| Title: | Building a multi-disease prognostic panel toward individualized preventive medicine |
| Lead Institution: | Lund University |
| Principal investigator: |
Application 105777
WARNING: the interactive features of this website use CSS3, which your browser does not support. To use the full features of this website, please update your browser.
About
Many age-related disease progress silently in the body without generating any symptoms. Once symptoms do appear, the disease are often in a more advanced stage and is harder to treat. In theory, being able to recognize, diagnose and track diseases early on would give health professionals a better chance at successfully treating diseases in a cost effective way. Unfortunately, we do not have good ways of tracking and diagnosing most age-related diseases. Even if we did, it would not be financially possible to run hundreds of diagnostic and pre-screening tests on every healthy adult coming in for an annual check-up. This proposal seeks to address both of these issues by building a panel of accurate diagnostic tests that can be obtained through a single blood test. This kind of test would be able to diagnose multiple diseases that would often require consultation and authorization from a specialist (which rarely happens before presence of symptoms). The test would also be inexpensive enough that it could be performed on healthy adults on a regular (i.e. annual) basis in order to monitor possible changes in health. Execution of our proposal relies on new data collected by the UK BioBank (and by our lab) the measures the concentration of thousands of proteins in the blood. Data from our own lab suggests that we can train artificial intelligence to learn how to spot diseases like Alzheimer's disease just by scanning the concentration of different proteins. The main focus of our proposal will be to train many different artificial intelligence programs to recognize and differentiate many types of diseases. This will be made possible for the first time thanks to the enormous size and high quality of UK BioBank data. We will train the programs such that, when they fail, they are unlikely to falsely signal the presence of the disease when there is none (at the cost of potentially missing some positive cases). Importantly, we will test whether these programs can also predict changes to other known medical markers that might become abnormal before diseases begin to show symptoms. As a final step, we will test these programs in a real clinical setting to see if they can diagnose diseases, or whether they can forecast future disease in healthy people. This will be an important test to see if a single blood test can predict the presence of multiple diseases.1 Publication
| Pub ID | Title | Author(s) | Year | Journal |
|---|---|---|---|---|
| 17639 | Identification of distinct and shared biomarker panels in different manifestations of cerebral small-vessel disease through proteomic profiling | Ines Hristovska (+14) | 2026 | Nature Aging |
Enabling scientific discoveries that improve human health