Model-based detection of spiculated lesions in mammograms.

Computer-aided mammographic prompting systems require the reliable detection of a variety of signs of cancer. In this paper we concentrate on the detection of spiculated lesions in mammograms. A spiculated lesion is typically characterized by an abnormal pattern of linear structures and a central mass. Statistical models have been developed to describe and detect both these aspects of spiculated lesions. We describe a generic method of representing patterns of linear structures, which relies on the use of factor analysis to separate the systematic and random aspects of a class of patterns. We model the appearance of central masses using local scale-orientation signatures based on recursive median filtering, approximated using principal-component analysis. For lesions of 16 mm and larger the pattern detection technique results in a sensitivity of 80% at 0.014 false positives per image, whilst the mass detection approach results in a sensitivity 80% at 0.23 false positives per image. Simple combination techniques result in an improved sensitivity and specificity close to that required to improve the performance of a radiologist in a prompting environment.

The influence of previous films on screening mammographic interpretation and detection of breast carcinoma.

AIM: To establish if the availability of previous mammograms improves the detection rate of carcinomas and reduces supplementary examinations in the National Health Service Breast Screening Programme (NHSBSP). METHOD: Eight radiologists with varying experience but an interest in reporting mammograms reported 100 mammograms on two occasions, at least 1 month apart. In this selection of 100 mammograms we randomly inserted 12 abnormal films, with both benign and malignant changes present. These were chosen retrospectively with histological correlation of the abnormality. On the first occasion only the current films were reviewed, however, on the second occasion previous films were available for comparison. The films were viewed under standard viewing conditions, and the participants were asked to comment if they required further views or would refer the patient for either an ultrasound examination or surgical referral. RESULTS: Receiver operating curves (ROC) were constructed for the group's overall performance on each of the occasions and there was no difference in the curves. This indicates that the presence of previous films did not improve diagnostic accuracy. The presence of previous films did, however, lead to a significant reduction in the number of additional views and ultrasound examinations that would have been requested. The most significant reductions occurred in the group of normal mammograms. CONCLUSION: The presence of previous mammograms does not lead to an improvement in diagnostic accuracy. However, the presence of previous films led to significant reduction in additional examinations and ultrasound examinations.