A regional registration technique for automated interval change analysis of breast lesions on mammograms

Analysis of interval change is a useful technique for detection of abnormal ities in mammographic interpretation. Interval change analysis is routinely used by radiologists and its importance is well-established in clinical pr actice. As a first step to develop a computerized method for interval chang e analysis on mammograms, we are developing an automated regional registrat ion technique to identify corresponding lesions on temporal pairs of mammog rams. In this technique, the breast is first segmented from the background on the current and previous mammograms. The breast edges are then aligned u sing a global alignment procedure based on the mutual information between t he breast regions in the two images. Using the nipple location and the brea st centroid estimated independently on both mammograms, a polar coordinate system is defined for each image. The polar coordinate of the centroid of a lesion detected on the most recent mammogram is used to obtain an initial estimate of its location on the previous mammogram and to define a fan-shap ed search region. A search for a matching structure to the lesion is then p erformed in the fan-shaped region on the previous mammogram to obtain a fin al estimate of its location. In this study, a quantitative evaluation of re gistration accuracy has been performed with a data set of 74 temporal pairs of mammograms and ground-truth correspondence information provided by an e xperienced radiologist. The most recent mammogram of each temporal pair exh ibited a biopsy-proven mass. We have investigated the usefulness of correla tion and mutual information as search criteria for determining correspondin g regions on mammograms for the biopsy-proven masses. In 85% of the cases ( 63/74 temporal pairs) the region on the previous mammogram that corresponde d to the mass on the current mammogram was correctly identified. The region centroid identified by the registration technique had an average distance of 2.8+/-1.9 mm from the centroid of the radiologist-identified region. The se results indicate that our new registration technique may be useful for e stablishing correspondence between structures on current and previous mammo grams. Once such a correspondence is established an interval change analysi s could be performed to aid in both detection as well as classification of abnormal breast densities. (C) 1999 American Association of Physicists in M edicine. [S0094-2405(99)00612-4].