The objective of this study was to determine whether linear discrimina
nt analysis of different independent features of MR images of breast l
esions can increase the sensitivity and specificity of this technique.
For MR images of 23 benign and 20 malignant breast lesions, three ind
ependent classes of features, including characteristics of Gd-DTPA-upt
ake curve, boundary, and texture were evaluated. The three classes inc
luded five, four and eight features each, respectively. Discriminant a
nalysis was applied both within and across the three classes, to find
the best combination of features yielding the highest classification a
ccuracy. The highest specificity and sensitivity of the different clas
ses considered independently were as follows: Gd-uptake curves, 83% an
d 70%; boundary features, 86% and 70%; and texture, 70% and 75%, respe
ctively, A combination of one feature each from the first two classes
and age yielded a specificity of 79% and sensitivity of 90%, whereas h
ighest figures of 93% and 95%, respectively, were obtained when a tota
l of 10 features were combined across different classes. Statistical a
nalysis of different independent classes of features in MR images of b
reast lesions can improve the classification accuracy of this techniqu
e significantly.