ECCO 2024
Abstract Submission

76. Approaches to the use of machine learning for medical problems

Contributed abstract in session FB-2: Machine learning, stream Machine learning.

Friday, 10:30 - 12:00
Room: M228

Authors (first author is the speaker)

1. Piotr Lukasiak
Institute of Computing Science, Poznan University of Technology
2. Maciej Majchrzak
Poznan University of Technology

Abstract

The ECBIG-MOSAIC project has established a unique form of collaboration that enables the application of artificial intelligence in conducting innovative research that integrates multidimensional biomedical and clinical data. The project lays the groundwork for the practical application of the latest scientific advancements in medicine, particularly in therapy, diagnosis, and disease prevention. As part of the project, the data collection and analysis process included civilization diseases selected based on national and international epidemiological, social, and economic conditions. These diseases include COVID-19, oncological diseases (such as thyroid cancer, lung cancer, and breast cancer), and cardiological diseases (such as premature coronary heart disease, resistant arterial hypertension, and atrial fibrillation). These activities could potentially help to identify the key processes involved in tumour development, which could lead to more personalized medical care for oncology and cardiovascular patients. The use of artificial intelligence in the diagnostic and therapeutic process could also be explored further.
AI is a technique used in medicine to analyze medical data using algorithms and statistical models. Its applications include predicting disease outcomes, diagnosing conditions, personalizing treatments, and enhancing patient care. By leveraging large datasets, machine learning can assist healthcare professionals in making informed decisions and improving overall health outcomes. In our research, we focused on head injuries and liver fibrosis.
Machine learning algorithms are increasingly being used in the context of head injuries to improve patient outcomes and guide clinical decisions. These algorithms analyze various factors to predict prognosis, risk stratification, and early detection of complications. Their application holds promise for personalized medicine and innovative protocols. With the help of acquired medical imaging data, a range of machine learning methods were applied to create classifiers that can assist radiologists in categorizing and determining the relevance of injuries. The results obtained have the potential to speed up the treatment of patients who require immediate medical attention.
Machine learning algorithms are also widely used in assessing and managing liver fibrosis and steatosis. They aid in diagnosis, staging, predicting disease progression, assessing treatment response, and risk stratification. These algorithms enhance patient care and contribute to personalized medicine. Data obtained from more than 100 patients were characterized by a set of broad-spectrum laboratory tests. Based on the learning set, an attempt was made to identify individuals who should be referred for detailed medical tests related to liver fibrosis and steatosis. The computational studies carried out and rules generated on the basis of various machine learning models make it possible to automate the diagnostic process.

Keywords

Status: accepted

Back to the list of papers