Stroke is a cerebral vascular disease, which has become the highest cause of death and disability in Indonesia. In dealing with stroke, thrombolysis is an implemented method to break blood clots that clog blood vessels in the brain. In an effort to shorten the time of handling stroke, a lecturer of the Faculty of Medicine (FK) of Universitas Indonesia (UI), also a student of the Doctoral Program of FKUI, dr. Reyhan Eddy Yunus, Sp.Rad, Subsp.NKL(K), M.Sc., developed an artificial intelligence model to predict the benefits of thrombolysis therapy in stroke patients.
“This research is the first in Indonesia, which utilizes artificial intelligence to determine therapy in patients with hyperacute occlusion stroke using local data. The research findings open new opportunities in handling stroke and is hoped to increase the patient’s life quality and decrease the burden on the national health system,” said dr. Reyhan.
Based on the process of occurrence, stroke is differentiated into two, namely hemorrhagic stroke and non-hemorrhagic or occlusion stroke. Hyperacute occlusion stroke happens when blood flow to the brain network is disturbed due to blockage within six hours of stroke onset. This onset time needs to be known in order to determine whether intravenous thrombolysis therapy or mechanical thrombectomy can be performed.
Through the dissertation, titled “Pengembangan Model Kecerdasan Buatan Pembelajaran Mesin untuk Prediksi Keberhasilan Terapi Trombolisis Intravena pada Stroke Iskemik Hiperakut Sirkulasi Anterior dengan Menggunakan CT scan Otak, Data Klinis, dan Laboratorium Darah,” dr. Reyhan involved 145 samples for machine learning algorithm development. Samples were taken retrospectively based on stroke code registration at Dr. Cipto Mangunkusumo Hospital (RSCM) from November 2014 to February 2023.
From the entirety of samples, patients with hyperacute occlusion stroke who are given thrombolysis treatment, have non-contrast brain CT scan data, have clinical data on hospital admission and 24 hours post thrombolysis, and have blood laboratory data related to stroke, are included in this study. Data variables are used as trial input in the development of a machine-learning model to predict the clinical improvement of patients after thrombolysis therapy.
Data processing and model development were done using Random Forest (RF) and Convolutional Neural Network (CNN) machine learning algorithms. This artificial intelligence model is able to help predict the outcome of hyperacute occlusion stroke patients after thrombolysis by utilizing clinical, laboratory, and brain CT scan data. This model can be used as a tool for clinicians, especially in hospitals with limited specialist doctors, but can perform thrombolysis therapy.
“With this research, it is hoped that there will be new breakthroughs in the management of occlusion stroke to reduce patient mortality and disability rates. However, this artificial intelligence model still requires testing and applied more widely to other stroke center hospitals that perform thrombolysis therapy,” said dr. Reyhan, who is also the Head of the Department of Radiology, FKUI-RSCM.
Owing to his research, dr. Reyhan managed to attain a Doctorate from FKUI at the doctoral promotion session held at the Auditorium, 3rd Floor of IMERI-FKUI Building, Jakarta, on Tuesday, July 9. The Dean of FKUI, Prof. Ari Fahrial hoped that there would be more research on utilizing Artificial Intelligence (AI) in the world of medicine and health in the future, just as what dr. Reyhan has done. “I see that dr. Reyhan is consistent with the topic raised as he is also involved in multiple AI studies, both for handling COVID-19 and tuberculosis, and now for stroke. I hope that he stays consistent and that there will be new research on AI in the field of Radiology,” said Prof. Ari Fahrial.
