Computer-implemented method for reducing the risk of interrupting a radiation therapy session due to deviations from planned values ​​of operating parameters of a particle acceleration system

Abstract

The present invention relates to a computer-implemented method for optimizing tolerance values of operating parameters of a particle acceleration system allowing a plurality of particle pencil beams accelerated along an irradiation axis (Z) to deposit a dose into a structure of interest of a patient according to a treatment plan (= TP) by means of a pencil beam scanning (= PBS). The method calculates a dose (rate) volume histogram (cD(R)VH) of a statistically representative number N of values randomly selected within a predefined confidence level (CLj) of a preselected tentative statistical distribution of these operating parameters and compares the thus obtained cD(R)VH with an acceptable variation band (BV) of a target D(R)VH. Once the tentative statistical distribution (Tj) yields N calculated cD(R)VHs all falling within the acceptable variation band (BV), the tentative statistical distribution is set as a final statistical distribution (Tf) and the particle acceleration system can be programmed with the final statistical distribution (Tf).

Description

Technical Field

[0001] The present invention belongs to the general field of treating tumor cells by irradiating them with accelerated particles, such as protons. In particular, the present invention relates to a method for setting the operating parameters of a given particle acceleration system to ensure that the given particle acceleration system can deliver accelerated particle beamlets that meet the requirements of a treatment plan (TP) included in an acceptable band of variation (BV) within a predefined confidence level. The advantage of this method is that it reduces the risk of having to abruptly interrupt a patient's irradiation treatment due to the given particle acceleration system failing to deliver one or more beamlets that meet the TP within the acceptable band of variation to within a predefined confidence level. Background Art

[0002] Radiation therapy using particles or waves (such as proton beams, electron beams, heavy ion beams, X-rays, gamma rays, etc.) has become an essential tool for treating cancer patients.

[0003] Pencil beam scanning (PBS) is a technique that involves directing a charged particle beamlet toward a target that includes tumor cells that define a structure of interest. PBS reduces unnecessary radiation exposure to surrounding non-cancerous cells by shaping the treated area to reflect the geometry of the tumor at the structure of interest. Pencil beam scanning can treat tumors with a single beam consisting of individual beamlets or with multiple beams (each consisting of individual beamlets) with different orientations, sometimes referred to as intensity modulated proton therapy (IMPT). In addition to the geometry of the target, PBS allows for local adjustment of the parameters of the beamlets based on the position within the target. These parameters can include the position and monitoring unit of each beamlet, as well as the scanning sequence of the beamlets, and the start and end times of each beamlet.

[0004] Since both tumor cells and healthy cells are damaged by this type of radiation, an important challenge in cancer treatment is to define a treatment plan (TP) that ensures that tumor cells are effectively destroyed or killed while preserving healthy cells as much as possible, especially those adjacent to these tumor cells. The first step in treatment planning is to capture images of the tumor area through CT scanning. Based on these images, oncologists identify the correct target and determine the location and dose for deposition to kill tumor cells. Such plans must meet multiple, often competing parameters and are therefore very complex. For this reason, treatment planning is usually performed by computer.

[0005] A treatment plan (TP) typically includes the definition of an array of n beamlets (bi), including values ​​for planning parameters including:

[0006] the planned position of each beamlet, which defines the position of the spot at which each beamlet is aimed, depending inter alia on the position, size and geometry of the tumor cells within the structure of interest,

[0007] • The planned monitoring units (MUpi) per beamlet, which is related to the number of particles that pass through the nozzle of the particle acceleration system and must reach a given spot in the structure of interest.

[0008] The planned beamlet scanning sequence at these planned positions (Xpi); this is important because the beamlets will deposit a certain dose into the corresponding spot, but may also deposit a lower dose into adjacent spots, which cannot be ignored when calculating the total dose deposited in each spot; this is especially true for rapid treatments as described in more detail below.

[0009] The treatment plan must ensure that at the end of treatment a total target dose greater than or equal to the minimum target dose that effectively destroys / kills the tumor cells has been delivered to the tumor cells. This can be defined by a target dose volume histogram (=tDVH) of the structure of interest. An example of a tDVH is represented by a solid line in FIG1( a), which shows a graph plotting the volume (%) of the structure of interest that must receive at least the target dose defined by the abscissa of the curve of FIG1( a). Oncologists also define an acceptable variation band (BV) within which the DVH can deviate from the target tDVH, and this variation band is represented by a dotted line in FIG1( a).

[0010] Historically, treatment plans for radiation therapy have involved delivering radiation doses to therapeutic cells at a conventional dose deposition rate (CDR) of less than 1 Gy / s. With very few exceptions, current radiation therapy devices deliver doses at a rate of approximately 0.1 Gy / s, and most clinical protocols involve delivering several target fractions of 2 to 15 Gy per day, cumulatively reaching a total target dose that typically exceeds the tolerance limit of normal tissues located in the radiation field, thereby damaging these normal tissues along with tumor cells. Recently, it has been observed that the same dose has different effects on healthy cells, but not on tumor cells, when deposited at a conventional dose deposition rate (CDR) or an ultra-high dose deposition rate (HDR); HDR can be one or more orders of magnitude greater than the commonly applied conventional dose deposition rate (CDR). Charge deposition at an ultra-high dose deposition rate (HDR) is also known as FLASH-radiotherapy (=FLASH-RT). Experiments in animals and various organs have demonstrated that ultra-high rate dose deposition at HDR can significantly spare healthy tissue compared to conventional deposition of the same dose at CDR, while tumor cells respond equally or better to HDR deposition than to CDR deposition. For example, FLASH-RT has been reported to cause a significant reduction in the incidence of pulmonary fibrosis, post-irradiation brain memory loss, and small intestinal necrosis in mice, while maintaining antitumor efficacy. Such specific normal tissue sparing has been demonstrated in large animals, and a patient with cutaneous lymphoma has been treated with FLASH-RT.

[0011] The dose rate distribution in the tissue can be defined by a target dose rate volume histogram (=tDRVH) of the structure of interest. An example of tDRVH is represented by a solid line in FIG1(b), which shows a graph plotting the volume (%) of the structure of interest that must receive a target dose rate or a higher dose rate of the tDRVH defined by the abscissa of the curve of FIG1(b). Oncologists also define an acceptable variation band (BV) within which the DRVH can deviate from the target tDRVH, and this variation band is represented by a dashed line in FIG1(b).

[0012] DVH and DRVH are cumulative histograms. However, there are other ways of representing the distribution of dose or dose rate. For example, Figure 1(c) shows a differential dose rate histogram (DDRH), which indicates the number of voxels in the structure of interest that receive dose at the corresponding dose rate indicated in the abscissa. The acceptable variation band (BV) is represented by a dotted line, and the long dashed line is the actual value measured during the treatment course (=aDDRH). Other representations are possible. All types of representations of the desired dose and dose rate distribution in the structure of interest that the treatment plan must achieve are collectively referred to herein as "dose distribution histograms (DDH)" and "dose rate distribution histograms (DRDH)". For the sake of brevity, the expression "dose (rate) distribution histogram (D(R)DH)" is also used herein to include both DDH and DRDH.

[0013] Implementing a planned beamlet scanning sequence may require defining a planned start time and a planned end time for each beamlet. This is particularly important for FLASH-RT.

[0014] The treatment planning system (TPS) defines the beamlet parameters including positions (Xj), monitoring units (MUj) and spot sequences to achieve the treatment plan (TP). The conversion system (=TS) determines the operating parameters of a given particle acceleration system required to implement the beamlet parameters, taking into account the limitations of the particle acceleration system. This is described in, for example, US20200298020, EP3932482 and EP3932481A1. The operating parameters are determined to ensure that the beamlet delivered by a given particle acceleration system will deposit the dose into the structure of interest according to the target dose (rate) distribution histogram (D(R)DH) within an acceptable variation band (BV). The conversion of TP to machine operating parameters is very important to ensure that the treatment course can be completed within the treatment plan and will not be interrupted because, at some point, the random variability of some operating parameters causes the beamlet actually delivered by the particle acceleration system to produce a D(R)DH that falls outside the acceptable variation band (BV).

[0015] A given particle accelerator system may not be able to meet the nominal values ​​of the operating parameters defined in the plan exactly. Instead, the particle accelerator system will actually operate at operating parameter values ​​that follow the Figure 2 The presented graph consists of a mean value (μj) which is a nominal value defined by the treatment plan and a variance (σj) which represents the random variability of the treatment machine. 2 ) is characterized by a specific statistical distribution (Tj). This means that even if the TS correctly converts the treatment plan (TP) into operating parameters to ensure that a given particle acceleration system delivers a beamlet characterized by a mean value (μj), the actual values ​​of the operating parameters during a particular treatment session will be distributed around the mean value (μj) on the statistical distribution curve (Tj) (refer to Figure 2 ). Therefore, it is clear that during a treatment session, the operating parameters will deviate from the mean (μi) following their distribution. In some cases, the actual operating parameter values ​​of the beamlets may produce a D(R)DH that extends beyond (exceeds) the corresponding acceptable variation band (BV), necessitating interruption of the treatment session despite correct TP switching by the TS. If the treatment session does not proceed as planned, it may pose a risk to the patient.

[0016] Some particle accelerator systems are equipped with monitoring equipment that measures the actual operating parameters of the beamlets as they are delivered through the nozzle. EP2116277, EP3375484, US10456598, EP3222322, WO 2020249565 and EP2833970 describe examples of equipment for in situ monitoring and verification of the selection of operating parameters for the beamlets delivered by a particle accelerator system. If the operating parameters of one or more beamlets so monitored differ from the planned values, which is likely to happen, there is a risk that the corresponding D(R)DH will extend beyond the acceptable band of variation (BV). If this happens, the treatment session must be stopped. This is very uncomfortable for the patient, who may have to return later to complete the treatment session according to the typically tight schedule of particle accelerator systems. It is therefore important to consider the distribution of the operating parameters to ensure that a treatment session can be completed within a predefined confidence level (CLj) with all beamlets producing a D(R)DH that conforms to the treatment plan.

[0017] Furthermore, the fact that an actual value of an operating parameter differs from a corresponding planned value does not necessarily mean that the corresponding D(R)DH extends beyond the acceptable variation band. Given the computing power of processors available to date, it is not conceivable to calculate a corresponding calculated D(R)DH for each measurement of an actual value of an operating parameter in order to decide whether to interrupt a treatment course because an actual value differs from a planned value of the operating parameter.

[0018] EP3498336 describes a system and method for treating a dummy (mannequin) and assessing the DVH by dosimetry before administering treatment to a patient. This technique significantly reduces the risk of having to interrupt a treatment course, but it also requires blocking the particle accelerator system for the time required to perform the test, during which time the particle accelerator system cannot be used to treat the patient. In addition, dosimetry tolerances can translate into different machine tolerance levels for spots at different locations or with different MUs (e.g., a spot on the edge of a structure of interest may have higher constraints on position accuracy than a spot at the center of the structure). It is also not obvious how to translate dose rate tolerances into tolerances for each spot in the spot map. In some places, it is more important to check whether the irradiation is in FLASH-RT mode than in other places. The tolerance on dose rate may also differ between locations in the tissue. For example, FLASH-RT may be required at the edge of a structure of interest, where tumor cells are flanked by healthy cells that must be preserved.

[0019] To date, most methods for ensuring that operating parameters will produce the desired D(R)DH have been a posteriori, i.e., by measuring the therapeutic properties of the beamlets delivered by the particle accelerator system, or at best based on dosimetry testing on dummies (mannequins) prior to treating patients. There remains a need in the art for a method for computationally determining (i.e., without having to use precious accelerator time and energy) a set of operating parameters of a particle accelerator system to be used that will produce the desired D(R)DH within an acceptable band of variation (BV) within a predefined confidence level.

[0020] The present invention proposes a computer-implemented method for optimizing the tolerance values ​​of the operating parameters of a particle acceleration system that allows a beam formed from a plurality of accelerated particle beamlets to deposit a dose to a patient according to a treatment plan (=TP) by pencil beam scanning (=PBS). The method allows determining a confidence level (CLj) that the beamlets delivered by a given particle acceleration system according to a set of operating parameters will meet the TP. If the confidence level thus obtained is too low, an alternative set of operating parameters needs to be evaluated. These and other advantages of the present invention continue to emerge. Summary of the Invention

[0021] The invention is defined in the accompanying independent claims. Preferred embodiments are defined in the dependent claims. In particular, the invention relates to a computer-implemented method for optimizing tolerance values ​​of operating parameters of a particle acceleration system, which particle acceleration system allows a beam formed by a plurality of particle beamlets accelerated along an irradiation axis (Z) to deposit a dose into a structure of interest of a patient by pencil beam scanning (PBS) according to a treatment plan (=TP), the computer-implemented method comprising,

[0022] (a) providing input, which includes

[0023] o The treatment plan (=TP), which comprises the definition of an array of beamlets (bi) characterized by planning parameters including:

[0024] the planned position (Xpi) of each beamlet (bi) on the plane (X, Y) normal to the irradiation axis (Z),

[0025] The planned monitoring units (MUpi) for each beamline,

[0026] The planned beamlet scan sequence at these planned positions (Xpi),

[0027] o the planned start time (t0pi) and end time (t1pi) at which each beamlet is delivered,

[0028] o the definition of the structure of interest, which defines the tissue or tissues through which the plurality of beamlets (bi) pass,

[0029] ○ Values ​​of one or more target dose (rate) distribution histograms (=tD(R)DH) of a structure of interest obtained by treatment performed using these planning parameters, the target dose (rate) distribution histogram comprising a dose distribution histogram (=tDDH) and / or a target dose rate distribution histogram (=tDRDH), wherein the dose distribution histogram (DDH) is preferably a dose distribution volume histogram (DDVH), and the dose rate distribution histogram (DRDH) is preferably a dose rate distribution volume histogram (DRDVH) or a differential dose rate histogram (DDRH).

[0030] o the acceptable band of variation (BV) within which one or more tD(R)DHs are allowed to vary,

[0031] (b) providing a tentative statistical distribution (Tj) of operating parameters of the particle acceleration system centered about corresponding mean values ​​(μj) representing the performance of the particle acceleration system, and defining a confidence level (CLj) of the tentative statistical distribution (Tj), wherein the operating parameters include:

[0032] ○ Monitoring unit (MUj) for each beamlet,

[0033] ○ The position of each beamlet (Xj), and

[0034] o the start time (t0j) and end time (t1j) of delivering each beamlet,

[0035] (c) randomly selecting the value of the monitoring unit (MUij), the value of the position of the beamlet (Xij), and the value (t0ij, t1ij) of each of the start time (t0ij) and the end time (t1ij) from the corresponding tentative statistical distribution (Tj) within a predefined confidence level (CLj),

[0036] (d) using the thus randomly selected values ​​to calculate one or more calculated dose (rate) distribution histograms (=cD(R)DHj-run x),

[0037] (e) repeating the last two steps (c), (d) a statistically representative number of times (N) to produce calculated distributions (CDj) characterizing one or more so calculated cD(R)DHj-runs x for all so randomly selected values ​​of the operating parameters,

[0038] (f) comparing the calculated distribution (CDj) of the one or more cD(R)DHj with the corresponding acceptable variation band (BV), and determining whether the calculated distribution (CDj) of the one or more cD(R)DHj is included in the corresponding acceptable variation band within the predefined confidence level (CLj).

[0039] The final statistical distribution (Tf) can be set by a human operator or a processor as the corresponding operating parameter with a final confidence level (CLf) as follows,

[0040] if for a given treatment plan (TP), the calculated distributions (CDj) of the one or more cD(R)DHj are all included within the corresponding acceptable variation band (BV) with the predefined confidence level (CLj), setting the tentative statistical distribution (Tj) to the final statistical distribution (i.e., Tf=Tj) and setting the confidence level (CLj) to the corresponding final confidence level (CLf=CLj) to define corresponding operating parameters,

[0041] If any of the one or more cD(R)DHj calculated with a set of randomly selected values ​​within the corresponding confidence level (CLj) of the statistical distribution of the operating parameters (Tj) extends beyond the corresponding acceptable variation band (BV), then steps (b) to (f) as defined above are repeated according to the following conditions,

[0042] o Use a new tentative statistical distribution (T(j+k)) of these operating parameters and / or

[0043] ○ Choose a new, less demanding confidence level (CL(j+k)),

[0044] Until the calculated distributions (CD(j+k)) of the one or more cD(R)DH(j+k) are all included in the corresponding acceptable variation band (BV), and the tentative statistical distribution (T(j+k)) is set to the final statistical distribution (i.e., Tf=T(j+k)) and the corresponding confidence level (CL(j+k)) is set to the final confidence level (CLf=CL(j+k)) to define the corresponding operating parameters.

[0045] The new tentative statistical distribution (T(j+k)) may have a lower standard deviation (σj) than the corresponding tentative statistical distribution (Tj) defined above. The tentative statistical distribution (Tj) for each of these operating parameters is preferably a Gaussian distribution, and the value of the confidence level (CLj) may be comprised between 68% and 99.7% of the tentative statistical distribution, preferably between 95.5% and 99%. Note that a 68% confidence level (CLj) corresponds to μj±σj, a 95% confidence level (CLj) corresponds to μj±2σj, and a 99.7% confidence level of the tentative statistical distribution corresponds to μj±3σj, where μj is the mean value and σj is the standard deviation of the corresponding tentative statistical distribution. The mean value (μj) and standard deviation (σj) of each operating parameter may be different for each beamlet (bi).

[0046] The particle acceleration system can be equipped with a cyclic checking module configured to measure actual values ​​of the operating parameters at different intervals or continuously, the operating parameters including monitoring units (MUai), positions (Xai), and start and end times (t0ai, t1ai) of the beamlets emitted by the particle acceleration system. The particle acceleration system can also be equipped with a processor configured to compare the actual values ​​of the operating parameters with corresponding confidence levels (CLj) and to stop the treatment course if an actual value of the operating parameter falls outside the corresponding confidence level (CLj).

[0047] The planning parameters also include a planned beamlet size (dj) and a planned beam current (Ij), the respective values ​​(dij, Iij) of which are used to calculate the one or more calculated dose (rate) distribution histograms (=cD(R)DH), which are randomly selected within the corresponding tentative statistical distribution (Tj) of the beamlet size and the planned beam current (Ij).

[0048] The calculated distribution (CDj) of the one or more cD(R)DH is defined by a corresponding area included in an envelope defined between a minimum calculated dose (rate) distribution histogram (=cD(R)DHj0) on the one hand and a maximum calculated dose (rate) distribution histogram (=cD(R)DHj1) on the other hand. cD(R)DHj0 is defined by the lowest value of cD(R)DHj calculated with a predefined confidence level (CLj) from N randomly selected values ​​of the monitoring unit (Muij), the position of the beamlets (Xij), and the start and end times (t0ij, t1ij), and cD(R)DHj1 is defined by the highest value of cD(R)DHj calculated with the predefined confidence level (CLj) from N randomly selected values ​​of the monitoring unit (Muij), the position of the beamlets (X0i), and the start and end times (t0ij, t1ij).

[0049] The treatment plan may include depositing a dose into at least a portion of the structure of interest at an ultra-high deposition rate (UHDR), which is defined as a deposition rate greater than or equal to 1 Gy / s.

[0050] The invention also relates to an error prediction module configured to implement the method as defined above, the error prediction module comprising,

[0051] a memory comprising, for each operating parameter, a plurality of tentative statistical distributions (Tj) centered around a plurality of corresponding mean values ​​(μj),

[0052] A user interface configured to:

[0053] o Input a treatment plan (TP) comprising one or more target dose (rate) distribution histograms (= tD(R)DH) and corresponding acceptable variation bands (BV), the target dose (rate) distribution histograms comprising a target dose distribution histogram (= tDDH) and / or a target dose rate distribution histogram (= tDRDH),

[0054] o Select from the memory or enter the planned start time (t0pi) and end time (t1pi) at which each beamlet is to be delivered,

[0055] o selecting from the memory or inputting for each beamlet a first tentative statistical distribution (Tj) of each operating parameter including the monitoring unit (MUj), the position (Xj) of the beamlet, and the start and end times (t0i, t1i),

[0056] ○ Enter the confidence levels (CLj) for these operating parameters,

[0057] a processor configured to:

[0058] (i) randomly selecting the value of the monitoring unit (MUij), the value of the position of the beam (Xij), the value of the start time and the value of the end time (t0ij, t1ij) included in the predefined confidence level (CLj) of the corresponding tentative statistical distribution (Tj),

[0059] (ii) calculating one or more of a calculated dose distribution histogram (=cDDHj) and a calculated dose rate distribution histogram (=cDRDHj) using the thus randomly selected values,

[0060] (iii) The last two steps are repeated a statistically representative number of times (N) to produce the calculated distribution (CDj) of the cDDHj and cDRDHj thus calculated for each beamlet.

[0061] The processor can be further configured to, in the event that any one of the one or more calculated distributions (CDj) of cDDHj and cDRDHj is not included in the corresponding acceptable variation band with a predefined confidence level (CLj), repeat the above steps (i) to (iii) using a new tentative statistical distribution (T(j+k)) of these operating parameters until the calculated distributions (CDj) of one or more of cDDHj and cDRDHj are included in the corresponding acceptable variation band with a predefined confidence level (CLj). BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more fully understand the essence of the present invention, reference is made to the following detailed description in conjunction with the accompanying drawings, in which:

[0063] Figure 1(a) to Figure 1(c) : Shows examples of tDVH, tDRVH and tDDRH curves and their corresponding acceptable variation bands (BV).

[0064] Figure 2 : shows the Gaussian distribution of the operating parameters of a given particle acceleration system.

[0065] Figure 3(a) to Figure 3(c) : Various steps of the method of the present invention with positive results are shown.

[0066] Figure 4(a) to Figure 4(c) : Various steps of the method of the invention with negative results are shown.

[0067] Figure 5(a) to Figure 5(d) : shows N computational runs that resulted in the calculated distributions (CDj) of cD(R)DHj thus calculated as shown in FIG3( b ) and FIG3( c ).

[0068] Figure 6: A flow chart showing the various steps of the method of the present invention is shown. DETAILED DESCRIPTION

[0069] The present invention relates to a computer-implemented method and an error prediction module that significantly reduces the risk of performing a treatment session that does not adhere to the corresponding treatment plan due to device reasons.

[0070] Method for optimizing tolerance values ​​of operating parameters

[0071] The present invention relates to a computer-implemented method for optimizing the tolerance values ​​of operating parameters of a particle acceleration system, which allows a beam formed by a plurality of particle beamlets accelerated along an irradiation axis (Z) to deposit a dose to a patient according to a treatment plan (=TP) by pencil beam scanning (=PBS). The particles are preferably protons, but they can also be electrons, heavy ion beams or waves formed by the interaction of accelerated particles with a conversion material, such as x-rays (or gamma rays). Figure 6 The computer-implemented method presented requires input of values ​​for a number of planning parameters. These values ​​may be provided by a treatment plan (TP) comprising the following planning parameters,

[0072] the planned position (Xpi) of each beamlet (bi) on the plane (X, Y) normal to the irradiation axis (Z),

[0073] The planned monitoring units (MUpi) for each beamline,

[0074] The planned beamlet scan sequence at these planned positions (Xpi),

[0075] The method also requires the planned start time (t0pi) and end time (t1pi) of each beamlet to be delivered. This is particularly important in the case of FLASH-RT. The start and end times may or may not be part of the TP.

[0076] A structure of interest must be defined to characterize one or more tissues that are traversed or interacted with by the beamlet (bi). The structure of interest includes the target, which includes the tumor cells to be killed, and also includes healthy tissue that is traversed or in some way contacted by the beamlet (bi). For example, this healthy tissue includes tissue upstream of the target along the irradiation axis (Z), i.e., tissue between the particle accelerator nozzle and the target, or tissue adjacent to and surrounding the target.

[0077] The goal of treatment is to produce a given value of the target dose (rate) distribution histogram (tD(R)DH) of the structure of interest at the end of the treatment course within an acceptable variation band (BV) that allows for variations in tD(R)VH. Figures 1(a) to 1(c)As shown, tD(R)DH can include, for example, a target dose volume histogram (=tDVH) (see FIG1(a)), a target dose rate volume histogram (=tDRVH) (see FIG1(b)), or a target differential dose rate histogram (tDDRH) (see FIG1(c)). Note that other histogram representations are possible and are included in the term tD(R)DH to display information about dose and dose rate distribution. Figure 1(a) to Figure 1(c) Examples of tDVH, tDRVH, and tDRDH represented by solid lines, and examples of corresponding acceptable variation bands (BVs) represented by the areas included between the dashed lines are shown. Similar acceptable variation bands can be defined for any type of alternative D(R)DH histogram representation. As shown in Figures 3(b) and 3(c), if the beamlet delivered by the particle accelerator system achieves a DVH and DRVH that are included within the acceptable variation band, then the treatment session is successful. As shown in Figures 4(b) and 4(c), if the DVH or DRVH falls outside the acceptable variation band (BV), the treatment session must be interrupted, which can be very uncomfortable for the patient, who must wait for a free time slot in the particle accelerator system schedule to resume the treatment session.

[0078] FIG1( c ) includes the actual values ​​of the dose rate distribution histogram (aDDRH) obtained when measuring the beamlet parameters during the treatment course (referenced by the long dashed line). As can be seen, the curve aDDRH in FIG1( c ) is completely enclosed within the acceptable variation band (BV). Therefore, the treatment course was successful. However, if the curve aDDRH extends beyond the acceptable variation band (BV), the treatment course would have to be interrupted. In order to reduce the risk of having to interrupt the treatment course due to the actual D(R)DH falling outside the acceptable variation band, the present invention proposes to implement the following measures using a computer.

[0079] First, as shown in Figures 3(a) and 4(a), a tentative statistical distribution (Tj) of the operating parameters (MUj, Xj, t0j, t1j) of the particle accelerator system is selected, centered around the corresponding mean value (μj) representing the performance of the particle accelerator system. A confidence level (CLj) for each of the tentative statistical distributions (Tj) of these operating parameters is defined. These operating parameters include:

[0080] ○ Monitoring unit (MUj) for each beamlet,

[0081] ○ The position of each beamlet (Xj), and

[0082] o The start time (t0j) and end time (t1j) of delivery of each beamlet.

[0083] The selection of the tentative statistical distribution (Tj) is based on the performance of the particle acceleration system and the desire to generate a beam that delivers a sequence of beamlet scans and generates a calculated dose volume histogram (=cDVHj) and a calculated dose rate volume histogram (=cDRVHj) in the structure of interest that is within an acceptable band of variation (BV). For example, the treatment planning system (TPS) may determine an average value (μi) achievable by the particle acceleration system that will produce the desired cDVH and cDRVH. However, the actual operating parameters of the treatment machine on a particular day are not limited to the corresponding average value (μi), but are typically distributed on a Gaussian curve that varies from day to day or throughout the day (see Figure 2 ). Therefore, the resulting cD(R)DH histogram will vary depending on the actual value of the operating parameter at a particular time of treatment and must therefore be calculated with the distribution in mind. It is impractical to consider all possible values ​​of each operating parameter defined by the corresponding distribution curve. A confidence level (CLj) can be defined to constrain the distribution to what is considered to be acceptable boundaries. For example, the confidence level can be defined relative to the standard deviation (σj), for example, μj ± nσj, where n = 1 to 3.

[0084] The values ​​of the monitoring unit (MUij), the position of the beamlet (Xij), and the start time (t0ij) and end time (t1ij) are randomly selected from the corresponding tentative statistical distribution (Tj) within the corresponding confidence level (CLj); these values ​​are shown as black circles in the Gaussian distribution curve (Tj) in Figures 3(a) and 4(a). These values ​​are selected within the previously defined confidence level (CLj).

[0085] As shown in FIG5(a), in the first run (=Run 1), the calculated dose (rate) distribution histogram (=cD(R)DHj) is calculated using the randomly selected values. If the calculated cD(R)DHj is contained within the corresponding acceptable variation band, new operating parameter values ​​are randomly selected from the tentative statistical distribution (Tj) and the corresponding cD(R)DHj is calculated in the second run (=Run 2) as shown in FIG5(b). As long as the calculated cD(R)DH-Run x is contained within the acceptable variation band (BV), these operations are repeated a statistical number of times (N) as shown in FIG5(c), i.e., "Run N". After N runs, the combination of N curves "cD(R)DH-Run x" with x = 1 to N defines the calculated distribution (CDj) as shown in FIG3(b) and FIG3(c), FIG4(b) and FIG4(c), and FIG5(d) as represented by the shaded area bounded by the dotted line.

[0086] On the other hand, if any of the cD(R)DHj-runs x thus calculated extends beyond the acceptable variation band (BV), it can be concluded that there is a risk, higher than a predefined confidence level (CLj), that the treatment run using the operating parameters of the particle acceleration system according to the tentative statistical distribution (Tj) will have to be interrupted. A new tentative statistical distribution (T(j+1)) of the operating parameters is then selected, and cD(R)DH(j+1) is calculated in a new series of N runs using the values ​​of the new tentative statistical distribution (T(j+1)) randomly selected in each successive run. This operation is repeated using the new tentative statistical distribution (T(j+k)) until the corresponding calculated distribution (CD(j+k)) is completely included in the acceptable variation band (BV).

[0087] In addition to the planned position (Xpi), the planned monitoring unit (MUpi), the planned beamlet scanning sequence at the planned position (Xpi), and the planned start time (t0pi) and end time (t1pi), the planning parameters may also include the planned beamlet size (dj) and beam current (Ij), the values ​​of which (dij, Iij)) used to calculate the calculated dose (rate) distribution histogram (=cD(R)DHj) are randomly selected within the corresponding tentative statistical distribution (Tj) of the beamlet size (dj) and beam current (Ij).

[0088] The calculated distribution (CDj) of cD(R)DHj can be defined by the corresponding region included between:

[0089] On the one hand, the minimum calculated dose (rate) distribution histogram (=cD(R)DHj0), and on the other hand,

[0090] Maximum calculated dose (rate) distribution histogram (=cD(R)DHj1),

[0091] in,

[0092] cD(R)DHj0 is defined by the lowest value of cD(R)DHj calculated with the predefined confidence level (CLj) from N randomly selected values ​​of the monitoring unit (Muij), the positions of the beamlets (Xij), and the start and end times (t0ij, t1ij), and where,

[0093] cD(R)DHj1 is defined by the highest value of cD(R)DHj calculated with the predefined confidence level (CLj) based on N randomly selected values ​​of the monitoring unit (Muij), the positions of the beamlets (X0i), and the start and end times (t0ij, t1ij).

[0094] By comparing the calculated distribution (CDj) of cD(R)DHj with the acceptable variation band (BV), it can be determined whether the calculated distribution (CDj) of cD(R)DHjj is included in the corresponding acceptable variation band for a predefined confidence level (CLj). Figures 3(b) and 3(c) illustrate examples of provisional statistical distributions (Tj) with associated confidence levels (CLj) that result in the calculated distribution (CDj) of cD(R)DHj being included within the corresponding acceptable variation band (BV). In contrast, Figures 4(b) and 4(c) illustrate examples of provisional statistical distributions (Tj) with associated confidence levels (CLj) that result in the calculated distribution (CDj) of cD(R)DHj extending beyond the corresponding acceptable variation band (BV), as indicated by the shaded area and black arrows. For example, in Figure 4(b), it can be seen that cD(R)DHj-Run 1 has extended beyond the acceptable variation band (BV). Calculations using the corresponding tentative statistical distribution may be stopped, and a new tentative statistical distribution (T(j+1)) may be selected to repeat the run N times using values ​​of operating parameters randomly selected from the new tentative statistical distribution (T(j+1)).

[0095] The method of the present invention can be concluded by setting the final statistical distribution (Tf) to the corresponding operating parameters with a final confidence level (CLf). For a given treatment plan (TP), if N runs of the calculated distribution (CDj) of cD(R)DHj are included in the corresponding acceptable variation band (BV) with a predefined confidence level (CLj), then the tentative statistical distribution (Tj) can be set to the final statistical distribution with the final confidence level (CLf=CLj) (i.e., Tf=Tj) to define the corresponding operating parameters. It can be concluded that by implementing the final statistical distribution (Tf) of the operating parameters, the particle acceleration system has a probability equal to the final confidence level (CLf) of delivering a fine beam that meets the treatment plan (TP), wherein the corresponding actual dose (rate) distribution histogram (aD(R)DH) is included in the acceptable variation band (BV), for example, as shown in Figure 1(c).

[0096] On the other hand, if any of the N runs of cD(R)DHj extends beyond the boundaries of the corresponding acceptable variation band (BV) with the predefined confidence level (CLj) (e.g., as shown in FIG4(b)), a new tentative statistical distribution (T(j+1)) and / or a new confidence level (CL(j+1)) may be defined for one or more of the operating parameters, and cD(R)DHj may be calculated N times using N randomly selected values ​​of each of the operating parameters. These operations may be repeated as needed using the new tentative statistical distribution (T(j+k)) and / or alternative confidence level (CL(j+k)) of the operating parameters until the calculated distribution (CD(j+k)) of cD(R)DHj(j+k) is included within the corresponding acceptable variation band (BV) of the predefined confidence level (CLj) or (CL(j+k)). Therefore, the final statistical distribution (Tf) can be set to the tentative statistical distribution (T(j+k)) (i.e., Tf=T(j+k)), and the corresponding confidence level (CL(j+k)) can be set to the final confidence level (CLf=CL(j+k)) to define the corresponding operating parameters.

[0097] A new tentative statistical distribution (T(j+k)) defined in the event that a previous tentative statistical distribution j+(k-1) does not produce a calculated distribution included within the band of variation (BV) of tD(R)DH around a predefined confidence level (CLj) may be selected to have a lower standard deviation (σj) (or variance (σj)) than the corresponding previous tentative statistical distribution (Tj+(k-1)). 2 ))'s distribution.

[0098] Setting the final statistical distribution (Tf) for the corresponding operating parameters may be performed by a human operator or automatically by a processor.

[0099] The tentative statistical distribution (Tj) of each of the operating parameters is preferably a Gaussian distribution. The value of the confidence level (CL) is preferably comprised between 68% and 99.7%, preferably between 95.5% and 99%, of the tentative statistical distribution. Figure 2 As shown, a 68% confidence level (CLj) corresponds to μj ± σj, a 95% confidence level (CLj) corresponds to μj ± 2σj, and a 99.7% confidence level of the tentative statistical distribution corresponds to μj ± 3σj, where σj is the standard deviation of the corresponding tentative statistical distribution. For example, in terms of absolute deviation, the position of a spot (Xj) may differ from the mean value μj of Xj by ± 1 mm. The monitoring unit (MUj) may, for example, vary by about 0.5% around the mean value (μj) of MUj. Note that the mean value (μj) and standard deviation (σj) of each operating parameter may be different for each beamlet (bi).

[0100] The particle acceleration system can be equipped with a cyclic checking module that is configured to measure the actual values ​​of the monitoring units (MUai), positions (Xai), and start and end times (t0ai, t1ai) of the fine beam emitted by the particle acceleration system at different intervals or continuously. The processor can be configured to determine whether any actual operating parameter falls outside the corresponding final confidence level (CLf) (refer to Figures 3(a) and 4(a), which show confidence levels (CLj) that only span a portion of the statistical distribution (Tj)). In this case, an alarm can be triggered to notify the operator of the event. As a safety measure, the processor can also be configured to stop treatment once it is detected that the operating parameter falls outside the confidence level (CLj).

[0101] The fact that one or more values ​​of the actual operating parameters fall outside the corresponding confidence level (CLj) does not necessarily mean that the resulting aD(R)DH falls outside the acceptable band of variation (BV). Therefore, interrupting a treatment session simply because one actual value of any operating parameter falls outside the corresponding confidence level (CLj) may be excessive, as it could perfectly well produce a D(R)DH that is included within the acceptable band of variation (BV). The processor can be configured to calculate the cD(R)DH whenever the measured actual values ​​of the operating parameters fall outside the corresponding confidence level (CLj) to determine whether they are included within the corresponding acceptable band of variation (BV). The cD(R)DH can be calculated based on the actual values ​​of the operating parameters measured on already delivered beamlets, including the parameters that fall outside the confidence level (CLj), and based on the mean value (μj) of the operating parameters of the final statistical distribution (Tf) of the beamlets still to be delivered to conclude the treatment session. If the cD(R)DH so calculated falls outside the acceptable band of variation (BV), the treatment session is stopped. On the other hand, if the calculated cD(R)VH is within the acceptable band of variation (BV), the treatment session can proceed further. The risk of stopping the treatment session is further reduced with this functionality of the processor.

[0102] This function does not require much computing power, since it only involves 100%-CLj% of the actual value of the operating parameter. For a 95% confidence level (CLj), i.e., μj ± 2σj, there is only a 5% probability that the value of the operating parameter will fall outside the confidence level (CLj) for which the D(R)DH calculation will have to be performed. For a 99.7% confidence level of the tentative statistical distribution, i.e., μj ± 3σj, there is only a 0.3% probability that such a calculation will have to be performed.

[0103] In a preferred embodiment, the treatment plan comprises a FLASH-RT, as the dose will be deposited in at least a portion of the structure of interest with an ultra-high deposition rate (UHDR), defined as a deposition rate greater than or equal to 1 Gy / s. In this embodiment, the plan parameters further comprise a beam current (I), the value of which (Ij) is used to compute the computed dose distribution histogram (= cDDHj) and, in particular, the computed dose rate distribution histogram (= cDRDHj), is randomly selected within the corresponding tentative statistical distribution (Tj) of the beam current.

[0104] error prediction module

[0105] The present application also relates to an error prediction module configured to implement the method as described above. The error prediction module comprises a memory comprising a plurality of tentative statistical distributions (Tj) centered on a plurality of corresponding average values (μj) of each operating parameter. The error prediction module further comprises a user interface configured to input,

[0106] • a planning operator comprising a treatment plan (TP) comprising a planned position (Xpi), a planned monitoring unit (MUpi) and a planned pencil beam scan sequence, a planned start time (t0pi) and an end time (t1pi),

[0107] • a target dose (rate) distribution histogram (= tD(R)DH) and a corresponding acceptable variation band (BV),

[0108] • for each pencil beam, a first tentative statistical distribution (Tj) selected from one or more of the plurality of tentative statistical distributions of each operating parameter comprising at least a monitoring unit (MUj), a position of the pencil beam (Xj), and a start time and an end time (t0i, t1i), or defined by inputting the value of its average (μj) and standard deviation (σj),

[0109] The error prediction module comprises a processor configured to,

[0110] • randomly select a value of the monitoring unit (MUij), a value of the position of the pencil beam (Xij), a value of the start time and the end time (t0ij, t1ij) comprised within a predefined confidence level (CLj) of the corresponding tentative statistical distribution (Tj),

[0111] • compute a computed dose (rate) distribution histogram (= cD(R)Dj) using the values thus randomly selected,

[0112] • The last two steps are repeated a statistically representative number of times (N) to produce the calculated distribution (CDj) of the cD(R)Dj thus calculated for each beamlet.

[0113] The processor can be further configured to, in a case where the calculated distribution (CDj) of cD(R)Dj is not included within the corresponding acceptable variation band with a predefined confidence level (CLj), repeat the aforementioned three steps using a new tentative statistical distribution (T(j+k)) of these operating parameters until the calculated distribution (CDj) of cD(R)Dj is included within the corresponding acceptable variation band with a predefined confidence level (CLj).

[0114] flow chart( Figure 6 )

[0115] Figure 6 A flow chart showing the various steps of the method of the present invention is shown. Figure 6 , step (1)) into the planning parameters of all thin beams (refer to Figure 6 , step (2)). One or more theoretical dose (rate) distribution histograms (tD(R)DH) are defined (refer to Figure 6 , step (3)) and its corresponding acceptable variation band (BV) (refer to Figure 6 , step (4)). As defined above, the dose (rate) distribution histogram (D(R)DH) may non-exhaustively include a dose volume histogram (DVDH), a dose rate volume histogram (DRVH) or a differential dose rate distribution histogram (DDRH), etc.

[0116] By defining the first tentative statistical distribution (Tj) and the corresponding confidence level (CLj) (refer to Figure 6 , steps (6) and (7)) simulate the particle acceleration system (refer to Figure 6 , step (5)). The value of each operating parameter is randomly selected within the confidence level (CLj) of the tentative statistical distribution (Tj) (refer to Figure 6 , step (8)), and calculate the cD(R)DHj-run (x=1) (refer to Figure 6 , step (9)). If the calculated CD(R)DH-run (x=1) is not within the acceptable variation band (BV), a new tentative statistical distribution (T(j+1)) with a corresponding confidence level (CL(j+1)) is selected (refer to Figure 6, steps (10) and (11)), and repeat steps (7) to (10). On the other hand, if the calculated CD(R)DH-run (x=1) is within the acceptable variation band (BV), then a new value for each operating parameter is randomly selected a statistically representative number of times (N) within the confidence level (CLj) of the tentative statistical distribution (Tj) (refer to Figure 6 , steps (11), (12) and (8)), and repeating steps (8) to (10). When all N runs x that produce the calculated CD(R)DH-runs (x=1 to N) are within the acceptable variation band (BV), the corresponding tentative statistical distribution (Tj) can be set as the final statistical distribution (Tf) (refer to Figure 6 , step (13). The particle accelerator system can be programmed with operating parameters according to the corresponding final statistical distribution (Tf). With such programming, the probability that the particle accelerator system delivers a beamlet that meets the treatment plan without interrupting the treatment course is CLj%.

[0117] The method of the present invention can be implemented with a single calculated dose (rate) histogram (cD(R)DH) or with several histograms, which must all be within the corresponding acceptable variation band (BV) to set the final statistical distribution (Tf). If a first tentative statistical distribution (Tj) produces a histogram that is contained within the corresponding acceptable variation band (BV) (e.g., cDVH), but produces another histogram that extends outside the acceptable variation band (BV) (e.g., cDRVH), a new tentative statistical distribution (T(j+1)) must be selected and the method performed again until a tentative statistical distribution (T(j+k)) is found that fits all the required histograms into the corresponding acceptable variation band (BV). Dose-rate dependent histograms (e.g., DRVH, DDRH) are particularly important for treatment plans that include beamlets to be emitted in FLASH mode in order to deposit dose in selected spots on the structure of interest at an ultra-high deposition rate.

[0118]

[0119]

Claims

1. A computer-implemented method for optimizing tolerance values ​​of operating parameters of a particle acceleration system that allows a beam formed of a plurality of particle beamlets accelerated along an irradiation axis (Z) to deposit a dose to a structure of interest in a patient by pencil beam scanning (PBS) according to a treatment plan (TP), the method comprising: (a) providing input, which includes o The treatment plan (TP), which includes the definition of beamlet bi arrays, each defined by planning parameters, including: The planned position Xpi of each beamlet bi in the plane (X, Y) normal to the irradiation axis (Z), ■ Planned monitoring unit MUpi for each beam, ■ The planned beamlet scanning sequence at these planned positions Xpi, o The planned start time t0pi and end time t1pi at which each beamlet is delivered, o the definition of the structure of interest, which defines the tissue or tissues through which the plurality of beamlets bi pass, o the value of one or more target dose distribution histograms tDDH or dose rate distribution histograms tDRDH for the structure of interest obtained by treatment using these planning parameters, and o an acceptable variation band BV within which the one or more target dose distribution histograms or target dose rate distribution histograms are allowed to vary, (b) providing a provisional statistical distribution Tj of operating parameters of the particle acceleration system centered about corresponding mean values ​​μj representing the performance of the particle acceleration system, and defining a confidence level CLj of the provisional statistical distribution Tj, wherein the operating parameters include: ○ Monitoring unit MUj of each beamlet, ○ The position Xj of each beamlet, and o the start time t0j and end time t1j of delivering each beamlet, (c) randomly selecting the value MUij of the monitoring unit, the value Xij of the position of the beam, and the values ​​t0ij, t1ij of each of the start time t0ij and the end time t1ij from the corresponding tentative statistical distribution Tj within a predefined confidence level CLj, (d) using the values ​​thus randomly selected to calculate one or more calculated dose distribution histograms cDDHj for run x, or dose rate distribution histograms cDRDHj for run x, where x = 1 to N is the number of calculation runs, (e) repeating the last two steps (c), (d) a statistically representative number N of times to produce calculated distributions CDj defining one or more thus calculated dose distribution histograms cDDHj or dose rate distribution histograms cDRDHj for all values ​​thus randomly selected of the operating parameters, (f) comparing the calculated distribution CDj of the one or more calculated dose distribution histograms cDDHj or dose rate distribution histograms cDRDHj with the corresponding acceptable variation band BV, and determining whether the calculated distribution CDj of the one or more calculated dose distribution histograms cDDHj or dose rate distribution histograms cDRDHj is included in the corresponding acceptable variation band within the predefined confidence level CLj.

2. The method according to claim 1 , comprising setting the final statistical distribution Tf with the final confidence level CLf as the corresponding operating parameter as follows, if, for a given treatment plan (TP), the calculated distributions CDj of the one or more calculated dose distribution histograms cDDHj or dose rate distribution histograms cDRDHj are all included within the corresponding acceptable variation band BV with the predefined confidence level CLj, then setting the tentative statistical distribution Tj to the final statistical distribution Tf=Tj and setting the predefined confidence level CLj to the corresponding final confidence level CLf=CLj to define corresponding operating parameters, if any of the one or more calculated dose distribution histograms cDDHj or dose rate distribution histograms cDRDHj calculated with a set of randomly selected values ​​within the corresponding confidence level CLj of the statistical distribution Tj of the operating parameters extends beyond the corresponding acceptable variation band BV, then repeating steps (b) to (f) of claim 1 according to the following conditions, o A new tentative statistical distribution T(j+k) using these operating parameters and / or ○ Choose a new, less demanding confidence level CL(j+k), Until the calculated distribution CD(j+k) of the one or more calculated dose distribution histograms cDDH(j+k) or dose rate distribution histograms cDRDH(j+k) are all included in the corresponding acceptable variation band BV, and the provisional statistical distribution T(j+k) is set to the final statistical distribution Tf=T(j+k), and the corresponding confidence level CL(j+k) is set to the final confidence level CLf=CL(j+k) to define the corresponding operating parameters.

3. The method according to claim 2, wherein: Setting the final statistical distribution Tf as the corresponding operating parameter is performed by a human operator or a processor.

4. The method according to claim 2, wherein: In the case where any one of the one or more calculated dose distribution histograms cDDHj or dose rate distribution histograms cDRDHj extends beyond the corresponding acceptable variation band having the predefined confidence level CLj, the new provisional statistical distribution T(j+k) defined in claim 2 has a lower standard deviation σj than the corresponding provisional statistical distribution Tj defined in step (b) of claim 1.

5. The method according to claim 3, wherein In the case where any one of the one or more calculated dose distribution histograms cDDHj or dose rate distribution histograms cDRDHj extends beyond the corresponding acceptable variation band having the predefined confidence level CLj, the new provisional statistical distribution T(j+k) defined in claim 2 has a lower standard deviation σj than the corresponding provisional statistical distribution Tj defined in step (b) of claim 1.

6. The method according to any one of claims 1 to 5, wherein The provisional statistical distribution Tj of each of these operating parameters is a Gaussian distribution, and wherein the value of the predefined confidence level CLj is included between 68% and 99.7% of the provisional statistical distribution, wherein the 68% confidence level corresponds to μj±σj, the 95% confidence level corresponds to μj±2σj, and the 99.7% confidence level of the provisional statistical distribution corresponds to μj±3σj, wherein μj is the mean and σj is the standard deviation of the corresponding provisional statistical distribution.

7. The method according to claim 6, wherein: The value of the predefined confidence level CLj is comprised between 95.5% and 99% of the provisional statistical distribution.

8. The method according to claim 6, wherein: The mean value μj and the standard deviation σj of each operating parameter are allowed to be different for each beamlet bi.

9. The method according to claim 7, wherein: The mean value μj and the standard deviation σj of each operating parameter are allowed to be different for each beamlet bi.

10. The method according to any one of claims 1 to 5, 7 to 9, wherein The particle acceleration system is equipped with a cyclic checking module, which is configured to measure the actual values ​​of these operating parameters at different intervals or continuously. These operating parameters include the actual monitoring unit MUai, the actual position Xai, and the actual values ​​of the start time t0ai and the end time t1ai of the fine beam emitted by the particle acceleration system.

11. The method according to claim 6, wherein: The particle acceleration system is equipped with a cyclic checking module, which is configured to measure the actual values ​​of these operating parameters at different intervals or continuously. These operating parameters include the actual monitoring unit MUai, the actual position Xai, and the actual values ​​of the start time t0ai and the end time t1ai of the fine beam emitted by the particle acceleration system.

12. The method according to claim 10, wherein: The particle acceleration system is equipped with a processor configured to compare the actual values ​​of the operating parameters with corresponding confidence levels CLj and to stop the treatment session if an actual value of an operating parameter falls outside the corresponding final confidence level CLf.

13. The method according to claim 11, wherein The particle acceleration system is equipped with a processor configured to compare the actual values ​​of the operating parameters with corresponding confidence levels CLj and to stop the treatment session if an actual value of an operating parameter falls outside the corresponding final confidence level CLf.

14. The method according to any one of claims 1 to 5, 7 to 9 and 11 to 13, wherein The planning parameters also include the planning beamlet size dj, the value of which dij used to calculate the one or more calculated dose distribution histograms cDDHj or dose rate distribution histograms cDRDHj being randomly selected within the corresponding tentative statistical distribution Tj of the planning beamlet size dj.

15. The method according to claim 6, wherein The planning parameters also include the planning beamlet size dj, the value of which dij used to calculate the one or more calculated dose distribution histograms cDDHj or dose rate distribution histograms cDRDHj being randomly selected within the corresponding tentative statistical distribution Tj of the planning beamlet size dj.

16. The method according to claim 10, wherein The planning parameters also include the planning beamlet size dj, the value of which dij used to calculate the one or more calculated dose distribution histograms cDDHj or dose rate distribution histograms cDRDHj being randomly selected within the corresponding tentative statistical distribution Tj of the planning beamlet size dj.

17. The method according to any one of claims 1 to 5, 7 to 9, 11 to 13 and 15 to 16, wherein The calculated distribution CDj of the one or more calculated dose distribution histograms cDDHj or dose rate distribution histograms cDRDHj is defined by a corresponding area comprised between a minimum calculated dose distribution histogram cDDHjO or a minimum calculated dose rate distribution histogram cDRDHjO on the one hand and a maximum calculated dose distribution histogram cDDHj1 or a maximum calculated dose rate distribution histogram cDRDHj1 on the other hand, wherein The minimum calculated dose distribution histogram cDDHj0 or the minimum calculated dose rate distribution histogram cDRDHj0 is defined by the lowest value of the calculated dose distribution histogram cDDHj or the calculated dose rate distribution histogram cDRDHj, which is calculated with the predefined confidence level CLj based on the monitoring unit MUij, the positions Xij of the beamlets, and N randomly selected values ​​of the start time t0ij and the end time t1ij, and wherein, ●The maximum calculated dose distribution histogram cDDHjO or the maximum calculated dose rate distribution histogram cDRDHjO is defined by the highest value of the calculated dose distribution histogram cDDHj or the calculated dose rate distribution histogram cDRDHj, which is calculated based on the monitoring unit MUij, the position X0i of these fine beams, and N randomly selected values ​​of the start time t0ij and the end time t1ij using the predefined confidence level CLj.

18. The method according to any one of claims 1 to 5, 7 to 9, 11 to 13 and 15 to 16, wherein The treatment plan includes depositing a dose into at least a portion of the structure of interest at an ultrahigh deposition rate, which is defined as a deposition rate greater than or equal to 1 Gy / s.

19. The method according to any one of claims 1 to 5, 7 to 9, 11 to 13 and 15 to 16, wherein The planning parameters also include the planning beam current Ij, whose values ​​Iij are used to calculate the one or more calculated dose distribution histograms cDDHj or dose rate distribution histograms cDRDHj, being randomly selected within the corresponding tentative statistical distribution Tj of the planning beam current Ij.

20. The method according to any one of claims 1 to 5, 7 to 9, 11 to 13 and 15 to 16, wherein The dose distribution histogram DDH selected from the target dose distribution histogram or the calculated dose distribution histogram is a dose volume histogram DVH, and wherein the dose rate distribution histogram DRDH selected from the target dose rate distribution histogram or the calculated dose rate distribution histogram is a dose rate volume histogram DRVH or a differential dose rate histogram DDRH.

21. An error prediction system configured to implement the method according to any one of claims 1 to 20, the error prediction system comprising, a memory comprising, for each operating parameter, a plurality of tentative statistical distributions Tj centered around a plurality of corresponding mean values ​​μj, A user interface configured to: o Input a treatment plan (TP) comprising one or more target dose distribution histograms tDHDH or target dose rate distribution histograms tDHRDH and corresponding acceptable variation bands BV, o Selecting from the memory or inputting the planned start time t0pi and the planned end time t1pi at which each beamlet is to be delivered, o selecting from the memory or inputting for each beamlet a first tentative statistical distribution Tj of each operating parameter, including the monitoring unit MUj, the position Xj of the beamlet, and the start and end times t0j and t1i, and ○ Enter the confidence levels CLj(OLj) for these operating parameters, A processor configured to: (i) randomly selecting values ​​of the value MUij of the monitoring unit, the value Xij of the position of the beam, the start time t0ij and the end time t1ij that are included in a predefined confidence level CLj of the corresponding tentative statistical distribution Tj, (ii) calculating one or more calculated dose distribution histograms cDDHj and calculated dose rate distribution histograms cDRDHj using the values ​​thus randomly selected, and (iii) The last two steps are repeated a statistically representative number N in order to produce the calculated distribution CDj of the dose distribution histogram cDDHj and dose rate distribution histogram cDRDHj thus calculated for each beamlet.

22. The error prediction system according to claim 21, wherein: The processor is further configured to, in a case where any one of the one or more calculated distributions CDj of the corresponding calculated dose distribution histogram (cDDHj) or dose rate distribution histogram (cDRDHj) is not included in the corresponding acceptable variation band with the predefined confidence level CLj, repeat steps (i) to (iii) of claim 21 using a new provisional statistical distribution T(j+k) of these operating parameters until all of the calculated distributions CDj of one or more of the calculated dose distribution histograms cDDHj or dose rate distribution histograms cDRDHj are included in the corresponding acceptable variation band with the predefined confidence level CLj.

Images