Hereditary non-polyposis colorectal cancer risk assessment based on AI analysis of pedigree data

Colorectal cancer (CRC) is one of the most common fatal cancers in developed countries and represents a significant public-health issue. About 3-5% of patients with CRC have hereditary non-polyposis colorectal cancer (HNPCC). Cancer morbidity and mortality can be reduced if early and intensive screening is pursued. But, despite advances in screening, population-wide genetic screening for HNPCC is not currently considered feasible due to its complexity and expense. If we can identify/assess the risk of a family having HNPCC, then only a fraction of the population will undergo intensive screening. This identification is currently performed by a genetic counsellor/physician who makes the decision based on some pre-defined criteria. The risk estimation by employing some mathematical methods, such as logistic regression, has also been reported. Our aim is to investigate the use of artificial intelligence techniques for genetic risk assessment. In this paper we summarize current knowledge on HNPCC and introduce the pedigree database used. Then we describe the system developed for HNPCC-risk assessment, which is based on analysing the pedigree data using self-organizing maps. The experimental evaluation shows good classification results.