The application of machine learning algorithms in the prognosis of liver cancer is investigated

At the next department seminar held at the Institute of Information Technology of the Ministry of Science and Education, PhD student Lala Garayeva gave a presentation titled "Application of machine learning algorithms for prognosis of Hepatocellular Carcinoma."

The speaker noted that, according to the World Health Organization, more than 14 million people worldwide suffer from cancer. Liver cancer ranks second in mortality rates, with hepatocellular carcinoma (HCC) accounting for 80% of cases. She stated that the disease can be treated more effectively if diagnosed at an early stage.

L. Garayeva highlighted that during the diagnosis and prognosis of HCC, numerous unstructured indicators can increase the likelihood of medical errors, emphasizing the necessity of building intelligent systems to support clinical decision-making, and noted, "The application of machine learning and deep learning methods plays a crucial role in developing such systems."

The speaker noted that, within the framework of the study, the Kaggle HCC Dataset—compiled from data on 204 patients with HCC at the University Hospital of Portugal—was selected and the clinical features in the dataset were analyzed. The target class was defined as a binary classification—"deceased" or "alive"—based on the patients’ one-year outcomes.

She noted that, to address missing data in the national HCC clinical patient database—developed through a collaboration between Azerbaijan Medical University and the Inonu Liver Transplantation Institute in Malatya, Türkiye—the MICE (Multiple Imputation by Chained Equations) method was applied. The advantages of this method are based on its ability to work with incomplete data, to preserve the natural variability in the data, and to obtain more reliable results.

At the conclusion of her presentation, L. Garayeva emphasized that the newly developed model enables the prediction of patients’ conditions based on real clinical data.

Then, extensive discussions were held on the topic, and questions were answered.

© All rights reserved. Citing www.iсt.az is necessary when using news.