WO2023148427A1

WO2023148427A1 - Method for training computing arrangement to provide prognosis of progression of tissue condition

Info

Publication number: WO2023148427A1
Application number: PCT/FI2023/050052
Authority: WO
Inventors: Mika MONONEN; Mikael TURUNEN; Rami KORHONEN
Original assignee: Aikoa Technologies Oy
Priority date: 2022-02-03
Filing date: 2023-01-25
Publication date: 2023-08-10

Abstract

Disclosed is a method of training a computing arrangement to provide a prognosis of a progression of a tissue condition. The method comprises deriving physical measurements of dimensions of the tissue and/or surrounding relevant tissue and organ structures from a first set of patients, deriving physiological data from the first set of patients, using the derived physical measurements of dimensions of the tissue and the physiological data as input parameters for a simulation model, wherein the simulation model simulates the progression of the tissue condition as function of time, and wherein the simulation model is executed for dimensions and data of each of the first set of patients and using results of the executed simulations and used respective input parameters to train the computing arrangement to provide the prognosis of progression of the tissue condition for an input physiological data for at least one patient different from the first set of patients.

Description

METHOD FOR. TRAINING COMPUTING ARRANGEMENT TO PROVIDE

PROGNOSIS OF PROGRESSION OF TISSUE CONDITION

TECHNICAL FIELD

The present disclosure relates generally to automated prognostic systems and methods and more specifically, to a method for training a computing arrangement to provide a prognosis of a progression of a tissue condition.

BACKGROUND

Treatment of different diseases is traditionally focused on alleviating symptoms rather than treatment of the disease or condition itself. A typical example is pain-management, wherein there may be no cure for the actual disease or condition. Moreover, since many diseases or medical conditions develop slowly and inevitably, wherein the symptoms emerge only after the disease or condition has already reached a point when preventive actions are no longer applicable and thus, if the disease or condition could be predicted early enough, the preventive and/or precautionary actions may be applied.

Generally, the precautionary actions are high beneficial in terms of increasing a patient's quality of life as well as reducing healthcare related direct and indirect costs. For example, osteoarthritis (OA) is the most common joint disease and the main cause for knee pain, which approximately affects one in every fourth person. It has been estimated that globally there are over 600 million people suffering from knee osteoarthritis. However, since there is no cure for the damaged knee cartilage, the most beneficial treatment would simply be to prevent the disease from developing. Thus, the key for prevention is early detection of the disease, wherein the available treatments are the most beneficial. Moreover, knee osteoarthritis (KOA) is a harmful joint disease as it usually develops to a stage when the only solution to relieve pain is a total knee joint replacement surgery. To avoid surgery, preventive actions should be promoted. However, prevention is possible only if the disease progression can be predicted and if personalized treatment plans can be provided for preventing or slowing down the progression of KOA.

Currently, there exist different methods for prediction of diseases and medical conditions based on biological samples, clinical image analysis, and population-based databases. One such utilized approach is a biomechanical modelling approach, wherein the typical mechanical loading of a biological system is used to simulate the mechanical responses, e.g., stresses, strains, and loads, in the associated tissue. This information together with the knowledge regarding the capabilities of the tissue or organ to withstand the loads, stresses, and strains, etc., can be used to predict the pathological alterations in the tissue, i.e., the disease or medical condition. Typically, biomechanical modelling is done computationally, i.e., biomechanical computational modelling, using the actual or approximate two or three dimensional (2D or 3D) geometry of the tissue and/or organ where each tissue type is given material properties that reflect the mechanical properties of the tissue. However, biomechanical computational modelling usually requires heavy computational power, energy, and is highly time-consuming. For example, considering the prediction of progression of KOA using traditional methods, i.e., manual segmentation of the knee joint tissues from a clinical image, building a biomechanical computational model, and simulation of the mechanical response of the tissues, prediction of the progression of OA, even with computational resources and utilized algorithms, takes around 2 days when done by a domain expert. There have been attempts to speed up the workflow, however there still exists a dire need for a breakthrough accounting each of the aforementioned problems. In outline, there is no general methodology which can be implemented to predict the progression of OA for all patients on an individual level.

In recent years, increasing growth in artificial intelligence technology has led to rapid development of various fields, such as software, medical, sports, telecommunication, networking industry, entertainment, epidemiology, material science, computational neuroscience, economics, and the like. In terms of biomechanical computational modelling, artificial intelligence technology can be utilized to develop faster, more robust, and energy-saving solutions, if there is enough teaching data.

Therefore, in light of the foregoing discussion, there exists a need to overcome the aforementioned drawbacks associated with the conventional systems and provide techniques for a prognosis of a progression of a tissue condition, specifically predicting the progression of osteoarthritis.

SUMMARY OF THE INVENTION

The present disclosure seeks to provide a method of training a computing arrangement to provide a prognosis of a progression of a tissue and/or organ condition, e.g., disease. The present disclosure also seeks to provide a method for providing a prognosis of a progression of tissue and/or organ condition, e.g., disease. An aim of the present disclosure is to provide a solution that overcomes at least partially the problems encountered in prior art.

In one aspect, an embodiment of the present disclosure provides a method of training a computing arrangement to provide a prognosis of a progression of a tissue condition, comprising: deriving physical measurements of dimensions of the tissue and/or surrounding relevant tissue and organ structures from a first set of patients; deriving physiological data from the first set of patients; using the derived physical measurements of dimensions of the tissue and/or surrounding relevant tissue and organ structures and the physiological data as input parameters for a simulation model, wherein the simulation model simulates the progression of the tissue condition as function of time, and wherein the simulation model is executed for dimensions and data of each of the first set of patients; and using results of the executed simulations and used respective input parameters to train the computing arrangement to provide the prognosis of progression of the tissue condition for an input physiological data for at least one patient different from the first set of patients.

In another aspect, an embodiment of the present disclosure provides a method for providing a prognosis of a progression of tissue condition, comprising: deriving physical measurements of dimensions of the tissue and/or surrounding relevant tissue and organ structures; deriving physiological data of a person; providing the physical measurements of dimensions and the physiological data as an input data to a computing arrangement, wherein the computing arrangement has been trained to provide the prognosis based on the input data; receiving the prognosis of the progression of the tissue from the computing arrangement; and presenting the prognosis to a user.

In yet another aspect, an embodiment of the present disclosure provides a computer program comprising computer executable program code which when executed by a processor causes an apparatus to perform the method of any one of the abovementioned claims. Embodiments of the present disclosure substantially eliminate or at least partially address the aforementioned problems in the prior art and enable automation of the provision of the prognosis of the tissue and/or organ condition, e.g., disease and thereby enable the person to remedy the associated risks preventively in an effective and efficient manner.

Additional aspects, advantages, features, and objects of the present disclosure would be made apparent from the drawings and the detailed description of the illustrative embodiments construed in conjunction with the appended claims that follow.

It will be appreciated that features of the present disclosure are susceptible to being combined in various combinations without departing from the scope of the present disclosure as defined by the appended claims.

BRIEF DESCRIPTION OF THE DRAWINGS

The summary above, as well as the following detailed description of illustrative embodiments, is better understood when read in conjunction with the appended drawings. For the purpose of illustrating the present disclosure, exemplary constructions of the disclosure are shown in the drawings. However, the present disclosure is not limited to specific methods and instrumentalities disclosed herein. Moreover, those skilled in the art will understand that the drawings are not to scale. Wherever possible, like elements have been indicated by identical numbers.

Embodiments of the present disclosure will now be described, by way of example only, with reference to the following diagrams wherein:

FIG. 1 is an illustration of a flowchart listing steps involved in a method for training a computing arrangement to provide a prognosis of a progression of tissue condition, in accordance with an embodiment of the present disclosure; FIG. 2 is an illustration of a flowchart listing steps involved in a method for providing a prognosis of a progression of tissue condition, in accordance with an embodiment of the present disclosure;

FIG. 3 is an exemplary graphical representation of tissue condition prediction curves depicting values of a quantitative condition parameter, in accordance with an embodiment of the present disclosure;

FIG. 4 is an exemplary interface depicting different views of the tissue and corresponding physiological data of the person, in accordance with various embodiments of the present disclosure;

FIG. 5 is a block diagram of a system for providing a prognosis of a progression of tissue condition, in accordance with an embodiment of the present disclosure.

In the accompanying drawings, an underlined number is employed to represent an item over which the underlined number is positioned or an item to which the underlined number is adjacent. A non-underlined number relates to an item identified by a line linking the non-underlined number to the item. When a number is non-underlined and accompanied by an associated arrow, the non-underlined number is used to identify a general item at which the arrow is pointing.

DETAILED DESCRIPTION OF EMBODIMENTS

The following detailed description illustrates embodiments of the present disclosure and ways in which they can be implemented. Although some modes of carrying out the present disclosure have been disclosed, those skilled in the art would recognize that other embodiments for carrying out or practising the present disclosure are also possible. In one aspect, the method of the present disclosure provides a method of training a computing arrangement to provide a prognosis of a progression of a tissue condition, comprising: deriving physical measurements of dimensions of the tissue and/or surrounding relevant tissue and organ structures from a first set of patients; deriving physiological data from the first set of patients; using the derived physical measurements of dimensions of the tissue and/or surrounding relevant tissue and organ structures and the physiological data as input parameters for a simulation model, wherein the simulation model simulates the progression of the tissue condition as function of time, and wherein the simulation model is executed for dimensions and data of each of the first set of patients; and using results of the executed simulations and used respective input parameters to train the computing arrangement to provide the prognosis of progression of the tissue condition for an input of physiological data for at least one patient different from the first set of patients.

Beneficially, utilization of biomechanical computational modelling-based training enables the generation of practically unlimited amount of training data for various subjects with different characteristics, e.g., age, gender, weight, height, and dimensions of the tissue and/or surrounding relevant tissue and organ structures.

In another aspect, the method of the present disclosure provides a method for providing a prognosis of a progression of tissue condition, e.g., disease, comprising: deriving physical measurements of dimensions of the tissue and/or surrounding relevant tissue and organ structures; deriving physiological data of a person; providing the physical measurements of dimensions and the physiological data as an input data to a computing arrangement, wherein the computing arrangement has been trained to provide the prognosis based on the input data; receiving the prognosis of the progression of the tissue from the computing arrangement; and presenting the prognosis to a user.

Beneficially, utilization of a trained computing arrangement in providing the prognosis of a progression of tissue condition, e.g., disease, decreases the time required (i.e., from days or hours to less than a second), energy, and memory (from gigabytes to megabytes) compared to biomechanical computational modelling-based approach. This is achieved by training a computing arrangement based on simulations done based on measurements of the anatomical dimensions of the tissue and/or surrounding relevant tissue and organ structures and physiological data from a first set of patients. Example for measurements of joint dimensions can be used. The measurements can be done for example from native x-ray images. Further, a set of physiological data is derived from the first set of patients. The measurements and physiological data of the first set of patients may be actual patient data or arbitrary generated data with variation of realistic values, enabling generation of practically unlimited training data. These measurements and data are used in a simulation model to predict progression of tissue (such as a joint) as a function of time. The simulations can be performed for example using finite element modelling (FEM), which is one type of biomechanical computational modelling. Technical problem of using such modelling is heavy computational complexity of the modelling. It might take several hours or days to complete one or set of simulations. Also, the simulation models error margin may increase over the time i.e., if one simulates duration of 1-2 years the error margins are small compared to for example 5-10 years of simulations. The set of simulations done for the first set of patients are used to train computing arrangement. Training refers to training of artificial intelligence (Al) or machine learning (ML) system. The trained computing arrangement is then validated using actual patient-derived data from the first set of patients with known outcome, which were not used in the training of the computing arrangement. This way we will obtain a trained computing arrangement which can provide progression prognosis of a tissue condition for a real patient which is different from the first set of patients. This solves the technical problem of FEM (related to computational issues and errors) as the trained model can be used to derive "end result" of the simulation in fast manner. End result refers to answer to question such as "what is the condition of the tissue in 5 years time based on provided input of measurement from a native x-ray image or other measurement" or "how will the condition of the knee develop over time for the next 10 years with the patients' current weight or if the patient would gain or lose for example 10 kg of weight".

The present disclosure provides a method of training a computing arrangement to provide a prognosis of a progression of a tissue condition, e.g., disease. Typically, the method is configured to process medical data related to a patient to analyse and thereby determine the prognosis of the progression of the tissue condition associated with the patient. Optionally, the method may be configured to determine a prognosis based on a current state of the tissue condition. The term "tissue condition" as used herein relates to a state of at least one tissue, organ, bone, or joint in any part of the body of the patient. Currently, existing workflows methods for prediction of the progression of tissue conditions utilizing biomechanical computational models, e.g., osteoarthritis, are computational expensive and time consuming (such as, around 2 days by a trained expert). Thus, to overcome the aforementioned problem, the method of the present disclosure is trained to enable prediction of the progression of tissue conditions such as, progression of osteoarthritis, in a faster and more efficient manner (for example, a few seconds/minutes vs several hours) via utilization of a customized computational model employing information derived from clinical imaging data (such as, but not limited to, Magnetic Resonance Imaging (MRI), Computed tomography (CT), native X-ray and the like) and other patient physiological data. Additionally, the prognosis of the progression of the tissue condition provided via the computing arrangement trained by the method further comprises preventive proposals, wherein the preventive proposals are provided to the patient to potentially decrease the rate of progression of the tissue condition. In an example, the preventive proposals may be in the form of "reduce weight" for a patient to reduce risk of further progression of osteoarthritis, reduce injuries, etc.

Optionally, the tissue is a joint. Herein, the joint refers to a human body part that connects at least two bones and allows movement therebetween in a specified manner. In an example, the tissue condition relates to the tissue condition of a joint of the patient, wherein the joint may be at least one of ball and socket joint, hinge joint, condyloid joint, pivot joint, gliding joint, saddle joint. Moreover, optionally, the joint is a knee joint. In an example, the tissue condition relates to a type of hinge joint i.e., the knee joint (or tibiofemoral joint). The knee joint connects the shinbone (tibia) and thighbone (femur) allowing mainly flexion-extension rotational movement between the bones. Moreover, smaller movements may be present in other degrees of freedom i.e., internal-external rotation, abduction-adduction rotation, anterior-posterior translation, medial- lateral translation, and distal-proximal translation. It will be appreciated that the method of the present disclosure may be implemented on any body part, tissue, or organ of a patient, wherein the patient may be a human, or an animal without limiting the scope of the present disclosure.

The term "computing arrangement" as used herein, refers to a structure and/or module that includes programmable and/or non-programmable components configured to store, process and/or share information and/or signals relating to the provision of prognosis of the progression of the tissue condition. The computing arrangement may be a controller having elements, such as a display, control buttons or joysticks, processors, memory and the like. Typically, the computing arrangement is operable to perform one or more operations for providing the prognosis of the progression of the tissue condition. In the present examples, the computing arrangement may include components such as a memory, a processor, a data communication interface, a network adapter and the like, to store, process and/or share information with other computing devices, such as the user devices, database arrangements, medical servers, hospital networks and so forth. Optionally, the computing arrangement includes any arrangement of physical or virtual computational entities capable of enhancing information to perform various computational tasks. Further, it will be appreciated that the computing arrangement may be implemented as a hardware server and/or plurality of hardware servers operating in a parallel or in a distributed architecture. Optionally, the computing arrangement is supplemented with additional computation system, such as neural networks, and hierarchical clusters of pseudo-analog variable state machines implementing artificial intelligence algorithms.

Optionally, the computing arrangement is implemented as a computer program that provides various services (such as database service) to other devices, modules or apparatuses. Moreover, the computing arrangement refers to a computational element that is operable to respond to and processes instructions to provide the prognosis of the progression of the tissue condition. Optionally, the computing arrangement includes, but is not limited to, a microprocessor, a microcontroller, a complex instruction set computing (CISC) microprocessor, a reduced instruction set (RISC) microprocessor, a very long instruction word (VLIW) microprocessor, Field Programmable Gate Array (FPGA) or any other type of processing circuit, for example as aforementioned. Additionally, the computing arrangement is arranged in various architectures for responding to and processing the instructions for the provision of prognosis of the progression of the tissue condition via the method or system. Herein, the computing arrangement elements may communicate with each other using a communication interface. The communication interface includes a medium (e.g., a communication channel) through which the system components communicates with each other. Examples of the communication interface include, but are not limited to, a communication channel in a computer cluster, a Local Area Communication channel (LAN), a cellular communication channel, a wireless sensor communication channel (WSN), a cloud communication channel, a Metropolitan Area Communication channel (MAN), and/or the Internet. Optionally, the communication interface comprises one or more of a wired connection, a wireless network, cellular networks such as 2G, 3G, 4G, 5G mobile networks, and a Zigbee connection.

In one or more embodiments, the training is done using gaussian regression (GPR) model utilizing exponential GPR algorithm. The gaussian process regression (GPR.) model is a nonparametric kernel-based probabilistic model with a set of random variables having a multivariate distribution. Typically, the method is trained utilizing GPR models, wherein the GPR model utilizes the exponential GPR algorithm. Beneficially, the training of the method using GPR models is time-efficient and computationally inexpensive compared to other regression models. Additionally, the training of the method using exponential GPR in comparison to other types of GPR models is more accurate. Moreover, optionally, a prior function (on the function space) may be defined to determine a posterior probability distribution using the training data, and thereby compute the predictive distribution on a region of interest based on the implementation of the method. Beneficially, the predicted distribution thus incorporates information from both the prior distribution and the dataset, wherein predictions at unseen points of interest, may be calculated via weighting of all possible predictions by their calculated posterior distribution.

The method further comprises deriving physical measurements of dimensions of the tissue and surrounding relevant tissue and organ structures from a first set of patients. Typically, the physical measurements are derived via the method utilizing the clinical imaging data of the tissue or body part being prognosed, wherein the clinical imaging data comprises, but is not limited to, imaging data via magnetic resonance imaging (MR.I), computed tomography (CT), native X-ray and the like. Herein, the physical measurements of the tissue and the surrounding relevant tissue and organ structures are derived from the clinical imaging data, wherein the physical measurements relate to the region of the interest (of the body part or organ of each patient from the first set of patients). Generally, the relevant physical dimensions (or the anatomical dimensions) of each of the tissues and/or organs, deformities, or cavities in and around the area of interest (i.e., the tissues, organs, or body parts in consideration) may be measured. The physical measurements comprise at least one of, but are not limited to, organ size, organ shape, bone length, bone thickness, bone curvature, bone shape, tissue thickness, tissue curvature, tissue shape, cavity, joint space, or other space between tissues and/or organs and the like that are required to determine the prognosis of progression of the tissue condition. The method may be configured to, predict, diagnose or evaluate at least on of, but are not limited to, pain, weakness, swelling, bleeding, degradation, rupture or other breakdown, or failure in the tissues or organs. For example, for the knee-joint these might be, damaged or degraded cartilage, meniscus, ligaments or tendons of the patient. In an exemplary scenario, the method of the present disclosure is implemented on a knee of a patient, for determining the prognosis of progression of the tissue condition associated with the knee. Herein, at least a mean right knee medial and lateral joint space width (JSW) and a mean left knee medial and lateral JSW may be measured. Generally, the measurements for the mean knee medial and lateral joint space width lie in the region of 3 mm to 8 mm. In an example, the measurements for the mean right medial and lateral joint space width lie in the 4.74 mm ± 0.75 and 5.63 mm ± 0.86, respectively and measurements for the mean left medial and lateral joint space width lie in the range of 4.74 mm ± 0.76 and 5.66 mm ± 0.87, respectively.

In one or more embodiments, the physical measurements of dimensions are derived from data acquired with an imaging device. The imaging device refers to an instrument or device configured for capturing different images of the tissue in consideration for further analysis. The imaging device is selected from at least one of a camera, an MRI scanner, an X- ray scanner, a CT scanner and the like. Herein, the imaging device is utilized to determine the physical measurements of the tissue that are derived from the images captured via the imaging device. In an example, the physical measurements of dimensions of the tissue are derived using images captured via an MRI scanner. In an exemplary scenario, the measured physical measurements such as, the measurement of the distance between the distal femur and the proximal tibia (i.e., the JSW), is an indirect way of measuring cartilage thickness and a reproducible tool for the assessment of progressive knee cartilage degenerative conditions. In other examples, the method may employ other imaging techniques to determine the physical measurements including, but not limited to, fluoroscopy, ultrasound, echocardiography. Beneficially, the method enables utilization of a wide variety of devices to provide a more comprehensive and accurate prognosis of progression of the tissue condition. Additionally, optionally, the method may employ molecular imaging techniques such as bone scintigraphy (or gamma scan), PET scan to additionally provide a broad spectrum of diagnosis in a more detailed and comprehensive manner. In one or more embodiments, the imaging device is an x-ray device. The X-ray device refers to a medical instrument configured to capture X-ray images of body parts associated with the tissue in consideration for measuring the physical measurements. Herein, the X-ray device is used for obtaining projection images i.e., 2-dimensional images for determining the physical measurements of the tissue and the surrounding body parts in consideration. Beneficially, X-ray measurements require lesser time, are inexpensive and have higher availability in comparison to other imaging devices and thus improves the cost effectiveness and utilization of the method.

In one or more embodiments, the x-ray device measurements are used to form simulated dimensions of a simulated tomographic image. Typically, the native x-ray device provides the projection images from imaged object, i.e., for the third orthogonal direction along which the projection is taken (which cannot be measured directly). Thus, based on the trained computing arrangement system, the simulated third orthogonal direction anatomical dimensions of a simulated tomographic image of a tissue acquired with a tomographic device (i.e., MRI or CT) can be estimated from the x-ray projection image. Beneficially, since the x-ray device is sufficient to formulate the dimensions of the simulated tomographic image of a tissue with acquired with a tomographic device, the need for tomographic imaging is beneficially removed to reduce the associated costs and power consumption.

In one or more embodiments, the imaging device is a tomographic imaging device. Typically, a clinical tomographic imaging device, e.g., MRI or CT, is configured to provide a 3-dimensional image of the imaged body part or tissue to enable the method to accurately determine the prognosis of progression of the tissue condition. Beneficially, the tomographic imaging via, e.g., the MRI or CT device provides 3- dimensional comprehensive information related to the tissue condition and enables to perform analysis for determining the prognosis thereafter.

In one or more embodiments, an accelerometer sensor is used to collect the physiological data of the person. Typically, the derived physiological data is collected via an accelerometer sensor configured for collecting the acceleration data of the human body part in consideration or being monitored for input to the prediction model. Beneficially, the accelerometer sensor provides the acceleration data in addition to the measured physical dimensions and thus enables to derive the physiological data and thereby accurately determine the prognosis of progression of the tissue condition.

In one or more embodiments, the derived physical measurements of the knee joint dimensions comprise one or more of Med_AP, Lat_AP, Med_JS, Lat_JS, Cond_dist, and wherein the dimensions are derived from a tomographic image taken from the first set of patients. Herein, the 'Med_AP' refers to a maximum anterior-posterior dimension in medial femoral condyle (in sagittal view), the 'Lat_AP' refers to a maximum anterior-posterior dimension in lateral femoral condyle (in sagittal view), the 'Med_JS' refers to a tibiofemoral joint space in medial compartment of the knee (in sagittal view), the 'Lat_JS' refers to a tibiofemoral joint space in lateral compartment of a knee (in sagittal view) and the 'Cond_dist' refers to a distance between medial and lateral condyle centre-points (in coronal view). In addition, Med_JS and Lat_JS dimensions can consist the separated tibia and femur articular cartilage thicknesses when visible, e.g., in MRI or contrast agent enhanced CT imaging data. It has been found out that knee joint is particularly suitable for the discussed methods of this disclosure.

In one or more embodiments of the invention the derived physical measurements of dimensions comprises one or more of maximum anterior-posterior dimension in medial femoral condyle, maximum anterior-posterior dimension in lateral femoral condyle, Tibiofemoral joint space in medial compartment of the knee, Tibiofemoral joint space in lateral compartment of a knee, distance between medial and lateral condyle centre-points, and wherein the dimensions are derived from a tomographic image taken from the first set of patients.

In one or more embodiments of the invention the derived physical measurements dimensions comprises one or more of: maximum medial- lateral dimension of distal femur, maximum anterior-posterior dimension in medial femoral condyle, maximum anterior-posterior dimension in lateral femoral condyle, tibiofemoral joint space in medial compartment of the knee, tibiofemoral joint space in lateral compartment of a knee, distance between medial and lateral condyle centre-points, Tibiofemoral alignment, varus or valgus angle between femur and tibia, medial condyle width of femur in medial-lateral direction, lateral condyle width of femur in medial-lateral direction, and wherein the dimensions are derived from x-ray image taken from the first set of patients.

The method comprises utilizing the clinical imaging data comprising the tomographic image taken from the first set of patients, wherein the tomographic image refers to the scanning done via a computed tomography (CT),MRI, or other clinical tomographic scanner for providing images of tissues and skeletal structure in the area of interest. Beneficially, the physical measurements are derived in a faster and timeefficient manner using the tomographic images. Typically, the knee geometry is quantified using a defined set of physical measurements that may be considered as unique for each knee. Moreover, to reduce the time for generating the biomechanical computational model, e.g., through manual segmentation of the tissues, the subject specific biomechanical computational model is generated based on the derived physical measurements via the method, an existing geometry of the knee is scaled to match with the physical measurements, e.g., using an atlas-based method. This reduces the time for generation of the data for the training of the computing arrangement.

In one or more embodiments, the derived physical measurements for the knee joint dimensions comprises one or more of: Med_Lat, Med_AP, Lat_AP, Med_JS, Lat_JS, Cond_dist, FT_angle, VV_angle, Med_CW, Lat_CW, and wherein the dimensions are derived from a native X-ray image taken from the first set of patients. Typically, the X-ray images taken from the first set of patients comprises knee alignment information (FT_angle, VV_angle) that may be further utilized to estimate personalized joint loads through tibiofemoral joint during a physical activity. Herein, the 'Med_Lat' refers to a maximum medial-lateral dimension of distal femur, the 'FT_angle' refers to a Tibiofemoral alignment, the 'VV_angle' refers to a varus or valgus angle between femur and tibia, the 'Med_CW' refers to medial condyle width of femur in medial-lateral direction and the 'Lat_CW' refers to a lateral condyle width of femur in medial-lateral direction. Beneficially, each of the parameters (i.e., the derived physical measurements) are utilized in an optimized manner to beneficially provide the prognosis via a computationally inexpensive, faster and power-efficient method. In addition, Med_JS and Lat_JS dimensions can consist the separated tibia and femur articular cartilage thicknesses when visible, e.g., contrast agent enhanced X-ray imaging data. The method further comprises deriving physiological data from the first set of patients. Herein, the derived physiological data refers to a subject specific data i.e., a patient specific data required for determining the prognosis of progression of the tissue condition associated with the patient. Typically, the physiological data comprises medical data (or input parameters) based on which the computing arrangement is trained and thereby the prognosis may be determined. For example, the physiological data comprises, but is not limited to, age, height, weight, gender, inflammation biomarkers, or pain indicators or measures including measures reported by mobile applications or, in case of osteoarthritis, clinical questionnaires such as Knee injury and Osteoarthritis Outcome Score (KOOS) or Western Ontario and McMaster Universities Arthritis Index (WOMAC) including three primary measures including pain (such as, during walking, using stairs, in bed, sitting or lying, and standing upright), stiffness (such as, after waking up or any time later in the day) and physical function (such as, using stairs, rising from sitting, standing, bending, walking, getting in or out of a car, shopping, putting on or taking off socks, rising from bed, lying in bed, getting in or out of bath, sitting, getting on or off toilet, doing heavy domestic duties or light domestic duties and the like). Moreover, the physiological data may further comprise data monitored via monitoring devices, systems such as, medical sensors, medical equipment, imaging devices and the like configured to perform one or more medical and/or monitoring operations. The clinical imaging data may be derived via performing the one or more operations including, but not limited to, monitoring user activity, employing motion analysis, employing imaging analysis, and other operations that may be required to be performed to comprehensively analyse and thereby provide an accurate prognosis of the progression of the tissue condition via the method. In an example, the one or more operations include, monitoring daily activity via wearable or non-wearable devices such as smart watches, monitoring sensors, medical devices and the like. In another example, the one or more operations include performing motion analysis via a smart phone, camera or the like; or other wearable devices, such as motion sensors, shoe sensors or the like.

In an exemplary implementation of the method, medical data such as, physiological data from several existing imaging datasets containing walking trials from human subjects or patients (N=4000 trials from 196 subjects) were analysed with musculoskeletal modelling software (such as, OpenSim® 4.1) to obtain compartmental joint loading (medial and lateral compartments of the knee) during the loading response (1st peak) and terminal knee extension (2nd peak) phases of walking. Additionally, the knee static abduction-adduction angles (estimate of frontal joint alignment) of each subject and the walking speed for each trial were obtained from motion capture data. The computing arrangement is trained via a musculoskeletal model based on the Rajagopal full-body model, wherein the musculoskeletal model is modified to reduce analysis time such that the arms of the subject were removed, and their mass added to the torso. Further, the mass, height, age and gender of each patient or subject available in any of the datasets is combined into a predictor dataset along with knee abduction-adduction angles and walking speeds, whereas joint contact forces (JCFs) is implemented into a response dataset. As a result, 9 different response variables i.e., medial, lateral, total JCF peaks at loading response, terminal knee extension, peak values over the whole stance phase and a high number of predictor combinations formulated from the knee abduction-adduction angle, mass, height, age, gender, and walking speed of the subjects are obtained.

Furthermore, the method comprises generating shallow neural networks trained separately for each response variable having different predictor combinations to determine the optimum regression model amongst each of the trained neural networks for predicting knee joint loading for each response variable. Herein, the physiological data is split into at least one of a training, a validation, and a testing subset randomly. Moreover, upon collection and analysis of multiple patients such as, from the first set of patients, wherein data from an individual patient from the first set of patients is present only in one of the subsets; the correlation coefficients (R) and the root mean square errors (RMSEs) of the resulting models are determined for each of the training, validation, and testing subset subsets. In one or more embodiments, the derived physiological data comprises at least one of: weight, height, age, type of exercise, acceleration data of a body, e.g., feet in respect to the ground, during at least one of walking or running. Herein, the type of exercise may be selected from at least one of walking, running, flexion, extension, bending, sitting, standing, stretching or any other physical or mental exercise. Further, the acceleration data of the body, e.g. feet in respect to the ground, may be determined via any conventional monitoring or imaging device to be analysed along with the other derived physiological data. Typically, the derived physiological data is used for training the method and thereby be utilized to determine an accurate prognosis via the method. Such a comprehensively derived physiological data complimented with the determined physical measurements enables to determine the prognosis in a more effective and efficient manner and the same time removes the need for several X-ray and/or tomographic images utilized otherwise, and thus saving time and energy.

In one or more embodiments, the acceleration data is used to estimate average impact to the joint of interest. Typically, the acceleration data comprises data received from a monitoring sensor or device such as, an accelerometer, a motion sensor, an imaging device and the like, configured to measure the acceleration data. Further, based on the measured acceleration data, the method is configured to be trained to generate a machine learning model to estimate forces within the joint, e.g., the tibiofemoral forces through the knee medial and lateral compartments. Beneficially, the tibiofemoral forces measured such as, via the accelerometer (e.g., located in a shoe of a patient) enables to analyse the results (shown as a graphical representation in FIG. 3) and compared to the prognosis for improving the accuracy of the method. Moreover, beneficially, the acceleration data may be collected in shortintervals of time and due to long lasting data collection (e.g., after collecting sample one to five times a day), the present method may re- utilize the derived physiological data in future prognosis to save memory and processing power.

The method further comprises using the derived physical measurements of dimensions and the physiological data as input parameter for a simulation model, wherein the simulation model simulates the progression of the tissue condition as function of time, and wherein the simulation model is executed for dimensions and data of each of the first set of patient. Typically, the method comprises prediction of progression of tissue condition (or disease progression) utilizing medical data i.e., the derived physical measurements and the derived physiological data as input for simulation of the disease progression while simultaneously providing an easy-to-use and clear user-interface. Herein, the results of the executed simulations and the respective medical data or clinical image data is utilized to train the computing arrangement to provide the prognosis of progression of the tissue condition. Moreover, the method also relates to automated image-analysis tools for measurement of anatomical dimensions required by the method for atlas-based modelling using artificial intelligence-based models to combine an atlas-based modelling workflow and a degeneration algorithm. Herein, via combining the degeneration algorithm (or a progression classification algorithm) with an image analysis algorithm (i.e., for measurement of anatomical dimensions in the body part associated with the tissue) enables to efficiently and effectively predict progression of the tissue condition. For example, the method enables provision of prognosis of progression of KOA (or knee osteoarthritis) in a comprehensive, user friendly manner and the same time enables presentation of medical proposals to avoid potential further injuries (for example, knee osteoarthritis injuries such as, if weight of a patient increases, medical proposals or suggestions are provided to perform physical activities to reduce weight). In one or more embodiments, the simulation model is a finite element model. Typically, the finite element model (FEM) represents a real-world object in a digitized form that may be made by drawing manually such as, via sketching/drawing or digitally via scanning or via processing (such as, segmentation) from a captured image. The FEM comprises of a plurality of elements, i.e., smaller areas or volumes having one or more features, e.g., mechanical properties and boundary conditions, defining the material (for example, of the tissue in consideration) of the simulation model. Beneficially, the FEM simulation enables to determine an accurate and comprehensive prognosis of progression of tissue condition over time based on one or more parameters (or influences).

The method further comprises using results of the executed simulations and used respective input parameters to train the computing arrangement to provide the prognosis of progression of the tissue condition for an input physiological data for at least one patient different from the first set of patients. Herein, the computing arrangement is trained using the input medical data having derived inputs and outputs i.e., results of the executed simulations and used respective input parameters (i.e., the derived physical measurements of dimensions and the physiological data) for enabling the trained computing arrangement to provide an accurate output (or result) with a set of realistic input medical data (or clinical imaging data). Additionally, the trained computing arrangement is tested and/or validated against a set of data with verified inputs and outputs, not utilized during the training to resolve any potential bias issues, e.g., using cross-validation. Further, upon training the computing arrangement, the method comprises receiving the input physiological data for at least one patient different from the first set of patients for which the prognosis of progression of tissue condition is determined. Typically, the method provides the prognosis of progression of the tissue condition based on current information (i.e., the input medical data) for providing a prediction (i.e., the output) of the tissue condition. In an example, the method may prognose the tissue condition (or organ condition) as normal (or healthy condition) or as a pathological state(s) of the tissue or organ in consideration, wherein the progression of the tissue condition may be predicted (output) based on the physiological data at the current state (input).

In the exemplary implementation scenario, the method further comprises generating shallow neural networks trained separately for each response variable having different predictor combinations to determine the optimum regression model amongst each of the trained neural networks for predicting knee joint loading for each response variable. Herein, the physiological data is split into at least one of a training, a validation, and a testing subset randomly, e.g., using cross-validation. Moreover, upon collection and analysis of multiple patients such as, from the first set of patients, wherein data from an individual patient from the first set of patients is present only in one of the subsets; the correlation coefficients (R) and the root mean square errors (RMSEs) of the resulting models are determined for each of the training, validation, and testing subsets. As a result, each of the R values and RMSE values between musculoskeletal modelling based JCFs and neural network predicted JCFs are utilized, wherein mass, height, gender, age, knee abduction-adduction angle, and walking speed are used as predictors. Herein, generally, the R value ranges from 0.51 to 0.80. Notably, the results are similar for most predictor combinations that contained mass, walking speed, and knee abduction-adduction angle.

In another aspect, the present disclosure provides a method for providing a prognosis of a progression of tissue condition. The method comprising deriving physical measurements of dimensions of the tissue and surrounding relevant tissue and organ structures, deriving physiological data of a person, providing the physical measurements of dimensions and the physiological data as an input data to a computing arrangement, wherein the computing arrangement has been trained to provide the prognosis based on the input data.

The method further comprises receiving the prognosis of the progression of the tissue from the computing arrangement and presenting the prognosis to a user. Typically, the prognosis of progression is based on current information i.e., the physical measurements of dimensions and the physiological data used to provide a prediction of the tissue or organ condition at a later point in time. In an example, the tissue and organ condition may be prognosed as a healthy state. In another example, the tissue or organ condition may be prognosed as a pathological state or states of the tissue or organ in consideration. Moreover, the method comprises presenting the prognosis of progression of tissue condition in an illustrative manner that enables to effectively understand the risk (if any) and possible remedial measures for reducing the associate risk. Technical effect of using the trained computing arrangement to provide the prognosis is that it requires less computing power than using simulations to provide prognosis. Indeed, the computing arrangement is trained using physical measurements, physiological data and a simulation model related to the first set of patients (or users / subjects) as discussed before. This method has been found out to reduce complexity of calculations and provide fast result

In one or more embodiments, the training has been carried out using any of the aforementioned embodiments of the method of the present disclosure.

In one or more embodiments, the provided prognosis is further used to provide a signal to configure a device. Typically, upon providing or presenting the prognosis to the user, the method further comprises providing the signal (such as, a control signal) to configure the device. The device may be used to further prognose the tissue condition or as a remedial measure to alleviate the potential risk of disease such as, osteoarthritis by re-configuring the device associated with the user to administer sufficient exercise to the user as proposed herein during prognosis. Conventionally, the user is required to configure the device on their volition and thus may be insufficient and/or inaccurate for alleviating and/or preventing progression of the tissue condition. Thus, the method enables automated configuration of the device based on the requirement of the user i.e., based on the prognosis to alleviate and/or prevent progression of the tissue condition more accurately and efficiently. In one or more embodiments, the device is at least one of: bike cycle, treadmill, rowing machine, a fridge, electric bike, cross trainer, shopping cart, brace, shoe insole, wearable, implants, prosthesis, mobile phone, tablet, eyeglasses, bracelet, or other smart technology.

In an exemplary scenario, the method provides the patient with proposals based on risk factors such as, but not limited to, gender and activity level contributing to the progression of the tissue condition. Generally, the proposals can be one of, but not limited to, reducing weight (e.g., less eating or dieting), getting more exercise, overall dietary change (e.g., reduction of cholesterol levels), and psychology counselling. For example, for osteoarthritis, the proposal can include weight-loss, using different shoes while walking or running, or using softer surface for training such as, on a treadmill. Moreover, the proposals may be used with a bicycle (for example, use of a smaller gear for lesser strain to make cycling easier), rowing machine (similar strain adjustment), a fridge with a display (for example, indicating what to eat and not to eat, propose to eat vegetables only, or propose to eat intermittently), an electric bike (providing more electric power provided to make it easier), a cross trainer (similar strain adjustment), a shopping cart with some indicator like a display (for example, indicating a route to healthy shelves only) and so forth. In another example, the method may be employed in sport activities, such as javelin throwing, wherein the height, muscle power, running speed of the person is used to train the computing arrangement. Further, the angle, speed of javelin may also be employed by the method to determine or predict the distance travelled by the javelin.

In one or more embodiments, an updated prognosis is provided after providing the signal. Upon configuring the device, the updated prognosis may be provided at a later point in time to check on the progression of the tissue condition. Beneficially, the updated prognosis enables the user to keep track of progression of tissue condition and employ appropriate remedial measures (such as, exercise, physiotherapy, rehabilitation, surgery etc.) to alleviate any associated risks. Additionally, the updated prognosis significantly reduces the number of medical check-ups for the user and consequently reduces the associated time and resources consumed.

Furthermore, the simulations for the generation of the training data from the first set of patients are based on finite element model (FEM) simulations. There is an atlas of FE-models from which the best one is selected based on a pre-defined criteria, for example the closest match to the to-be-simulated geometry, wherein the atlas of FE-models includes the same information from input parameters that defines a geometry or a shape when the selected FE-model is scaled to match the geometry from the input. Optionally, the geometry is scaled using scaling factors defined from the difference between the input data and respective FE- model data, e.g., dimensions or angles or shapes. Beneficially, creation of the FE-model is very robust and takes only seconds, compared to manual segmentation which requires up to a working day, and saving processing capacity, memory and power. Optionally, FE-models may include also other simulation parameters that are scaled based on the input data. For example, in case of a knee joint, the force pattern through the joint during walking or running is scaled using the input data from subject characteristics, e.g., weight, age, height, gender, joint alignment. Moreover, the input parameters may vary and may form from an infinite number of combinations, e.g., a random set of multiple dimensions, weight, and age. The FE-model gives an output, e.g., force, stress, strain, pressure, etc., for each set of input combinations and the computing arrangement is trained using the same set of input combinations wherein the target is the simulated FE-model output. Said FE-model simulations are used then for training of a computing arrangement. The trained computing arrangement is thus faster (in comparison of for running just FE simulations) and consumes less power when there is no need to use any real measured data, e.g., from clinical images, or any real-life data, e.g., patients, subjects, analysis of the movement, gait analysis, or any force analysis. Thus, the training data can be completely arbitrary, and only the validation data, i.e., when the actual outcome, e.g. tissue generation, is known, requires real-life follow-up patient data. When coupled with an algorithm (formulation, equation) relating the simulated outputs to a changing variable, e.g., age or weight, the progression of disease can be predicted faster.

DETAILED DESCRIPTION OF THE DRAWINGS

Referring to FIG. 1, illustrated is a flowchart listing steps involved in a method 100 of training a computing arrangement to provide a prognosis of a progression of a tissue condition, in accordance with an embodiment of the present disclosure. At a step 102, the method 100 comprises deriving physical measurements of dimensions of the tissue and/or surrounding relevant tissue and organ structures from a first set of patients. At a step 104, the method 100 further comprises deriving physiological data from the first set of patients. Upon deriving the physical measurements of dimensions of the tissue and surrounding bone structures, the method 100 further comprising deriving the physiological data from the first set of patients. At a step 106, the method 100 further comprises using the derived physical measurements of dimensions and the physiological data as input parameter for a simulation model, wherein the simulation model simulates the progression of the tissue condition as function of time, and wherein the simulation model is executed for dimensions and data of each of the first set of patient. Herein, the derived physical measurements of dimensions and the physiological data are collectively utilized as input parameter for the simulation model to simulate the progression of the tissue condition as function of time for each of the first set of patients. And, at a step 108, the method 100 further comprises using results of the executed simulations and used respective input parameters to train the computing arrangement to provide the prognosis of progression of the tissue condition for an input physiological data for at least one patient different from the first set of patients. Herein, the executed simulations and used respective input parameters are utilized to train the computing arrangement to provide the progression of tissue condition based on the input physiological data for the at least one different patient.

It may be appreciated that the steps 102 to 108 are only illustrative, and other alternatives can also be provided where one or more steps are added, one or more steps are removed, or one or more steps are provided in a different sequence without departing from the scope of the present disclosure.

Referring to FIG. 2, illustrated is a flowchart listing steps involved in a method 200 for providing a prognosis of a progression of tissue condition, in accordance with an embodiment of the present disclosure. At a step 202, the method 200 comprises deriving physical measurements of dimensions of the tissue and/or surrounding relevant tissue and organ structures. For example, the method 200 comprises deriving physical measurements of dimensions of the tissue and surrounding bone structures from a first set of patients. At a step 204, the method 200 further comprises deriving physiological data of a person. Upon deriving the physical measurements of dimensions of the tissue and/or surrounding relevant tissue and organ structures, the method 100 further comprises deriving the physiological data of the person whose tissue condition is being analysed via the method 200. At a step 206, the method 200 further comprises providing the physical measurements of dimensions and the physiological data as an input data to a computing arrangement, wherein the computing arrangement has been trained via the method 100 to provide the prognosis based on the input data. Herein, the computing arrangement trained based on the method 100 is utilized to determine and thereby provide the prognosis based on the input data i.e., the physical measurements of dimensions and the physiological data. At a step 208, the method 200 further comprises receiving the prognosis of the progression of the tissue from the computing arrangement. And, at a step 210, the method 200 further comprises presenting the prognosis to a user. Upon receiving the prognosis of the progression of the tissue from the computing arrangement, the prognosis is presented to the user. For example, the user may be a doctor, health personnel or the person.

It may be appreciated that the steps 202 to 210 are only illustrative, and other alternatives can also be provided where one or more steps are added, one or more steps are removed, or one or more steps are provided in a different sequence without departing from the scope of the present disclosure.

Referring to FIG. 3, illustrated is an exemplary graphical representation 300 of tissue condition prediction curves depicting values of quantitative condition parameter, in accordance with an embodiment of the present disclosure. Herein, the x-axis 300A depicts a time range of 10 years and the y-axis 300B depicts a risk index comprising a low, medium, or high- risk regions therein. As shown, a first curve 302 represents a prediction curve based on the derived physiological data of a person, wherein the first curve 302 reaches a medium risk index after 4 years of time. Herein, the derived physiological data comprises at least one of: weight, height, age, gender, type of exercise, acceleration data of a feet in respect to the ground during at least one of walking or running, based on which prediction curves are generated. In an example, the derived physiological data is measured via an accelerometer (for e.g., in a shoe) and wherein the generated prediction curves may be compared to determine the prognosis of progression of the tissue condition of the person. Notably, the accelerometer is configured to determine the acceleration data that is used to estimate average impact to the joint of interest. Further, as shown, a second curve 304 represents a modified (or improved) prediction curve based on a first status change, e.g., weight change, over a period of time as suggested, wherein the second curve 304 reaches medium index at approximately after 6 years as compared to 4 years based on the current status. For example, the first status change may be weight change of 10 kilograms. Further, as shown, a third curve 306 represents another modified (or improved) prediction curve based on a second status change, e.g., weight change, over a period of time as suggested, wherein the third index 304 stays in the low-risk index as compared to 4 years and 6 years based on the current status or the first status change, respectively. Herein, for example, the second status change may be weight change of 15 kilograms. Beneficially, such a representation enables the person to formulate a plan to possibly alter the status, e.g., lose weight (such as, an exercise plan or diet plan), and alleviate the risk of diseases such as, osteoarthritis.

Referring to FIG. 4, illustrated is an exemplary depiction of an interface 400 depicting different views of the tissue and corresponding physiological data of a person, in accordance with an embodiment of the present disclosure. As shown, the three different views 402A, 402B and 402C are clinical images of a knee of the person. Generally, the clinical images (or the clinical imaging data) are tomographic images captured via an imaging device configured to capture clinical imaging data of the body part or tissue in consideration or an X-ray image taken by a native X-ray device. The imaging device may be a camera, an X-ray machine, an MRI scanner, a CT scanner, and the like, configured to provide the clinical imaging data for further analysis. Typically, the different views 402A, 402B and 402C enable the creation of the biomechanical computational model such as, the FE simulation model, to simulate the forces in the tissues to estimate the progression of the tissue condition. Further, the interface 400 provides the physiological data of the person, as depicted in a first column 404. Furthermore, the exemplary depiction of an interface 400 provides a contrast tab 406 to optimize the contrast of the shown clinical image views an information tab 408 for efficient categorization and faster recall of the clinical imaging data to collectively improve the efficiency of the system. As a result, the simulation model and the clinical imaging data are collectively utilized to determine the physical measurements of the dimensions of the tissue and the surrounding body parts.

Thus, based on the presented prognosis, the user may determine the possible course of action for the person and beneficially alleviate any risks associated with progression of the tissue condition. Moreover, optionally, based on the provided prognosis, the method 100 and/or 200 is further configured to provide a signal to configure a device, wherein the device is at least one of: bike cycle, treadmill, rowing machine, a fridge, electric bike, cross trainer, shopping cart, brace, shoe insole, wearable, implants, prosthesis, mobile phone, tablet, eyeglasses, bracelet, or other smart technology.

Referring to FIG. 5, illustrated is a block diagram of a computing arrangement configured to provide a prognosis of a progression of a tissue condition, in accordance with an embodiment of the present disclosure. As shown, the computing arrangement 500 comprises components including a memory 502 including computer program code, a processor 504, a data communication interface 506, to store, process and/or share information with other computing devices, such as the imaging devices, the database arrangements, medical servers, hospital networks and so forth. The memory 502 and the computer program code are configured to, with the processor 504, cause the apparatus to perform the method 100 or 200. Modifications to embodiments of the present disclosure described in the foregoing are possible without departing from the scope of the present disclosure as defined by the accompanying claims. Expressions such as "including", "comprising", "incorporating", "have", "is" used to describe and claim the present disclosure are intended to be construed in a non- exclusive manner, namely allowing for items, components or elements not explicitly described also to be present. Reference to the singular is also to be construed to relate to the plural.

Claims

1. A method of training a computing arrangement to provide a prognosis of a progression of a tissue condition, comprising: deriving physical measurements of dimensions of the tissue and/or surrounding relevant tissue and organ structures from a first set of patients; deriving physiological data from the first set of patients; using the derived physical measurements of dimensions of the tissue and the physiological data as input parameters for a simulation model, wherein the simulation model simulates the progression of the tissue condition as function of time, and wherein the simulation model is executed for dimensions and data of each of the first set of patients; and using results of the executed simulations and used respective input parameters to train the computing arrangement to provide the prognosis of progression of the tissue condition for an input physiological data for at least one patient different from the first set of patients.

2. A method according to claim 1, wherein the simulation model is a finite element model.

3. A method according to claim 1 or 2, wherein the tissue is a joint.

4. A method according to claim 3, wherein the joint is knee joint.

5. A method according to any of preceding claims, wherein the derived physical measurements of dimensions comprises one or more of maximum anterior-posterior dimension in medial femoral condyle, maximum anterior-posterior dimension in lateral femoral condyle,

Tibiofemoral joint space in medial compartment of the knee, Tibiofemoral joint space in lateral compartment of a knee, distance between medial and lateral condyle centre-points, and wherein the dimensions are derived from a tomographic image taken from the first set of patients.

6. A method according to any of claims 1-4, wherein the derived physical measurements dimensions comprises one or more of: maximum medial-lateral dimension of distal femur, maximum anterior-posterior dimension in medial femoral condyle, maximum anterior-posterior dimension in lateral femoral condyle, tibiofemoral joint space in medial compartment of the knee, tibiofemoral joint space in lateral compartment of a knee, distance between medial and lateral condyle centre-points, Tibiofemoral alignment, varus or valgus angle between femur and tibia, medial condyle width of femur in medial-lateral direction, lateral condyle width of femur in medial-lateral direction, and wherein the dimensions are derived from x-ray image taken from the first set of patients.

7. A method according to any of the preceding claims, wherein the derived physiological data comprises at least one of: weight, height, age, type of exercise, acceleration data of a feet in respect to the ground during at least one of walking or running.

8. A method according to claim 7, wherein the acceleration data is used to estimate average impact to the joint of interest.

9. A method according to any of the preceding claims, wherein the training is done using gaussian regression (GPR) model utilizing exponential GPR. algorithm.

10. A method for providing a prognosis of a progression of tissue condition, comprising: deriving physical measurements of dimensions of the tissue and/or surrounding relevant tissue and organ structures; deriving physiological data of a person; providing the physical measurements of dimensions and the physiological data as an input data to a computing arrangement, wherein the computing arrangement has been trained to provide the prognosis based on the input data; receiving the prognosis of the progression of the tissue from the computing arrangement; and presenting the prognosis to a user.

11. A method according to claim 10, wherein the physical measurements of dimensions are derived from data acquired with an imaging device.

12. A method according to claim 11, wherein the imaging device is an x-ray device.

13. A method according to claim 12, wherein the x-ray device measurements are used to form a simulated dimensions of a simulated tomographic image.

14. A method according to any of the claims 10-13, wherein the imaging device is an MR.I device.

15. A method according to any of the claims 10-14, wherein an accelerometer sensor is used to collect the physiological data of the person.

16. A method according to any of the claims 10-15, wherein the training has been carried out using method of claims 1-9.

17. A method according to any of the claims 10-16, wherein the provided prognosis is further used to provide a signal to configure a device.

18. A method according to claim 17, wherein the device is at least one of: bike cycle, treadmill, rowing machine, a fridge, electric bike, cross trainer, shopping cart, brace, shoe insole, wearables, implants, prosthesis, mobile phone, smart technology such as clock, tablet, eyeglasses, bracelets.

19. A method according to claim 17, wherein an updated prognosis is provided after providing the signal.