A team of researchers from Ludwig-Maximilians-Universität München (LMU) and the Max Planck Institute of Quantum Optics (MPQ) in Germany has developed a new way to analyse a drop of blood to provide healthy insights within minutes, according to a study published in the scientific journal Cell Reports Medicine.
Using infrared light and machine learning, researchers have developed a method to screen small blood samples to detect multiple health conditions with just one measurement.
Infrared spectroscopy has been used to analyse molecular composition of substances for many years. When applied to blood, this technique can identify different molecular signals which can be used in medical diagnosis. However, despite being routinely used in research, this procedure has not really been established in medical diagnosis.
In this study, the team analysed samples from thousands of participants in the KORA study, a comprehensive health research project in Augsburg, Germany. Random participants were selected for a representative sample and were recruited for medical examinations and blood donations.
More than 5,000 blood samples were analysed using Fourier transform infrared (FTIR) spectroscopy. The team then applied machine learning to analyse the molecular fingerprints and correlate them with the patient’s medical data. The results showed that these blood samples contain valuable information for a quick health screening, including detecting abnormal levels of blood lipids and various changes in blood pressure, as well as spotting type-2 diabetes and even pre-diabetes. Curiously, the AI system could also detect the participants who were healthy and likely to stay healthy over the investigated years.
These results were important for two reasons: firstly, most people experience health changes during their lives, but finding healthy individuals is also relevant. Secondly, many individuals suffer from multiple conditions. To diagnose these patients, doctors need a test for each disease. However, this method identifies a wide range of health issues and spots complex situations involving multiple illnesses simultaneously. In addition, it can predict the development of metabolic syndrome years before symptoms appear, providing a window for interventions.
The authors believe this new system has the potential to become a routine part of health screenings. This would be especially valuable to detect cholesterol problems and diabetes, where a timely and effective intervention significantly improves outcomes. However, potential applications extend even further. It’s likely that with further developments, this technique could identify even more health conditions. The aim is to develop personalized health monitoring, where individuals can check their health status and catch potential issues long before they become serious.
Eissa T, Leonardo C, Kepesidis KV, Fleischmann F, Linkohr B, Meyer D, Zoka V, Huber M, Voronina L, Richter L, Peters A, Žigman M. Plasma infrared fingerprinting with machine learning enables single-measurement multi-phenotype health screening. Cell Rep Med. 2024 Jul 16;5(7):101625.