Clinical Trial of a CAD System for Microcalcification Detection and Classification
Antonis Frigas, George M. Spyrou, George Zografos, Dimitra P. Koulocheri, John Mantas, Panos A. Ligomenides · Medinfo 2007: Proceedings of the 12th World Congress on Health (Medical) Informatics; Building Sustainable Health Systems · 2007
According to the World Health Organisation, breast cancer is the second leading cause of cancer deaths in women today and is the most common cancer among women, excluding nonmelanoma skin cancers. Several studies have proved that the early detection of breast cancer can reduce mortality rates and improve the chances that the patient is diagnosed at an early stage and treated successfully. Mammography is widely used as the primary method for breast cancer detection. Breast microcalcifications usually appear in the form of clusters and sometimes can be easily detected on mammografic films due to their high clustering density. An approach in the classification of microcalcifications has been presented by Ligomenides et al. and it is based on detailed analysis and evaluation of seven features of individual microcalcifications and of formed clusters. The proposed system is now in clinical trial in an Athens University Hospital's Breast Unit. The results from the first 51 cases (all biopsy tested) that were processed are presented in this paper and indicate that all malignant cases were classified correctly.