Offline quality control of beam shaping devices for radiation therapy

Abstract

The invention relates to a method for evaluating the quality of a beam shaping device (11) manufactured according to a planning device design (11d) for shaping an accelerated particle beam (100.i) emitted by a particle accelerator system, the method comprising: (a) establishing a planning device design (11d) of the beam shaping device using a treatment planning system (TPS), (b) manufacturing the beam shaping device according to the planning device design, (c) establishing a CT scan of the beam shaping device to produce an actual CT image (11a), (d) determining dimensions and local material densities from the actual CT image, (e) determining a calculated dose distribution in a treatment volume (V) by virtually irradiating the treatment volume using the beam with a virtual beam shaping device having a geometry and densities defined by the device actual CT image, (f) comparing the calculated dose distribution (cDD) with a reference dose distribution (rDD).

Description

TECHNICAL FIELD

[0001] The present invention relates to a quality assessment (QA) method for ensuring that an actual beam shaping device has been manufactured satisfactorily according to a plan device determined by a treatment planning system (TPS) to deposit a planned dose distribution (pDD) in a treatment volume comprising tumor cells. The present method does not require the use of a particle accelerator system. The present method is further advantageous in that, in many cases, it can identify geometrical or density imperfections in the actual beam shaping device that potentially cause deviations from the planned dose distribution (pDD). These manufacturing-induced imperfections can be corrected in subsequent manufacturing process improvements. BACKGROUND

[0002] Radiation therapy with particles or waves such as electron beams, proton beams, heavy ion beams, x-rays, gamma rays, etc. has become an essential tool for the treatment of patients with tumors.

[0003] The first challenge in cancer treatment is to limit the treatment plan so as to ensure that the defined dose is deposited into the tumor cells to effectively destroy or kill the tumor cells, while limiting the dose deposited into healthy cells to harm the healthy cells as little as possible. The second challenge is to actually deposit the defined dose into the tumor cells, while actually depositing the limited dose into the healthy cells as planned.

[0004] Different radiation modalities deposit their energy in different patterns. For example, x-rays deposit most of their energy near the skin level, and the deposited energy decreases as the depth of penetration into the tissue increases. Thus, healthy tissue located upstream of the treatment volume of tumor cells receives a higher dose than cells located in the treatment volume. In contrast, charged particle beams such as protons and carbon ions deposit most of their energy near the end of their beam path, forming a so-called Bragg peak.

[0005] Charged particles coming out of the nozzle of a particle accelerator form a narrow pencil beam. In order to cover a treatment volume of actual size, the pencil beam has to be scattered by a foil or the pencil beam has to be scanned. Pencil beam scanning (PBS) and dual scattering (DS) proton therapy are two techniques that allow the physician to deliver radiation with a dose that precisely and efficiently covers the tumor while minimizing the radiation exposure to healthy tissue. This can be achieved by shaping the beam(s) delivered towards the volume to be treated by a beam shaping device. A beam shaping device is a device positioned in or fixed to the nozzle of a particle accelerator, between the source of accelerated particles and the treatment volume, which changes the energy profile and / or the geometry of the beam.

[0006] Penumbra Beam Scanning (PBS) is an active scanning technique that consists in diverting charged particle PBS beams (or pencil beams) along corresponding beam axes (Xi) toward individual beam spots of a beam spot matrix that defines a treatment zone including tumor cells. Thus, a predetermined target dose is deposited into cells covered by individual beam spots. PBS beams are diverted along corresponding beam axes (Xi) and dose deposition is performed according to a treatment plan that defines a dose (Dij) to be deposited into each cell covered by a given beam spot along a beam axis (Xi), as well as a scanning order of beam spot irradiation. PBS reduces unnecessary radiation exposure to surrounding non-cancerous cells by shaping the region to be treated to reflect the geometry of the tumor. In addition to the geometry of the target zone, PBS allows local tuning of the intensity of each PBS beam to achieve the required dose distribution within the target zone.

[0007] PBS is very advantageous because it optimizes the geometrical distribution of dose deposition to match the geometry of the treatment zone (V) that encompasses the tumor. However, PBS treatment time can be long because PBS beams have to scan each beam spot (Sij) and each energy layer (Tj). Moving a PBS beam from a beam axis (Xi) to a different beam axis (X(i+1)) takes a few milliseconds. Changing the energy of a given PBS beam parallel to a given beam axis (Xi) to deposit a desired dose (Dij) into cells (Cij) of a different layer (Tj) is even more time consuming, requiring about 500 ms. Thus, the number of layers (Tj) has a strong impact on treatment duration.

[0008] Saving treatment time reduces the occupancy time of the particle accelerator per patient. It is also more comfortable for the patient. It is also advantageous when the treatment plan includes FLASH irradiation, where a dose is deposited into cells at a high dose rate (HDR) of at least 1 Gy / s or even up to at least 40 Gy / s. The same dose deposited at HDR has been shown to be non-damaging to healthy cells relative to the same dose deposited at a lower conventional dose rate (CDR). FLASH irradiation is of particular interest when a given dose deposited into tumor cells has the same killing effect whether it is deposited at HDR or at CDR.

[0009] Depositing a predetermined dose (Dij) into the treatment volume by PBS can be achieved by using a ridge filter and a PBS beam of single energy layer. The ridge filter requires the beam spot of each layer to be aligned along the corresponding beam axis (Xi). Ridge filters comprising energy degrading units in the form of smooth pins or stepped pyramids or wave crest have been described in the art. For example, EP21208699 describes a ridge filter comprising a plurality of energy degrading units in the form of apertures or pins arranged side by side in a support base according to the beam spot array. Each energy degrading unit is formed by one or more sub-degrading units in the form of apertures or pins having a generalized cylindrical geometry with a cross-sectional area (Ai) and extending from the support block along the corresponding beam axis (Xi). For example, the sub-degrading units of the same energy degrading unit can be stacked one on top of the other along the corresponding irradiation axis (Xi). The superposition of the sub-degrading units forming each energy degrading unit allows shaping and increasing the width of the spread Bragg peak (SOBP) along the corresponding beam axis (Xi).

[0010] The principle of the ridge filter is that portions of the beam (100.i) having a given energy oriented along the corresponding beam axis (Xi) pass through different material thicknesses of the filter, thereby producing Bragg peaks of different ranges that superimpose to produce a uniform SOBP extending along the corresponding beam axis (Xi) from the upstream boundary of the treatment volume (V) to the downstream boundary of the volume, thereby spanning the entire depth of the volume along Xi. The ridge filter can be integrated or combined with other beam shaping devices including range shifters and range compensators.

[0011] Dual scattering proton therapy is a passive scattering technique in which a wide beam passes through a uniform first scatterer and produces a Gaussian beam profile on a second scatterer which must be somewhat non-uniform in order to modify the Gaussian distribution and the beam energy. Beam shaping devices used in dual scattering techniques include range compensators (to control the energy profile) and apertures (to control the shape of the irradiation cross section). Hereafter, unless otherwise clearly defined, the term "beam" is used to refer to both PBS beams (or pencil beams) for PBS applications and wide beams for dual scattering applications.

[0012] A CT scan image of the patient is produced, the values of which are converted into proton stopping power. A planned device design of one or more beam shaping devices is produced by a treatment planning system (TPS), the planned device design shaping the beam to match the geometry of the treatment volume (V) including the tumor cells (3t) to deposit a specific dose (Dij) into a specific location within the treatment volume (V). A planned dose distribution (pDD) can thus be generated to meet the objectives of dose deposition in the treatment volume and surrounding healthy tissue set by the physician.

[0013] One or more beam-shaping devices are designed and manufactured according to the corresponding planning device. Different techniques can be used, including machining, but nowadays 3D printing has advantages in terms of cost-effectiveness, speed and its ability to generate fine and complex geometries, and a wide variety of materials are suitable for this technology.

[0014] However, neither machining nor 3D printing guarantees that the beam-shaping device is generated exactly according to the planning device design. Several manufacturing errors can occur, which can include errors in the width and height of the tower or cavity of the beam-shaping device, or the axis of the energy degrading unit of the ridge filter can deviate from the corresponding beam axis. The density of the material can be non-uniform throughout the volume of the device, and there can be air bubbles in the body of the device, and the surfaces of the beam-shaping device designed as flat bases can be warped or curved from the 3D printer.

[0015] Depending on their size, these printing errors can cause the dose distribution to differ from the planned dose distribution that the TPS tried to achieve. For example, Monte Carlo simulations have shown that the deviation between the nominal width or height of the tower and the printed width or height should be less than 1 mm to avoid significant deviations in the dose distribution. Therefore, after the beam-shaping device is manufactured, a quality control procedure is needed to identify and quantify the defects in the manufacturing process and to quantify the resulting deformation of the dose distribution in the treatment volume caused by these defects.

[0016] One accepted QA procedure is to install the beam-shaping device(s) on the nozzle, then deliver the treatment plan using the therapy machine, and then measure the 3D dose distribution in a water phantom. The measured 3D dose distribution can be compared to the planned 3D dose distribution (pDD) of the TPS to evaluate the quality of the beam-shaping device(s). However, this method is not optimal because installing the dosimetry equipment and acquiring the data is both time-consuming and cumbersome. In a busy PT center, the time available for performing QA measurements is limited, and the cost of operating the proton beam can be quite high. Therefore, a QA method that does not require occupying the proton beam time would have advantages in terms of time, availability, and cost. In addition, any reason for not matching the planned dose distribution cannot be evaluated by this method.

[0017] Heng Li et al. (A CT-based software tool for evaluating range compensator quality in passively scattered proton therapy. Phys. Med. Biol. 55, 6759-6771 (2010)) and Yoon M. et al.

[11] (Computerized tomography-based quality assurance tool for proton range compensators. Med. Phys. 35, (2008)) disclose alternative methods based on a geometrical comparison of the surface of the beam-shaping device(s). The reference surface is the ideal shape resulting from the TPS optimization, and this reference surface is compared to the surface estimated from the CT scan of the beam-shaping device(s). The advantage of these methods is that they do not use beam time in the treatment room, but they also have the limitation that they only consider geometrical deviations from the reference, and ignore any impact of the inhomogeneity of the material density inside the beam-shaping device on the dose distribution inside the treatment volume.

[0018] The problem addressed by the present invention is to provide a fast quality assurance method for a beam-shaping device that does not require the use of a beam and that allows a quantitative analysis of the functioning of the beam-shaping device, including a quantitative analysis of the beam-stopping power, of the dose distribution in the volume downstream of the beam-shaping device, and allows the identification of defects in the beam-shaping device in case the quantitative analysis of the dose distribution is not satisfactory. These and other advantages of the present invention will be described later. SUMMARY

[0019] The present invention is defined in the appended independent claims. Preferred embodiments are defined in the dependent claims. In particular, the present invention relates to a method for evaluating the quality of a beam-shaping device manufactured according to a planned device design, the beam-shaping device being used for shaping one or more accelerated particle beams emitted by a particle accelerator system. The method comprises:

[0020] (a) establishing a planning device design of a beam shaping device using a treatment planning system (TPS), which planning device design is adapted to shape one or more particle beams to match the geometry of a treatment volume (V) comprising tumor cells so as to deposit a certain dose (Dij) to a certain location within the treatment volume (V) and thus to define a planning dose distribution (pDD) in order to meet a target for dose deposition in the treatment volume set by a physician,

[0021] (b) manufacturing the beam shaping device according to the planning device design,

[0022] (c) establishing a CT scan of the beam shaping device to produce an actual CT image,

[0023] (d) determining dimensions and local material densities from the actual CT image,

[0024] (e) extending the actual CT image to include the treatment volume (V), preferably to include one or more of ridge filters, range shifters and / or range compensators not included in the beam shaping device,

[0025] (f) determining a calculated dose distribution in the treatment volume (V) using a dose engine, preferably using the same dose engine as in the TPS used in step (a), which calculated dose distribution is obtained by virtually irradiating the treatment volume (V) using one or more beams through a virtual beam shaping device having the geometry and densities defined by the device actual CT image,

[0026] (g) comparing the calculated dose distribution (cDD) with a reference dose distribution (rDD).

[0027] The reference dose distribution (rDD) can be:

[0028] • a calculated planning dose distribution (cpDD) obtained by:

[0029] o first, creating a high resolution planning device design (1 1hrd) using the TPS, which high resolution planning device design corresponds to the planning device design but has a higher resolution matching the resolution of the actual CT image, and

[0030] o second, determining the calculated planning dose distribution (cpDD) using a dose engine by virtually irradiating the treatment volume (V) using one or more beams through the high resolution planning device; or

[0031] • the planning dose distribution (pDD) or a function thereof.

[0032] In particular, the reference dose distribution (rDD) can be any of:

[0033] • the computed planned dose distribution (cpDD) as defined above, and wherein the high resolution planning device design is created with a voxel size equal to the CT voxel size used for the establishing the actual CT image, with a tolerance of ± 20%, preferably of ± 10%, or

[0034] • a function of the planned dose distribution (pDD) which is the result of a transformation of the planned dose distribution (pDD) computed at a given spatial resolution and changing the voxel size to match the CT voxel size with a tolerance of ± 20%, preferably of ± 10%, preferably using linear interpolation or nearest neighbor interpolation on the planned dose distribution (pDD).

[0035] In a preferred embodiment, the CT scan of the beam shaping device is performed with a CT voxel size of not more than 0.5 mm, preferably of not more than 0.2 mm. The planning device design can be created using a TPS, preferably the same TPS as in step (f) defined above, from the treatment plan (TP) with a voxel size similar to or smaller than the CT voxel size.

[0036] The comparison of the computed dose distribution (cDD) with the reference dose distribution (rDD) as defined above in step (g) can be performed using gamma evaluation. The beam shaping device can be considered to be in conformity with the planning device design if a gamma value (γ) lower than or equal to a reference gamma value (γr) (i.e. γ≤ γr) is obtained in a predetermined percentage of the voxels of the reference dose distribution (rDD). The gamma value (γ) is defined as the minimum of the following function:

[0037]

[0038] wherein |d(cDD)-d(rDD)| is the distance between the analysis points; |D(cDD)-D(rDD)| is the dose difference, DTaand ΔDare scaling factors.

[0039] If the gamma value (γ) is higher than the reference gamma value (γr) (i.e. γ> γr) in a predetermined number of voxels of the reference dose distribution (rDD), the beam shaping device is not in conformity and the dimensions and local material densities determined from the actual CT image can be compared to the planning device design or to the computed high resolution planning device design as defined above to identify deviations of the dimensions and local material densities of the beam shaping device from the planning device design.

[0040] • if one or more deviations are identified, the method can comprise investigating the manufacturing parameters which can have caused the one or more deviations and correcting these manufacturing parameters accordingly, and

[0041] • If no deviation is identified, the method can comprise repeating step (a) as defined above and establishing an alternative planning device design of the beam-shaping device.

[0042] The actual CT image of the device, obtained by CT scanning of the beam-shaping device, is preferably in the form of a grey-scale image characterized by CT Hounsfield Units (HU). Thus, the local material density can be determined by transforming the HU into a corresponding density.

[0043] The beam-shaping device can be produced by 3D printing or machining in a material selected from a polymer, or a metal, or any combination thereof.

[0044] The beam-shaping device can be a ridge filter comprising a set of energy degrading units, wherein each energy degrading unit is configured to reduce the initial energy (E0) of a corresponding charged particle beam to a reduced energy (Ei) such that the dose is deposited into the treatment volume (V) according to a planned dose distribution (pDD). For example, the energy degrading units can be in any of the following forms:

[0045] • apertures arranged side-by-side in a support base having a thickness measured along the beam axis (X), each aperture extending from an aperture opening at a surface of the support base and penetrating to a given depth measured along the beam axis (X), or

[0046] • pins arranged side-by-side and supported on a support base, each pin extending from the support base along the beam axis (X).

[0047] In a preferred embodiment, the energy degrading units are formed by one or more sub-degrading units. BRIEF DESCRIPTION OF DRAWINGS

[0048] For a more complete understanding of the nature of the present application, reference is made to the following detailed description taken in connection with the accompanying drawings in which:

[0049] Figure 1 (a): shows a depth dose curve of a typical Bragg peak of a beam having a given energy. The abscissa axis represents depth in water. The maximum range in water (W0) is defined as the depth beyond the Bragg peak maximum and the corresponding energy is equal to 80% of the Bragg peak maximum, wherein W0 must be at least equal to or greater than dij (dij < W0).

[0050] Figure 1 (b) shows the depth dose curve of the Bragg peak of the beam of figure 1 (a) wherein an energy degrading unit intersects the path of the beam.

[0051] Figure 2(a) shows beams (100.i, 100.(i+1)...) extending along respective beam axes (Xi, X(i+1)...) parallel to each other and to an irradiation axis (X) through the beam shaping device comprising a range shifter, a ridge filter, and a range compensator.

[0052] Figure 2(b) shows beams (100.i, 100.(i+1)...) extending along respective beam axes (Xi, X(i+1)...) fanning out around an irradiation axis (X) through the range shifter, the ridge filter, and the range compensator.

[0053] Figure 2(c) shows two energy degrading units, wherein the beams (100.i, 100.(i+1)) extend along respective beam axes (Xi, X(i+1)) not parallel to each other ((Xi, X(i+1))-angle exaggerated).

[0054] Figures 3(a) to 3(c) Figure 3(a) shows three embodiments of a ridge filter.

[0055] Figure 3(d) shows an embodiment of a range shifter,

[0056] Figure 3(e) shows an embodiment of a range compensator.

[0057] Figure 4 Figure 4 shows a flow chart illustrating steps of a method of the present application. DETAILED DESCRIPTION

[0058] The present application relates to a method for evaluating the quality of a beam shaping device (11) manufactured according to a planned device design (11d). The beam shaping device is used for shaping one or more accelerated particle beams (100.i) emitted by a particle accelerator system. The beam shaping device (11) is a device positioned in a nozzle of a particle accelerator, between an accelerated particle source and a treatment zone, which changes the energy and fluence profile as well as the geometry of the beam. The beam shaping device (11) can be, for example, a ridge filter (11f), a range compensator (11c), a range shifter (11s), or a combination of two or more of the aforementioned devices. If the beam shaping device comprises a combination of two or more of the aforementioned devices, the two or more devices can be in the form of separate modules arranged in series along an irradiation axis (Xi), or alternatively, the two or more devices can be integrated in a single device. For example, the base of a ridge filter can form a range shifter. In the following, the present application is described with respect to a ridge filter (11f) alone, for the sake of clarity and because the geometry of a ridge filter is more complex than the geometry of a range shifter. The method comprises the following steps.

[0059] First, a planned device design (1 Id) of the beam shaping device (11) is established using a treatment planning system (TPS). The planned device design (1 Id) has to be adapted to shape one or more particle beams (100.i) to match the geometry of a treatment volume (V) comprising tumor cells (3t) so as to deposit a specific dose (Dij) to a specific location within the treatment volume (V) thus defining a planned dose distribution (pDD) to meet the goals for dose deposition in the treatment volume and in the surrounding healthy tissue set by the physician. A CT image (CT-V) of the patient is acquired, which is used by the TPS to calculate the planned dose distribution. The voxel size of the CT image (CT-V) is typically 1 mm to 3 mm.

[0060] The beam shaping device (11) is manufactured according to the planned device design (1 Id). For example, the beam shaping device (11) can be produced by 3D printing in any suitable material, preferably a polymer or a combination of polymers. A CT scan of the beam shaping device (11) is established to produce an actual CT image (11a) of the beam shaping device (11). The resolution of the voxels of the CT image (11a) is chosen to correctly resolve the smallest geometric features of the beam shaping device (11). The dimensions and the local material density are determined from the actual CT image (11a).

[0061] The gist of the present invention is to calculate a calculated dose distribution in the treatment volume (V) obtained by virtually irradiating the treatment volume (V) using the beam shaping device (11) according to the actual CT image (11a). To this end, the method comprises the following steps.

[0062] The actual CT image (11a) is extended to include the treatment volume (V). CT-V is interpolated and inserted into the extended CT image (11a). If the voxel size of CT-V is different from the voxel size of the CT image (11a), interpolation is required. If the particle accelerator system comprises one or more additional beam shaping devices selected from ridge filters (11f), range shifters (11s) and / or range compensators (11c) not comprised in the beam shaping device (11) according to the planned device design (1 Id), the actual CT image (11a) is extended to include the corresponding additional beam shaping devices.

[0063] A dose engine (preferably the same dose engine as used in the TPS used in step (a) of the present method) is used to determine a calculated dose distribution in the treatment volume (V) obtained by virtually irradiating the treatment volume (V) using one or more beams by the virtual beam shaping device having the geometry and density defined by the device actual CT image (11a) including any additional beam shaping devices present in the particle accelerator system.

[0064] A "dose engine" is an algorithm that computes a dose distribution from a description of the geometry and beam characteristics. There are different types of dose engines. A Monte Carlo dose engine is commonly used in the art.

[0065] At this stage, the computed dose distribution (cDD) can be compared to the reference dose distribution (rDD) to assess the quality of the manufactured beam-shaping device (11). The comparison of the computed dose distribution (cDD) to the reference dose distribution (rDD) is more instructive than the comparison of the planned device design (1 Id) to the actual CT image (11a) because the differences between the planned device design (1 Id) and the actual CT image (11a) do not necessarily result in an unacceptable deviation from the treatment plan. This comparison gives specific information on whether the manufactured beam-shaping device (11) is suitable for treatment according to the treatment plan. If the comparison reveals that the manufactured beam-shaping device is not suitable, the reasons can be understood by comparing the planned device design (1 Id) to the actual CT image (11a).

[0066] Reference dose distribution (rDD)

[0067] In one embodiment, the reference dose distribution (rDD) is a computed planned dose distribution (cpDD) obtained by:

[0068] • first, a high-resolution planned device design (1 1hrd) is created using the TPS, which corresponds to the planned device design (1 Id) but with a higher resolution, matching the resolution of the actual CT image, and

[0069] • second, the computed planned dose distribution (cpDD) is determined using the dose engine by virtual irradiation of the treatment volume (V) by the high-resolution planned device (1 1hrd) using one or more planned pencil beams (in PBS) or using one large beam (in dual scatter)

[0070] For example, the high-resolution planned device design (1 1hrd) can be created with a voxel size equal to the CT voxel size used in the actual CT image, with a tolerance of ±20%, preferably of ±10%. This way, the resolution of the high-resolution planned device design (1 1hrd) and the resolution of the actual CT image (11a) are such that the resolutions are comparable, resulting in comparable dose distribution computations.

[0071] In an alternative embodiment, the reference dose distribution (rDD) is the planned dose distribution (pDD) or a function thereof. The function of the planned dose distribution (pDD) can be the result of a transformation of the planned dose distribution (pDD) computed at a given spatial resolution and changed in voxel size to match the CT voxel size with a tolerance of ± 20%, preferably of ± 10%, preferably using linear interpolation or nearest neighbor interpolation on the planned dose distribution (pDD). In this way, the dose distributions pDD and cDD thus obtained can be compared at a similar resolution.

[0072] Typically, the CT scan of the beam shaping device (11) can be performed with a CT voxel size of not more than 0.5 mm, preferably of not more than 0.2 mm. In contrast, usually, the planning device design (1 Id) is created from the treatment plan (TP) using a TPS at a voxel size similar to or smaller than the CT-V voxel size. The voxel size used for establishing the planning device design (1 Id) can be about 1.0 mm or less.

[0073] Thus, the ratio of the resolutions between the actual CT image (1 la) and the planning device design (1 Id) can be 1 to 5. In case of such a ratio of the sizes, it is preferred to reduce the ratio of the resolutions by any of the preceding embodiments to yield a reference dose distribution (rDD) which is more directly comparable to the computed dose distribution (cDD).

[0074] The actual CT image (1 la) of the device obtained by the CT scan of the manufactured beam shaping device (11) is preferably in the form of a gray scale image characterized by CT Hounsfield units (HU). The local material density can be determined by transforming the HU into a corresponding density.

[0075] In case of a Monte Carlo dose engine, the dose deposited by the protons can be statistically simulated on a grid of voxels of a different size than the expanded CT image (1 la). Thus, while the proton trajectories through the beam shaping device (11) are properly described in the high resolution expanded CT image (1 la), for reducing the memory usage and improving the signal-to-noise ratio statistics of the Monte Carlo simulation, the dose distribution is statistically computed on a map of lower resolution. Thus, the computed dose distribution (cDD) and the reference dose distribution (rDD) can have a larger voxel than the expanded CT image (1 la).

[0076] Quality assessment of the beam shaping device (11)

[0077] The method of the invention comprises a step of comparing the calculated dose distribution (cDD) with a reference dose distribution (rDD), the reference dose distribution being defined as above. In a preferred embodiment, the comparison of the calculated dose distribution (cDD) with the reference dose distribution (rDD) is performed using a gamma evaluation. The beam shaping device (11) is considered to be in conformity with the planning device design (11d) if a gamma value (y) lower than or equal to a reference gamma value (yr) (i.e. y < yr) is obtained in a predetermined percentage of voxels of the reference dose distribution (rDD). The gamma value (y) is defined as the minimum of the following function:

[0078]

[0079] where |d(cDD)-d(rDD)| is the distance between the analysis point and the closest point of the same dose in the reference image; |D(cDD)-D(rDD)| is the dose difference of the same point in the two images, DTA and AD are scale factors. For example, if yr = 1, the predetermined percentage of voxels satisfying this criterion y < yr can be 90%, preferably 95%. For example, the scale factors can be DTA = 3 mm and AD = 3%.

[0080] If the gamma value (y) is higher than the reference gamma value (yr) (i.e. y > yr) in a predetermined percentage of voxels, the beam shaping device (11) is not in conformity. The dimensions and local material density of the beam shaping device can be determined from the actual CT image (11a) and compared with the planning device design (11d) or the calculated high resolution planning device design (11hrd), as defined above. Deviations of the dimensions and local material density of the beam shaping device (11) from the planning device design (11d) can thus be identified. Depending on experience, possibly with the help of artificial intelligence, different types of deviations can be attributed to a specific influence on the calculated dose distribution (cDD).

[0081] If one or more deviations of the dimensions and / or local material density of the beam shaping device (11) from the planning device design (11d) are identified, the method can thus comprise investigating manufacturing parameters that can have caused the deviations and correcting these manufacturing parameters accordingly. If no deviations are identified, the method can comprise repeating the steps of establishing an alternative planning device design (11d) of the beam shaping device (11) using a treatment planning system (TPS).

[0082] The beam shaping device (11)

[0083] As Figures 3(a) to 3(e) illustrated, the beam shaping device (11) can be, for example, a ridge filter (11f), a range compensator (11c), a range shifter (11s), or a combination of two or more of the aforementioned devices.

[0084] Figure 1(a) plots the dose (Dij) deposited by a beam (100.i) having a given energy (E0) as a function of the water equivalent thickness (WET) along the beam axis (Xi). The maximum beam range in water (WET) is defined as Wo, corresponding to the depth beyond the Bragg peak maximum and the corresponding energy equal to 80% of the Bragg peak maximum. This means that the beam (100.i) having a given energy cannot penetrate deeper into the tissue than the corresponding water equivalent thickness Wo (i.e., target depth dij < Wo). If the target depth (dij) to which the treatment volume (V) extends is deeper than the Bragg peak (i.e., dij > Wo), a beam having a higher energy must be used. Figure 1(b) shows the shift of the Bragg peak upon insertion of a beam-shaping device (11) or elements thereof in the path of the same beam (100.i) having a given energy (E0). It can be seen that the target depth (dij) at which the WET of the Bragg peak now lies is less than Wo from the patient's skin (3s) (i.e., dij < Wo). This is because a portion of the energy of the beam (100.i) (E0 - Eij) is absorbed by the beam-shaping device (11) or elements thereof through which the beam must pass. Figure 1(b) shows how one Bragg peak can be shifted from WET = Wo to target WET = dij by interposing the shaping device (11) or elements thereof. Since a SOBP can be formed by superimposing several Bragg peaks distributed along the irradiation axis (X) over a range of depths, several sub-degrading units (11.ij) can be superimposed to form an energy degrading unit (11.i) as shown in Figures 3(b) and 3(c) in order to deposit the required dose (Dij) into the entire sub-volume of the treatment volume (V) by a PBS having a single energy (or plot) layer or by double scattering. The size of each sub-degrading unit (11.ij) is determined to shift the Bragg peak from WET = Wo to the target depth (dij) to ensure that the required dose (Dij) is deposited in each cell (Cij) of the sub-volume irradiated by the beam (100.i). The number of Bragg peaks required to produce the desired SOBP according to the TP and their corresponding target depths (dij) must be determined, which defines the number of energy degrading units (11.i) and sub-degrading units (11.ij) that make up these energy degrading units. The geometry, material, and material density for the ridge filter (11f) also need to be determined to establish the planning device design (11d).

[0085] Range shifter (11s)

[0086] The range shifter (11s) illustrated in Fig. 3(d) generally has a simple parallelepiped geometry and is used to absorb a controlled fraction of the energy of the beam (100.i) thus reducing the depth of penetration into the patient to match the depth of the treatment volume (V) comprising the tumor cells. Due to its simple geometry, the range shifter (11) is generally less problematic in terms of quality except that the material density can present local inhomogeneities depending on the process used for its manufacture.

[0087] The range compensator (11c)

[0088] The range compensator (11c) illustrated in Fig. 3(e) is formed by a block comprising a cavity reflecting the distal boundary of the treatment volume (V). The range compensator conforms the dose deposition pattern to the distal edge of the treatment volume (V). The geometry of the range compensator (11c) is more complex than the geometry of the range shifter (11s) discussed above. Therefore, both the geometry and the density profile of the range compensator (11c) are subject to quality assessment to ensure the success of the planned treatment.

[0089] The ridge filter (11f)

[0090] Figures 3(a) to 3(c) Three different embodiments of the ridge filter (11f) are illustrated. The three different embodiments have in common that they comprise a set of energy degrading units (11.i) wherein each energy degrading unit (11.i) is configured to reduce the initial energy (E0) of the corresponding charged particle beam (100.i) to a reduced energy (Ei) and a corresponding number of particles such that the dose prescribed by the physician is deposited into the treatment volume (V) according to the planned dose distribution (pDD).

[0091] As illustrated in Fig. 3(c), the energy degrading units (11.i) can be in the form of apertures arranged side by side in a support base (11fb) having a thickness (xb) measured along the beam axis (X). Each aperture extends from an aperture opening at the surface of the support base (11fb) and penetrates to a given depth measured along the beam axis (X). Alternatively, as illustrated in Figs. 3(a) and 3(b), the energy degrading units (11.i) can be in the form of pins arranged side by side and supported on a support base (11fb). Each pin extends from the support base along the beam axis (X). As illustrated in Fig. 3(a), all pins can be parallel to a single irradiation axis (X). Alternatively, as illustrated in Figs. 2(b) and 2(c), considering that in a pencil beam scanning (PBS) the beam scanning causes angles to be formed between the individual beams, the pins can be parallel to different irradiation axes (Xi).

[0092] As shown in Fig. 2(c), Fig. 3(b) and Fig. 3(c), the energy degrading unit (11.i) can be formed by one or more sub-degrading units (11.ij). For example, a method for designing a ridge filter with sub-degrading units (11.ij) according to a treatment plan is described in EP 2021 / 0208699.

[0093] The beam shaping device (11) can also be a combination of two or more of the ridge filter (11f), the range compensator (11c), the range shifter (11s). As shown in Fig. 2(a) and Fig. 2(b), the two or more devices can be separated from each other. Alternatively, they can be integrated in a single device. For example, the range shifter (11s) can form the support base (11fb) of the ridge filter (11f), or can be added to the thickness of the range compensator (11c).

[0094] Figure 4 Flowchart of the method

[0095] Figure 4 A flowchart illustrating the method of the present invention is shown. A medical practitioner establishes a treatment plan (TP) (see Figure 4 in (A)). The TP includes dose objectives to be reached at the end of the treatment, prescribed by the physician. These dose objectives can include one or more of the following:

[0096] • a minimum dose that should be deposited in a predetermined proportion of the treatment zone or sub-zone comprising tumor cells,

[0097] • a maximum dose that should not be exceeded in a sub-zone comprising healthy cells,

[0098] • a dose that should be deposited in a sub-zone comprising both tumor cells and healthy cells at a minimum dose deposition rate,

[0099] • etc.

[0100] In (B), a plan device design (11d) is generated using a treatment planning system (TPS). A first resolution (defined by the voxel size) is applied to design the geometry and density of the beam shaping device (11). In order to save computing resources and time, the first resolution can be moderate, about 1 mm per voxel.

[0101] In block (C), the planned dose distribution (pDD) is calculated by positioning the beam shaping device (11) with the plan device design (11d) on one or more irradiation beams (100.i). The pDD is required to meet the requirements defined in the TP.

[0102] As indicated in block (D), the beam shaping device (11) can be manufactured according to the planning device design (11d). The beam shaping device can be manufactured by 3D printing or machining. In block (E), an actual CT image (11a) of the thus manufactured beam shaping device (11) is generated. The actual CT image (11a) is typically in the form of a gray scale image characterized by CT Hounsfield units (HU) which represent the local density of the beam shaping device (11) material.

[0103] In block (F), a calculated dose distribution (cDD) is calculated from the actual CT image (11a) of the beam shaping device. In block (M), this calculated dose distribution (cDD) is compared to a reference dose distribution (rDD) defined in block (L). As examples for determining rDD, two alternative approaches are proposed: Option 1 (= OPT. 1) defined in blocks (G) to (I) and (L), and Option 2 (= OPT. 2) defined in blocks (J) to (L).

[0104] In Option 1, in block (H), a high resolution planning device design (11hrd) is created using a TPS, which corresponds to the planning device design (11d) but with a higher resolution. The resolution of the actual CT image is increased to the higher CT resolution of the planning device design via interpolation. Then, in block (I), a calculated planning dose distribution (cpDD) is determined using a dose engine by virtually irradiating the treatment volume (V) by the high resolution planning beam shaping device (11hrd) using one or more beams. In block (L), the reference dose distribution (rDD) is defined as the calculated planning dose distribution (cpDD) (i.e. rDD = cpDD).

[0105] In Option 2, if the resolution of the planning device design (11d) is compatible with the resolution of the actual CT image, in block (L), the reference dose distribution (rDD) can be defined as the planning dose distribution (pDD). If the resolution of the planning device design (11d) is lower than and incompatible with the resolution of the actual CT image, in block (L), the reference dose distribution (rDD) can be defined as a function (f(pDD)) of the planning dose distribution (pDD) which is the result of a transformation of the planning dose distribution (pDD).

[0106] The calculated dose distribution (cDD) determined in block (F) can be compared with the reference dose distribution (rDD) defined in block (L). The comparison of the calculated dose distribution (cDD) with the reference dose distribution (rDD) is preferably performed using a gamma evaluation, such that if a gamma value (γ) lower than or equal to a reference gamma value (γr) (i.e. γ≤ γr) is obtained in a predetermined percentage of the voxels of the reference dose distribution (rDD), the beam shaping device (11) is considered to be in conformity with the planning device design (11d) and thus the quality assessment operation is successfully ended (see arrow "yes" from block (M) to block "end"). If the gamma value (γ) is not lower than or equal to the reference gamma value (γr) (i.e. γ> γr) in a predetermined percentage of the voxels of the reference dose distribution (rDD), the actual CT image of the device (11a) is compared with the planning device design (11d) in terms of geometry and local density in block (N).

[0107] If a geometrical and / or density mismatch between the actual CT image (11a) and the planning device design (11d) is identified, which can be attributed to the manufacturing of the beam shaping device (11), a new beam shaping device (11d) can be manufactured using different process parameters in order to reduce, preferably eliminate, the mismatch thus identified (see arrow "yes" from block (O) to block (D)). On the other hand, if no mismatch can be identified or the mismatch identified cannot be corrected in the manufacturing process, a new planning device design (11d) can be established. For example, a ridge filter (11f) initially designed with pins as illustrated in Figs. 3(a) and 3(b) can be difficult to produce due to the required accuracy (e.g. long pins and / or thin pins). The corresponding ridge filter (11f) can be designed with apertures instead of pins as illustrated in Fig. 3(c), which can be manufactured without defects.

[0108] The present method is very advantageous for the following reasons:

[0109] • The present method does not require at all the use of a particle accelerator system, which is usually very tight in its schedule and which is also expensive in terms of energy usage,

[0110] • The present method can be applied to any type of beam shaping device (11),

[0111] • The present method compares the calculated dose distribution (cDD) with the reference dose distribution (rDD), which is a direct information on whether the beam shaping device (11) is suitable for delivering the treatment plan,

[0112] • only in case the calculated dose distribution (cDD) differs from the reference dose distribution (rDD) by more than a predetermined level (e.g. gamma evaluation), the analysis of the mismatch of geometry and / or density between the actual CT image (11a) and the planned device design (11d) is performed,

[0113] • in many cases, the present method allows to assign a given mismatch of geometry and / or density between the actual CT image (11a) and the planned device design (11d) to a corresponding mismatch between the calculated dose distribution (cDD) and the reference dose distribution (rDD), whereby the present method can be implemented under new processing conditions to eliminate the latter mismatch,

[0114] • in some cases, the present method can teach that the planned device design (11d) cannot be easily manufactured, while an alternative planned device design can be advantageous.

[0115]

[0116]

Claims

1. A method for assessing the quality of a beam-shaping device (11) manufactured according to a planned device design (11d) of the beam-shaping device for shaping one or more accelerated particle beams (100.i) emitted by a particle accelerator system, the method comprising the steps of: (a) establishing a planned device design (11d) of the beam-shaping device (11) using a treatment planning system, TPS, which is adapted to shape one or more accelerated particle beams (100.i) so as to match the geometry of a treatment volume (V) of tissue comprising tumor cells (3t) for depositing a specific dose (Dij) to a specific location within the treatment volume (V) so as to define a planned dose distribution (pDD) that fulfills a target for dose deposition in the treatment volume set by a physician, (b) manufacturing the beam-shaping device (11) according to the planned device design (11d), (c) establishing a CT scan of the beam-shaping device (11) to produce an actual CT image (11a), (d) determining the dimensions and local material density of the beam-shaping device (11) from the actual CT image (11a), characterized in that the method comprising the steps of: (e) extending the actual CT image (11a) to include the treatment volume (V), (f) determining a calculated dose distribution in the treatment volume (V) using a dose engine, which is an algorithm that calculates dose distributions from a description of geometry and beam properties, by virtually irradiating the treatment volume (V) using one or more beams through a virtual beam-shaping device having a geometry and density defined by the actual CT image (11a), (g) comparing the calculated dose distribution (cDD) to a reference dose distribution (rDD).

2. The method of claim 1, wherein, Step (e) further comprises extending the actual CT image (11a) to include one or more of ridge filters (11f), range shifters (11s) and / or range compensators (11c) not included in the beam-shaping device (11).

3. The method of claim 1, wherein, The method further comprises calculating the planned dose distribution (pDD) using a dose engine in the TPS used in step (a), and wherein the dose engine used in step (f) is the same as the dose engine used in step (a).

4. The method of any one of claims 1-3, wherein, The reference dose distribution (rDD) is: • a calculated planned dose distribution (cpDD) obtained by: o first, creating a high-resolution planned device design (11hrd) using the TPS, which corresponds to the planned device design (11d) but has a higher resolution matching the resolution of the actual CT image, and o second, determining the calculated planned dose distribution (cpDD) using the dose engine by virtually irradiating the treatment volume (V) using the high-resolution planned device design (11hrd) using the one or more beams; or • the planned dose distribution (pDD) or a function thereof.

5. The method of claim 4, wherein, Said reference dose distribution (rDD) is either: • said calculated planned dose distribution (cpDD) and wherein said high resolution planning device design (11hrd) is created with a voxel size equal to the CT voxel size used to acquire said actual CT image, with a tolerance of ± 20%, or • a function of said planned dose distribution (pDD) which is the result of a transformation of said planned dose distribution (pDD) calculated at a given spatial resolution and which changes the voxel size of said planned dose distribution (pDD) to match said CT voxel size, with a tolerance of ± 20%.

6. The method of claim 5, wherein, Said tolerance is ± 10%.

7. The method of claim 5, wherein, Changing the voxel size of said planned dose distribution (pDD) comprises using linear interpolation or nearest neighbor interpolation on said planned dose distribution (pDD).

8. The method of any one of claims 1-3 and 5-7, wherein, The CT scan of said beam shaping device (11) is performed with a CT voxel size not exceeding 0.5 mm.

9. The method of claim 8, wherein, The CT scan of said beam shaping device (11) is performed with a CT voxel size not exceeding 0.2 mm.

10. The method of claim 8, wherein, Said planning device design (11d) is created using a TPS from said treatment plan (TP) with a voxel size similar to or smaller than said CT voxel size.

11. The method of any one of claims 1-3, 5-7, and 9-10, wherein, Said comparison of said calculated dose distribution (cDD) with said reference dose distribution (rDD) is performed using a gamma evaluation and wherein said beam shaping device (11) is considered to be in conformity with said planning device design (11d) if a gamma value (γ) lower than or equal to a reference gamma value (γr) is obtained in a predetermined percentage of voxels of said reference dose distribution (rDD), wherein γ is defined as the minimum value of the following function: wherein |d(cDD)-d(rDD)| is the distance between analysis points; |D(cDD)-D(rDD)| is the dose difference, DTA and ΔD are scaling factors.

12. The method of claim 4, wherein, Said comparison of said calculated dose distribution (cDD) with said reference dose distribution (rDD) is performed using a gamma evaluation and wherein said beam shaping device (11) is considered to be in conformity with said planning device design (11d) if a gamma value (γ) lower than or equal to a reference gamma value (γr) is obtained in a predetermined percentage of voxels of said reference dose distribution (rDD), wherein γ is defined as the minimum value of the following function: wherein |d(cDD)-d(rDD)| is the distance between analysis points; |D(cDD)-D(rDD)| is the dose difference, DTA and ΔD are scaling factors.

13. The method of claim 11, wherein, If said gamma value (γ) is higher than said reference gamma value (γr) in a predetermined number of voxels of said reference dose distribution (rDD), said beam shaping device (11) is not compliant and said dimensions and local material density determined from said actual CT image (11a) are compared to said planning device design (11d) to identify deviations of said dimensions and local material density of said beam shaping device (11) from said planning device design (11d). If said gamma value (γ) is higher than said reference gamma value (γr) in a predetermined number of voxels of said reference dose distribution (rDD), said beam shaping device (11) is not compliant and said dimensions and local material density determined from said actual CT image (11a) are compared to said planning device design (11d) to identify deviations of said dimensions and local material density of said beam shaping device (11) from said planning device design (11d).

14. The method of claim 12, wherein, If the gamma value (g) is higher than the reference gamma value (gr) in a predetermined number of voxels in the reference dose distribution (rDD), the beam-shaping device (11) is not qualified and the dimensions and local material density determined from the actual CT image (11a) are compared to the computed high-resolution plan device design (11hrd) to identify deviations of the dimensions and local material density of the beam-shaping device (11) from the plan device design (11d).

15. The method according to claim 13 or 14, wherein, • if one or more deviations are identified, the method comprises investigating manufacturing parameters that can have caused the one or more deviations and correcting the manufacturing parameters accordingly, and • if no deviations are identified, the method comprises repeating step (a) in claim 1 and establishing an alternative plan device design for the beam-shaping device (11).

16. The method of any one of claims 1-3, 5-7, 9-10, and 12-14, wherein, The actual CT image (11a) of the device obtained by CT scanning of the beam-shaping device (11) is in the form of a gray-scale image characterized by CT Hounsfield units, HU, and wherein the local material density is determined by transforming the HU into a corresponding density.

17. The method of any one of claims 1-3, 5-7, 9-10, and 12-14, wherein, The beam-shaping device (11) is produced by 3D printing or machining in a material selected from a polymer, or a metal, or any combination thereof.

18. The method of any one of claims 1-3, 5-7, 9-10, and 12-14, wherein, The beam-shaping device (11) is a ridge filter (11f) comprising a set of energy degrading units (11.i), wherein each energy degrading unit (11.i) is configured to reduce an initial energy (E0) of a corresponding charged particle beam (100.i) to a reduced energy (Ei) such that a dose is deposited into the treatment volume (V) according to the planned dose distribution (pDD).

19. The method of claim 18, wherein, The energy degrading unit (11.i) is in any one of: • apertures arranged side-by-side in a support base (11fb) having a thickness (xb) measured along the beam axis (X), each aperture extending from an aperture opening at a surface of the support base (11fb) and penetrating to a given depth measured along the beam axis (X), or • pins arranged side-by-side and supported on the support base (11fb), each pin extending from the support base along the beam axis (X).

20. The method of claim 19, wherein, The energy degrading unit (11.i) is formed by one or more sub-degrading units (11.ij). The energy degrading unit (11.i) is formed by one or more sub-degrading units (11.ij).

Images