The first few hours following a stroke are crucial. To be able to treat a patient effectively, doctors must rapidly localize the damaged blood vessel and determine what kind of stroke occurred. In most cases, either a ruptured blood vessel releases blood into the brain, or a blood clot obstructs a blood vessel in the brain. Patients who experience the second type of stroke are prescribed medication to dissolve the blood clot. If this medication is given to patients of the first type, however, it will fluidify the blood and only make the hemorrhaging worse. Yet doctors must take action quickly because the faster a stroke is treated, the lower the likelihood of severe consequences. \u201cStroke patients are given an MRI [magnetic resonance imaging] scan immediately upon arriving at the hospital,\u201d says Antoine Madrona, who is completing a Master\u2019s degree in life sciences. \u201cThe scan is used to confirm that it\u2019s indeed a stroke and to identify what type, in order to prescribe the right treatment.\u201d<\/p>\n
Applying AI to medical imaging<\/strong><\/p>\n Deep-learning algorithms that sort through vast data sets to generate predictions can help doctors make the right diagnosis. They can show radiologists where they should focus their attention, provide quantitative information on the position, size and number of blood vessel injuries, and speed the process of selecting a treatment. As part of his Master\u2019s project, Madrona helped develop an algorithm based on data from diffusion-weighted magnetic resonance imaging (DW-MRI) \u2013 a form of MRI that uses the diffusion of water molecules to generate contrast in images. The DW-MRI data sets that Madrona obtained from various Swiss hospitals were extremely heterogeneous. \u201cI was surprised to see so much variation in the images from different hospitals,\u201d he says. \u201cI didn\u2019t expect to see such little standardization.\u201d<\/p>\n In other words, the same patient undergoing an MRI in two different hospitals will end up with two very different images. That\u2019s due to variations in the machine model and the image acquisition protocol. \u201cThere\u2019s no national or even international standard,\u201d says Madrona. \u201cEach hospital uses its own magnetic pulse sequence and sets its own pulse directions and intensities. All that affects the image contrast and overall appearance.\u201d<\/p>\n Data security <\/strong><\/p>\n To make sure patients\u2019 data remain confidential, Madrona used a federated learning method for his algorithm. This method entails training an algorithm across several data sets but without exchanging any data between them. None of the medical imaging systems currently on the market employs a federated method. \u201cI\u2019m particularly interested in how decentralized algorithms can be used to protect patient data,\u201d he says. \u201cThis gets to an important ethical issue that I believe needs to be addressed more widely in the healthcare industry. Patient confidentiality shouldn\u2019t be the price to pay for more efficient diagnostics.\u201d Madrona, now 25, intends to pursue a career in this direction.<\/p>\n His Master\u2019s project isn\u2019t quite over, but Madrona already views it as a positive experience. The algorithm could eventually help radiologists in analyzing strokes, regardless of the hospital, MRI machine or image acquisition protocol used. \u201cWhat really motivates me about my project is the concrete benefits it can deliver directly to clinical applications,\u201d he says.<\/p>\n Another aspect that Madrona appreciates was the opportunity to work with many different institutes. \u201cMy project is being supervised jointly by EPFL and the CHUV,\u201d he explains. \u201cI\u2019ve received guidance from Prof. Jean-Philippe Thiran at EPFL\u2019s Signal Processing Laboratory 5 from the School of Engineering, from Jonathan Pati\u00f1o, a postdoc at the lab, and from Jonas Richiardi, the principal investigator at UNIL\u2019s Translational Machine Learning Laboratory, which is affiliated with the CHUV\u2019s Department of Medical Radiology.\u201d Madrona\u2019s research took place under the umbrella of the Advanced Stroke Analytics Platform (ASAP), a project funded in part by Innosuisse. He also interacted with two other project partners: <\/strong>the Inselspital university hospital in Bern as well as Siemens Healthineers, a pioneer in advanced healthcare technology. A prototype of the new system will be tested at the CHUV and Inselspital in the next six months.<\/p>\n","protected":false},"featured_media":827,"template":"","project":[],"faculty":[32,31],"public":[27,28,30,25,29,26],"themes":[24],"news-category":[23],"class_list":["post-826","news","type-news","status-publish","has-post-thumbnail","hentry","faculty-sti","faculty-sv","public-collaborators","public-industries-partners","public-media","public-prospective-students","public-public","public-students","themes-health","news-category-research"],"_links":{"self":[{"href":"https:\/\/neuro-x.epfl.ch\/en\/wp-json\/wp\/v2\/news\/826","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/neuro-x.epfl.ch\/en\/wp-json\/wp\/v2\/news"}],"about":[{"href":"https:\/\/neuro-x.epfl.ch\/en\/wp-json\/wp\/v2\/types\/news"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/neuro-x.epfl.ch\/en\/wp-json\/wp\/v2\/media\/827"}],"wp:attachment":[{"href":"https:\/\/neuro-x.epfl.ch\/en\/wp-json\/wp\/v2\/media?parent=826"}],"wp:term":[{"taxonomy":"project","embeddable":true,"href":"https:\/\/neuro-x.epfl.ch\/en\/wp-json\/wp\/v2\/project?post=826"},{"taxonomy":"faculty","embeddable":true,"href":"https:\/\/neuro-x.epfl.ch\/en\/wp-json\/wp\/v2\/faculty?post=826"},{"taxonomy":"public","embeddable":true,"href":"https:\/\/neuro-x.epfl.ch\/en\/wp-json\/wp\/v2\/public?post=826"},{"taxonomy":"themes","embeddable":true,"href":"https:\/\/neuro-x.epfl.ch\/en\/wp-json\/wp\/v2\/themes?post=826"},{"taxonomy":"news-category","embeddable":true,"href":"https:\/\/neuro-x.epfl.ch\/en\/wp-json\/wp\/v2\/news-category?post=826"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}