Dr Vassilia Orfanou, PhD, Post Doc
Writes for the Headline Diplomat Magazine, LUDCI.eu
Introduction
According to the latest research data, Luxembourg is 69% ready for the artificial intelligence (AI) revolution. That’s 9% behind the United Kingdom – the most AI-ready economy in Europe. Already, the landlocked country in the middle of Western Europe has a burgeoning innovation hub in HealthTech.
Luxembourg is actively integrating AI into its healthcare landscape, supported by a vibrant ecosystem comprising startups, SMEs, research laboratories, and government initiatives. As the country positions itself as a major player in the field of HealthTech, the integration of AI technologies in healthcare settings is driving advancements in patient care while presenting unique challenges.
In this article, we explore how Luxembourg is leveraging AI for diagnostic support, personalized treatment recommendations, and predictive analytics, while also addressing the associated challenges and opportunities.
How AI-integrated is Luxembourg’s Health Care?
In medical diagnosis
Luxembourg’s HealthTech sector is taking steps to integrate AI, with a focus on “AI-supported diagnosis and treatment procedures.” Companies, like Bionext LAB are utilizing AI algorithms for non-invasive tests and diagnosis of conditions like endometriosis, enabling faster and more accurate analysis of medical tests.
Artificial intelligence is not new in Luxembourg’s healthcare. Back in 2019, Hanalytics, a health diagnostic company, launched an AI-based radiology image analysis model, called BioMind to address the “shortage of image specialists and radiologists” in Luxembourg.
Certified Conformité Européenne (CE), BioMind is a diagnostic system that uses deep learning for analyzing neurological disorders like vascular diseases, brain tumors and stroke.
According to LuxInnovation, BioMind is the “first CE-certified AI application for brain diagnosis in the world.” It is integrated with a cutting-edge deep learning tech that mimics the way real doctors diagnose their patients’ medical images.
Following years of research, some AI software products are now used in patients after promising test phases. The area most impacted today is medical imaging. Deep learning – a branch of AI, works well with image-like data such as X-rays or MRI scans.
Similarly, Icometrix, a health tech company in Luxembourg, specializes in AI-powered imaging solutions for neurology, providing personalized treatment recommendations based on advanced brain imaging techniques.
According to Amazon, “more than 500 clinical practices employ icometrix’s artificial intelligence (AI) imaging solutions to aid in detecting and treating MS and Alzheimer’s.”
In cardiology
In the domains of cardiology and autoimmune and autoinflammatory diseases, a field significantly impacted by the recent pandemic, AI algorithms are being developed to analyze heart scans, predict potential issues, and personalize treatment plans. The pandemic caused a surge in heart-related complications; traditional diagnostic methods were overwhelmed.
One of the interventions of AI in the Luxembourg healthcare as regarding cardiology issues is in the innovation of an artificial intelligence and machine learning-based diagnostic project called “COVIRNA.” The cardiological project is funded by the European Commission and led by the Luxembourg Institute of Health’s Cardiovascular Research Unit (CVRU), the project was developed to predict how severe COVID-19 was while identifying patients that have high risk of post-COVID complications. The project helped in tracking individuals with cardiovascular effects such as the “long-COVID” syndrome.
“Although we were not experts in COVID at that time, we leveraged our prior expertise in RNAs and cardiology to design novel methods which could predict the severity of COVID-19 disease”, explains COVIRNA project coordinator Dr Yvan Devaux. “We found that when we measure RNA molecules in the blood, we can predict whether the patient will develop severe problems or will eventually succumb to the disease in the long run.”
Autoimmune diseases
Research is also ongoing in other areas impacted by the pandemic, such as autoimmune diseases like Type-1 Diabetes. A project by Dulce Canha, a PhD student at the Luxembourg Institute of Health (LIH), is exploring how AI can be used to understand the psychological impact of Type-1 Diabetes.
Canha’s research, titled “Development of a neuro-symbolic AI approach to characterize diabetes distress profiles in people with type-1 diabetes,” focuses on the challenges faced by Type-1 Diabetes patients in managing their condition, which can lead to psychological distress. This distress can then have a negative impact on their overall health outcomes.
Canha’s project utilizes “deep digital phenotyping” and “deep immuno-phenotyping” to analyze data collected from patients. They then leverage “neuro-symbolic AI techniques” to identify patterns in this data that can help to understand the factors contributing to psychological stress in these patients.
The goal of this research is to develop “clinically relevant clusters” that can be used by both doctors and patients. This will allow for the creation of “personalized health strategies” to improve the quality of life for people with Type-1 Diabetes.
Canha emphasizes the importance of “explainability and trust” in AI for medical applications. Her research focuses on developing AI models that are “interpretable to both clinicians and patients.”
Another specific example of popular tech-based diagnosis using AI is diabetic retinopathy. This is a common complication of diabetes that can lead to partial or complete vision loss if not treated early enough. AI imaging software approved by official authorities can automatically diagnose diabetic retinopathy based on a funduscopy (an ophthalmological examination).
To consolidate these developments and ensure AI use is as ethical as possible, the Luxembourg Institute of Health (LIH) announced in March 2024, the Dataspace4Health to ensure secure and compliant health data based on a governed framework.
The benefits are numerous. Seeing that AI integration is exploding here, especially when it comes to working with large amounts of patient data or biological samples, it is now possible to evaluate huge amounts of data that could not be processed.
According to the LIH, the Dataspace4Health project will ensure enhanced patient care, foster health innovation and research and help build a secure ecosystem for collaborative institutional data sharing.
Integration of AI in Luxembourg’s Healthcare Strategy
A core focus of Luxembourg’s tech-focused health care strategy is personalized medicine, particularly in light of its growing elderly population. This necessitates healthcare providers and institutions to leverage the potential of AI for preventative measures. Taking advantage of AI’s analytical prowess means healthcare professionals can map diseases and identify individual risk factors. This allows for proactive interventions and improved patient outcomes.
The Ministry of Health envisions a use case for its upcoming e-health strategy – leveraging AI-powered search engines to empower patients and offer personalized medicine service:
“You have been feeling sick for days and it is only getting worse. You visit your doctor and, although he or she is not immediately sure how to help, a quick global search of your symptoms reveals a similar case in Southeast Asia. More detailed tests reaffirm the suspicion and within a few hours, you begin your personalized treatment, which successfully provides relief. Every year, hundreds of thousands of medical research papers, case studies and articles are released in dozens of languages all over the world, bringing a wealth of knowledge. AI-powered search engines make that knowledge accessible to you and your doctor, improving diagnoses and treatment.”
As part of its national economic diversification strategy, Luxembourg also prioritizes the digitization of healthcare to address contemporary health challenges, while capitalizing on economic opportunities. The country’s emphasis on advancing the digitization of health and the patient journey underscores its commitment to innovation and resilience.
Luxembourg’s HealthTech companies, along with public research laboratories like LIH and LSCB, are collaborating to develop data spaces for shared strategies, framework, as well as cutting-edge solutions.
Challenges and Opportunities
With the digitization of many medical practices, the amount of health data generated in healthcare is growing exponentially.
However, one of the challenges this innovation presents is in the loss of jobs that are easily automated but were previously done manually by humans. Before the pandemic, 41% of CEOs surveyed in Luxembourg had expressed that AI will displace more jobs in the long run. And this became even truer when the pandemic hit and there had to be more use of technology to work remotely during the lockdowns.
According to the Luxembourg Times, bankruptcies recorded in post-pandemic Luxembourg reached a six-year high in 2023. There has also been a continuous shift in the skills demanded for in the healthcare industry, like in many others. There is a spike in the demand for AI or deep learning-related skills on the job market.
One of the primary challenges is bridging the skills gap in AI-related disciplines and a potential polarization of the labor market, as noted earlier and by Xavier Bettel, the Deputy Prime Minister of Luxembourg.
“Integrating AI into existing workflows, for example, may raise new questions related to working conditions, employment law and labor relations,” noted Xavier, in an interview with Lena Martensson, the Senior Marketing & Communication Officer at Luxinnovation GIE.
According to him, “The increased automation of previously manual tasks brings challenges, but at the same time, there is a great potential to create new jobs. These will probably require a completely different skill set, however, which could potentially lead to a polarization of the labor market.”
The application of artificial intelligence represents an essential lever for modernizing healthcare, providing added value for both healthcare professionals and patients. In fact, artificial intelligence systems can be used in healthcare as a tool for diagnosis, prevention, monitoring, prediction, prognosis for treatment or alleviation of a disease.
However, big data and the use of artificial intelligence in healthcare face several challenges. According to Martha Wyatt, an Associate Consultant at IDR Medical, the three common barriers cited as reasons for slowing or hindering development include: lack of trust, regulatory compliance and privacy concerns.
Therefore, in order for the use of artificial intelligence systems to be beneficial, they must be sufficiently understandable, transparent, and reliable to gain the trust of patients and healthcare professionals. The goal is to better manage healthcare resources, while protecting privacy.
With proactive measures and strategic investments in education, training, and regulatory frameworks, these challenges can be addressed, paving the way for a more sustainable and efficient healthcare system.
Future Development Strategies
Luxembourg is poised to consolidate and expand its leadership in the HealthTech sector through strategic initiatives, such as public-private partnerships, academic collaborations, and robust support for research and development. For instance, establishing the Campus HE:AL (Health And Lifescience Innovation) planned for 2024 will focus on data-driven personalized medicine shows Luxembourg’s commitment to pioneering advancements in digital health.
“It will attract companies active in medical devices, in vitro diagnostics, and digital health tools and services,” according to Luxembourg economy minister, Franz Fayot, as quoted by Silicone Luxembourg.
AI applications are currently being developed in many areas of medicine in Luxembourg and across Europe. Advances can be expected in oncology. among others. This will in turn analyze images, e.g. mammography data for the detection and analysis of tumors. In addition, AI will also be used to predict possible cancers using biomarkers. Biomarkers are measurable parameters of biological processes used as indicators of pathology in blood samples.
Advantages and/or Disadvantages of AI in Medical Diagnostics
In the field of medical imaging, the advantage of AI over an expert is often the speed, but it is also the accuracy and reproducibility of the analyses. AI will systematically produce the same results, but this is not infallible either.
Biases in AI algorithms can always occur and are often due to incorrect or incomplete training. To put it simply, if an algorithm is trained only on the data of men around the age of 50, its application to young women will not work properly and may produce incorrect results. Therefore, it is crucial that the data used to train the algorithms is of high quality.
Will AI replace doctors?
AI will not replace doctors. Medical staff will continue to be essential for communicating with patients during consultation hours. What one can imagine is that doctors who use AI will replace those who do not. The application of AI will help them improve their diagnoses and save valuable time.
“We do have to mould ourselves to technology and accept what is new for us. Doctors using AI may replace doctors who are reluctant to use AI or adopt it. That certainly could happen in future,” said Dr Harsh Mahajan, a radiologist and one of the pioneers of medical imaging in India. “AI and technology are going to be transformative. The only way we can provide quality health care to the masses of the country will be through technology.’
The integration of AI into the consultation, be it for administrative tasks or for faster diagnosis, should help doctors to spend more time on the human relationship with the patient. So, it’s a very positive impact and AI could help bring the human aspect back into medical advice.
Conclusion
Luxembourg is a key player in HealthTech in Europe, driven by its AI-readiness, dynamic ecosystem, and forward-thinking initiatives. The integration of AI technologies in healthcare reflects the country’s vision for shaping the future of medicine through innovation and collaboration. As Luxembourg continues to invest in HealthTech and AI, it is poised to lead the way in revolutionizing patient care and driving sustainable growth in the digital health sector.
Call to Action
Stakeholders in Luxembourg’s healthcare ecosystem, including businesses, research institutions, and policymakers, are encouraged to collaborate closely to harness the full potential of AI in healthcare. By fostering partnerships, supporting innovation, and addressing regulatory challenges, Luxembourg can solidify its position as a leader in HealthTech and pave the way for a healthier and more technologically advanced future.
References
Canha, D. (2024). Development of a neuro-symbolic AI approach to characterize diabetes distress profiles in people with type-1 diabetes, Tailor Network. Available at https://tailor-network.eu/development-of-a-neuro-symbolic-ai-approach-to-characterize-diabetes-distress-profiles-in-people-with-type-1-diabetes/. Retrieved on April 13, 2024.
European Health Management Association (2021). AI and Machine Learning at the Base of the COVIRNA Diagnostic Kit. Available at https://ehma.org/ai-and-machine-learning-at-the-base-of-the-covirna-diagnostic-kit/. Retrieved on April 13, 2024.
Luxembourg Times (2023). More bankruptcies in Luxembourg, job losses at six-year high. Available at https://www.luxtimes.lu/luxembourg/more-bankruptcies-in-luxembourg-job-losses-at-six-year-high/4419284.html. Retrieved on April 13, 2024.
Thormundsson, B. (2023). Western European governments’ artificial intelligence (AI) readiness index rankings in 2023, Statista. Available at https://www.statista.com/statistics/1231714/western-europe-government-artificial-intelligence-readiness-index/. Retrieved on April 10, 2024.
LIH (2023). A cardiovascular diagnostic test for the surveillance and care of long COVID-19 patients. Available at https://www.lih.lu/en/article/a-cardiovascular-diagnostic-test-for-the-surveillance-and-care-of-long-covid-19-patients/. Retrieved on April 13, 2024.
Lovell, T. (2019). Luxembourg hospital group to pilot AI radiological image analysis solution, HealthcareIT News. Available at https://www.healthcareitnews.com/news/emea/luxembourg-hospital-group-pilot-ai-radiological-image-analysis-solution. Retrieved on April 13, 2024.
Martensson, L. (2019). Interview with Xavier Bettel in Happen: “Artificial intelligence for the benefit of all,” The Luxembourg Government. Available at https://gouvernement.lu/en/actualites/toutes_actualites/interviews/2019/12-decembre/20-bettel-happen.html. Retrieved on April 10, 2024.
Lux Innovation (2019). Hanalytics launches AI-based image analysis in Luxembourg. Available at https://www.luxinnovation.lu/news/hanalytics-launches-ai-based-image-analysis-in-luxembourg/. Retrieved on April 10, 2024.
Lim, J.I. et al. (2023). Artificial Intelligence Detection of Diabetic Retinopathy, PubMed Central. Available at https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9636573/. Retrieved on April 10, 2024.
Luxembourg Institute of Health (2024). Luxembourg launches Dataspace 4 Health. Available at https://www.lih.lu/en/article/luxembourg-launches-dataspace-4-health-a-pioneering-dataspace-and-governance-framework-for-secure-and-compliant-health-data-exchange/. Retrieved on April 10, 2024.
Digital Luxembourg (N.D.). Artificial Intelligence : a strategic vision for Luxembourg. Available at https://gouvernement.lu/dam-assets/fr/publications/rapport-etude-analyse/minist-digitalisation/Artificial-Intelligence-a-strategic-vision-for-Luxembourg.pdf. Retrieved on April 10, 2024.
Wyatt, M. (2021). Barriers to AI in Healthcare, Greenbook. Available at https://www.greenbook.org/insights/barriers-to-ai-in-healthcare. Retrieved on April 10, 2024.
The Pool (2022). HE:AL: A Campus Dedicated To Health Technologies, Silicone Luxembourg. Available at https://www.siliconluxembourg.lu/heal-a-campus-dedicated-to-health-technologies/. Retrieved on April 10, 2024.
Plus One (2022). Artificial intelligence supporting cancer patients across Europe—The ASCAPE project, Pub Med Central. Available at https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9022843/. Retrieved on April 10, 2024.
Smeets, D. (2023). Icometrix helps to detect and treat neurological diseases with AI imaging, powered by AWS, Amazon Web Services. Available at https://aws.amazon.com/blogs/publicsector/icometrix-helps-to-detect-and-treat-neurological-diseases-with-ai-imaging-powered-by-aws/. Retrieved on April 13, 2024
Zhang, B. (2023). Machine Learning and AI in Cancer Prognosis, Prediction, and Treatment Selection: A Critical Approach, Pub Med Central. Available at https://www.ncbi.nlm.nih.gov/pmc/articles/PMC10312208/. Retrieved on April 10, 2024.
ANI (2024). Doctors using AI may replace those who are reluctant to use it, Economic Times. Available at https://health.economictimes.indiatimes.com/news/industry/doctors-using-ai-may-replace-those-who-are-reluctant-to-use-it-health-expert/108179400. Retrieved on April 10, 2024.
