Method and device for assisting the determination of an antioxidant dosage for a person suffering from oxidative stress

Abstract

The invention relates to a method (100) for assisting the determination of a dosage of antioxidants to be administered to a patient suffering from oxidative stress, said method (100) comprising at least one iteration of a prediction phase (108) comprising the following steps: - measurement (110), on a blood sample previously taken from said patient, of a level of at least the following input parameters: cholesterol level, zinc level, copper level, vitamin C level, vitamin E level, and selenium level; - estimation (130) of a dosage for at least one antioxidant to be administered to said patient, via a previously trained estimation model, executed by a calculation unit, said estimation model taking as input said at least one input parameter. The invention also relates to a device which implements such a method.

Description

DESCRIPTION Title: Method and device for assisting in determining an antioxidant dosage for a person suffering from oxidative stress

[0001] The present invention relates to a method for assisting in determining a dosage of antioxidants to be administered to a person suffering from oxidative stress, and in particular a person suffering from facioscapulohumeral dystrophy (FSHD). It also relates to a device implementing such a method.

[0002] The field of the invention is generally the field of assistance in determining a dosage of antioxidants to be administered to a person suffering from oxidative stress, and more specifically to a person suffering from FSHD. State of the art

[0003] The use of antioxidants is known for the treatment of FSHD. Currently, the dosages of antioxidants to be administered to a person with FSHD are standardized. There are no predetermined rules for quickly determining the most appropriate dosage. Specifically, for each antioxidant, a standardized starting amount is administered to the patient. Then, the patient's response to the starting amount is monitored. Depending on the patient's response, the starting amount is adjusted in predetermined increments, and in particular, constant values. This adjustment operation is repeated as often as necessary depending on the patient's condition.

[0004] It is understood that this way of determining and adjusting the dosage of antioxidants to be administered to a patient with FSHD is not optimized and does not allow for the individual determination of the dosage that is best suited to each patient.

[0005] Moreover, with the current solution, determining a suitable dosage for a patient is done by trial and error, which can be detrimental to the patient.

[0006] Furthermore, with the current solution, determining a suitable dosage for a patient requires a significant amount of observation time, which makes the current solution long and unergonomic, or even ineffective in the event of a change in the patient's condition, this change being able to be monitored and quantified with scores such as Brooke and Vignos scores for FSHD, and more generally with an MFM score (for "Motor Function Measurement").

[0007] An aim of the present invention is to remedy at least one of the drawbacks of the state of the art.

[0008] Another aim of the invention is to propose a solution for determining a dosage of antioxidants to be administered to a patient suffering from oxidative stress, and in particular FSHD, in an individualized manner.

[0009] Another aim of the invention is to propose a solution for determining a dosage of antioxidants to be administered to a patient suffering from oxidative stress, and in particular FSHD, in a faster and more ergonomic manner for the patient.

[0010] It is also another aim of the present invention to propose a solution for determining a dosage of antioxidants to be administered to a patient suffering from oxidative stress, and in particular FSHD, more precisely.

[0011] It is also another aim of the present invention to propose a solution for determining a more effective dosage of antioxidants to be administered to a patient suffering from oxidative stress, and in particular FSHD. Statement of the invention

[0012] The invention proposes to achieve at least one of the aforementioned aims by a method for assisting in determining a dosage of antioxidants to be administered to a patient suffering from oxidative stress, and in particular a patient affected by FSHD, said method comprising at least one iteration of a prediction phase comprising the following steps: - measurement, on a blood sample previously taken from said patient, of a level of at least the following input parameters: ■ cholesterol level, ■ zinc level, ■ copper level, ■ vitamin C level, ■ vitamin E level; and ■ selenium level; - estimation of a dosage for at least one antioxidant to be administered to said patient, by a previously trained model, executed by a calculation unit, said model taking as input said at least one input parameter.

[0013] Thus, the invention provides assistance in determining an antioxidant dosage for a person suffering from oxidative stress, and more specifically FSHD, with an estimation model previously trained and executed by a computing unit. The estimation of the dosage is carried out according to at least one input parameter, specific to the patient, and the value of which is measured on a blood sample of said patient so that the estimation model allows an estimation of an individualized dosage for said patient, unlike the current solution which sets a dosage corresponding to standardized doses, and therefore not personalized, regardless of the patient.

[0014] In addition, the estimation of the dosage by the present invention is carried out by a pre-trained estimation model for estimating a dosage and taking as input input parameters measured on a blood sample of the patient. Thus, the invention makes it possible to estimate a dosage of antioxidant(s) for a patient suffering from oxidative stress, and in particular FSHD, in a faster and more ergonomic manner for the patient, compared to the current solution, because it does not require successive iterations to determine, by trial and error, the appropriate dosage for a patient.

[0015] Furthermore, the invention makes it possible to determine a dosage of antioxidant(s) on the upper limits of each antioxidant of a healthy population more precisely, and more effectively, for a patient suffering from oxidative stress, and in particular from FSHD, because it is not based on the use of predetermined levels, but on precise values ​​depending on the input parameters measured on a blood sample from the patient.

[0016] In a non-limiting manner, this solution can be adapted to a patient suffering from oxidative stress specific to a pathology where the intervals and types of vitamins and / or types of trace elements will be specific to the targeted pathology.

[0017] Preferably, the estimation model makes an estimation of a dosage for several antioxidants.

[0018] Preferably, the estimation model can perform an estimation of a dosage for at least one, and in particular for each, of the following antioxidants: - zinc, - copper, - vitamin C, - vitamin E, and - selenium.

[0019] According to a preferred embodiment, the dosage may include a value for each of the parameters listed above.

[0020] For at least one antioxidant, the estimation model may be trained to provide a value within a predetermined range of values ​​for said antioxidant. For example, the estimation model may be trained to estimate a value within said range of values, such that the provided value may be any value within said range of values.

[0021] Alternatively, or in addition, for at least one antioxidant, the estimation model can be trained to choose one value from several predefined discrete values ​​for said antioxidant. In this case, the estimation model necessarily returns one of these predefined values.

[0022] According to embodiments, for zinc, the estimation model can be trained to provide a dosage between 15 mg / day and 45 mg / day. According to a non-limiting example embodiment, for zinc, the estimation model can be trained to choose a dosage from the following predefined discrete values: 15 mg / day, 30 mg / day, 45 mg / day.

[0023] According to embodiments, for copper, the estimation model can be trained to provide a dosage between 0 and 0.3 mg / day. According to a non-limiting example embodiment, for copper, the estimation model can be trained to choose a dosage from the following predefined discrete values: 0 mg / day and 0.3 mg / day.

[0024] According to embodiments, for vitamin E, the estimation model can be trained to provide a dosage between 125 mg / day and 500 mg / day. According to a non-limiting example embodiment, for vitamin E, the estimation model can be trained to choose a dosage from the following predefined discrete values: 125 mg / day, 250 mg / day, 500 mg / day.

[0025] According to embodiments, for vitamin C, the estimation model can be trained to provide a dosage between 125 mg / day and 500 mg / day. According to a non-limiting example embodiment, for vitamin C, the estimation model can be trained to choose a dosage from the following predefined discrete values: 125 mg / day, 250 mg / day, 500 mg / day.

[0026] As indicated above, the prediction phase also includes a measurement of a selenium level in the patient's blood, the estimation of the dosage by the estimation model being further carried out as a function of said measured selenium level.

[0027] The selenium level is therefore an additional input parameter given as input to the estimation model.

[0028] Selenium level can be measured from the blood sample previously taken from the patient.

[0029] The estimation model can further provide a dosage for selenium to be administered to the patient.

[0030] For selenium, the estimation model can be trained to provide a value within a predetermined range of values. For example, the estimation model can be trained to estimate a value within said range of values, so that the provided value can be any value within said range of values.

[0031] Alternatively, or in addition, for Selenium, the estimation model can be trained to choose one value from several predefined discrete values. In this case, the estimation model necessarily returns one of these predefined values.

[0032] According to embodiments, for selenium, the estimation model can be trained to provide a dosage between 0 and 200 pg / day. According to a non-limiting example embodiment, for selenium, the estimation model can be trained to choose a dosage from the following predefined discrete values: 0 and 200 pg / day.

[0033] According to embodiments, the prediction phase may further comprise a measurement of an iron level in the patient's blood, the estimation of the dosage by the estimation model being further carried out as a function of said measured iron level.

[0034] In this case, the iron level is an additional input parameter given as input to the estimation model.

[0035] The iron level can be measured from the blood sample previously taken from the patient.

[0036] Vitamin C allows the absorption of iron in the human body. Therefore, the indication of a vitamin C dosage should preferably take into account the iron level in the patient's body to avoid causing an excessive increase in the iron level in the patient's body. Thus, taking into account the iron level in the patient's blood allows the vitamin C dosage to be adjusted.

[0037] The method according to the invention may further comprise a measurement of a level of diabetes in the patient's blood, the estimation of the dosage by the estimation model being further carried out as a function of said level of diabetes.

[0038] The level of diabetes can be measured by measuring a level of glycated hemoglobin.

[0039] The level of diabetes can be measured from the blood sample previously taken from the patient.

[0040] The measurement of the level of diabetes can be carried out prior to the first iteration of the prediction phase to determine whether or not the patient suffers from diabetes.

[0041] Alternatively, or in addition, measurement of diabetes level can be performed during each iteration of the prediction phase.

[0042] Whether or not the patient has diabetes may result in an adjustment of the selenium dosage. When the patient has diabetes, the selenium dosage can be between 0.1g / day and 100pg / day. When the patient does not have diabetes, the selenium dosage can be between 0.1g / day and 200pg / day.

[0043] Additionally, the estimation model can be trained to determine an antioxidant dosage that achieves the following ratios in the patient's blood, for example at the end of a treatment period: 0.7 < (Vitamin C) / (Vitamin E) <0.9 0.8 < (Copper) / (Zinc) <1 7 mg / g < (Vitamin E) / (Cholesterol) < 10 mg / g

[0044] According to embodiments, the estimation model may be a decision tree.

[0045] Alternatively, the estimation model may be a neural network performing classification or regression, such as, for example, a convolutional neural network.

[0046] In this case, the estimation model can be trained with a training base comprising a multitude of training sets, each training set comprising: - an input vector comprising a value for at least one parameter; and - an output vector comprising a dosage for at least one antioxidant. Training games can be measured on a multitude of patients with oxidative stress, and more particularly with FSHD.

[0047] Training can be performed using any known training algorithm, such as gradient backpropagation.

[0048] The method according to the invention can advantageously be implemented for determining a dosage of antioxidants to be administered to a patient suffering from facioscapulohumeral dystrophy, FSHD.

[0049] According to another aspect of the present invention, there is provided a computer-implemented estimation model comprising executable instructions which, when executed by a computing device, implement the estimation step of the method according to the invention.

[0050] The estimation model can be in any computer language, such as machine language, C, C++, JAVA, Python, etc.

[0051] According to another aspect of the present invention, there is provided a device for assisting in determining a dosage of antioxidants to be administered to a patient suffering from oxidative stress, and in particular FSHD, comprising: - at least one means of measuring, on a blood sample previously taken from said patient, a level of at least the following input parameters: ■ cholesterol level, ■ zinc level, ■ copper level, ■ vitamin C level, ■ vitamin E level, ■ selenium level, and ■ possibly an iron level, and / or a glycated hemoglobin level; - a previously trained estimation model for estimating a dosage for at least one antioxidant to be administered to said patient based on said measured values.

[0052] The device according to the invention may comprise a computing unit for executing the estimation model. The computing unit may be any type of device such as a server, a computer, a tablet, a calculator, a processor, a computer chip, etc.

[0053] The means of measuring at least one input parameter may be any known means.

[0054] For example, at least one input parameter can be measured in the laboratory.

[0055] The device according to the invention can advantageously be used for determining a dosage of antioxidants to be administered to a patient suffering from facioscapulohumeral dystrophy, FSHD. Description of figures and embodiments

[0056] Other advantages and characteristics will appear on examining the detailed description of non-limiting embodiments, and the attached drawings in which: - FIGURE 1 is a schematic representation of a non-limiting exemplary embodiment of a method according to the invention; - FIGURE 2 is a schematic representation of a non-limiting exemplary embodiment of a device according to the invention; and - FIGURE 3 is a schematic representation of a non-limiting exemplary embodiment of a decision tree that can be used as an estimation model within the framework of the present invention.

[0057] It is understood that the embodiments which will be described below are in no way limiting. In particular, we can imagine variants of the invention comprising only a selection of features described below isolated from the other features described, if this selection of features is sufficient to confer a technical advantage or to differentiate the invention from the state of the prior art. This selection includes at least one preferably functional feature without structural details, or with only a part of the structural details if it is this part which is only sufficient to confer a technical advantage or to differentiate the invention from the state of the prior art.

[0058] In particular, all the variants and embodiments described can be combined with each other if there is no technical obstacle to this combination.

[0059] In the figures and in the rest of the description, the elements common to several figures retain the same reference.

[0060] FIGURE 1 is a schematic representation of a non-limiting exemplary embodiment of a method according to the present invention.

[0061] The method 100 of FIGURE 1 may be used to assist a practitioner in determining the dosage of antioxidants to be administered to a patient suffering from oxidative stress, and in particular FSHD, with an estimation model executed by a computing unit and previously trained / determined for the estimation of the dosage for at least one antioxidant.

[0062] In a preliminary step 102, not forming part of the method 100, a blood sample is taken from the patient. This blood sample can be taken in a conventional manner.

[0063] The method 100 comprises an optional step 104 of measuring a glycated hemoglobin level. Measuring the glycated hemoglobin level is a well-known operation and can be carried out in a conventional manner, for example in a laboratory.

[0064] The hemoglobin level is compared to a predetermined threshold in an optional step 106 to determine whether the patient also suffers from diabetes, or not. The output of this comparison can be used as an input parameter signaling whether or not the patient has diabetes. For example, this input parameter can take a binary value: - one, for example “0”, indicating that the patient does not suffer from diabetes, and - the other, for example 1, indicating that the patient suffers from diabetes.

[0065] The method 100 comprises a phase 108 of predicting dosage from biological input parameters of the patient.

[0066] The prediction phase 108 comprises a step 110 of measuring at least one biological parameter on the blood sample, called input parameter.

[0067] In particular, the measurement step 110 comprises: - a step 112 of measuring the cholesterol level, - a step 114 of measuring the zinc level, - a step 116 of measuring the copper level, - a step 118 of measuring the vitamin C level, - a step 120 of measuring the vitamin E level, and - a step 122 of measuring the level of selenium in the blood sample.

[0068] Optionally, the measuring step 110 may further comprise a step 124 of measuring the iron level in the blood sample.

[0069] Each of the values ​​measured in the measuring step 110 constitutes an input parameter for a pre-trained, or predetermined, estimation model for estimating a dosage of antioxidants to be administered to the patient.

[0070] For example, taking into account the optional steps, the set of input parameters can form an input vector, denoted E, of 8 values: E = {D, Ch, Zc, Cu, C, E, Se, Fe} in which D indicates whether the patient suffers from diabetes or not, and Ch, Zc, Cu, C, E, Se and Fe are, respectively, the measured values ​​of the levels of cholesterol, zinc, copper, vitamin C, vitamin E, selenium and iron in the blood sample taken from the patient.

[0071] The prediction phase 108 comprises a step 130 when which the input vector V eis given as input to an estimation model, predetermined for estimating a dosage for at least one antioxidant, and in particular for a combination of at least two antioxidants, to be administered to the patient.

[0072] Preferably, the estimation model is a decision tree. Alternatively, the estimation model can be a neural network or a polynomial model.

[0073] In response to the input vector V e , the estimation model provides an output vector, for example denoted V s , comprising at least one dosage value. According to a non-limiting exemplary embodiment, the output vector comprises: - a dosage for zinc, - a dosage for copper, - a dosage for vitamin C, - a dosage for vitamin E, and - a dosage for selenium.

[0074] The estimation model can be trained, or determined, to determine an antioxidant dosage that achieves the following ratios in the patient's blood, for example at the end of a treatment period: 0.7 < (Vitamin C) / (Vitamin E) <0.9 0.8 < (Copper) / (Zinc) <1 7 mg / g < (Vitamin E) / (Cholesterol) < 10 mg / g

[0075] Additionally, the estimation model can be determined, or trained, to provide an antioxidant dosage, such as: - the dosage of Zinc is chosen from the following predefined discrete values: 15mg / day, 30mg / day, 45mg / day. - the Cu dosage is chosen from the following predefined discrete values: 0 mg / day and 0.3 mg / day. - the dosage of vitamin E is chosen from the following predefined discrete values: 125mg / day, 250mg / day, 500mg / day. - the dosage of Vitamin C is chosen from the following predefined discrete values: 125mg / day, 250mg / day, 500mg / day

[0076] Additionally, the estimation model can be determined, or trained, to provide a selenium dosage as follows: - when the patient does not suffer from diabetes: the dosage of Selenium is chosen between 0 and 200pg / day, but can be either 0 pg / day or 200pg / day. - when the patient suffers from diabetes: the dosage of Selenium is chosen between 0 and 100pg / day, but can be either 0 pg / day or 100pg / day.

[0077] Of course, all of the above values ​​are given as non-limiting examples.

[0078] The prediction phase 108 can be repeated as many times as desired for the same patient in order to always adapt the dosage of antioxidants to the patient's condition.

[0079] For example, the prediction phase 108 may be repeated at a predetermined frequency. Alternatively, or in addition, the prediction phase 108 may be repeated upon request of the patient or upon decision of a practitioner.

[0080] Thus, the patient's condition can be monitored and the dosage of antioxidants can be adjusted to the evolution of their condition immediately.

[0081] In the method 100, the steps 104 and 106 relating to diabetes are carried out prior to the first iteration of the prediction phase 108. Alternatively, these steps may be part of the prediction phase 108 and reiterated at each iteration of the prediction phase 108.

[0082] FIGURE 2 is a schematic representation of a non-limiting exemplary embodiment of a device according to the present invention.

[0083] The device 200 of FIGURE 2 may be used to assist a practitioner in determining the dosage of antioxidants to administer to a patient. suffers from oxidative stress, and in particular to a patient suffering from FSHD, with an estimation model executed by a computing unit and previously trained / determined for the estimation of the dosage for at least one antioxidant.

[0084] The device 200 of FIGURE 2 can be used to implement a method according to the invention, and in particular the method 100 of FIGURE 1.

[0085] The device 200 comprises at least one means 202 for measuring at least one biological parameter on a blood sample from the patient, such as for example the parameters described above with reference to the method 100, i.e. a zinc level, a copper level, a vitamin C level, a vitamin E level, and a selenium level, a glycated hemoglobin level, and an iron level. At least one of these parameters can be measured by a dedicated measuring means. At least two of these parameters can be measured by the same measuring means.

[0086] The device 200 further comprises a calculation unit 204 for executing a pre-trained, or predetermined, estimation model 206 for estimating, as a function of the input parameter(s) measured by the at least one measuring means, a dosage of at least one antioxidant, such as for example the antioxidants described above with reference to the method 100, i.e. a dosage for Zinc, a dosage for Copper, a dosage for vitamin C, a dosage for vitamin E, and a dosage for Selenium.

[0087] The computing unit 204 may be any type of device such as a server, a computer, a tablet, a calculator, a processor, a computer chip, etc.

[0088] The estimation model 206 can be a decision tree or a neural network.

[0089] The input parameters may be presented to the estimation model 206 in the form of an input vector, denoted E, comprising a value for each input parameter. The dosage value(s) may be provided by the estimation model in the form of an output vector, denoted Vs, comprising a dosage value for each antioxidant.

[0090] FIGURE 3 is a schematic representation of a non-limiting schematic example of a decision tree that may be used as an estimation model within the scope of the present invention.

[0091] The decision tree shown in FIGURE 3 can be used in a method, respectively a device, according to the invention and in particular in the method 100 of FIGURE 1 or in the device 200 of FIGURE 2.

[0092] The decision tree 300 of FIGURE 3 first takes as input the measured value of vitamin E. The measured value of vitamin E is used to choose a branch of the decision tree, from among several branches each corresponding to a value, or a range of values, of vitamin E.

[0093] Then, the decision tree takes into account the measured cholesterol value. The measured cholesterol value allows you to choose a branch of the decision tree, from several branches each corresponding to a cholesterol value, or a range of values. And so on with the measured value of vitamin C, then copper, then zinc, and finally selenium to end up with a terminal branch 302i, from several terminal branches 302i-302 n . Each terminal branch 302i corresponds to a dosage for at least one, and in particular each of the following parameters: - zinc, - copper, - vitamin C, - vitamin E, and - selenium.

[0094] Of course, the decision tree 300 can take into account other input parameters such as iron level, glycated hemoglobin level.

[0095] Additionally, the dosage corresponding to each terminal branch of the decision tree may include a dosage for parameters other than those indicated, such as, for example, the selenium level.

[0096] Further, the decision tree 300 may be organized differently such that the input parameters are considered in an order other than those indicated above.

[0097] Of course, the invention is not limited to the examples which have just been described.

Claims

CLAIMS 1. Method (100) for assisting in determining a dosage of antioxidants to be administered to a patient suffering from oxidative stress, said method (100) comprising at least one iteration of a prediction phase (108) comprising the following steps: - measurement (110), on a blood sample previously taken from said patient, of a level of at least the following input parameters: ■ cholesterol level, ■ zinc level, ■ copper level, ■ vitamin C level, ■ vitamin E level, and ■ selenium level; - estimation (130) of a dosage for at least one antioxidant to be administered to said patient, by a previously trained estimation model (206), executed by a calculation unit (204), said estimation model (206) taking as input said at least one input parameter.

2. Method (100) according to the preceding claim, characterized in that the estimation model (206) carries out an estimation of a dosage for at least one, and in particular for each, of the following antioxidants: - zinc, - copper, - vitamin C, - vitamin E, and - selenium.

3. Method (100) according to the preceding claim, characterized in that the estimation model (206) is trained to provide: - a dosage for zinc between 15mg / day and 45mg / day; - a dosage for copper between 0 and 0.3 mg / day; - a dosage for vitamin E between 125mg / day and 500mg / day; - a dosage for vitamin C between 125mg / day and 500mg / day; - a dosage for selenium between 0 and 200pg / day 4. Method (100) according to any one of the preceding claims, characterized in that the prediction phase (108) further comprises a measurement (124) of an iron level in the patient's blood, the estimation of the dosage by the estimation model (206) being further carried out as a function of said measured iron level.

5. Method (100) according to any one of the preceding claims, characterized in that it further comprises a measurement (104) of a level of diabetes in the patient's blood, and in particular of a level of glycated hemoglobin, the estimation of the dosage by the estimation model (206) being further carried out as a function of said level of diabetes.

6. Method (100) according to any one of the preceding claims, characterized in that the estimation model (206) is a decision tree.

7. Method (100) according to any one of the preceding claims, characterized in that it is implemented for determining a dosage of antioxidants to be administered to a patient suffering from facioscapulohumeral dystrophy, FSHD.

8. A computer-implemented estimation model (206) comprising executable instructions which, when executed by a computing device, implement the step (130) of estimating the method (100) of any preceding claim.

9. Device (200) for assisting in determining a dosage of antioxidants to be administered to a patient suffering from oxidative stress, comprising: - at least one measuring means (202), on a blood sample previously taken from said patient, of a level of at least the following input parameters: ■ cholesterol level, ■ zinc level, ■ copper level, ■ vitamin C level, ■ vitamin E level, ■ selenium level, and “possibly an iron level, and / or a glycated hemoglobin level; and - an estimation model (206), previously trained, for estimating a dosage of antioxidants to be administered to said patient based on said measured values.

10. Use of the device (200) according to the preceding claim for determining a dosage of antioxidants to be administered to a patient suffering from facioscapulohumeral dystrophy, FSHD.

Images