Unifying Laboratory and Point-of-Care Test Methods towards an eHealth Platform

Franziska Dinter¹; Peter Schierack¹; Gregory Dame²; Michał Burdukiewicz³; Stefan Rödiger¹*
¹ Brandenburg Technical University Cottbus – Senftenberg
² Brandenburg Medical School Theodor Fontane
³ Warsaw University of Technology
* stefan.roediger@b-tu.de

Challenges in modern medicine and patient care in rural regions

Cardiovascular diseases are one of the major causes of adult mortality in the western world. Lifestyle, genetic predisposition, false diagnoses and other factors all contribute to this and impose a heavy burden on the health system. Demographic and structural changes pose additional challenges for patients in rural areas with low population density 1Conventional diagnostic methods are primarily based on anamnesis and qualitative and quantitative measurements of of biomarkers. All techniques frequently used to detect biochemical or genetic biomarkers as immunoassays, bioimage informatics, flow cytometry, PCR-based methods and Next Generation Sequencing require expensive and immobile analystical devices 2.
Addressing the challenge of the growing number of cardiovascular diseases requires a two-step solution involving the detection of prognostic or diagnostic biomarkers on site in combination with automatic data exchange. Since the 2000s, the volume of digital information stored in the electronic health record (EHR) or available in the literature has also increased rapidly. The EHR is intended for use of patient information for administrative tasks in health care (e.g., billing), clinical informatics and bioinformatics. Such large volumes of data have to be analyzed with more sophisticated techniques based on machine learning. The self-learning algorithms can be employed to evaluate the patient-specific information and possibly return suggestions related to possible diagnosis or therapy. Prerequisites, however, are availability of homogenous data and model interpretability 3The former provides all the information necessary for the learning process. The latter is necessary to allow physicians to understand reasons behind the decision. The combination of elements of these techniques has a high potential for the development of patient-oriented and patient-personalized diagnostic systems 4.

Combining laboratory and point-of-care test methods

In our work we investigate how biomarkers, concepts from laboratory devices, bioimage informatics, medical OMICS data and bioinformatics can be combined into a system for patient-oriented data acquisition. Initially, we used only three biomarkers for cardiovascular diseases. The protein-based biomarkers c-reactive protein (CRP) and brain natriuretic peptide (BNP) were covalently bound to microbead surfaces as described elsewhere 1Cell-free mitochondrial DNA (cfmDNA) served as DNA-based biomarker. This was captured by a complementary capture probe on the microbead surface. Fluorescence- and size-coded microparticles were used to distinguish the biomarkers. The microparticles were prepared and immobilised in a flow cell on a microfluidic chip (bi.flow Systems GmbH). cfmDNA was detected by a complementary fluorescence-labeled probe. Specific fluorescence-labelled antibodies were used to detect the protein biomarkers. In combination with our VideoScan technology 2 the samples can be analyzed fully automatically.
We identified a buffer in which both DNA and protein biomarkers can be detected simultaneously. It was possible to detect the biomarkers both in independent reactions and at the same time specifically. The antibody detection achieved a signal-to-noise ratio (SNR = 10 log10 (signal-to-noise)[dB]) and a relative fluorescence intensity (refMFI) of 9 dB and 0.8 for BNP, and 14 dB and 2.8 refMFI for CRP. The DNA-based biomarker cfmDNA was detected with an SNR of 20 dB and a refMFI signal of 0.7 in the presence of 1ng/µL cfmDNA. Kinetic experiments indicate that 50% of the fluorescence intensity is achieved within 7 minutes. In addition to the hardware we have developed digilogger, an interactive open source web application. As already described elsewhere3 we use the shiny technology to easily merge underlying data analysis framework and user-accessible web interface. Open source software can contribute to reproducibility and exchange of data 4digilogger can be used for data evaluation, visualisation and in the future for classification and machine learning (→ decision support system). The overall system 5 is to be further expanded in the future as a point-of-care technology.


Stefan Rödiger et al., “Fluorescence Dye Adsorption Assay to Quantify Carboxyl Groups on the Surface of Poly(Methyl Methacrylate) Microbeads,” Analytical Chemistry 83, no. 9 (May 1, 2011): 3379–85, https://doi.org/10.1021/ac103277s.
Stefan Rödiger et al., “VideoScan – A Microscope Imaging Technology Platform for the Multiplex Real-Time PCR,” March 18, 2013, www.gene-quantification.de.
Stefan Rödiger et al., “Enabling Reproducible Real-Time Quantitative PCR Research: The RDML Package,” Bioinformatics, August 26, 2017, doi.org/.
Stefan Rödiger et al., “R as an Environment for the Reproducible Analysis of DNA Amplification Experiments,” The R Journal 7, no. 2 (2015): 127–50, journal.r-project.org.
Franziska Dinter et al., “Integration of Cardiovascular Disease Biomarkers in a Microfluidic Microbead Chip,” in New Challanges & Perspectives for IVD in the Aging Society (Potsdam Days on Bioanalysis 2017, Potsdam: F1000Research [Poster], 2017), https://doi.org/DOI: 10.7490/f1000research.1115161.1; Claudia Deutschman et al., “Comparison of Lab and Point of Care (POC) Technologies – Case Study for CHI3L1,” in New Challanges & Perspectives for IVD in the Aging Society (Potsdam Days on Bioanalysis 2017, Potsdam, 2017), doi.org.

de_DEGerman en_GBEnglish