Codes to better see breast cancer tumours
Various descriptive criteria are being used in assessment to determine when a breast cancer is enhancing. Lesions can thus be described either as a focus, a mass, or a non-mass-like. Features like the distribution pattern, shape, and internal enhancement patterns, all help determine whether a non-mass-like breast lesion is benign or malignant. Within the EU-funded project MAMMA (Spatio-temporal modeling for enhanced automated detection and classification of non-mass lesions in breast MRI), scientists developed novel computer-assisted diagnosis techniques to help determine with a higher degree of specificity and sensitivity, whether a lesion is cancerous or benign. The project team developed improved spatio-temporal recognition techniques that overcame difficulties of existing image analysis techniques. Their validation in three specific experiments revealed substantial improvement in diagnostic accuracy and efficiency. Addition of new algorithms in breast computer-assisted diagnosis systems led to the creation of a flexible toolbox. Only minimal modification is required for application to other cases such as the identification of different lesion types or monitoring the response to chemotherapy. The EU has identified the overarching issues in cancer treatment – prevention, early detection, diagnosis and optimal treatment. They have supported numerous cancer research initiatives to improve the way cancers are tackled. With MAMMA’s improved image analysis methods, radiologists can now enhance interpretation of MRIs or mammograms to accurately detect and classify tumours.
Keywords
Breast cancer, imaging, non-mass-like, MAMMA, computer-assisted diagnosis
