Novel Biomarkers from genes in the apoptotic pathway for Prediction of HCC Progression using Association Rule Mining

Abstract

Liver cancer, a main cause of death, is extremely difficult to be diagnosed at its early stages. On a positive side, predicting the disease development or progression by analyzing medical data can be helpful for the future early diagnosis and accordingly the increase of the patients' survival. Medical investigation and researchers raise that Single nucleotide polymorphisms in certain apoptosis-related genes are related to the cancer development. The objective of this paper is to find quantitative associations between apoptotic gene-related polymorphisms and the progression level of the liver cancer. To find these associations, Association rule mining is applied using the Frequent Pattern algorithm. An experimental study on an Egyptian cohort of 1246 patients with advanced cirrhosis and liver cancer resulted in associations which can serve as novel biomarkers. It has been found that CDKN2A and HLA-DP genes have relation to the HCC development with a confidence value 0.55, and CDKN1B and Il28b, are related to the liver cancer progression with a confidence value 0.54.