Seeing More Than the Doctor
The volume of the hippocampus is the key marker. If it has degenerated to a greater extent than is normal, it may indicate the start of Alzheimer’s. ARDX supports the diagnosis of the status quo. However, it is not yet possible to predict dementia. “ARDX is a low-threshold test offered to people with memory impairment who fear they are suffering from early Alzheimer’s dementia,” explains Dr. Lothar Spies, managing director of jung diagnostics. The Hamburg company has specialized in medical image analysis services. It is the first company worldwide to have developed an image-based Alzheimer’s risk diagnostic method and has marketed this as an approved medical product.
Every specialist, neurologist, psychiatrist and general practitioner can offer ARDX. Patients are first subjected to a memory test and examined radiologically with a conventional MRI system. At jung diagnostics, the MRI image data is then checked to ensure that the quality is good enough and analyzed with the aid of the computer. This means that a patient’s data set is deformed in such a way that it fits optimally onto a standard head, a so-called atlas. However, the individual characteristics of the brain are also retained.
The computer program then analyzes the images of the brain regions pixel by pixel and recognizes them as gray or white substance or cerebral fluid. With the aid of mathematical and statistical methods, the distribution of the gray substance is then compared to a collective of healthy test persons. Finally, the result is documented in a report that is automatically generated by the program. However, no report leaves the lab without being checked first. “We subject every report to a comprehensive final check and review the results for consistency,” emphasizes Dr. Spies. “In doing so, we use an extensive list of plausibility criteria.”
Exceptional analytical performance thanks to mathematical algorithms
The algorithm, which forms the basis for the exceptional analytical performance of ARDX and which jung diagnostics has refined for use in routine clinical work, originates from the Fraunhofer Institute for Medical Image Computing MEVIS in Lübeck. This is where Prof. Dr. Herbert Thiele and his team work in the image registration project group on new possibilities of computer-assisted analysis of medical image data. “Our research is practice-oriented,” says the project coordinator, emphasizing the significance of the close collaboration between science and industry. The Lübeck project group is well known for its expertise in image registration. Together with the Bremen parent institute, it is, for example, researching into and developing interactive assistance systems for daily use in clinics – including algorithms that can help doctors evaluate computer medical image data and identify risk factors.
The cooperation with jung diagnostics started with a personal meeting at an event. “At that time, we very quickly decided to pool our know-how in a joint project,” recalls Prof. Thiele, who sees huge potential in collaborating with small and mid-sized firms. “If firms have small or no development departments, we can jointly extend their innovative lead. The types of collaboration are very individual and range from providing advice to product component development.”
The Fraunhofer researchers cover a huge range. The current projects range from algorithms aimed at improving tumor diagnosis in digital pathology to the optimized positioning of brain stimulation electrodes for Parkinson’s patients. To bring together the right partners in these complex subjects, the professor says that the good network of clinics, business and research that Life Science Nord Managemen t GmbH has helped to create is especially valuable. “The support the agency gives us is very effective,” says the professor. “And this is a key factor in scientific and commercial success.”
Early Alzheimer diagnosis with the hippocampus Volumetry method
ARDX shows what can be created when the right partners come together. Even though the method of hippocampus volumetry is still too complex and costly to be available as a standard procedure in general healthcare, the market is growing. “At present, we are pioneers in this field,” tells Dr. Spies. “However, the technology is also increasingly being used in specialized clinics and experts to diagnose dementia. This is arousing the interest of other firms that now also see opportunities.” A further positive signal is that the assessment of the hippocampus as a biomarker of Alzheimer’s disease has been incorporated into the European guidelines on the diagnosis of Alzheimer’s. While patients still have to pay for ARDX themselves, the managing director of jung diagnostics is convinced that the method’s economic benefits will be accepted in the future. “If Alzheimer’s can be diagnosed earlier with our methods, this will help not only patients, but will also lower the costs for the health insurance carriers in the long term.” Until then, the research team headed by Prof. Thiele and jung diagnostics are already working on the next joint project: “We are developing a method to measure the spinal cord in the region of the cervical spine,” says Dr. Spies. In the case of multiple sclerosis patients, pathological changes that could lead to paralysis occur there. Once again, the computer can extract more from the image data than a human, and accordingly these changes can be detected at an early stage and therapy can be adapted accordingly. Does that represent a further step in computer-assisted image analysis – namely that that the doctor will be completely replaced eventually? “I can’t see that happening,” says Prof. Thiele, and Dr. Spies also shakes his head. “Our technology aims to assist health professionals in time-consuming routine tasks and give them more time for demanding diagnoses. The human factor will still play a role even in medical image analysis.”
Source / Copyright: Life Sciences Nord / www.lifesciencenord.de