AI-based diagnosis of Lyme disease
Lyme borreliosis or Lyme disease is a vector-borne infection that spreads to humans by the bite of certain types of ticks. Current serological tests offer a sensitivity below 40 % in early stages and are highly dependent on the immune response against the bacteria Borrelia burgdorferi that cause the infection. These tests focus on the antibodies produced to fight off the disease. However, in many cases the host’s immune reaction is delayed, weak or even absent.
An improved diagnostic assay for Lyme disease
To address these shortcomings, the EU-funded DualDur project developed an in vitro diagnostic (IVD) method that can directly identify pathogens. “DualDur® reagent and method was developed in 1986 in Hungary and was routinely used on thousands of samples as a confirmatory test. With the help of the Horizon 2020 grant we have been able to transform it into a standard testing procedure that is compatible with the most modern laboratory systems,” explains Andras Pal Bozsik, DualDur coordinator and founder of Lyme Diagnostics Ltd. The method entails the use of the special cell technology medium DualDur®, followed by a preparation and concentration method of bacteria from blood. The medium maintains the stability of samples for over a week at room temperature. The samples are subsequently examined using a uniquely designed automated microscope combined with an AI module that identifies the pathogens while minimising human error. Sample concentration ensures sensitivity even when bacteria numbers in the blood are low. Thus, the DualDur® method can detect even the lowest concentration of B. burgdorferi that exists in humans during an infection. Importantly, it does so at all stages of the disease, offering a more certain diagnosis in the early stage of the infection, as well as in the more advanced phase of the disease with chronic symptoms. Moreover, it can be used as a tool for monitoring treatment efficacy.
Clinical efficacy and impact
In vitro tests on hundreds of artificially inoculated samples confirmed the DualDur® method’s level of detection. Using this method, an average operator can differentiate B. burgdorferi from similar spirochaetes with great certainty. After these encouraging results, DualDur® entered a multi centre prospective clinical study of 400 patients, where it was tested against standard methods. “The automated AI-based DualDur® system outperformed all other tested laboratory methods, reaching a sensitivity that is by 60 % higher than the first-choice serological method,” emphasises Bozsik. DualDur partners have completed the CE registration process according to the EU regulation on IVDs. The automated system is ready to be installed at selected laboratories, while the development of new features and stability solutions is ongoing. The system will soon be launched on the market and further investment will be necessary to carry out the envisaged European expansion plan. Lyme disease is the most common zoonotic disease in Europe with 650-850 000 estimated cases every year. Late-stage disseminated disease poses a heavy burden on patients in terms of quality of life, direct and indirect costs. It also incurs a significant economic burden on the European healthcare system. Implementation of the DualDur® system is expected to improve early diagnosis and treatment of patients suffering from Lyme disease.
Keywords
DualDur, Lyme disease, bacteria, IVD, Borrelia burgdorferi, in vitro diagnostic
