Device and method for providing a treatment recommendation based on a prediction of at least one impusle control disorder occurrence for a patient undergoing a treatment with dopamine agonists

Abstract

The present invention relates to methods and devices for providing a treatment recommendation based on a prediction of at least one Impulse Control Disorder occurrence for a patient undergoing a treatment with Dopamine Agonists.

Description

DEVICE AND METHOD FOR PROVIDING A TREATMENT RECOMMENDATION BASED ON A PREDICTION OF AT LEAST ONE IMPUSLE CONTROL DISORDER OCCURRENCE FOR A PATIENT UNDERGOING A TREATMENT WITH DOPAMINE AGONISTSFIELD OF INVENTION

[0001] The present invention relates to the technical field of disease prediction and prevention. Notably, the invention relates to a device for providing a treatment recommendation based on a prediction of at least one Impulse Control Disorder occurrence for a patient undergoing a treatment with dopamine agonists for any indication including but not restricted to, Parkinson Disease (PD), restless leg syndrome, depression or apathy, pituitary adenoma. The invention also relates to a method for providing a treatment recommendation for a patient having PD and to a method for treating a patient having Parkinson Disease. The invention further relates to a method for treating a patient having PD.BACKGROUND OF INVENTION

[0002] Impulse control disorders (ICDs) represent a frequent and serious complication in Parkinson’s disease (PD) patients receiving dopamine agonist (DA) therapy. These disorders include pathological gambling, compulsive shopping, hypersexuality, and binge eating, which can significantly impair quality of life..

[0003] While several clinical scales exist to assess non-motor symptoms and behavioral disturbances in PD, current practice relies heavily on retrospective identification of ICDs, often after significant damage has occurred. There is therefore a strong need for proactive strategies capable of anticipating ICD onset..

[0004] Machine learning techniques have recently been explored in the medical field to identify patients at risk of adverse drug reactions, including ICDs..

[0005] In particular, Faouzi et al. (2022), Machine-Learning-Based Prediction of Impulse Control Disorders in Parkinson ’s Disease From Clinical and Genetic Data,Movement Disorders 37(6): 920-932, discloses a research study in which machine learning algorithms were trained on clinical and genetic variables from Parkinson’s cohorts to predict the risk of ICD occurrence in patients undergoing DA therapy. This work highlights the potential of predictive modeling in this context..

[0006] However, this prior study is limited to the presentation of predictive results in a research setting. It does not describe a dedicated device, any mechanism for integrating patient follow-up data into structured scores, nor an automated process for generating treatment recommendations based on risk thresholds..

[0007] Consequently, there remains a need for a solution that not only predicts the risk of ICDs but also translates this prediction into actionable treatment guidance, enabling physicians to adjust DA therapy proactively in order to reduce the likelihood of ICD onset..

[0008] The present invention addresses this need by providing a device and method that integrate patient data collected during follow-up sessions, process these data into structured scores, and apply a trained machine learning model to predict ICD risk. The prediction is then compared to predefined thresholds to classify the patient’s risk and to output a concrete treatment recommendation..SUMMARY

[0009] This invention thus relates to a device for providing a treatment recommendation based on a prediction of at least one Impulse Control Disorder occurrence for a patient undergoing a treatment with Dopamine Agonists, said prediction being provided by a trained machine learning prediction model configured to receive as input for said patient at least one data ensemble comprising: demographic and anthropomorphic data, mental health data, data relative to said treatment with Dopamine Agonists, each data ensemble being collected during a follow-up session, said follow-up session being part of a treatment process,said machine learning model being previously trained on a training dataset comprising, for each subject of a plurality of subjects undergoing a treatment with Dopamine Agonists, multiple training samples, wherein each training sample is obtained during a follow-up session, said follow-up session being part of a treatment process undergone by said subject, wherein each training sample comprises at least one demographic and anthropomorphic score derived from demographic and anthropomorphic data collected on said subject, at least one mental health score derived from mental health data collected on said subject and at least one treatment score derived from data relative to said treatment with Dopamine Agonists for undergone by said subject, said device further comprising: at least one processor configured to: o calculate at least one demographic and anthropomorphic score from said demographic and anthropomorphic data collected on said patient, at least one mental health score from said mental health data collected on said patient and at least one treatment score from said data relative to a said treatment with Dopamine Agonists undergone by said subject, o feed said demographic and anthropomorphic score, mental health score and treatment score to said trained machine learning model so as to output said prediction, and o compare said prediction with at least one threshold so as to determine said treatment recommendation based on the comparison, and at least one output configured to output said treatment recommendation.

[0010] Advantageously, predicting the occurrence of ICDs, in patient to which DAs treatment is indicated, and providing a treatment recommendation allows to at least delay the occurrence of ICDs or even avoid the occurrence of ICDs. This proactive approach, rather than reactive, enhances patient safety by minimizing the risks associated with ICDs, such as severe financial, social, and emotional consequences. It further allows for a more tailored treatment regimen, ensuring that patients receive the therapeutic benefits of DAs medications, without the associated risks of ICDs. This balance improves patient adherence to medication, thereby enhancing the overall efficacy of the underlying diseasemanagement. This approach also mitigates the legal and ethical concerns associated with prescribing DAs, as clinicians can make more informed decisions, thereby reducing the likelihood of malpractice claims.

[0011] According to one embodiment: said demographic and anthropomorphic data include an age, a gender, a weight and a duration since start of the treatment process, said mental health data include data representative of an anxiety level, a depression level, an impulsivity level, a sleep quality level and an indicator of presence of a dopamine dysregulation syndrome, said data relative to said treatment with Dopamine Agonists undergone by said patient include at least one treatment name and a treatment dosage.

[0012] According to one embodiment, at least part of said mental health data is collected using at least question comprised in at least one of a MDS-UPDRS I questionnaire, a QUIP SHORT questionnaire and an ASBPD / ECMP questionnaire.

[0013] According to one embodiment, said machine learning prediction model is a XGBoost.

[0014] According to one embodiment, said trained machine learning prediction model configured to receive as input for said patient multiple data ensembles collected during a plurality of follow-up sessions, said at least one demographic and anthropomorphic score, said at least one mental health score and said at least one treatment score being obtained using the demographic and anthropomorphic data, the mental health data and the data relative to said treatment with Dopamine Agonists collected during the plurality of follow-up sessions already undergone by said patient.

[0015] Advantageously, the demographic and anthropomorphic scores, the mental health scores and the treatment scores take into account a temporality of the treatment process, by encompassing data from the past follow-up sessions of the patient.

[0016] According to one embodiment, said prediction outputted by said trained machine learning prediction model is compared to two thresholds, including a first threshold and a second threshold, the first threshold being inferior to the second threshold, wherein:- when said prediction is inferior to the first threshold and the second threshold, the patient is classified as a patient with a low risk of developing at least one Impulse Control Disorder, and the treatment recommendation is to continue the treatment with Dopamine Agonists;- when said prediction is superior to the first threshold and inferior to the second threshold, the patient is classified as a patient with a medium risk of developing at least one Impulse Control Disorder, and the treatment recommendation is to reduce the dosage of the treatment with Dopamine Agonists;- when said prediction is superior to the first threshold and superior to the second threshold, the patient is classified as a patient with a high risk of developing at least one Impulse Control Disorder, and the treatment recommendation is to stop the treatment with Dopamine Agonists.

[0017] According to one embodiment, said trained machine learning prediction model is obtained using a 5-fold cross-validation scheme.

[0018] The present invention further relates to a method for providing a treatment recommendation for a patient undergoing a treatment with Dopamine Agonists, said method comprising: a) predicting the occurrence of at least one Impulse Control Disorder (ICD), b) assigning the patient to a risk group of developing at least one ICD based on the comparison of the prediction determined in step a) with two thresholds, including a first threshold and a second threshold, the first threshold being inferior to the second threshold, wherein: o said patient is assigned to a low risk of developing at least one ICD when the prediction is inferior to the first threshold and the second threshold;o said patient is assigned to a medium risk of developing at least one Impulse Control Disorder when the prediction is superior to a first threshold and inferior to the second threshold; and o said patient is assigned to a high risk of developing at least one Impulse Control Disorder when the prediction is superior to the first threshold and the second threshold; and c) providing a treatment recommendation depending on the identified risk.

[0019] The treatment recommendations described above are derived from statistical analysis performed on different cohorts of patients treated by DAs, on which it was deduced that the treatment recommendation by the approach, in the right patients, at the right time, of stopping the Dopamine Agonists treatment, decreases the risk of developing ICDs by up to 90%, and reducing the Dopamine Agonists treatment by 50% or less decreases the risk by up to 70% in the worst-case scenario. Thus the objective of the present method of preventing the occurrence of ICDs in at least 50% of patients is far reached.

[0020] According to one embodiment, said method for providing a treatment recommendation for a patient undergoing a treatment with Dopamine Agonists is a computer-implemented method and the step of predicting the occurrence of at least one Impulse Control Disorder (ICD) is performed using a machine learning prediction model previously trained on a training dataset comprising, for each subject of a plurality of subjects, at least one demographic and anthropomorphic score derived from demographic and anthropomorphic data collected on said subject, at least one mental health score derived from mental health data collected on said subject and at least one treatment score derived from data relative to said treatment with Dopamine Agonists, said method further comprising: receiving as input: o demographic and anthropomorphic data, o mental health data, and o data relative to said treatment with Dopamine Agonist taken by said patient,calculating at least one demographic and anthropomorphic score from said demographic and anthropomorphic data collected on said patient, at least one mental health score from said mental health data collected on said patient and at least one treatment score from said data relative to said treatment with Dopamine Agonists,- feeding said demographic and anthropomorphic score, mental health score and treatment score to said trained machine learning model so as to output said prediction.

[0021] According to one embodiment, the patient suffers from a disease or condition associated with dopamine dysregulation, and wherein the treatment recommendation comprises adapting the treatment dosage.

[0022] According to one embodiment, wherein the treatment recommendation is:- to continue the treatment with Dopamine Agonists, when the patient is assigned to a low risk of developing at least one Impulse Control Disorder, and / or- to reduce the dosage of the treatment with Dopamine Agonists, when said patient is assigned to a medium risk of developing at least one Impulse Control Disorder, and / or- to stop the treatment with Dopamine Agonists, when said patient is assigned to a high risk of developing at least one Impulse Control Disorder.

[0023] The present invention further relates to a method for treating a patient having a disease or condition associated with dopamine dysregulation, preferably Parkinson’s Disease, and undergoing a treatment with Dopamine Agonists, said method comprising: a) predicting the occurrence of at least one Impulse Control Disorder (ICD), b) assigning the patient to a risk group of developing at least one Impulse Control Disorder based on the comparison of the prediction determined in step a) with two thresholds, including a first threshold and a second threshold, the first threshold being inferior to the second threshold, wherein:o said patient is assigned to a low risk of developing at least one Impulse Control Disorder when the prediction is inferior to the first threshold and the second threshold; o said patient is assigned to a medium risk of developing at least one Impulse Control Disorder when the prediction is superior to a first threshold and inferior to the second threshold; and o said patient is assigned to a high risk of developing at least one Impulse Control Disorder when the prediction is superior to the first threshold and the second threshold; c) treating said patient depending on the identified risk.

[0024] According to one embodiment, the patient suffers from Parkinson’s disease and wherein the step c) (i.e. step c) of the method for treating a patient having a disease or condition associated with dopamine dysregulation) comprises treating the patient with one or more of the following medications: Amantadine, levodopa, Dopamine Agonists, catechol-O-methyltransferase (COMT) inhibitors and Monoamine Oxydase Inhibitor B (IMAO-B), depending on the identified risk.

[0025] According to one embodiment, when the patient is assigned to a low risk of developing at least one Impulse Control Disorder, the step c) (i.e. step c) of the method for treating a patient having a disease or condition associated with dopamine dysregulation) comprises continuing, reducing, or increasing the dosage of the Dopamine Agonists treatment, as well as combining it with any other medication for treating the disease or condition.

[0026] According to one embodiment, when the patient is assigned to a medium risk of developing at least one Impulse Control Disorder, the step c) (i.e. step c) of the method for treating a patient having a disease or condition associated with dopamine dysregulation) comprises : i) replacing the treatment with Dopamine Agonists with another medication for the disease or condition, preferably replacing the treatment with Dopamine Agonists with levodopa in Parkinson’s disease patient, or ii) reducing the dosage of the treatment with Dopamine Agonists and supplementing it with anothermedication for the disease or condition, preferably supplementing it with levodopa in Parkinson’s disease patient.

[0027] According to one embodiment, when the patient is assigned to a high risk of developing at least one Impulse Control Disorder, the step c) (i.e. step c) of the method for treating a patient having a disease or condition associated with dopamine dysregulation) comprises stopping the treatment with Dopamine Agonists and replacing it with another medication for the disease or condition, preferably replacing it with levodopa in Parkinson’s disease patient.

[0028] Advantageously, the method as described herein may enable to prevent or reduce the risk of developing at least one Impulse Control Disorder in a patient undergoing a treatment with Dopamine Agonists.

[0029] In addition, the disclosure relates to a computer program product comprising instructions which, when the program is executed by a computer, cause the computer to carry out a computer-implemented method for providing a treatment recommendation for a patient undergoing a treatment with Dopamine Agonists according to any of the herein disclosed embodiments.

[0030] The disclosure also relates to a computer-readable storage medium comprising instructions which, when executed by a computer, cause the computer to carry out the computer-implemented method for providing a treatment recommendation for a patient undergoing a treatment with Dopamine Agonists according to any of the herein disclosed embodiments.

[0031] The present disclosure further pertains to a non-transitory program storage device, readable by a computer, tangibly embodying a program of instructions executable by the computer to perform the computer-implemented method compliant with the present disclosure.

[0032] Such a non-transitory program storage device can be, without limitation, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor device, or any suitable combination of the foregoing. It is to be appreciated that the following, whileproviding more specific examples, is merely an illustrative and not exhaustive listing as readily appreciated by one of ordinary skill in the art: a portable computer diskette, a hard disk, a ROM, an EPROM (Erasable Programmable ROM) or a Flash memory, a portable CD-ROM (Compact-Disc ROM).DEFINITIONS

[0033] In the present invention, the following terms have the following meanings:

[0034] The term “Active surveillance”, as used herein, mean closely monitoring a patient’s condition without giving any treatment until symptoms appear or change.

[0035] The terms “adapted” and “configured” are used in the present disclosure as broadly encompassing initial configuration, later adaptation or complementation of the present device, or any combination thereof alike, whether effected through material or software means (including firmware).

[0036] The term “processor” should not be construed to be restricted to hardware capable of executing software, and refers in a general way to a processing device, which can for example include a computer, a microprocessor, an integrated circuit, or a programmable logic device (PLD). The processor may also encompass one or more Graphics Processing Units (GPU), whether exploited for computer graphics and image processing or other functions. Additionally, the instructions and / or data enabling to perform associated and / or resulting functionalities may be stored on any processor- readable medium such as, e.g., an integrated circuit, a hard disk, a CD (Compact Disc), an optical disc such as a DVD (Digital Versatile Disc), a RAM (Random-Access Memory) or a ROM (Read-Only Memory). Instructions may be notably stored in hardware, software, firmware or in any combination thereof.

[0037] “Machine learning (ML)” designates in a traditional way computer algorithms improving automatically through experience, on the ground of training data enabling to adjust parameters of computer models through gap reductions between expected outputs extracted from the training data and evaluated outputs computed by the computer models.

[0038] A “hyper-parameter” presently means a parameter used to carry out an upstream control of a model construction, such as a remembering-forgetting balance in sample selection or a width of a time window, by contrast with a parameter of a model itself, which depends on specific situations. In ML applications, hyper-parameters are used to control the learning process.

[0039] “Datasets” are collections of data used to build an ML mathematical model, so as to make data-driven predictions or decisions. In “supervised learning” (i.e. inferring functions from known input-output examples in the form of labelled training data), three types of ML datasets (also designated as ML sets) are typically dedicated to three respective kinds of tasks: “training”, i.e. fitting the parameters, “validation”, i.e. tuning ML hyperparameters (which are parameters used to control the learning process), and “testing”, i.e. checking independently of a training dataset exploited for building a mathematical model that the latter model provides satisfying results.

[0040] A “neural network (NN)” designates a category of ML comprising nodes (called “neurons”), and connections between neurons modeled by “weights”. For each neuron, an output is given in function of an input or a set of inputs by an “activation function”. Neurons are generally organized into multiple “layers”, so that neurons of one layer connect only to neurons of the immediately preceding and immediately following layers.

[0041] “Treating” or “treatment”, as used herein, refers to alleviating a specified condition, eliminating or reducing the symptoms of a condition, slowing or eliminating the progression of a condition, and preventing or delaying the initial occurrence of a condition in a subject, or preventing or delaying the reoccurrence of a condition in a previously afflicted subject.

[0042] “Subject”, as used herein, refers to individuals already diagnosed with the condition (ICD) that may be used to generate a training dataset for training the machine learning segmentation model for pulvinar segmentation from the invention.

[0043] “Patient”, as used herein, refers to an individual to be diagnosed or treated according to the methods of the present invention. Patients include, but are not limited to,mammals (e.g., murines, simians, equines, bovines, porcines, canines, felines, and the like), preferably to primates, and most preferably to humans.

[0044] The term “Dopamine Agonists” refers to compounds activating dopamine receptors, which are G protein-coupled receptors. Two families of dopamine receptors exist: the Di-like family comprising Di and Ds receptors and the D2 family comprising D2, D3 and D4 receptors.

[0045] “Therapeutically effective amount” refers to the level or amount of a treatment as described herein that is aimed at, without causing significant negative or adverse side effects to the target, (1) delaying or preventing the onset of a disease, disorder, or condition; (2) slowing down or stopping the progression, aggravation, or deterioration of one or more symptoms of the disease, disorder, or condition; (3) bringing about ameliorations of the symptoms of the disease, disorder, or condition; (4) reducing the severity or incidence of the disease, disorder, or condition; or (5) curing the disease, disorder, or condition. A therapeutically effective amount may be administered prior to the onset of the disease, disorder, or condition, for a prophylactic or preventive action. Alternatively or additionally, the therapeutically effective amount may be administered after initiation of the disease, disorder, or condition, for a therapeutic action.BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The present disclosure will be better understood, and other specific features and advantages will emerge upon reading the following description of particular and non-restrictive illustrative embodiments, the description making reference to the annexed drawings wherein:

[0047] Figure 1 is a block diagram representing schematically a particular mode of a device for obtaining a trained machine learning prediction model for prediction of at least one Impulse Control Disorder occurrence compliant with the present disclosure;

[0048] Figure 2 is a flow chart showing successive steps executed with the device of Figure 1;

[0049] Figure 3 is a block diagram representing schematically a particular mode of a device for providing a treatment recommendation based on a prediction of at least one Impulse Control Disorder occurrence, said prediction being provided by a trained machine learning prediction model obtained from the device of Figure 1;

[0050] Figures 4 is a flow chart showing successive steps executed with the device from Figure 3;

[0051] Figure 5 shows an apparatus integrating the functions of the device for obtaining a trained machine learning prediction model for prediction of at least one Impulse Control Disorder occurrence of figure 1 and of the device for providing a treatment recommendation based on a prediction of at least one Impulse Control Disorder occurrence of figure 3.ILLUSTRATIVE EMBODIMENTS

[0052] The present description illustrates the principles of the present disclosure. It will thus be appreciated that those skilled in the art will be able to devise various arrangements that, although not explicitly described or shown herein, embody the principles of the disclosure and are included within its scope.

[0053] All examples and conditional language recited herein are intended for educational purposes to aid the reader in understanding the principles of the disclosure and the concepts contributed by the inventor to furthering the art, and are to be construed as being without limitation to such specifically recited examples and conditions.

[0054] Moreover, all statements herein reciting principles, aspects, and embodiments of the disclosure, as well as specific examples thereof, are intended to encompass both structural and functional equivalents thereof. Additionally, it is intended that such equivalents include both currently known equivalents as well as equivalents developed in the future, i.e., any elements developed that perform the same function, regardless of structure.

[0055] Thus, for example, it will be appreciated by those skilled in the art that the block diagrams presented herein may represent conceptual views of illustrative circuitryembodying the principles of the disclosure. Similarly, it will be appreciated that any flow charts, flow diagrams, and the like represent various processes which may be substantially represented in computer readable media and so executed by a computer or processor, whether or not such computer or processor is explicitly shown.

[0056] The functions of the various elements shown in the figures may be provided through the use of dedicated hardware as well as hardware capable of executing software in association with appropriate software. When provided by a processor, the functions may be provided by a single dedicated processor, a single shared processor, or a plurality of individual processors, some of which may be shared.

[0057] It should be understood that the elements shown in the figures may be implemented in various forms of hardware, software or combinations thereof. Preferably, these elements are implemented in a combination of hardware and software on one or more appropriately programmed general-purpose devices, which may include a processor, memory and input / output interfaces.

[0058] The present disclosure will be described in reference to a particular functional embodiment of a device 1 for obtaining a trained machine learning prediction model 30 for prediction of at least one Impulse Control Disorder (ICD) occurrence for a patient undergoing a treatment with Dopamine Agonists, as illustrated on Figure 1.

[0059] According to a first embodiment, the device 1 is adapted to receive as an input for the training: demographic and anthropomorphic data 21, mental health data 22 and data relative to at least one treatment with a Dopamine Agonist 23, collected on a plurality of subjects, at different points in time (e.g. during successive follow-up sessions with a physician, as part of a treatment process). In this first embodiment, the input for the training do not comprise genetic data collected on a plurality of subjects.

[0060] According to a second embodiment, the device 1 is adapted to receive as an input for the training: demographic and anthropomorphic data 21, mental health data 22 and data relative to at least one treatment with Dopamine Agonists 23, collected on a plurality of subjects, at different points in time, and genetic data collected on a plurality of subjects. Said genetic data may be genetic variant(s) of one or more of the following genes for instance but not restricted to this list: DRD2, DRD3, DAT1, COMT, DDC, GRIN2B,ADRA2C, SERT, TPH2, HTR2A, OPRK1 and OPRM1. The genetic characteristics may be the genetic variant(s) of DRD2 gene. The genetic characteristics may be the genetic variant(s) of DRD3 gene. The genetic characteristics may be the genetic variant(s) of DAT1 gene. The genetic characteristics may be the genetic variant(s) of COMT gene. The genetic characteristics may be the genetic variant(s) of DDC gene. The genetic characteristics may be the genetic variant(s) of GRJN2B gene. The genetic characteristics may be the genetic variant(s) of ADRA2C gene. The genetic characteristics may be the genetic variant(s) of SERT gene. The genetic characteristics may be the genetic variant(s) of TPH2 gene. The genetic characteristics may be the genetic variant(s) of HTR2A gene. The genetic characteristics may be the genetic variant(s) of OPRKlgene. The genetic characteristics may be the genetic variant(s) of 0PRM1 gene. The genetic characteristics may be the genetic variant(s) of multiple and numerous genes associated in a combination of variants to inform treatment recommendation. Therefore, according to these two embodiments the training may be performed with training data comprising or not the genetic data.

[0061] The device 1 for training the untuned machine learning prediction model 20 is associated with a device 2 for providing a treatment recommendation based on a prediction of at least one Impulse Control Disorder occurrence provided by the trained machine learning prediction model 20 obtained using device 1, as represented on Figure 3, which will be subsequently described.

[0062] Though the presently described devices 1 and 2 are versatile and provided with several functions that can be carried out alternatively or in any cumulative way, other implementations within the scope of the present disclosure include devices having only parts of the present functionalities.

[0063] Each of the devices 1 and 2 is advantageously an apparatus, or a physical part of an apparatus, designed, configured and / or adapted for performing the mentioned functions and produce the mentioned effects or results. In alternative implementations, any of the device 1 and the device 2 is embodied as a set of apparatus or physical parts of apparatus, whether grouped in a same machine or in different, possibly remote, machines. The device 1 and / or the device 2 may have functions distributed over a cloudinfrastructure and be available to users as a cloud-based service, or have remote functions accessible through an API.

[0064] The device 1 and the device 2 may be integrated in a same apparatus or set of apparatus, and intended to same users. In other implementations, the structure of the device 2 may be completely independent of the structure of the device 1, and may be provided for other users. For example, the device 2 may have a trained machine learning prediction model 30 available to operators for prediction of at least one Impulse Control Disorder (ICD) occurrence, wholly set from previous training effected upstream by other players with the device 1.

[0065] In what follows, the modules are to be understood as functional entities rather than material, physically distinct, components. They can consequently be embodied either as grouped together in a same tangible and concrete component, or distributed into several such components. Also, each of those modules is possibly itself shared between at least two physical components. In addition, the modules are implemented in hardware, software, firmware, or any mixed form thereof as well. They are preferably embodied within at least one processor of the device 1 or of the device 2.

[0066] The device 1 comprises a module 11 configured to receive as input, for a plurality of subjects, a data ensemble comprising the demographic and anthropomorphic data 21, the mental health data 22 and the data relative to said treatment with Dopamine Agonists 23 undergone by each subject, collected during at least one follow-up session, together with the untuned machine learning prediction model 20.

[0067] Alternatively, the device 1 may comprise a module 11 configured to receive as input, for a plurality of subjects, a data ensemble comprising the demographic and anthropomorphic data 21, the mental health data 22, the data relative to said treatment with Dopamine Agonists 23 undergone by each subject, and the genetic data collected during at least one follow-up session, together with the untuned machine learning prediction model 20.

[0068] The input data may be stored in one or more local or remote database(s) 10. The latter can take the form of storage resources available from any kind of appropriate storage means, which can be notably a RAM or an EEPROM (Electrically-ErasableProgrammable Read-Only Memory) such as a Flash memory, possibly within an SSD (Solid-State Disk).

[0069] More precisely, the demographic and anthropomorphic data 21 , the mental health data 22 and the data relative to said treatment with Dopamine Agonists 23 undergone by each subject may be collected by a physician (or by the subject himself or by its caregiver if applicable) during a plurality of successive follow-up sessions (e.g. medical visits). The subjects may have their data collected during, for instance, two successive follow-up sessions, three successive follow-up sessions, four successive follow-up sessions, five successive follow-up sessions and between 6 follow-up sessions and 20 follow-up sessions. The elapsed time between two successive follow-up sessions may be consistently the same. Alternatively, the elapsed time between two successive follow-up sessions may vary. The time between two successive follow-up sessions may be comprised between a few weeks and a few years. The subjects may have their data collected as long as they are under treatment with Dopamine Agonists. The subjects may have their data collected as long as they are alive. The subjects may have their data collected as long as they want to be followed-up.

[0070] The mental health data 22 and the data to said treatment with Dopamine Agonists 23 undergone by each subject may be collected using questions and scales extracted from widely accepted and used questionnaires in both clinical practice and research, including motor and non-motor symptoms, disease progression, cognitive function, and overall quality of life. Such questionnaires may include:- the MDS-UPDRS (Movement Disorder Society-Sponsored Revision of the Unified Parkinson's Disease Rating Scale) assessment tool, that evaluates nonmotor and motor experiences of daily living, motor examination, and motor complications;- the PDQ-39 (Parkinson's Disease Questionnaire-39): a quality-of-life questionnaire specifically designed for people with Parkinson's Disease, covering 39 items across eight domains: mobility, activities of daily living, emotional wellbeing, stigma, social support, cognition, communication, and bodily discomfort;- the QUIP -RS (Questionnaire for Impulsive-Compulsive Disorders in Parkinson's Disease - Rating Scale): a validated tool specifically designed to screen for impulse control disorders (ICDs) and related behaviors in Parkinson's Disease patients. The short form of this questionnaire, QUIP-S, helps in quickly assessing the presence of ICD symptoms such as compulsive gambling, eating, shopping, and sexual behaviors;- the NMSQuest (Non-Motor Symptoms Questionnaire): a questionnaire designed to identify the range and impact of non-motor symptoms in Parkinson's Disease patients;- the ECMP (Behavioral Evaluation in Parkinson's Disease): a questionnaire designed to assess non-motor symptoms in Parkinson's patients. The ECMP specially focuses on evaluating behaviors related to mood (depression, anxiety), apathy, fluctuations ("on" and "off' states), impulse control disorders (hyperdopaminergic behaviors) like hypersexuality, impulse buying, punding (repetitive, purposeless behavior), gambling etc.;- the MoCA (Montreal Cognitive Assessment): a cognitive screening tool that assesses different cognitive domains including attention, concentration, executive functions, memory, language, visuoconstructional skills, conceptual thinking, calculations, and orientation; and / or- the SCOPA-SLEEP (Scales for Outcomes in Parkinson's Disease-Sleep): a scale used to evaluate sleep disturbances in Parkinson's Disease patients.

[0071] The demographic and anthropomorphic data 21 may include the age, the gender, the weight and the duration since start of the treatment process (e.g. diagnosis of any disease at a stage requiring Dopamine Agonists treatment such as for instance Parkinson’s Disease or Restless Leg Syndrome, Hyperprolactinemia, depression or apathy, acromegaly, Tourette Syndrome, pituitary adenoma, Attend on-Deficit / Hyperactivity Disorder (ADHD) or narcolepsy). At each follow-up session, the physician (or the subject himself or its caregiver if applicable) may collect these data if necessary. For instance, the birth date and date of disease diagnosis may be collected at the first follow-up session and then automatically calculated during the subsequent follow-up sessions, the gendergenerally remains unchanged and the weight may be collected at each follow-up session by weighting the subject on a scale.

[0072] The mental health data 22 may include information allowing to estimate an anxiety level, a depression level, an impulsive level, a sleep quality level and an indicator of presence of a dopamine dysregulation syndrome.

[0073] For instance, the information allowing to estimate the anxiety level may be collected by asking the subject to answer at least one question from at least one questionnaire.

[0074] For instance, such question (extracted from the MDS-UPDRS questionnaire) may include: “Over the past week have you felt nervous, worried, or tense? If yes, was this feeling for longer than one day at a time? Did it make it difficult for you to follow your usual activities or to be with other people? [If yes, examiner asks patient or caregiver to elaborate and probes for information. '

[0075] Depending on the subject’s response, the physician (or the patient himself or his caregiver if applicable) may score the level of anxiety with the following scale:0: Normal: No anxious feelings;1 : Slight: Anxious feelings present but not sustained for more than one day at a time. No interference with patient’s ability to carry out normal activities and social interactions;2: Mild: Anxious feelings are sustained over more than one day at a time, but without interference with patient’s ability to carry out normal activities and social interactions;3: Moderate: Anxious feelings interfere with, but do not preclude, the patient’s ability to carry out normal activities and social interactions.4: Severe: Anxious feelings preclude patient’s ability to carry out normal activities and social interactions.

[0076] Alternatively or cumulatively, at least one question of the Hospital Anxiety and Depression Scale (HADS), specifically the Anxiety subscale (HADS-A) may be used. It comprises 7 questions that may be answered by a value comprised between 0 and 3. Thequestions for instance are the following: “I feel tense or 'wound up ’", wherein the response can for instance be “0: Not at all, 1: From time to time, occasionally, 2: A lot of the time, 3: Most of the lime". Another example of question is: “I get a sort of frightened feeling as if something awful is about to happen”,' or “Worrying thoughts go through my mind' etc.

[0077] The information allowing to estimate the depression level may be collected by asking the subject to answer at least one question from at least one questionnaire.

[0078] Such question (extracted from the MDS-UPDRS questionnaire) may include: “Over the past week have you felt low, sad, hopeless, or unable to enjoy things? If yes, was this feeling for longer than one day at a time? Did it make it difficult for you carry out your usual activities or to be with people? [If yes, examiner asks patient or caregiver to elaborate and probes for information.] ”

[0079] Depending on the subject’s response, the physician (or the patient himself or his caregiver if applicable) may score the level of depression with the following scale:0: Normal: No depressed mood;1 : Slight: Episodes of depressed mood that are not sustained for more than one day at a time. No interference with patient’s ability to carry out normal activities and social interactions;2: Mild: Depressed mood that is sustained over days, but without interference with normal activities and social interactions;3: Moderate: Depressed mood that interferes with, but does not preclude the patient’s ability to carry out normal activities and social interactions.4: Severe: Depressed mood precludes patient’s ability to carry out normal activities and social interactions.

[0080] Alternatively or cumulatively, the information allowing to estimate the depression level may be collected by asking the subject to answer at least one question from the Geriatric Depression Scale (GDS) and / or the Hospital Anxiety and Depression Scale (HADS), specifically the Depression subscale (HADS-D).

[0081] The GDS is a questionnaire comprising 30 questions that may be answered by either “Yes” or “No”. The questions relate to how the individual has been feeling over the past week. For instance, the questions include “Are you satisfied with your current life?”, “Are you often bored? ” etc.

[0082] The Hospital Anxiety and Depression Scale (HADS) and specifically the Depression subscale (HADS-D) is a questionnaire comprising 7 questions that may be answered by a value comprised between 0 and 3. The questions for instance are the following: “I still enjoy the things I used to enjoy”, wherein the response can either be “0: Definitely as much, 1: Not quite so much, 2: Only a little, 3: Hardly at all” . Other questions may be “I can laugh and see the funny side of things”, “I feel cheerful” etc.

[0083] The information allowing to estimate the sleep quality level may be collected by asking the subject to answer at least one question from at least one questionnaire.

[0084] Such question (extracted from the MDS-UPDRS questionnaire) may include: “Over the past week, have you had trouble going to sleep at night or staying asleep through the night? Consider how rested you felt after waking up in the morning. ”

[0085] Depending on the subject’s response, the physician (or the patient himself or his caregiver if applicable) may score the sleep quality level with the following scale:0: Normal: No problems;1 : Slight: Sleep problems are present but usually do not cause trouble getting a full night of sleep;2: Mild: Sleep problems usually cause some difficulties getting a full night of sleep;3 : Moderate: Sleep problems cause a lot of difficulties getting a full night of sleep, but I still usually sleep for more than half the night;4: Severe: I usually do not sleep for most of the night.

[0086] Alternatively or cumulatively, the following question (extracted from the MDS- UPDRS questionnaire) may be asked: “Over the past week, have you had trouble staying awake during the daytime?”

[0087] Depending on the subject’s response, the physician (or the patient himself or his caregiver if applicable) may score the level of sleep quality with the following scale:0: Normal: No daytime sleepiness;1 : Slight: Daytime sleepiness occurs, but I can resist and I stay awake;2: Mild: Sometimes I fall asleep when alone and relaxing. For example, while reading or watching TV;3 : Moderate: I sometimes fall asleep when I should not. For example, while eating or talking with other people;4: Severe: I often fall asleep when I should not. For example, while eating or talking with other people.

[0088] Alternatively or cumulatively, the following question (extracted from the MDS- UPDRS questionnaire) may be asked: “Over the pastweek, have you usually felt fatigued? This feeling is not part of being sleepy or sad "

[0089] Depending on the subject’s response, the physician (or the patient himself or his caregiver if applicable) may score the level of sleep quality with the following scale:0: Normal: No fatigue.1 : Slight: Fatigue occurs. However, it does not cause me trouble doing things or being with people.2: Mild: Fatigue causes me some troubles doing things or being with people.3 : Moderate: Fatigue causes me a lot of troubles doing things or being with people. However, it does not stop me from doing anything.4: Severe: Fatigue stops me from doing things or being with people.

[0090] Alternatively or cumulatively, the Parkinson's Disease Sleep Scale (PDSS) may be used. It comprises 15 questions that address specific aspects of sleep or sleep-related disturbance, such as nocturnal motor symptoms, nocturia, and vivid dreams or nightmares. Each item is scored on a visual analog scale ranging from 0 to 10, where 0 indicates the worst possible sleep disturbance and 10 indicates no disturbance. For instance, the questions encompass: Difficulty falling asleep (0: severe difficulty, 10: no difficulty), Difficulty staying asleep (0: severe difficulty, 10: no difficulty), Painful muscle cramps in the night (0: very frequent, 10: none).

[0091] The information allowing to estimate the indicator of presence of a dopamine dysregulation syndrome may be collected by asking the subject to answer at least one question from at least one questionnaire.

[0092] Such question (extracted from the MDS-UPDRS questionnaire) may include: “Over the past week, have you had unusually strong urges that are hard to control? Do you feel driven to do or think about something and find it hard to stop? [Give patient examples such as gambling, cleaning, using the computer, taking extra medicine, obsessing about food or sex, all depending on the patient. / ”

[0093] Depending on the subject’s response, the physician (or the patient himself or his caregiver if applicable) may score the level of presence of a dopamine dysregulation syndrome with the following scale:0: Normal: No problems present.1 : Slight: Problems are present but usually do not cause any difficulties for the patient or family / caregiver.2: Mild: Problems are present and usually cause a few difficulties in the patient’s personal and family life.3: Moderate: Problems are present and usually cause a lot of difficulties in the patient’s personal and family life.4: Severe: Problems are present and preclude the patient’s ability to carry out normal activities or social interactions or to maintain previous standards in personal and family life.

[0094] The information allowing to estimate the impulsive level may be collected by asking the subject to answer at least one question from at least one questionnaire.

[0095] For instance, the ECMP questionnaire may be used. This questionnaire comprises four categories of questions, including six psychic evaluation questions (depressive mood, hypomanic, manic mood, anxiety, irritability, aggressiveness, hyperactivity, psychotic symptoms), one functioning in apathetic mode question, two non-motor fluctuations questions (on, off), eleven hyperdopaminergic behaviors questions (nocturnal hyperactivity, daytime sleepiness, eating behavior, creativity, DIY (Do it yourself), punding (repetitive, purposeless activity), risky behavior, compulsiveshopping, pathological gambling, hypersexuality, dopaminergic addiction). A value comprised between 0 and 4 may be attributed to each question, wherein 0 corresponds to an absence of a symptom and 4 corresponds to a severe presence of the symptom. It is the French equivalent of the ASBPD (“Ardouin Scale of Behavior in Parkinson's Disease”).

[0096] Alternatively or cumulatively, the Impulsive-Compulsive Disorders in Parkinson’s Disease-Rating Scale (QUIP -RS) may be used. Only the questions related to specific impulsive-compulsive disorder behaviors may be used.

[0097] The data relative treatment for Parkinson Disease 23 undergone by each subject may include at least one treatment name and a treatment dosage. For instance, the subject may be treated with Dopamine Agonists, such as Pramipexole, Ropinirole or Rotigotine, with respectively a dosage of 0.18 mg, three times a day, 0.25mg, three times a day or 2mg / 24h, transdermal. The subject may be treated with a Dopamine Agonists as described herein.

[0098] The genetic data may be obtained by sequencing DNA from a sample of the subject during a follow-up session. In particular, the genetic data may be obtained by extracting DNA from a blood sample of the subject, and detecting single nucleotide polymorphisms (SNPs) associated with neurological disorders with genotyping arrays. Examples of genotyping arrays that may be used include, without limitation, neurochips that are custom-designed Illumina genotyping arrays, which are engineered to detect SNPs associated with neurological disorders. After several rounds of quality control, genome-wide association studies (GWAS) may be performed, incorporating imputation techniques to generate a comprehensive dataset containing millions of SNPs, ready for detailed analysis.

[0099] The device 1 further comprises optionally a module 12 for preprocessing the input data. For instance, module 12 may be configured to detect missing data, corrupted data or aberrant data and delete the corresponding data ensemble.

[0100] The device may further comprise a module 13 configured to calculate or aggregate scores based on the data received as input.

[0101] According to the invention, a score is a quantifiable or categorical measure representing an attribute of an entity. These scores serve as features (input variables) that a machine learning model can learn from, to make predictions or classifications.

[0102] In one embodiment, the data received as input may be directly exploitable as scores. Alternatively, the scores may be derived from at least one data received as input. The input scores serve as ground truth during the training stage; in other words, these scores are used as input features for the model to learn from.

[0103] For instance, the demographic and anthropomorphic scores may comprise an age score corresponding to the age data received as input or calculated (i.e. corresponding to the age at the last follow-up session), a gender score (e.g. male or female), that may for instance be represented as a binary value (e.g. “0” for male and “1” for female) or a categorical value (e.g. a string value “male” or “female”), a weight score corresponding to the weight data received as input (i.e. corresponding to the weight at the last follow-up session), and / or a duration score corresponding to the duration since start of the treatment process data received as input or calculated (i.e. duration between start of treatment and last follow-up session).

[0104] The mental health scores may comprise at least one anxiety score. Such score may correspond to the value (e.g. between 0 and 4) of the question about anxiety from the MDS-UPDRS I questionnaire mentioned above.

[0105] Alternatively, the anxiety score may be derived from the values obtained by answering the HADS-A subscale. The values obtained for each question may for example be summed to obtain the score or averaged to obtain the score.

[0106] In another example, the STAI (State-Trait Anxiety Inventory) scale could be used to calculate the anxiety score. The values obtained from such a scale could be summed or averaged to derive the anxiety score.

[0107] The mental health scores may further comprise at least one depression score. Such score may correspond to the value (e.g. between 0 and 4) of the question about depression from the MDS-UPDRS I questionnaire mentioned above.

[0108] Alternatively, the depression score may for example be derived from the values obtained by answering the GDS and / or the HADS-D subscale. The values obtained for each question may be summed to obtain the score or averaged to obtain the score.

[0109] The mental health scores may comprise at least one sleep quality score that may correspond to the value (e.g. between 0 and 4) of one of the questions about sleep quality from the MDS-UPDRS I questionnaire mentioned above. Alternatively, the sleep quality score may be obtained by combining the values obtained for the three questions about sleep quality from the MDS-UPDRS I questionnaire mentioned above. Typically, the at least one sleep quality score may be a sum of the values obtained for each of the three questions about sleep quality from the MDS-UPDRS I questionnaire mentioned above. Alternatively, the at least one sleep quality score may be a mean of the three values obtained for each of the three questions about sleep quality from the MDS-UPDRS I questionnaire mentioned above.

[0110] Alternatively, the at least one sleep quality score may be derived from the values obtained by answering the PDSS questionnaire, either by summing the values obtained for each question or by averaging these values.

[0111] The mental health scores may further comprise a dopamine dysregulation score. Such score may correspond to the value (e.g. between 0 and 4) of the question about dopamine dysregulation from the MDS-UPDRS I questionnaire mentioned above.

[0112] The at least one impulsive score may be determined using the values obtained by answering the ECMP questionnaire and / or the QUIP -RS questionnaire, either by summing the values obtained for each question or by averaging these values.

[0113] Overall, the scores may be derived from questions belonging to one questionnaire or multiple questionnaires.

[0114] The treatment score may comprise at least one dosage score. The at least one dosage score may correspond to the quantity of Dopamine Agonists taken by the subject (e.g. expressed in mg per day).

[0115] The treatment score may further comprise at least one treatment score (e.g. Dopamine Agonist, Levodopa, Monoamine Oxydase Inhibitor B (IMAO-B) etc.), that may for instance be represented as a numerical value (e.g. “0” for Dopamine Agonist, “1”for Levodopa, “2” for IMAO-B) or a categorical value (e.g. a string value “Dopamine Agonist”, “Levodopa”, “IMAOB”).

[0116] The genetic score may comprise at least one genetic score for each variant of the genes, that may for instance be represented as a numerical value. For instance, the genetic score may be binary, such as 0 or 1, indicating the presence or absence of a single variant. Additionally, more comprehensive numerical values, such as, but not limited to, polygenic risk scores (PRS) or polygenic scores, may be calculated. These scores aggregate the effects of multiple genetic variants across the genome to provide a global measure of genetic predisposition to certain traits or conditions.

[0117] All the above scores may be normalized, for example to be comprised between 0 and 1.

[0118] As mentioned above, the device is configured to receive the input data for a plurality of follow-up sessions as part of the treatment process. Some of the scores (called temporal scores) may therefore take into account the temporality of the multiple followup sessions. Such temporality may be used to train the machine learning predictive model.

[0119] For instance, such temporality may be conveyed to the machine learning predictive model by averaging the scores obtained over the past n follow-up sessions. Typically, the scores obtained over the past five follow-up sessions may be taken into account. For instance, the anxiety scores, the depression scores, the sleep quality scores, the dopamine dysregulation scores, the impulsive scores, the treatment scores and the dosages obtained during the past n follow-up sessions may be taken into account to calculate at least one temporal anxiety score, at least one temporal depression score, at least one temporal sleep quality score, at least one temporal dopamine dysregulation score, at least one temporal impulsive score, at least one temporal treatment score and at least one temporal dosages score. Temporal scores may be obtained by averaging the scores obtained during the past n follow-up sessions and / or by summing the scores obtained during the past n follow-up sessions and / or by calculating a minimum or a maximum score and / or by calculating a variance of the scores.

[0120] Typically, for the dosage scores, a cumulated dosage score, a maximum dosage score and a minimum dosage score may be calculated.

[0121] The device may further comprise a module 14 configured to generate a training dataset comprising the obtained scores.

[0122] The device may comprise a module 15 configured to train the untuned machine learning prediction model 20 using the training dataset constructed (or received) by module 14.

[0123] The untuned machine learning prediction model 20 may be an ensemble learning model such as a XGBoost that uses the principles of gradient boosting, where models are trained sequentially to minimize the residual errors of the previous models using gradient descent optimization. Notably, XGBoost operates by constructing a sequence of decision trees, where each tree is specialized to correct the previous mistakes made in the series.

[0124] Alternatively, the untuned machine learning prediction model 20 may be a recurrent neural network (RNN) with long short-term memory (LSTM) layers. LSTM layers are a type of recurrent neural network architecture designed to effectively capture and utilize long-range dependencies within sequential data. Unlike standard RNNs, which can suffer from vanishing or exploding gradient problems during training, LSTM networks address these issues with a unique cell state and gated mechanism. Notably, an LSTM network operates by maintaining a memory cell that updates through a series of gates: input gate, forget gate, and output gate. These gates regulate the flow of information, allowing the model to selectively remember or forget previous information, and thus make more accurate predictions over long sequences. The RNN with LSTM layers is effective in applications where the prediction task depends on understanding the temporal dynamics of the input data, such as in time series forecasting.

[0125] During the training, the model may be optimized by tuning its hyperparameters, such as learning rate, maximum depth of trees, and regularization terms.

[0126] A 5-fold grouped cross-validation may be done (to address the nuances of longitudinal data), followed by fine-tuning through a Gaussian Process.

[0127] Once the training completed, module 15 is configured to output the trained machine learning prediction model 30, The trained segmentation model 31 may then by stored in one or more local or remote database(s) 10. The latter can take the form of storage resources available from any kind of appropriate storage means, which can benotably a RAM or an EEPROM (Electrically-Erasable Programmable Read-Only Memory) such as a Flash memory, possibly within an SSD (Solid-State Disk).

[0128] In its automatic actions, the device 1 may for example execute the following process (Figure 2):- receiving for a plurality of subjects, the demographic and anthropomorphic data 21, the mental health data 22 and the data relative to said treatment with Dopamine Agonists undergone by each subject, collected at different points in time, together with the untuned machine learning prediction model 20 (step 41),- optionally preprocessing the input data (step 42),- calculating demographic and anthropomorphic scores from said demographic and anthropomorphic data 21, mental health scores from said mental health data 22 and treatment scores from said data relative said treatment with Dopamine Agonists 23 (step 43),- generating a training dataset comprising the calculated scores (step 44),- training the untuned machine learning prediction model using the generated training dataset so that the machine learning prediction model is configured to output a prediction of occurrence of at least one ICD for a patient receiving Dopamine Agonist treatment (step 45)

[0129] The present invention also relates to the device 2 for providing a treatment recommendation based on a prediction of at least one Impulse Control Disorder occurrence for a patient undergoing Dopamine Agonists treatment provided by a trained machine learning prediction model 30 obtained using the device 1. The device 2 will be described in reference to a particular function embodiment as illustrated in Figure 3.

[0130] The device 2 is adapted to receive as input the trained machine learning prediction model 30 and for a same patient: demographic and anthropomorphic data 121, mental health data 122, and data relative to at least one treatmentl23 taken by said patient.

[0131] Optionally, the device 2 is adapted to receive as input the training machine learning prediction model 30 as described herein and for a same patient: demographic and anthropomorphic data 121, mental health data 122, data relative to at least one treatment 123 taken by said patient and genetic data. Said genetic data may be as described herein.

[0132] The device 2 is adapted to use the trained machine learning prediction model 30 to obtain a prediction of occurrence of at least one Impulse Control Disorder for the patient and to compare the prediction to at least one threshold in order to provide a treatment recommendation 31 for said patient.

[0133] To that end, the device 2 may comprise a module 111 for receiving the trained machine learning prediction model 30 and for the patient: the demographic and anthropomorphic data 121, the mental health data 122, and the data relative to at least one treatment taken by said patient 123. The input data may be stored in one or more local or remote database(s) 10. The latter can take the form of storage resources available from any kind of appropriate storage means, which can be notably a RAM or an EEPROM (Electrically-Erasable Programmable Read-Only Memory) such as a Flash memory, possibly within an SSD (Solid-State Disk). In advantageous embodiments, the trained machine learning prediction model 30 and all its training parameters may have been previously generated by a system including the device 1 for training. Alternatively, the trained machine learning prediction model 30 and its training parameters may be received from a communication network.

[0134] Optionally, the device 2 may comprise a module 111 for receiving the trained machine learning prediction model 30 and for the patient: the demographic and anthropomorphic data 121, the mental health data 122, and the data relative to at least one treatment taken by said patient 123 and genetic data.

[0135] The demographic and anthropomorphic data 121, mental health data 122, and data relative to at least one treatment taken by said patient 123 may be collected by the physician during a medical visit. For instance, the physician may have access to a local or remote platform (e.g. web-based or cloud-based platform) that stores and manages patient health information. The platform may be accessible via the internet or via an application installed on a computer, a smartphone or a tablet. The platform may comprise an interface that leads the physician on the information to collect during the medical visit. For instance, the interface may include structured forms and prompts guiding the physician to input the demographic and anthropomorphic details such as the weight, as well as inquire about the patient's mental health status. To that end, the interface may display the questions to ask to the patient and a corresponding scale (e.g. discretecheckboxes scale or continuous cursor scale, for instance comprise between 0 and 5) with which the physician may give a score corresponding to the mental health status of the patient, depending on the responses provided by said patient to the question. Additionally, the interface may provide dropdown menus or checkboxes to record the specific type of treatment (e.g. for Parkinson's Disease) to be taken by the patient. Additionally, the interface may provide text input fields to manually enter the treatment dosage.

[0136] The genetic data may be obtained by sequencing DNA from a sample of the patient during a baseline or follow-up session.

[0137] The input data may be obtained in the same way (e.g. using the same questionnaires) as described in reference to module 11.

[0138] The device 2 further comprises optionally a module 112 for preprocessing the input data. For instance, module 112 may be configured to detect missing data, corrupted data or aberrant data and delete the corresponding data ensemble.

[0139] The device 2 may further comprise a module 113 configured to calculate scores based on the input data. To that end, the same calculation processes as described above, in reference to module 13.

[0140] The device 2 may further comprise a module 114 configured to feed the calculated scores to the trained segmentation model 30 to obtain prediction of an occurrence of said at least one ICD.The device 2 may also comprise a module 115 configured to compare the prediction to at least one threshold.

[0141] The device 2 may interact with a user interface 18, via which information can be entered and retrieved by a user. The user interface 18 includes any means appropriate for entering or retrieving data, information or instructions, notably visual, tactile and / or audio capacities that can encompass any or several of the following means as well known by a person skilled in the art: a screen, a keyboard, a trackball, a touchpad, a touchscreen, a loudspeaker, a voice recognition system.

[0142] In its automatic actions, the device 2 may for example execute the following process (Figure 4):- receiving the demographic and anthropomorphic data 121, the mental health data 122, the data relative to at least one treatment taken by said patient 123, together with the trained machine learning prediction model 30 (step 141),- optionally preprocessing the demographic and anthropomorphic data 121, the mental health data 122 and the data relative to at least one treatment taken by said patient 123 (step 142),- calculating at least one demographic and anthropomorphic score from said demographic and anthropomorphic data 121 collected on said patient, at least one mental health score from said mental health data 122 collected on said patient and at least one treatment score from said data relative to at least one treatment 123 taken by said subject (step 143),- feeding said at least one demographic and anthropomorphic score, said mental health score and said treatment score to said trained machine learning model 30 so as to output said prediction (step 144),- comparing said prediction to at least one threshold so as to determine a treatment recommendation (step 145),

[0143] A particular apparatus 9, visible on Figure 5, is embodying the device 1 as well as the device 2 described above. It corresponds for example to a workstation, a laptop, a tablet, a smartphone, or a head-mounted display (HMD).

[0144] That apparatus 9 is suited to IVF outcome predictions and to related ML training. It comprises the following elements, connected to each other by a bus 95 of addresses and data that also transports a clock signal:- a microprocessor 91 (or CPU);- a graphics card 92 comprising several Graphical Processing Units (or GPUs) 920 and a Graphical Random Access Memory (GRAM) 921;- a non-volatile memory of ROM type 96;- a RAM 97;- one or several I / O (Input / Output) devices 94 such as for example a keyboard, a mouse, a trackball, a webcam; other modes for introduction of commands such as for example vocal recognition are also possible;- a power source 98; and- a radiofrequency unit 99.

[0145] According to a variant, the power supply 98 is external to the apparatus 9.

[0146] The apparatus 9 also comprises a display device 93 of display screen type directly connected to the graphics card 92 to display synthesized images calculated and composed in the graphics card. The use of a dedicated bus to connect the display device 93 to the graphics card 92 offers the advantage of having much greater data transmission bitrates and thus reducing the latency time for the displaying of images composed by the graphics card. According to a variant, a display device is external to apparatus 9 and is connected thereto by a cable or wirelessly for transmitting the display signals. The apparatus 9, for example through the graphics card 92, comprises an interface for transmission or connection adapted to transmit a display signal to an external display means such as for example an LCD or plasma screen or a video-projector. In this respect, the RF unit 99 can be used for wireless transmissions.

[0147] It is noted that the word “register” used hereinafter in the description of memories 97 and 921 can designate in each of the memories mentioned, a memory zone of low capacity (some binary data) as well as a memory zone of large capacity (enabling a whole program to be stored or all or part of the data representative of data calculated or to be displayed). Also, the registers represented for the RAM 97 and the GRAM 921 can be arranged and constituted in any manner, and each of them does not necessarily correspond to adjacent memory locations and can be distributed otherwise (which covers notably the situation in which one register includes several smaller registers).

[0148] When switched-on, the microprocessor 91 loads and executes the instructions of the program contained in the RAM 97.

[0149] As will be understood by a skilled person, the presence of the graphics card 92 is not mandatory, and can be replaced with entire CPU processing and / or simpler visualization implementations.

[0150] In variant modes, the apparatus 9 may include only the functionalities of the device 1, and not those of the device 6. In addition, the device 1 and / or the device 6 may be implemented differently than a standalone software, and an apparatus or set ofapparatus comprising only parts of the apparatus 9 may be exploited through an API call or via a cloud interface.

[0151] According to another aspect, the invention relates to a method for providing a treatment recommendation for a patient undergoing a treatment with Dopamine Agonists, using a device as described herein.

[0152] According to one embodiment, the invention relates to a method for providing a treatment recommendation for a patient undergoing a treatment with Dopamine Agonists, said method comprising: a) predicting the occurrence of at least one Impulse Control Disorder (ICD), b) assigning the patient to a risk group of developing at least one ICD based on the comparison of the prediction determined in step a) with two thresholds, including a first threshold and a second threshold, the first threshold being inferior to the second threshold, wherein: o said patient is assigned to a low risk of developing at least one ICD when the prediction is inferior to the first threshold and the second threshold; o said patient is assigned to a medium risk of developing at least one Impulse Control Disorder when the prediction is superior to a first threshold and inferior to the second threshold; and o said patient is assigned to a high risk of developing at least one Impulse Control Disorder when the prediction is superior to the first threshold and the second threshold; and c) providing a treatment recommendation depending on the identified risk.

[0153] The patient undergoing a treatment by Dopamine Agonists may suffer from a disease or condition associated with dopamine dysregulation. Patients suffering such disease or condition requiring use of Dopamine Agonists may develop ICD. As used herein, a disease associated with dopamine dysregulation is a disease or condition associated with dysfunction in the dopaminergic system. Examples of such diseases include, without limitation, Parkinson’s Disease, Restless Leg Syndrome,Hyperprolactinemia, depression or apathy, acromegaly, Tourette Syndrome, pituitary adenoma, Attention-Deficit / Hyperactivity Disorder (ADHD) or narcolepsy.

[0154] Thus, the patient as described herein may suffer from Parkinson’s Disease, Restless Leg Syndrome, Hyperprolactinemia, depression or apathy, acromegaly, Tourette Syndrome, pituitary adenoma, ADHD or narcolepsy.

[0155] According to one specific embodiment, the patient suffers from Parkinson’s Disease (PD). As used herein, PD refers to a progressive, neurodegenerative disease characterized by a tremor that is maximal at rest, retropulsion (z.e. a tendency to fall backwards), rigidity, stooped posture, slowness of voluntary movements, and a masklike facial expression. Pathologic features include loss of melanin containing neurons in the substantia nigra.

[0156] According to one specific embodiment, the patient suffers from Restless Leg Syndrome. According to one specific embodiment, the patient suffers from Hyperprolactinemia. According to one specific embodiment, the patient suffers from depression or apathy. According to one specific embodiment, the patient suffers from acromegaly. According to one specific embodiment, the patient suffers from Tourette Syndrome. According to one specific embodiment, the patient suffers from pituitary adenoma. According to one specific embodiment, the patient suffers ADHD. According to one specific embodiment, the patient suffers from narcolepsy.

[0157] Dopamine Agonists as described herein may be ergot-derived Dopamine Agonists. As used herein, ergot-derived Dopamine Agonists refers to Dopamine agonists that are derived from ergot. In particular, ergot-derived Dopamine Agonists may be Dopamine Agonists selected from the group comprising or consisting of cabergoline (CAS number : 81409-90-7), bromocriptine (CAS number : 25614-03-3), pergolide (CAS number 66104-22-1), lisuride (CAS number : 8016-80-3) and dihydroergocryptine (CAS number : 25447-66-9).

[0158] Dopamine Agonists as described herein may be non-ergot-derived Dopamine Agonists. In particular, non-ergot-derived Dopamine Agonists may be Dopamine Agonists selected from the group comprising or consisting of pramipexole (CAS number : 104632-26-0), ropinirole (CAS number : 91374-21-9), rotigotine (CAS number : 92206-54-7), apomorphine (CAS number : 41372-20-7), dihydrexidine (CAS number : 137417- 08-4), Piribedil (CAS number : 3605-01-4), dopexamine (CAS number : 86484-91-5) and quinagolide (CAS number : 140630-79-1).

[0159] Preferably, Dopamine Agonists as described herein may be non-ergot-derived Dopamine Agonists, and, in particular, may be selected from the group comprising or consisting of pramipexole, ropinirole, rotigotine, apomorphine, dihydrexidine, Piribedil, dopexamine and quinagolide.

[0160] The method as described herein may be a computer-implemented method and the step of predicting the occurrence of at least one Impulse Control Disorder (ICD) may be performed using a machine learning prediction model previously trained on a training dataset comprising, for each subject of a plurality of subjects, at least one demographic and anthropomorphic score derived from demographic and anthropomorphic data collected on said subject, at least one mental health score derived from mental health data collected on said subject and at least one treatment score derived from data relative to said treatment with Dopamine Agonists, said method further comprising: receiving as input: o demographic and anthropomorphic data, o mental health data, and o data relative to said treatment with Dopamine Agonist taken by said patient, calculating at least one demographic and anthropomorphic score from said demographic and anthropomorphic data collected on said patient, at least one mental health score from said mental health data collected on said patient and at least one treatment score from said data relative to said treatment with Dopamine Agonists,- feeding said demographic and anthropomorphic score, mental health score and treatment score to said trained machine learning model so as to output said prediction.

[0161] Alternatively, the step of predicting the occurrence of at least one Impulse Control Disorder (ICD) may be performed using a machine learning prediction model previously trained on a training dataset comprising, for each subject of a plurality of subjects, at least one demographic and anthropomorphic score derived from demographic and anthropomorphic data collected on said subject, at least one mental health score derived from mental health data collected on said subject, at least one treatment score derived from data relative to said treatment with Dopamine Agonists, and at least one genetic score derived from genetic data, said method further comprising: receiving as input: o demographic and anthropomorphic data, o mental health data, o data relative to said treatment with Dopamine Agonist taken by said patient, and o genetic data calculating at least one demographic and anthropomorphic score from said demographic and anthropomorphic data collected on said patient, at least one mental health score from said mental health data collected on said patient, at least one treatment score from said data relative to said treatment with Dopamine Agonists, and at least one genetic score from genetic data, feeding said demographic and anthropomorphic score, mental health score, treatment score and genetic data to said trained machine learning model so as to output said prediction.

[0162] The genetic data may be as described herein. In particular, the genetic data may be genetic variants of one or more of the following genes for instance but not restricted to this list : DRD2, DRD3, DAT1, COMT, DDC, GRIN2B, ADRA2C, SERT, TPH2, HTR2A, OPRK1 and OPRMP The genetic characteristics may be the genetic variant(s) of multiple and numerous genes associated in a combination of variants to inform treatment recommendation.

[0163] Said at least one threshold may be set to detect subjects that will develop at least one Impulse Control Disorder and to prevent Impulse Control Disorder without needing a recommendation to stop treatment in most cases.

[0164] In one example, the least one threshold is set so as to detect at least 65% of all subjects that will develop Impulse Control Disorder, while less than half of the subjects receive a recommendation to stop the treatment.

[0165] In another example, the least one threshold is set so as to detect at least 55% of all subjects that will develop Impulse Control Disorder, while less than half of the subjects receive a recommendation to stop treatment.

[0166] It will be understood to the skilled artisan in the art that the method as described herein may enable to stratify patients according to the risk of developing at least one Impulse Control Disorder or to predict the risk of developing at least one Impulse Control Disorder in a patient, and to adapt the treatment depending on the identified risk.

[0167] The treatment recommendation may comprise adapting the treatment dosage.

[0168] The treatment recommendation may be:- to continue the treatment with Dopamine Agonists, when the patient is assigned to a low risk of developing at least one Impulse Control Disorder, and / or- to reduce the dosage of the Dopamine Agonists, when the patient is assigned to a medium risk of developing at least one Impulse Control Disorder, and / or- to stop the treatment with Dopamine Agonists, when the patient is assigned to a high risk of developing at least one Impulse Control Disorder.

[0169] It will be understood to the skilled artisan in the art that, when the patient is assigned to a low risk of developing at least one Impulse Control Disorder, the patient can continue the treatment with Dopamine Agonists as deemed appropriate. In particular, the treatment recommendation may allow continuation, reduction, or increase of the dosage of the Dopamine Agonists treatment as deemed appropriate, as well as combinations with any other medication for treating the disease or condition.

[0170] It will be understood to the skilled artisan in the art that, when the patient is assigned to a medium risk of developing at least one Impulse Control Disorder, the patient should decrease the dosage of Dopamine Agonists. In particular, the treatment recommendation may be to: i) replace the treatment with Dopamine Agonists with another medication for the disease or condition, or ii) reduce the dosage of the treatment with Dopamine Agonists and supplement it with another medication for the disease or condition.

[0171] When the patient is assigned to a medium risk of developing at least one Impulse Control Disorder, the treatment recommendation may be to reduce the dosage of the treatment with Dopamine Agonists to a dose equivalent to 40mg of levodopa, and supplement the reduction with another medication for the disease or condition. It will be understood that the equivalence doses between Dopamine Agonists and levodopa are well known by the skilled artisan in the art.

[0172] It will be understood to the skilled artisan in the art that, when the patient is assigned to a high risk of developing at least one Impulse Control Disorder, the patient should stop treatment with Dopamine Agonists. The treatment recommendation may be to replace treatment with Dopamine Agonists with another medication for the disease or condition.

[0173] When the patient suffers from restless leg syndrome, the treatment recommendation may be:- to continue the treatment with Dopamine Agonists, when the patient is assigned to a low risk of developing at least one Impulse Control Disorder, and / or- to reduce the dosage of the treatment with Dopamine Agonists, when said patient is assigned to a medium risk of developing at least one Impulse Control Disorder, and / or- to stop the treatment with Dopamine Agonists and replace it with another medication for treating restless leg syndrome, when said patient is assigned to a high risk of developing at least one Impulse Control Disorder.

[0174] When the patient suffers from restless leg syndrome, the treatment recommendation may be to treat the patient with one or more of the following medications: Dopamine Agonists such as pramipexole, ropinirole, rotigotine, and cabergoline, antiseizure / antiepileptic drugs such as gabapentine, gabaline, pregabaline and their derivatives, benzodiazepines, opioids such as fentanyl and prolonged-release oxycodone / naloxone, ferric carboxymaltose, ferrous sulfate, levodopa, vitamin C and E supplementation, depending on the identified risk.

[0175] When said patient suffering from restless leg syndrome is assigned to a low risk of developing at least one Impulse Control Disorder, the treatment recommendation may be to continue the treatment with Dopamine Agonists as deemed appropriate. In particular, the treatment recommendation may allow continuation, reduction, or increase of the dosage of the Dopamine Agonists treatment as deemed appropriate, as well as combinations with any other medication for treating restless leg syndrome.

[0176] Examples of treatments for treating restless leg syndrome are mentioned hereinabove. Thus, when the patient suffering from restless leg syndrome is assigned to a low risk of developing at least one Impulse Control Disorder, said patient may be treated with one or more of the following medications: Dopamine Agonists such as pramipexole, ropinirole, rotigotine, and cabergoline, antiseizure / antiepileptic drugs such as gabapentine, gabaline, pregabaline and their derivatives, benzodiazepines, opioids such as fentanyl and prolonged-release oxycodone / naloxone, ferric carboxymaltose, ferrous sulfate, levodopa, vitamin C and E supplementation.

[0177] When said patient is assigned to a medium risk of developing at least one Impulse Control Disorder, the treatment recommendation may be to: i) replace the treatment with Dopamine Agonists with other medications for treating restless leg syndrome, or ii) reduce the dosage of the treatment with Dopamine Agonists and supplement it with other medications for treating restless leg syndrome if necessary.

[0178] When said patient is assigned to a high risk of developing at least one Impulse Control Disorder, the treatment recommendation may be to treat the patient suffering from restless leg syndrome with medications other than Dopamine Agonists.

[0179] Medications other than Dopamine Agonists for treating restless leg syndrome are mentioned hereinabove. Medications other than Dopamine Agonists include, without limitation, antiseizure / antiepileptic drugs such as gabapentine, gabaline, pregabaline and their derivatives, benzodiazepines, opioids such as fentanyl and prolonged-release oxycodone / naloxone, ferric carboxymaltose, ferrous sulfate, levodopa, vitamin C and E supplementation.

[0180] When the patient suffers from pituitary adenoma, the treatment recommendation may be:- to continue the treatment with Dopamine Agonists, when the patient is assigned to a low risk of developing at least one Impulse Control Disorder, and / or- to reduce the dosage of the treatment with Dopamine Agonists, when said patient is assigned to a medium risk of developing at least one Impulse Control Disorder, and / or- to stop the treatment with Dopamine Agonists and replace it with another medication or treatment for treating pituitary adenoma, when said patient is assigned to a high risk of developing at least one Impulse Control Disorder.

[0181] When the patient suffers from pituitary adenoma, the treatment recommendation may be to treat the patient with one or more of the following medications: Dopamine Agonists such as cabergoline, bromocriptine, quinagolide, hormone replacement therapy, alkylating chemotherapeutic agents such as temozolomide, immunotherapy with checkpoint inhibitors such as ipilimumab and nivolumab, targeted oncological agents such as everolimus, bevacizumab and lapatinib, oestrogen receptor modulators such as tamoxifen and peptide receptor radionuclide treatment, depending on the identified risk.

[0182] The patient may be also treated with treatments other than medications, such as surgery, radiation or active surveillance.

[0183] When said patient suffering from pituitary adenoma is assigned to a low risk of developing at least one Impulse Control Disorder, the treatment recommendation may be to continue the treatment with Dopamine Agonists as deemed appropriate. In particular, the treatment recommendation may allow continuation, reduction, or increase of thedosage of the Dopamine Agonists treatment as deemed appropriate, as well as combinations with any other medication or treatment for treating pituitary adenoma.

[0184] Examples of medications or treatments for treating pituitary adenoma are mentioned hereinabove. Thus, when the patient suffering from pituitary adenoma is assigned to a low risk of developing at least one Impulse Control Disorder, said patient may be treated with one or more of the following medications: Dopamine Agonists such as cabergoline, bromocriptine, quinagolide, hormone replacement therapy, alkylating chemotherapeutic agents such as temozolomide, immunotherapy with checkpoint inhibitors such as ipilimumab and nivolumab, targeted oncological agents such as everolimus, bevacizumab and lapatinib, oestrogen receptor modulators such as tamoxifen and peptide receptor radionuclide treatment. The patient may be also treated with treatments other than medications, such as surgery, radiation or active surveillance.

[0185] When said patient is assigned to a medium risk of developing at least one Impulse Control Disorder, the treatment recommendation may be to: i) replace the treatment with Dopamine Agonists with other medications or treatments for treating pituitary adenoma, or ii) reduce the dosage of the treatment with Dopamine Agonists and supplement it (or compensate it) with other medications or treatments for treating pituitary adenoma if necessary.

[0186] When said patient is assigned to a high risk of developing at least one Impulse Control Disorder, the treatment recommendation may be to treat the patient suffering from pituitary adenoma with medications or treatments other than Dopamine Agonists.

[0187] Medications and treatments other than Dopamine Agonists for treating pituitary adenoma are mentioned hereinabove. Medications and treatments other than Dopamine Agonists include, without limitation, hormone replacement therapy, alkylating chemotherapeutic agents such as temozolomide, immunotherapy with checkpoint inhibitors such as ipilimumab and nivolumab, targeted oncological agents such as everolimus, bevacizumab and lapatinib, oestrogen receptor modulators such as tamoxifen and peptide receptor radionuclide treatment, surgery, radiation and active surveillance.

[0188] When the patient suffers from depression or apathy, the treatment recommendation may be:- to continue the treatment with Dopamine Agonists, when the patient is assigned to a low risk of developing at least one Impulse Control Disorder, and / or- to reduce the dosage of the treatment with Dopamine Agonists, when said patient is assigned to a medium risk of developing at least one Impulse Control Disorder, and / or- to stop the treatment with Dopamine Agonists and replace it with another medication or treatment for treating depression or apathy, when said patient is assigned to a high risk of developing at least one Impulse Control Disorder.

[0189] When the patient suffers from depression or apathy, the treatment recommendation may be to treat the patient with one or more of the following medications: Dopamine Agonists, serotonin reuptake inhibitors (SSRIs), serotonin and norepinephrine reuptake inhibitors (SNRIs), tricyclic antidepressants (TCAs), monoamine oxidase inhibitors (MAOIs) or psychedelics, depending on the identified risk. The patient may be also treated with treatments other than medications, such as Repetitive transcranial magnetic stimulation (rTMS).

[0190] When said patient suffering from depression or apathy is assigned to a low risk of developing at least one Impulse Control Disorder, the treatment recommendation may be to continue the treatment with Dopamine Agonists as deemed appropriate. In particular, the treatment recommendation may allow continuation, reduction, or increase of the dosage of the Dopamine Agonists treatment as deemed appropriate, as well as combinations with any other medication or treatment for treating depression or apathy.

[0191] Examples of medications or treatments for treating depression or apathy are mentioned hereinabove. Thus, when the patient suffering from depression or apathy is assigned to a low risk of developing at least one Impulse Control Disorder, said patient may be treated with one or more of the following medications: Dopamine Agonists, serotonin reuptake inhibitors (SSRIs), serotonin and norepinephrine reuptake inhibitors (SNRIs), tricyclic antidepressants (TCAs), monoamine oxidase inhibitors (MAOIs) orpsychedelics. The patient may be also treated with treatments other than medications, such as Repetitive transcranial magnetic stimulation (rTMS).

[0192] When said patient is assigned to a medium risk of developing at least one Impulse Control Disorder, the treatment recommendation may be to: i) replace the treatment with Dopamine Agonists with other medications or treatments for treating depression or apathy, or ii) reduce the dosage of the treatment with Dopamine Agonists and supplement it (or compensate it) with other medications or treatments for treating depression or apathy if necessary.

[0193] When said patient is assigned to a high risk of developing at least one Impulse Control Disorder, the treatment recommendation may be to treat the patient suffering from depression or apathy with medications or treatments other than Dopamine Agonists.

[0194] Medications and treatments other than Dopamine Agonists for treating depression or apathy are mentioned hereinabove. Medications and treatments other than Dopamine Agonists include, without limitation, serotonin reuptake inhibitors (SSRIs), serotonin and norepinephrine reuptake inhibitors (SNRIs), tricyclic antidepressants (TCAs), monoamine oxidase inhibitors (MAOIs), psychedelics, and repetitive transcranial magnetic stimulation (rTMS).

[0195] When the patient suffers from Parkinson’s Disease, the treatment recommendation may be:- to continue the treatment with Dopamine Agonists, when the patient is assigned to a low risk of developing at least one Impulse Control Disorder, and / or- to reduce the dosage of the treatment with Dopamine Agonists, when said patient is assigned to a medium risk of developing at least one Impulse Control Disorder, and / or- to stop the treatment with Dopamine Agonists and replace it with levodopa, when said patient is assigned to a high risk of developing at least one Impulse Control Disorder.

[0196] When the patient suffers from Parkinson’s Disease, the treatment recommendation may be to treat the PD patient with one or more of the following medications: Amantadine, levodopa, Dopamine Agonists, catechol-O-methyltransferase (COMT) inhibitors, and Monoamine Oxydase Inhibitor B (IMAO-B), depending on the identified risk.

[0197] When said patient suffering from PD is assigned to a low risk of developing at least one Impulse Control Disorder, the treatment recommendation may be to continue the treatment with Dopamine Agonists as deemed appropriate. In particular, the treatment recommendation may allow continuation, reduction, or increase of the dosage of the Dopamine Agonists treatment as deemed appropriate, as well as combinations with any other medication for treating Parkinson’s disease.

[0198] Examples of medications other than Dopamine Agonists for treating Parkinson’ s disease include, without limitation, levodopa, amantadine, catechol-O-methyltransferase (COMT) inhibitors or Monoamine Oxydase Inhibitor B (IMAO-B). Thus, when the patient is assigned to a low risk of developing at least one Impulse Control Disorder, said patient may be treated with one or more of the following medications: Amantadine, levodopa, Dopamine Agonists, catechol-O-methyltransferase (COMT) inhibitors and Monoamine Oxydase Inhibitor B (IMAO-B).

[0199] When said patient is assigned to a medium risk of developing at least one Impulse Control Disorder, the treatment recommendation may be to: i) replace the treatment with Dopamine Agonists with levodopa, or ii) reduce the dosage of the treatment with Dopamine Agonists and supplement it with levodopa if necessary, in PD patients.

[0200] When said patient is assigned to a medium risk of developing at least one Impulse Control Disorder, the treatment recommendation may be to reduce the dosage of the treatment with Dopamine Agonists to a dose equivalent to 40mg of levodopa, and supplement the reduction with levodopa in PD patient.

[0201] Thus, when said patient is assigned to a medium risk of developing at least one Impulse Control Disorder, the treatment recommendation may be to treat the PD patient with levodopa and / or Dopamine Agonists.

[0202] When said patient is assigned to a high risk of developing at least one Impulse Control Disorder, the treatment recommendation may be to treat the PD patient with medications other than Dopamine Agonists, and in particular, levodopa.

[0203] According to another aspect, the invention relates to a method for treating a patient having a disease or condition associated with dopamine dysregulation and undergoing a treatment with Dopamine Agonists, using a device as described herein.

[0204] The method as described herein may enable to prevent or reduce the risk of developing at least one Impulse Control Disorder. Thus, in other words, the present invention relates to a method for preventing or reducing the risk of developing at least one Impulse Control Disorder in a patient having a disease or condition associated with dopamine dysregulation and undergoing a treatment with Dopamine Agonists.

[0205] The disease or condition associated with dopamine dysregulation may be as described herein. In particular, the disease or condition associated with dopamine dysregulation may be Parkinson’s disease.

[0206] The Dopamine Agonist may be as described herein.

[0207] According to one embodiment, the method comprises : a) predicting the occurrence of at least one Impulse Control Disorder (ICD), b) assigning the patient to a risk group of developing at least one Impulse Control Disorder based on the comparison of the prediction determined in step a) with two thresholds, including a first threshold and a second threshold, the first threshold being inferior to the second threshold, wherein: o said patient is assigned to a low risk of developing at least one Impulse Control Disorder when the prediction is inferior to the first threshold and the second threshold; o said patient is assigned to a medium risk of developing at least one Impulse Control Disorder when the prediction is superior to a first threshold and inferior to the second threshold; ando said patient is assigned to a high risk of developing at least one Impulse Control Disorder when the prediction is superior to the first threshold and the second threshold; c) treating said patient with Dopamine Agonists, and / or any other medications for the disease or condition, depending on the identified risk.

[0208] The step (c) may comprise: continuing the treatment with Dopamine Agonists, when the patient is assigned to a low risk of developing at least one Impulse Control Disorder, and / or reducing the dosage of the treatment with Dopamine Agonists, when said patient is assigned to a medium risk of developing at least one Impulse Control Disorder, and / or stopping the treatment with Dopamine Agonists, when said patient is assigned to a high risk of developing at least one Impulse Control Disorder.

[0209] When the patient is assigned to a low risk of developing at least one Impulse Control Disorder, the step c) may comprise continuing the treatment with Dopamine Agonists as deemed appropriate. In particular, the step c) may comprises continuing, reducing, or increasing the dosage of the Dopamine Agonists treatment as deemed appropriate, as well as combining it with any other medication for treating the disease or condition.

[0210] When the patient is assigned to a medium risk of developing at least one Impulse Control Disorder, the step c) may comprise: i) replacing the treatment with Dopamine Agonists with another medication for the disease or condition, or ii) reducing the dosage of the treatment with Dopamine Agonists and supplementing it with another medication for the disease or condition.

[0211] When the patient is assigned to a medium risk of developing at least one Impulse Control Disorder, the step c) may comprise reducing the dosage of the treatment with Dopamine Agonists to a dose equivalent to 40mg of levodopa, and supplementing the reduction with another medication for the disease or condition.

[0212] When said patient is assigned to a high risk of developing at least one Impulse Control Disorder, the treatment may be to replace the treatment with Dopamine Agonists with another medication for the disease or condition.

[0213] When the patient suffers from restless leg syndrome, the step (c) may comprise: continuing the treatment with Dopamine Agonists, when said patient is assigned to a low risk of developing at least one Impulse Control Disorder, and / or reducing the dosage of the treatment with Dopamine Agonists, when said patient is assigned to a medium risk of developing at least one Impulse Control Disorder, and / or stopping the dosage of the treatment with Dopamine Agonists, and replace it with another medication for treating restless leg syndrome, when said patient is assigned to a high risk of developing at least one Impulse Control Disorder.

[0214] When the patient suffers from restless leg syndrome, the step c) may comprise treating the patient with one or more of the following medications: Dopamine Agonists such as pramipexole, ropinirole, rotigotine, and cabergoline, antiseizure / antiepileptic drugs such as gabapentine, gabaline, pregabaline and their derivatives, benzodiazepines, opioids such as fentanyl and prolonged-release oxycodone / naloxone, ferric carboxymaltose, ferrous sulfate, levodopa, vitamin C and E supplementation, depending on the identified risk.

[0215] When said patient suffering from restless leg syndrome is assigned to a low risk of developing at least one Impulse Control Disorder, the step c) may comprise continuing the treatment with Dopamine Agonists as deemed appropriate. In particular, the step c) may comprise continuing, reducing, or increasing the dosage of the Dopamine Agonists treatment as deemed appropriate, as well as combining it with any other medication for treating restless leg syndrome.

[0216] Examples of medications for treating restless leg syndrome are mentioned hereinabove.

[0217] When said patient is assigned to a medium risk of developing at least one Impulse Control Disorder, the step c) may comprise: i) replacing the treatment with DopamineAgonists with other medications for treating restless leg syndrome, or ii) reducing the dosage of the treatment with Dopamine Agonists and supplementing it with other medications for treating restless leg syndrome if necessary.

[0218] When said patient is assigned to a high risk of developing at least one Impulse Control Disorder, the step c) may comprise treating the patient suffering from restless leg syndrome with medications other than Dopamine Agonists.

[0219] Medications other than Dopamine Agonists for treating restless leg syndrome are mentioned hereinabove.

[0220] When the patient suffers from pituitary adenoma, the step (c) may comprise: continuing the treatment with Dopamine Agonists, when said patient is assigned to a low risk of developing at least one Impulse Control Disorder, and / or reducing the dosage of the treatment with Dopamine Agonists, when said patient is assigned to a medium risk of developing at least one Impulse Control Disorder, and / or stopping the dosage of the treatment with Dopamine Agonists, and replace it with another medication or treatment for pituitary adenoma, when said patient is assigned to a high risk of developing at least one Impulse Control Disorder.

[0221] When the patient suffers from pituitary adenoma, the step c) may comprise treating the patient with one or more of the following medications: Dopamine Agonists such as cabergoline, bromocriptine, quinagolide, hormone replacement therapy, alkylating chemotherapeutic agents such as temozolomide, immunotherapy with checkpoint inhibitors such as ipilimumab and nivolumab, targeted oncological agents such as everolimus, bevacizumab and lapatinib, oestrogen receptor modulators such as tamoxifen and peptide receptor radionuclide treatment, depending on the identified risk. The patient may be also treated with treatments other than medications, such as surgery, radiation or active surveillance.

[0222] When said patient suffering from pituitary adenoma is assigned to a low risk of developing at least one Impulse Control Disorder, the step c) may comprise continuing the treatment with Dopamine Agonists as deemed appropriate. In particular, the step c)may comprise continuing, reducing, or increasing the dosage of the Dopamine Agonists treatment as deemed appropriate, as well as combining it with any other medication or treatment for treating pituitary adenoma.

[0223] Examples of medications or treatments for treating pituitary adenoma are mentioned hereinabove.

[0224] When said patient is assigned to a medium risk of developing at least one Impulse Control Disorder, the step c) may comprise: i) replacing the treatment with Dopamine Agonists with other medications or treatments for treating pituitary adenoma, or ii) reducing the dosage of the treatment with Dopamine Agonists and supplementing it (or compensate it) with other medications or treatments for treating pituitary adenoma if necessary.

[0225] When said patient is assigned to a high risk of developing at least one Impulse Control Disorder, the step c) may comprise treating the patient suffering from pituitary adenoma with medications or treatments other than Dopamine Agonists.

[0226] Medications and treatments other than Dopamine Agonists for treating pituitary adenoma are mentioned hereinabove.

[0227] When the patient suffers from depression or apathy, the step (c) may comprise: continuing the treatment with Dopamine Agonists, when said patient is assigned to a low risk of developing at least one Impulse Control Disorder, and / or reducing the dosage of the treatment with Dopamine Agonists, when said patient is assigned to a medium risk of developing at least one Impulse Control Disorder, and / or stopping the dosage of the treatment with Dopamine Agonists, and replace it with another medication or treatment for depression or apathy, when said patient is assigned to a high risk of developing at least one Impulse Control Disorder.

[0228] When the patient suffers from depression or apathy, the step c) may comprise treating the patient with one or more of the following medications: Dopamine Agonists, serotonin reuptake inhibitors (SSRIs), serotonin and norepinephrine reuptake inhibitors(SNRIs), tricyclic antidepressants (TCAs), monoamine oxidase inhibitors (MAOIs) or psychedelics, depending on the identified risk. The patient may be also treated with treatments other than medications, such as Repetitive transcranial magnetic stimulation (rTMS).

[0229] When said patient suffering from depression or apathy is assigned to a low risk of developing at least one Impulse Control Disorder, the step c) may comprise continuing the treatment with Dopamine Agonists as deemed appropriate. In particular, the step c) may comprise continuing, reducing, or increasing the dosage of the Dopamine Agonists treatment as deemed appropriate, as well as combining it with any other medication or treatment for treating depression or apathy.

[0230] Examples of medications or treatments for treating depression or apathy are mentioned hereinabove.

[0231] When said patient is assigned to a medium risk of developing at least one Impulse Control Disorder, the step c) may comprise: i) replacing the treatment with Dopamine Agonists with other medications or treatments for treating depression or apathy, or ii) reducing the dosage of the treatment with Dopamine Agonists and supplementing it with other medications or treatments for treating depression or apathy if necessary.

[0232] When said patient is assigned to a high risk of developing at least one Impulse Control Disorder, the step c) may comprise treating the patient suffering from depression or apathy with medications or treatments other than Dopamine Agonists.

[0233] Medications and treatments other than Dopamine Agonists for treating depression or apathy are mentioned hereinabove.

[0234] When the patient suffers from Parkinson’s Disease, the step (c) may comprise: continuing the treatment with Dopamine Agonists, when said patient is assigned to a low risk of developing at least one Impulse Control Disorder, and / or reducing the dosage of the treatment with Dopamine Agonists, when said patient is assigned to a medium risk of developing at least one Impulse Control Disorder, and / orstopping the dosage of the treatment with Dopamine Agonists, and replace it with levodopa, when said patient is assigned to a high risk of developing at least one Impulse Control Disorder.

[0235] When the patient suffers from Parkinson’s Disease, the step c) may comprise treating the patient with one or more of the following medications: Amantadine, levodopa, Dopamine Agonists, catechol-O-methyltransferase (COMT) inhibitors and Monoamine Oxydase Inhibitor B (IMAO-B), depending on the identified risk.

[0236] When said patient is assigned to a low risk of developing at least one Impulse Control Disorder, the step c) may comprise continuing the treatment with Dopamine Agonists as deemed appropriate. In particular, the step c) may comprise continuing, reducing, or increasing the dosage of the Dopamine Agonists treatment as deemed appropriate, as well as combining it with any other medication for treating Parkinson’s disease.

[0237] Thus, when the patient is assigned to a low risk of developing at least one Impulse Control Disorder, the patient may be treated with one or more of the following medications: Amantadine, levodopa, Dopamine Agonists, catechol-O-methyltransferase (COMT) inhibitors and Monoamine Oxydase Inhibitor B (IMAO-B).

[0238] When said patient is assigned to a medium risk of developing at least one Impulse Control Disorder, the step c) may comprise: i) replacing the treatment with Dopamine Agonists with levodopa, or ii) reducing the dosage of the treatment with Dopamine Agonists and supplementing it with levodopa if necessary, in PD patients.

[0239] When said patient is assigned to a medium risk of developing at least one Impulse Control Disorder, the step c) may comprise reducing the dosage of the treatment with Dopamine Agonists to a dose equivalent to 40mg of levodopa, and supplement the reduction with levodopa in PD patients.

[0240] Thus, when said patient is assigned to a medium risk of developing at least one Impulse Control Disorder, the step c) may comprise treating the PD patient with levodopa and / or Dopamine Agonists.

[0241] When said patient is assigned to a high risk of developing at least one Impulse Control Disorder, the step c) may comprise treating the PD patient with medications other than Dopamine Agonists, and in particular, levodopa.

[0242] In cases where a change of treatment is advised, such as reduction or stopping of Dopamine Agonists, a transition phase to the new treatment regimen may be implemented. Said transition phase may be of 3 -month, for instance.

[0243] The treatment may be administered with a therapeutically effective amount.EXAMPLES

[0244] The present invention is further illustrated by the following example of a clinical trial to show its efficacy.Efficacy of a prediction model-based algorithm to prevent drug-induced Impulse Control Disorders in Parkinson’s diseaseMain objective and primary endpoint

[0245] Main objective: To assess the efficacy of ICD SHIELD app, a software with a computer-based algorithm assisting clinicians in the prescription of dopaminergic medications, on the prevention of ICDs within a 2 year follow up period in PD patients, compared to the standard of care.

[0246] Primary endpoint: Rate of patients with at least one clinically significant ICD over the 2 year-follow upSecondary objectives and endpoints

[0247] Secondary objectives:To assess the impact of a software with a computer-based algorithm assisting clinicians in the prescription of dopaminergic medications within a 2 year follow up period in PD patients, compared to the standard of care on global disease severity, medication adjustment, other indices of ICDs, and on other PD motor and non-motor symptoms:To describe the observance and acceptability of the software with a computer- based algorithm for clinicians in the prescription of dopaminergic medications, by type of recommendation.

[0248] Secondary endpoints:- Perception of global disease severity o by the patient: change in disease severity of PD on the Patient Global Impression of Improvement (PGI-I) scale filled by the patient over the 2 year-follow up o by the clinician: change in disease severity of PD on the Clinical Global Impression of Improvement (CGI-I) scale filled by the clinician over the 2 year-follow up- Effect on medication adjustment endpoints: o Total levodopa equivalent daily dose (LEDD) over the 2 year-follow up o Proportions of DA (Dopamine Agonists), levodopa, COMT inhibitors, MAO-B inhibitors, amantadine treatments out of total LEDD over the 2 year-follow up o Cumulative DA daily doses and cumulative levodopa daily dose, assessed by the sum of the daily doses at each visit (M6, M12, Ml 8, M24).- Efficacy secondary endpoints: o Time to onset of first occurrence of clinically significant (MILD or above) ICDRBs (at least one ASBPD “Ardouin Scale of Behavior in Parkinson's Disease” subscore in any of the subcategories 3 to 5 and 7 to 10 >2) over the 2 year-long-follow up. o Severity of ICDRBs at time of first occurrence assessed on the QUTP- RS score (Score ranges from 0 to 112, a higher score reflecting a more severe ICDRB) o Severity of ICDRBs at time of first occurrence assessed on ASBPD score (highest severity on part IV of the ASBPD, hyperdopaminergic behaviors, in any of the subcategories 3 to 5 and 7 to 10).o Global severity of ICDRBs at time of first occurrence assessed by the sum of ICDRBs items’ scores on ASBPD (subcategories 3 to 5 and 7 to 10).Impact on motor control of PD endpoints: o Change in motor score on the MDS-UPDRS (MDS-UPDRS III subscore) over the 2 years.(Score between 0 and 132, a higher score reflects a more severe motor state. The MDS-UPDRS III sub-score is the gold standard for measuring the clinical motor state of PD patients) o Change in the dyskinesia and motor fluctuations score on the MDS- UPDRS (MDS-UPDRS IV sub-score part A and B) over the 2 years.Impact on non-motor control of PD: o Change in depression score item 1.3 of the MDS-UPDRS scale over the 2 year-follow up o Change in HADS depression subscore over the 2 year-follow up o Change in anxiety score item 1.4 of the MDS-UPDRS scale over the 2 year-follow up o Change in HADS anxiety subscore over the 2 year-follow up o Change in somnolence score item 1.8 of the MDS-UPDRS scale over the 2 year-follow up o Change in sleep issues score item 1.7 of the MDS-UPDRS scale over the 2 year-follow up o Change in apathy score item 1.5 of the MDS-UPDRS scale over the 2 year-follow upObservance and acceptability o Percentage of visits in which the clinician decided not to follow the software and algorithm recommendation o Reasons for not following the software and algorithm recommendation: A short open-ended questionnaire to assess the acceptability of the intervention. These open-ended questions will be analyzed qualitatively.

[0249] Exploratory objectives: To assess the impact of the software with a computer- based algorithm (i.e. ICD shield app) assisting clinicians in the prescription of dopaminergic medications within a 2 year follow up period in PD patients, compared to the standard of care on:- Efficacy: o Cumulative severity of ICDs within the 2-year period o Subtypes of ICDs o Perception of global disease severity by the caregiver o Patients quality of life o Body weight changeSafety: o Adverse events causing DA decrease or stopping o Occurrence of Dopamine agonist withdrawal syndrome (DAWS) o Additional unscheduled visits due to adverse eventsTo describe the observance and acceptability of the computer-based software for clinicians in the prescription of dopaminergic medications on the percentage of visits not following ICD SHIELD app recommendation and reasons for not following algorithm recommendation, by type of recommendationSubstudy Exploratory objective : o To assess the impact of genetic factors on the efficacy of a computer- based algorithm assisting clinicians in the prescription of dopaminergic medications, on the prevention of ICDs within a 2 year follow up period in PD patients, compared to the SoC. o In the SoC arm, to assess the potential additional value of adding genetic factors in the existing algorithm to predict incidence of ICDs in PD patients.

[0250] Exploratory endpointsCumulative severity of ICDRBs within the 2-year period assessed by the sum of the QUIP -RS score at each visit (M6, M12, Ml 8, M24).Change in each subtype of ICDRBs within the 2-year period assessed by each respective QUIP -RS sub-score (sub-scores A to G). Score ranges from 0 to 16 for each category. A higher score reflects a more severe ICD.Change in the quality of life measured by PDQ-39 scale score over the 2 year- follow up. A higher score reflects a poorer quality of life.Change in body weight (measured at each patient visit) over the 2 year-long- -follow up- Rate of DA stopping or reducing for any other adverse events reason than ICD (lower limbs swelling, Excessive Daytime Sleepiness, fainting or dizziness due to orthostatic hypotension, hallucinations, DA withdrawal syndrome, . . .)- Number and type of serious adverse events- Rate of patients with a DA withdrawal syndrome- Number of additional unscheduled visits due to adverse events to the neurologistSub-study Exploratory endpoints : o Genetic characteristics: 13 candidate variants selected from the DRD2, DRD3, DAT1, COMT, DDC, GRIN2B, ADRA2C, SERT, TPH2, HTR2A, 0PRK1 and 0PRM1 genes. o In the standard of care arm, change in the standardized net benefit statistic, change in true positive rates, change in false positive rates, between the algorithm with and without genomic results, calculated at threshold that would be considered the most relevant for each algorithm.Design

[0251] This Clinical Investigation is a comparative randomized, controlled superiority trial, with 2 parallel groups. Patient and assessors for primary endpoint will be blind to treatment arm allocation. PD patients, treated by DA at inclusion, will be randomized either to the standard of care (SoC) arm, or to the Algorithm-guided arm. Patients allocated to standard of care (SoC) arm will be treated based on clinician’s evaluation applying treatment adaptation based on the 2024 expert consensus guidelines(Debove etal., 2024); and patients allocated to the algorithm -guided arm will be managed as per clinician’s recommendation guided by the ICD SHIELD app recommendation. Visits with the neurologist will be performed at MO, M6, M12, Ml 8, M24 with assessment of medical and treatment history, neurological examination, PGI and CGI scales, ASBPD and QUIP -RS scales, MDS-UPDRS sections I to IV, HAD scale, MoCA scale and PDQ39 questionnaire. A blood sample will be drawn at one of the visits for DNA extraction and biobanking.Population of study participants

[0252] 528 patients with a diagnosis of PD, recruited at 24 PD expert centers of the NSPARK / FCRIN network, treated by dopamine agonists at inclusion.Inclusion criteria

[0253] Inclusion criteria:- Diagnosis of PD according to the 2015 Movement Disorders Society criteria (Postuma et al., 2015), with bradykinesia AND at least ONE of the following: muscular rigidity, or resting tremor; with no other suspected cause of parkinsonism- Disease duration below 6 years included at baseline- No ongoing clinically significant (Mild or above) ICDRBs (any ASBPD sub score <2)- Patients currently treated with DA.Non-inclusion criteria

[0254] Non-inclusion criteria:Atypical or secondary parkinsonism such as supranuclear palsy, multisystem atrophy or drug-induced parkinsonism, etc...- Patients with a cognitive or psychiatric disorder preventing patient’s participation as per investigator’s judgement- Not willing to participate to the study or to sign the consent- Pregnant or lactating woman, or WOCBP tested positive in <serum or urine> pregnancy test- Participation in investigational drug trials within 30 days prior to screening or within 5 half-life of investigational product whatever the longest- Participant not affiliated or beneficiary of a social security system or beneficiary of such a regime.Device under investigation

[0255] The investigational medical device is based on a computer-based algorithm that utilizes a machine learning model, XGBoost to predict ICD and propose a strategy of dopaminergic treatment adaptation to prevent ICD. This model has been cross-validated on previous independent cohorts (DIGPD and PPMI cohorts) and tested on another cohort (ICEBERG). It leverages clinical input data to prevent ICDRBs and recommends treatment adaptation. No part of the medical device comes in contact with the patient.

[0256] This device is used by the clinician, but not by the patient himself / herself. The device will be used at the end of each patient visit, at the time of treatment prescription adaptation.

[0257] The computer-based algorithm takes clinical variables and medications info as inputs to predict ICDs incidence at the next visit and gives a recommendation for dopaminergic medication management.

[0258] The non-identifying clinical input data will be securely submitted through an online interface at each patient visit. This data will be immediately processed by the algorithm, and the interface will provide the clinician with feedback on the recommended therapeutic approach based on the algorithm's output: a Red signal (indicating to STOP DA treatment and replace it with Levodopa), an Orange signal (advising to DECREASE DA to a dose equivalent to 40mg of Levodopa and compensate the reduction with supplement Levodopa), or a Green signal (suggesting no specific recommendation, thereby allowing the continuation, reduction, or increase of DA treatment as deemed appropriate). In cases where a change of treatment is advised such as reduction or stopping of DA, a standardized 3 -month transition phase to the new treatment regimenwill be implemented, with a progressive decrease over the 3 -months period and a phone call visit at 3 months.

[0259] In the Algorithm-based group, after the evaluation of the patient and the clinical inputs entered in the ICD SHIELD app including the planned choice of prescription by the neurologist for the next period, the clinician will receive the therapeutic approach recommended by the ICD SHIELD app depending on the output given by the algorithm. The clinician can repeat the use of the app if he / she plans to try various choice of prescription in the app if deemed necessary, but a single use is recommended at each visit. The neurologist will have to follow the recommendation of the ICD SHIELD app as much as possible unless judged inappropriate. The neurologist makes the final decision.Comparator arm

[0260] The SoC group will follow the usual follow up with their neurologist, that will apply treatment adaptation based on the 2024 expert consensus guidelines (Debove et al., 2024).Additional interventions added for the study

[0261] The patients will undergo their usual every 6 months follow up with the neurologist, with a phone call follow up after 3 months after decrease or stop of DA in case of change of prescription, and no extra follow up visits will be added by the study.

[0262] The additional interventions will be:

[0263] The burden due to the Clinical Investigation could be related to additional interventions that will be:A blood sample drawn either at baseline, or at a follow-up visit if not possible at baseline auto-questionnaires and clinical examination scales: the PGI, CGI-I, MoCA scales, HAD scale, ASBPD and QUIP -RS scales, PDQ39 questionnaire, and the full MDS-UPDRS scale at each visit (some of which are sometimes already performed during regular follow up).So, the burden could be related to an increased duration of each follow up visit.The risks associated with optional blood sampling may include discomfort, moderate pain when the needle is inserted, or bruising of the arm where the blood is drawn. The genetic analysis will be optional, and no individual results will be provided. This sub-study will be for research purposes only.Clinical assessments are commonly used during normal care and do not carry any risk: they require a minimal physical effort, with possible breaks if you are tired. They could represent a minor burden as that could increase by 1 hour the duration each usual follow up visit.

[0264] The algorithm will be applied on the patient clinical data, but no medical device will be used on the body of the patient, so no foreseeable risk and burden will be added on the patient during the use of the algorithm.Duration of the study

[0265] Duration of the study: inclusion period: 2 years participation period (treatment + follow-up): 2 years- total duration: 4 yearsStatistical analysis

[0266] Analysis of effect of the randomization group on the rate of patients with at least one ICDs mild or above over the 2 year-follow up will use a mixed generalized linear model (GLMM) with a logit link adjusted for co-factors (multivariate analysis: primary analysis) and Chi2 test / Fisher's Exact Test (univariate analysis). Relative risk and 95% confidence intervals will be estimated. The primary analysis will be performed in intention to treat (ITT) population.

Claims

CLAIMS1. A device (1) for providing a treatment recommendation based on a prediction of at least one Impulse Control Disorder occurrence for a patient undergoing a treatment with Dopamine Agonists, said prediction being provided by a trained machine learning prediction model configured to receive as input for said patient at least one data ensemble comprising: demographic and anthropomorphic data, mental health data, data relative to said treatment with Dopamine Agonists, each data ensemble being collected during a follow-up session, said follow-up session being part of a treatment process, said machine learning model being previously trained on a training dataset comprising, for each subject of a plurality of subjects undergoing a treatment with Dopamine Agonists, multiple training samples, wherein each training sample is obtained during a follow-up session, said follow-up session being part of a treatment process undergone by said subject, wherein each training sample comprises at least one demographic and anthropomorphic score derived from demographic and anthropomorphic data collected on said subject, at least one mental health score derived from mental health data collected on said subject and at least one treatment score derived from data relative to said treatment with Dopamine Agonists for undergone by said subject, said device further comprising: at least one processor configured to: o calculate at least one demographic and anthropomorphic score from said demographic and anthropomorphic data collected on said patient, at least one mental health score from said mental health data collected on said patient and at least one treatment score from said data relative to a said treatment with Dopamine Agonists undergone by said subject,o feed said demographic and anthropomorphic score, mental health score and treatment score to said trained machine learning model so as to output said prediction, and o compare said prediction with at least one threshold so as to determine said treatment recommendation based on the comparison, and at least one output configured to output said treatment recommendation.

2. The device according to claim 1, wherein: said demographic and anthropomorphic data include an age, a gender, a weight and a duration since start of the treatment process, said mental health data include data representative of an anxiety level, a depression level, an impulsivity level, a sleep quality level and an indicator of presence of a dopamine dysregulation syndrome, said data relative to said treatment with Dopamine Agonists undergone by said patient include at least one treatment name and a treatment dosage.

3. The device according to any of claims 1 or 2, wherein at least part of said mental health data is collected using at least one question comprised in at least one of a MDS-UPDRS I questionnaire, a QUIP SHORT questionnaire and an ASBPD / ECMP questionnaire.

4. 1 to 2, wherein said machine learning prediction model is a XGBoost.

5. The device according to any of claims 1 to 3, wherein said trained machine learning prediction model configured to receive as input for said patient multiple data ensembles collected during a plurality of follow-up sessions, said at least one demographic and anthropomorphic score, said at least one mental health score and said at least one treatment score being obtained using the demographic and anthropomorphic data, the mental health data and the data relative to said treatment with Dopamine Agonists collected during the plurality of follow-up sessions already undergone by said patient.

6. The device according to any of claims 1 to 4, wherein said prediction outputted by said trained machine learning prediction model is compared to two thresholds,including a first threshold and a second threshold, the first threshold being inferior to the second threshold, wherein:- when said prediction is inferior to the first threshold and the second threshold, the patient is classified as a patient with a low risk of developing at least one Impulse Control Disorder, and the treatment recommendation is to continue the treatment with Dopamine Agonists;- when said prediction is superior to the first threshold and inferior to the second threshold, the patient is classified as a patient with a medium risk of developing at least one Impulse Control Disorder, and the treatment recommendation is to reduce the dosage of the treatment with Dopamine Agonists;- when said prediction is superior to the first threshold and superior to the second threshold, the patient is classified as a patient with a high risk of developing at least one Impulse Control Disorder, and the treatment recommendation is to stop the treatment with Dopamine Agonists.

7. The device according to any of claims 1 to 5, wherein said trained machine learning prediction model is obtained using a 5-fold cross-validation scheme.

8. Method for providing a treatment recommendation for a patient undergoing a treatment with Dopamine Agonists, said method comprising: a) predicting the occurrence of at least one Impulse Control Disorder (ICD), b) assigning the patient to a risk group of developing at least one ICD based on the comparison of the prediction determined in step a) with two thresholds, including a first threshold and a second threshold, the first threshold being inferior to the second threshold, wherein: o said patient is assigned to a low risk of developing at least one ICD when the prediction is inferior to the first threshold and the second threshold; o said patient is assigned to a medium risk of developing at least one Impulse Control Disorder when the prediction is superior to a first threshold and inferior to the second threshold; ando said patient is assigned to a high risk of developing at least one Impulse Control Disorder when the prediction is superior to the first threshold and the second threshold; and c) providing a treatment recommendation depending on the identified risk.

9. Method according to claim 7, wherein said method is a computer-implemented method and the step of predicting the occurrence of at least one Impulse Control Disorder (ICD) is performed using a machine learning prediction model previously trained on a training dataset comprising, for each subject of a plurality of subjects, at least one demographic and anthropomorphic score derived from demographic and anthropomorphic data collected on said subject, at least one mental health score derived from mental health data collected on said subject and at least one treatment score derived from data relative to said treatment with Dopamine Agonists, said method further comprising: receiving as input: o demographic and anthropomorphic data, o mental health data, and o data relative to said treatment with Dopamine Agonist taken by said patient, calculating at least one demographic and anthropomorphic score from said demographic and anthropomorphic data collected on said patient, at least one mental health score from said mental health data collected on said patient and at least one treatment score from said data relative to said treatment with Dopamine Agonists,- feeding said demographic and anthropomorphic score, mental health score and treatment score to said trained machine learning model so as to output said prediction.

10. Method according to either one of claims 7 or 8, wherein the patient suffers from a disease or condition associated with dopamine dysregulation, and wherein the treatment recommendation comprises adapting the treatment dosage.

11. Method according to claim 9, wherein the treatment recommendation is:- to continue the treatment with Dopamine Agonists, when the patient is assigned to a low risk of developing at least one Impulse Control Disorder, and / or - to reduce the dosage of the treatment with Dopamine Agonists, when said patient is assigned to a medium risk of developing at least one Impulse Control Disorder, and / or- to stop the treatment with Dopamine Agonists, when said patient is assigned to a high risk of developing at least one Impulse Control Disorder.

12. A computer program product comprising instructions which, when the program is executed by a computer, cause the computer to carry out the method according to any one of claims 9 to 11.

13. A computer-readable medium comprising instructions which, when executed by a computer, cause the computer to carry out the method according to any one of claims 9 to 11.

Images