Predictive Modeling
Principal investigators
Senior researchers & Experts
- Mikael Turunen
- Antti Heikkilä
Ph.D. students (primary supervisor)
Ph.D. students (co-supervisor)
- Sana Jahangir
- Tulashi Simkheada
Active collaborators
- Professor José Jaime García Alvarez – Universidad del Valle
- Professor Ilse Jonkers – KU Leuven
- Professor Jussi Tohka – University of Eastern Finland
- Hospital physicist Mimmi Liukkonen – Kuopio University Hospital
Overview
enabled by new technologies offered by the research results.
Problem
Osteoarthritis (OA) is a degenerative joint disease that causes pain and evokes functional disabilities, with knee osteoarthritis (KOA) having an estimated 14% lifetime risk. It affects approximately 100 million citizens in the US and EU and this number is estimated to increase by 25% during the next 20 years due to aging of the population. Moreover, it has also been reported that 43% of those suffering from OA are of working age in Western Europe. The current clinical method to diagnose or predict joint health status is through clinical imaging, which usually is only prescribed once the patient’s symptoms are unbearable, and the disease is already advanced. Therefore, OA is usually identified too late, which inevitably leads to surgical joint replacements being the only recommended clinical solution. This causes substantial and unnecessary costs to societies and healthcare systems as, for example, it can be estimated that a significant number of knee surgeries due to OA could be avoided by preventive actions. A total knee replacement (TKR) can cost up to 50 000€, and it has been reported that over 600 000 TKR procedures are performed each year in the US among patients aged 50–59. Currently, OA represents a $303.5 billion annual burden to US economy alone.
Solution and vision
Mika’s research group aims to revolutionize the current care guidelines for monitoring and management of OA and related conditions with the implementation of a proactive healthcare solution system. The long-term goal is to create a platform and an ecosystem in which users, physiotherapists, clinicians, and other stakeholders interact and adopt preventive healthcare technologies in an encouraging environment. The proposed platform would facilitate the radical transformation of the healthcare system to preventive healthcare and contribute to the EU’s leadership in emerging health technologies. Currently, we are developing novel methods and technologies for a clinical tool that will be capable to predict and monitor personalized progression of knee osteoarthritis. By utilizing machine learning, the outcomes from computational finite element (FE) models with subject-specific information (age, height, weight, gender, information of joint injuries) will be merged into a same software. This software with a user-friendly interface will enable to visualize quantitatively personalized progression of osteoarthritis as a function of aging and helps doctors in clinical decision making when optimizing adjustable measures (weight, gait, shoe type, knee brace) for preventative treatment.
Current projects
Development and validation of template-based modeling to predict patient-specific progression of knee osteoarthritis, and possibilities in economic and societal benefits
The aim of this project is to study the possibility to utilize computational modeling and clinical imaging data with subject information to make personalized prediction for the onset and progression of osteoarthritis years before. In this research project, novel and automatic methods are developed and validated to obtain subject-specific geometries from clinical images (image analysis/shape modeling) and subject-specific loading based on clinical images and subject information without motion analysis. Furthermore, previously published degeneration algorithm is developed further to predict the onset and progression of knee osteoarthritis by considering all age-related changes in cartilage. Secondary aim in the project is to study economical and societal benefits due to prevention in a real-world health care setting utilizing data from available real-world datasets with several years follow-up. Outcomes of these novel validated methods, with quantitative information about economic and societal possibilities, offer novel data for clinical applications in future.
Computational tool for predicting and monitoring personalized progression of knee osteoarthritis
The main aim of this high risk/high gain project is to develop clinical tool for predicting and monitoring personalized progression of knee osteoarthritis after joint traumas. Detailed aims can be divided in applied and experimental studies. While the applied studies are aiming to combine clinical imaging data (geometry and information of joint injuries), biochemical biomarkers and subject characteristics (age, height, weight, physical activity, muscle strength/physical activity and joint motion/loading) with computational modeling, the experimental studies are aiming to evaluate age and muscle strength/physical activity level related changes to site-specific tissue level material properties including failure mechanisms. Finally, the outcomes will be merged for a clinical tool utilizing AI, which will be based on the supervised learning. Developed novel clinical tool helps doctors in clinical decision making after knee joint traumas. The tool enables to visualize personalized progression of OA as a function of aging and to evaluate quantitatively effects between different treatments options (surgical or rehabilitation) before they are applied to the patients.
