Current breast cancer screening includes mammography and clinical examination. Findings from screening that are referred to biopsy after final mammographic workup are malignant in approximately 20-30% of the cases. The low specificity of diagnostic mammography has led to the development of alternative imaging tools for distinguishing benign from malignant lesions in the breast. Tools such as digital mammography, ultrasound, MRI, optical imaging, and PET have been suggested and preliminary data suggest promise in improving diagnostic accuracy over film screen mammography alone in all cases. There are few data on the relative diagnostic performance of these newer studies. Furthermore, little is known about the correlation between findings from the different imaging modalities or the value of combinations of diagnostic imaging modalities. Research teams at the University of Pennsylvania have been developing diagnostic imaging protocols in all of these technologies. This proposal represents a unique collaboration between the individual modality teams to study the relative performance and interaction between proposed diagnostic imaging undergo an imaging battery consisting of Digital Mammography, Ultrasound, Optical Imaging, MRI and in years 3-5 PET performed with a dedicated breast scanner. Patients will be characterized with respect to mammographic density, presenting mammographic or clinical finding, compliance with mammographic screening and other characteristics relevant to breast cancer risk and imaging performance. In addition to quantifying the relative diagnostic performance of these modalities, individual features extracted by human observers will be studied for reliability and predictive value. Interpretation models for individual imaging modalities and selected groups of imaging modalities will be developed. Strategies for multi- modality diagnostic work ups for different patient strata will be developed.
Showing the most recent 10 out of 14 publications