Automated regional registration and characterization of corresponding microcalcification clusters on temporal pairs of mammograms for interval change analysis

Abstract

A computerized regional registration and characterization system for analysis of microcalcification clusters on serial mammograms is being developed in our laboratory. The system consists of two stages. In the first stage, based on the location of a detected cluster on the current mammogram, a regional registration procedure identifies the local area on the prior that may contain the corresponding cluster. A search program is used to detect cluster candidates within the local area. The detected cluster on the current image is then paired with the cluster candidates on the prior image to form true (TP-TP) or false (TP-FP) pairs. Automatically extracted features were used in a newly designed correspondence classifier to reduce the number of false pairs. In the second stage, a temporal classifier, based on both current and prior information, is used if a cluster has been detected on the prior image, and a current classifier, based on current information alone, is used if no prior cluster has been detected. The data set used in this study consisted of 261 serial pairs containing biopsy-proven calcification clusters. An MQSA radiologist identified the corresponding clusters on the mammograms. On the priors, the radiologist rated the subtlety of 30 clusters (out of the 261 clusters) as 9 or 10 on a scale of 1 (very obvious) to 10 (very subtle). Leave-one-case-out resampling was used for feature selection and classification in both the correspondence and malignant/benign classification schemes. The search program detected 91.2% (238/261) of the clusters on the priors with an average of 0.42 FPs/image. The correspondence classifier identified 86.6% (226/261) of the TP-TP pairs with 20 false matches (0.08 FPs/image) relative to the entire set of 261 image pairs. In the malignant/benign classification stage the temporal classifier achieved a test A(z) of 0.81 for the 246 pairs which contained a detection on the prior. In addition, a classifier was designed by using the clusters on the current mammograms only. It achieved a test A(z) of 0.72 in classifying the clusters as malignant and benign. The difference between the performance of the temporal classifier and the current classifier was statistically significant (p=0.0014). Our interval change analysis system can detect the corresponding cluster on the prior mammogram with high sensitivity, and classify them with a satisfactory accuracy.

Figure 1

Visibility ratings of the microcalcification clusters on the current mammogram plotted against those on the prior mammogram for (a) malignant and (b) benign temporal pairs. The visibility was rated on a ten-point discrete scale (1=most obvious, 10=subtlest). The area of the circles is proportional to the number of data points with the same ratings. The smallest circles in both (a) and (b) represent one data point. The largest circle in (a) represents 15 data points, and the largest circle in (b) represents 29 data points. The solid diagonal line y=x represents equal visibility ratings on the current and prior mammograms. The dashed lines are the linear regression lines for the data fitted as y=0.782x+2.54 for (a) and as y=0.899x+0.925 for (b). The correlation coefficient is 0.593 for the malignant clusters and 0.791 for the benign clusters.

Figure 2

Temporal interval between the current and the prior mammograms for the 261 pairs in our data set.

Figure 3

Block diagram of the regional registration for temporal microcalcification clusters (stage 1).

Figure 4

Temporal pair of mammograms containing microcalcification cluster. (a) Current and prior mammograms with automatically detected breast boundaries, (b) current and prior microcalcification cluster. The current and prior images were obtained 2 years apart.

Figure 5

Initial estimation of the cluster centroid position on the prior mammogram based on the nipple-cluster distance and the angle between the nipple-cluster axis and breast periphery on the current mammogram.

Figure 6

Definition of an initial fan-shaped search region on the prior mammogram centered at the predicted centroid location (black dot). An automated microcalcification detection program was used to detect cluster candidates [true (TP) and false (FP)] within the search region on the prior. The cluster on the current image is paired with the detected candidates on the corresponding prior image to form true (TP-TP) or false (TP-FP) pairs.

Figure 7

Block diagram of the correspondence classifier used to reduce the false pairs (TP-FP).

Figure 8

Block diagram of classification stage for temporal microcalcification clusters (stage 2).

Figure 9

Block-diagram of the temporal classifier for classification of malignant and benign microcalcification clusters. Both TP-TP and TP-FP pairs can be input to the system.

Figure 10

ROC curves for the temporal malignant-benign classifier (A_z=0.81±0.03) and current malignant-benign classifier (A_z=0.72±0.04). The difference in A_z between the two classifiers was statistically significant (p=0.0014).