined. The ratio between the 50% of the prescription isodose volume and the prescription isodose volume (i.e., R50 ) was included in the objective function to explicitly optimize this important parameter. A ring structure was created outside the PTV to assist the minimization of R50 by minimizing the mean dose to voxels with doses greater than 50% of the prescription dose. CPLEX (Academic Research Edition 12.2) are used to solve the final linear optimization problem.

[0066] To compare with BROOMBA results, the VMAT treatment plans for the benchmark cases were developed as follows: The VMAT treatment plans were developed using a published technique on a commercial linear accelerator (Truebeam, Varian Oncology, Palo Alto). To compare with BROOMBA results, both coplanar and non- coplanar 6 MV flattening-filter-free beams were applied for VMAT treatment planning. For coplanar treatment planning, one transverse arc spanned 358° at the couch angle of 0° was used. For non-coplanar treatment planning, the above transverse coplanar arc and 2 parasagittal arcs were employed. The non-coplanar parasagittal arcs were both 179.9° arcs with the couch angle of +30°, respectively. To minimize the digitization effect of finite leave width for the multileaf collimator, the collimator was also rotated either 30° or 45° at non-zero couch angles. All the clinical treatment plans were optimized on a clinical treatment planning platform (Eclipse Progressive Resolution Optimizer Version 11.0, Varian Oncology, Palo Alto). The final treatment planning results of BROOMBA and VMAT were all exported via DicomRT protocol into the same dose analysis workstation for comparisons (Mim Vista, Cleveland). For consistent analysis, all the treatment plans were normalized such that a single fractional dose of 20 Gy was prescribed to cover at least 99% of each individual target volume.

[0067] The optimized BROOMBA beam orientation did not exhibit a trend of conforming to the beam path of the clinical VMAT technique. This is expected as the BROOMBA results were derived from patient- specific target locations while VMAT were fixed beam path configurations as discussed above.

[0068] For all the treatment plans, BROOMBA and VMAT produced satisfying dose coverage of individual targets (V100 > 99%) and achieved equivalent (p> 0.90 ; paired two-tailed t-test) Paddick dose conformity indices with the mean values of

0.84+0.04, 0.84+0.04, 0.80+0.04, 0.78+0.04 for N=3, 6, 9, and 12 target treatment plans respectively. However, when comparing the dose spillage to the normal brain, large discrepancy was noted between BROOMBA and VMAT results, where peripheral normal brain volumes enclosed by various isodose surfaces such as from 16 Gy (i.e. 80% of the prescription dose) to 4 Gy (20% of the prescription dose) were observed.

[0069] BROOMBA with 20 non-coplanar beams produced the lowest peripheral normal brain dose for all the cases. The improvements at the 8 Gy and 12 Gy isodose levels were particularly noteworthy with increasing number of targets. For example, when compared with the VMAT results using 6 MV beams, the 12-Gy normal brain isodose volumes were consistently lowered by 32.8%, 34.4%, 47.6% and 65.6% for N=3, 6, 9, and 12 targets respectively. Similarly, the 8-Gy normal brain isodose volumes were lowered by 37.9%, 42.2%, 54.7% and 54.8% for N=3, 6, 9, and 12 targets, respectively.

[0070] When further examining the dose interplay effects at the 8-Gy and 12-Gy isodose levels, the improvements with BROOMBA were evident especially with increasing number of targets. BROOMBA lowered the ambient background dose contribution to the 8-Gy and 12-Gy isodose volumes as indicated by the reduction in the y- intercept value of the curves when compared to the VMAT treatment plans. Furthermore, BROOMA also reduced the trending sloped of the plotted curves, especially for the 8-Gy isodose volumes, suggesting effectiveness of the technique in minimizing such an effect as zero gradient would indicate an ideal situation of zero dose interference among the targets at a given isodose volume.

[0071] Cross-firing fixed or arc beams from multiple directions is a well-known technique for radiosurgical treatment. However, when dealing with multiple (N>3) metastatic targets, the study is the first in demonstrating a need and potential benefits of employing a high number (e.g. N=20) of non-coplanar optimized beams for such a treatment. Our study has found that conventional VMAT technique based on preset configuration of limited arc beams is suboptimal in sparing the peripheral normal brain tissue for treating multiple brain metastases. A patient- specific beam optimization approach such as BROOMBA developed in the study is needed. Of note, cautions should be taken when applying such a technique for multiple brain metastases treatments, especially at the peripheral normal brain isodose volumes such at the 8-Gy or the 12-Gy volumes.

[0072] Although no clear guideline has been established on normal brain dose tolerance levels for multiple brain metastases treatments, several retrospective studies have correlated the peripheral normal brain dose such as the 8-Gy to 12-Gy isodose volumes with the incidence of symptomatic adverse radiation effects (AREs) with radiosurgery of brain lesions. Major cooperative group studies such as RTOG 90-05 study and the Quantec guideline have also suggested that normal brain complication is the major factor limiting a high dose to be delivered to a tumor. Therefore, peripheral normal brain dose has become a major parameter in measuring and scoring the treatment planning quality of multiple brain metastases treatments.

[0073] As a feasibility study of BROOMBA for planning multiple brain metastases, we have found that an increase in the total number of intensity-modulated beams such as N=20 can dramatically lower the peripheral normal brain dose and curtail the dose-interplay effects across the 8-Gy to 12-Gy isodose levels. Despite BROOMBA being capable of optimizing hundreds even thousands of beams, it was intentionally capped at a total number of beams under 20 for the illustrated case to ensure its clinical implementability. However additional BROOMBA optimization studies for the case have found that the largest gain for increased beam number tended to occur in the range of 10- 20 beams and additional beams continued to decrease the peripheral isodose volumes at a moderate rate. For example, using 40 beams versus 20 beams decreased the 8-Gy to 12 Gy isodose volumes by approximately 8% or 0.7 mL for the 12-target case. Furthermore, 6 MV FFF beams was found to produce better results than 10 MV FFF beams in agreement with the VMAT results. This was likely caused by the fact that 6 MV FFF beams possessed a narrower penumbra than the 10 MV FFF beams. Leaf Transmission and interleaf leakage are also lower for 6 MV FFF beams than 10 MV FFF beams.

[0074] With rapid advancements in the linear accelerator technologies such as the flattening-filer-free high output x-ray beams and digitally controlled linear accelerator maneuverability, delivering tens to hundreds of intensity modulated beams as optimized by BROOMBA or related approach is well supported. We expect that the total treatment for the illustrated 12-target case can be delivered within 20 minutes or less and entire treatment delivery is a simple turnkey operation.

[0075] In summary, we have developed and demonstrated a new treatment method of BROOMBA for managing multiple brain metastases. We found that optimizing a relatively large number (such as N=20) of non-coplanar intensity modulated beams can significantly decrease the dose interplay effects thus improve normal brain sparing and potentially outperform pre-configured VMAT delivery for such a treatment. With the encouraging results observed for the first time for BROOMBA, further technical implementation and clinical studies will be carried out to translate such a new approach for treating patients with multiple brain metastases. [0076] Example 3. Sharpening Peripheral Dose Gradient Via Beam Number

Enhancement From Patient Head Tilt For Stereotactic Brain Radiosurgery

Sharp dose fall-off is the hallmark of brain radiosurgery for the purpose of delivering a high dose of radiation to the target while minimizing the peripheral dosing of the normal brain tissue. In this study, a technique is developed to enhance the peripheral dose gradient by magnifying the total number of beams focused toward each isocenter through pre-programmed patient head tilting. This technique was tested in clinical settings on a dedicated brain radiosurgical system (GKPFX, Gamma Knife Perfexion, Elekta Oncology) by comparing dosimetry as well as delivery efficiency for 20 radiosurgical cases previously treated with the system. The 3-fold beam number enhancement (BNE) treatment plans were found to produce nearly identical target volume coverage (absolute value < 0.5%, P > 0.2) and dose conformity (BNE CI= 1.41+0.22 versus 1.41+0.11, P > 0.99) as the original treatment plans. The total beam-on time for the 3-fold BNE treatment plans were also found to be comparable (< 0.5 min or 2%) with those of the original treatment plans for all the cases. However, BNE treatment plans significantly improved the mean gradient index (BNE GI = 2.94+0.27 versus original GI =2.98+0.28 p<0.0001) and low-level isodose volumes such as 20-50% prescribed isodose volumes by 1.7% to 3.9% (p<0.03). With further 4 to 5-fold increase in the total number of beams, the absolute gradient index can decrease by as much as - 0.5 in absolute value or - 20% for a treatment. In conclusion, BNE via patient head tilt has been demonstrated to be a clinically effective and efficient technique for physically sharpening the peripheral dose gradient for brain radiosurgery.

[0077] Stereoetactic radiosurgery (SRS) typically employs a large number of photon beams to converge at an isocenter to create a sharp dose gradient. This allows a high dose to be delivered to the target while minimizing the dose to the surrounding normal brain. For example, if 100 beams converge at the isocenter, then the central dose contribution from a single beam would be approximately 1%, which is significantly lower than the central dose. Gamma Knife radiosurgery (GKSRS) is a dedicated SRS platform that directly employs such as a principle, where 192 confocal Co-60 beams are used to create a 3D dose distribution (called a "shot") at each isocenter, and depending on the complexity of the treatment, 1 to 30 isocenters or shots are typically used for a GKSRS treatment.

[0078] The question therefore arises as to whether significantly magnifying the total number of beams per isocenter would significantly affect the overall quality of a GKSRS treatment. Evidently, continuous elevating the total number of beams not only decreases the entrance or the exit dose, but also increases the internal scattering and interference among individual beams due to finite collimator size or accessible beam angles etc. For this reason, only 192 beams per isocenter are currently employed for the Leksell Gamma Knife Perfexion (PFX) system, where it is physically prohibitive to drill more apertures onto a universal collimator constructed from a single piece of tungsten metal.

[0079] In order to over the hardware limitation of the system, this study investigated a variable patient-head tilting technique to magnify the total number of beams per isocenter. The rationale for such an approach is that by tilting the patient head to different angle, confocal beams can be directed toward an isocenter from variable solid angles. The key advantage of such an approach is that it obviates the need of a large number of radioactive Co-60 sources and it also allows multiple folds of expansion of the total number of beams per isocenter.

[0080] For this study, clinical feasibility and delivery efficiency of the proposed method was investigated by comparing the dosimetric characteristics of the beam number enhancement (BNE) approach with the conventional GKSRS approach for 20 cases previously treated with GKSRS. Treatment delivery efficiency in terms of beam-on time and planning effort were also examined.

[0081] Twenty patient cases treated via the Leksell PERFEXION(TM) system (PFX) were randomly selected for the study. All cases are single-target treatments of variable target sizes from approximately 0.1 mL to 10 mL that included indications such as acoustic schwannoma, meningioma, brain metastases and mesial temporal lobe epilepsy (MTLE). The general characteristics of these cases are summarized in Table 1. For each treatment plan, all the shot coordinates were expanded via pre- set 3 head-tilt angles for PFX delivery. For cases with fewer than 10 shots, additional head tilt angles were also expanded via the previous 4C system delivery taking into consideration that the maximum number of shots is limited to 50 for the 4C delivery and no limit (up to 500 tested in the study) was found for the PFX delivery.

[0082] Technically, expanding individual shots can be manually performed by any user on the GKSRS treatment planning system (LGP version 10.2). For the current study, however, a shot scripting software was employed instead, which had significantly streamlined the process and made the BNE achievable with a couple of mouse clicks. [0083] For the data analysis, dose volume histograms were extracted for all the treatment plans and dosimetric indices such as conformity index (CI), gradient index (GI) plus isodose volume ratios for low isodose levels such as from 20% to 90% of the prescription dose were computed and compared between the original treatment plans and the beam- number-enhanced (BNE) treatment plans. In addition to conventional GI (defined as the volume ratio between 50% and 100% prescription isodose volumes), generalized GI values (defined as the volume ratio between any peripheral isodose volume at X% and 100% prescription) were also compared. Two-tail paired student t-tests were performed to determine differences in these mean values.

[0084] The linear dose fall-off curve for a single shot of variable collimator size (e.g. 4 mm, 8 mm and 16 mm) between the conventional and the BNE delivery was compared. The mean distance (D) for x-axis value is the fall-off distance averaged along all directions, which is calculated as D = 6 V 1/3 /□□ , where V is the volume for a percentage dose surface of the value on y-axis value. The two curves only start to exhibit small - 1% difference beyond 3 times the nominal distance or beyond 50% of the prescribed dose levels for a single shot. Note that the lowest 10% dose level was limited by the size of the dose calculation matrix box encompassing each shot for the calculations.

[0085] The mean BNE effect for the multi-shot treatment plans (n=20 cases), where generalized GI values were plotted against the peripheral isodose volume of interest. Logarithmic regression lines were also fitted for the original plans and the BNE treatment plans. Statistical significance (P <0.01) in the mean value difference was found between the BNE and original plans beyond the prescription isodose surface except at the 90% of the prescribe dose level (P=0.0791). Note the absolute value difference in the standard GI values of P< 0.0001. This result is further support by logarithmic regression curves, where the trending slope (p<0.01) in the logarithmic scale of 0.302 (original plans) versus 0.0297 (BNE plans) and FT = 0.958 were found for both fitted curves.

[0086] In terms of absolute isodose volumes (versus generalized GI values), difference was also noted at the dose levels beyond the prescription dose, e.g., P = 0.0483, 0.0399, 0.0251, 0.0144, and 0.0016 were observed at 60%, 50% 40% 30% and 20% of the prescription dose levels respectively versus P = 0.476, 0.137, and 0.0687 at 90%, 80% and 70% of the prescription dose levels respectively. Note that P values decreases with decreasing dose levels.

[0087] 3D isodose plots of an example case with small target volume (V=0.27 mL) and limited number of shots (n=9) are illustrative. Despite nearly identical prescription isodose curves (12.5 Gy for this case) between the BNE-plan and the original plan, lower level isodose volume such as the 4-Gy isodose curves (i.e., -30% isodose volumes) were less for the BNE plan versus the original plan, particularly on the top axial and lower sagittal plane.

[0088] The results for additional head tilt angles for the case were also compared. As the number of beams is increased from 2 to 5 fold from 9 (original plan) to 45 (BNE plan), large improvements as much as 50% were noted for the generalized GI. In addition, additional the overall effect tends to level at the 4 fold or 36 beams for majority of the dose levels for this case.

[0089] The difference in the overall beam-on time was negligible with a mean decrease of - 0.67+0.07 min for the BNE compared to the original treatment plans treatment plans. This is expected due to the dose contribution for the single shot of the original plan practically shared equally among the three or more tilt angles of the same isocenter. The difference observed is most likely contributed by the patient's skull size dependence on the head tilt angles.

[0090] In this study, a simple and effective means of physically sharpening the peripheral dose gradient is demonstrated for brain SRS. The method was tested and implemented for GKSRS and significant improvements (p< 0.01) in the peripheral dose volume just beyond target boundary or prescription isodose surface were observe and reported here for the first time.

[0091] Although the technique is demonstrating for GKSRS due to a large number of beams currently in practice for its treatment, the concept and general design of the technique is also applicable to linac or any brain SRS modalities that utilize patient head immobilizer or a couch. Regardless treatment modalities or the number of isocenter required for a treatment, variable patient head-tilting efficiently expands the beam- accessing solid angles. For example, it makes the para-coronal arc feasible for linac-based delivery or conversely, reduces the number of beams needed or the total number of radioactive sources required a standard treatment. In the context of GKSRS, this may lead to significant cost reduction if the total number of radioactive sources can be reduced by a factor 3 or more for the existing beam collimation schemes.

[0092] In the context of clinical relevance, creating rapid dose fall-off for brain radiosurgery is highly important for brain radiosurgery given that multiple studies have correlated the incidence of treatment complications such as symptomatic radionecrosis with increasing peripheral isodose volume. This issue has been also highlight in the treatment of multiple brain metastases where small worsening or improvement in the dose fall-off a single lesion may be doubly magnified for a multi-shot and/or multi-target treatments.

[0093] All patents, patent applications, and publications cited in this specification are herein incorporated by reference in their entirety to the same extent as if each independent patent, patent application, or publication was specifically and individually indicated to be incorporated by reference. The disclosed embodiments are presented for purposes of illustration and not limitation. While the invention has been described with reference to the described embodiments thereof, it will be appreciated by those of skill in the art that modifications can be made to the structure and elements of the invention without departing from the spirit and scope of the invention as a whole.

Claims

CLAIMS What is claimed is:

Claim 1. An apparatus for enabling the accurate targeting of therapeutic radiation beams to a target region within a patient comprising

a motorized, computer controlled patient positioning device, wherein the device is capable of rotation along two or more axes of a three dimensional coordinate system and can comfortably change the position of the patient's body and/or target body part with respect to one or more radiation beam emitters.

Claim 2. The apparatus of Claim 1, wherein

the patient immobilization device is a headrest.

Claim 3. The apparatus of Claim 1, wherein

the patient positioning device comprises a bed or couch.

Claim 4. The apparatus of Claim 1, wherein

the apparatus is integrated with a gamma knife radiation treatment system.

Claim 5. The apparatus of Claim 1, wherein

the apparatus is integrated with a LINAC radiation treatment system.

Claim 6. A system for the delivery of multiple radiation beams to a target within a patient, the system comprising

a shielding body;

a plurality of at least ten radioactive sources; for each radioactive source, a mechanical actuator which can move the radioactive source from a blocked position such that it emits no radiation through the shielding body to a position wherein the radioactive source is above or in proximity to a channel in the shielding body such that a collimated beam of radiation is emitted through the channel; and wherein

at least 10 of the radioactive sources can be moved independently of one another.

Claim 7. The system of Claim 6, wherein

at least twenty radioactive sources can be moved independently of one another.

Claim 8. The system of Claim 6, wherein

at least one hundred radioactive sources can be moved independently of one another.

Claim 9. The system of Claim 6, wherein

the radioactive source is cobalt-60.

Claim 10. The system of Claim 6, wherein

the shielding body is a dome.

Claim 11. The system of Claim 6, wherein

the shielding body comprises tungsten.

Claim 12. A LINAC therapeutic radiation delivery system, comprising

two or more separate radiation beam emitters.

Claim 13. The system of Claim 12, comprising

twenty or more separate radiation beam emitters.

Claim 14. A method of treating a target within a patient with a plurality of radioactive beams, comprising the steps of:

mapping the position of one or more targets in the patient;

performing a planning step which takes into account: (1) the location and shape of the one or more targets; (2) potential movement patterns of the patient; and (3) potential beams that can be delivered to the target by the selected treatment system at different points along potential movement patterns, wherein the output of the planning step is an optimized movement pattern and a series of associated beams to be delivered;

training the patient to perform the movement pattern;

positioning the patient within the treatment system; and

having the patient perform the movement pattern and delivering the series of associated beams to effectively treat each target with a sufficient dosage of radiation, wherein beam firing is guided by a system that maps the position of the one or more targets in real time as the patient performs the movement pattern.

Claim 15. The method of Claim 14, wherein

the target is a tumor.

Claim 16. The method of Claim 14, wherein

the target is a tumor within a patient's head.

Claim 17. The method of Claim 16, wherein

the movement pattern encompasses the patient moving their head in the optimized pattern.

Claim 18. The method of Claim 14, wherein the planning step comprises

a traveling salesman algorithm to determine the optimal movement pattern and series of beams to be delivered.

Claim 19. The method of Claim 14, wherein the treatment system is a gamma knife system.

Claim 20. The method of Claim 14, wherein the treatment system is a LINAC system.

Claim 21. A method of delivering therapeutic radiation to a target within a patient, comprising the delivery of 500 or more beams.

Claim 22. A method of delivering therapeutic radiation to a target within a patient, comprising the delivery of 2,000 or more beams.

Claim 23. A method of delivering therapeutic radiation to a target within a patient, comprising the delivery of 20,000 or more beams.