Brain tumors are notoriously hard to diagnose. Usually they are found with the help of neuroimaging studies, most commonly computed tomography (CT) imaging. These methods are costly, expose patients to ionizing radiation and must be repeated many times as they often give false positive results. The new method developed by a team led by Simon Podnar from the Neurology Division of the University Medical Center in Ljubljana can discover brain tumors and other diseases using a routine blood test. “Routine blood test results are assumed to contain much more information than is usually recognized even by the most experienced clinicians. Using routine blood tests from 15,176 neurological patients we built a machine learning predictive model for the diagnosis of brain tumors,” explain Slovenian scientists in an article, published by Scientific Reports Journal in October.
According to the authors, the accuracy of the new method is comparable to that of imaging studies, though they do not see the new method as an alternative but rather as complementary to the established diagnostic tools. The blood test method could be used as a cheap and efficient alternative approach that is useful for patients with headaches and other symptoms which could be a sign of tumors.
The machine learning algorithm is marketed through Swiss-based company Smart Blood Analytics. According to the authors, the brain tumor diagnostic method is only the first step. The new AI based approach could “predict hundreds of groups of diseases and medical conditions in the field of internal medicine”.