B. Sahiner et al., Improvement of mammographic mass characterization using spiculation measures and morphological features, MED PHYS, 28(7), 2001, pp. 1455-1465
Citations number
46
Categorie Soggetti
Radiology ,Nuclear Medicine & Imaging","Medical Research Diagnosis & Treatment
We are developing new computer vision techniques for characterization of br
east masses on mammograms. We had previously developed a characterization m
ethod based on texture features. The goal of the present work was to improv
e our characterization method by making use of morphological features. Towa
rd this goal, we have developed a fully automated, three-stage segmentation
method that includes clustering, active contour, and spiculation detection
stages. After segmentation, morphological features describing the shape of
the mass were extracted. Texture features were also extracted from a band
of pixels surrounding the mass. Stepwise feature selection and linear discr
iminant analysis were employed in the morphological, texture, and combined
feature spaces for classifier design. The classification accuracy was evalu
ated using the area A(z) under the receiver operating characteristic curve.
A data set containing 249 films from 102 patients was used. When the leave
-one-case-out method was applied to partition the data set into trainers an
d testers, the average test A(z) for the task of classifying the mass on a
single mammographic view was 0.83 +/- 0.02, 0.84 +/- 0.02, and 0.87 +/- 0.0
2 in the morphological, texture, and combined feature spaces, respectively.
The improvement obtained by supplementing texture features with morphologi
cal features in classification was statistically significant (p = 0.04). Fo
r classifying a mass as malignant or benign, we combined the leave-one-case
-out discriminant scores from different views of a mass to obtain a summary
score. In this task, the test A(z) value using the combined feature space
was 0.91 +/- 0.02. Our results indicate that combining texture features wit
h morphological features extracted from automatically segmented mass bounda
ries will be an effective approach for computer-aided characterization of m
ammographic masses. (C) 2001 American Association of Physicists in Medicine
.