Computerized Evaluation of Mammographic Lesions: What Diagnostic Role Does the Shape of the Individual Microcalcifications Play Compared with the Geometry of the Cluster?
I. Leichter1,
R. Lederman2,
S. S. Buchbinder3,4,
P. Bamberger5,
B. Novak1 and
S. Fields2
1 Department of Electro-Optics, Jerusalem College of Technology, Jerusalem,
Israel.
2 Department of Radiology, Hadassah University Hospital, Jerusalem,
Israel.
3 Department of Radiology, Montefiore Medical Center, Albert Einstein College of
Medicine, Bronx, NY.
4 Staten Island University Hospital, 475 Seaview Ave., Staten Island, NY
10305.
5 Department of Electronics, Jerusalem College of Technology, Jerusalem,
Israel.
View larger version (106K):
[in a new window]
|
Fig. 1. —Computer-processed image shows triangulation lines of
Delaunay [19] connecting each
microcalcification to its nearest neighbors, obtained by computerized analysis
of cluster of microcalcifications on digital image. Patient was asymptomatic
71-year-old woman.
|
|
View larger version (15K):
[in a new window]
|
Fig. 2. —Receiver operating characteristic curve shows performance of
classification scheme based on discriminant analysis, using weighted values
for average number of neighbors in cluster and individual microcalcification
shape. Area under curve is 0.87.
|
|
View larger version (16K):
[in a new window]
|
Fig. 3. —Bar graph shows distribution of average number of neighbors
(cluster geometry) per microcalcification in cluster, extracted automatically
by computerized analysis. Black bar = malignant, white bar = benign.
|
|
View larger version (15K):
[in a new window]
|
Fig. 4. —Receiver operating characteristic curve shows performance of
classification scheme based on discriminant analysis using only average number
of neighbors in cluster. Area under the curve is 0.84.
|
|
View larger version (112K):
[in a new window]
|
Fig. 5A. —Digital image shows malignant cluster in which both shape of
individual microcalcifications and cluster geometry indicate malignancy.
Patient was asymptomatic 47-year-old woman. Original digitized radiograph
shows cluster of microcalcifications.
|
|
View larger version (111K):
[in a new window]
|
Fig. 5B. —Digital image shows malignant cluster in which both shape of
individual microcalcifications and cluster geometry indicate malignancy.
Patient was asymptomatic 47-year-old woman. Computer-processed image shows
irregularly shaped microcalcifications and triangulation lines that indicate
tight cluster geometry.
|
|
View larger version (99K):
[in a new window]
|
Fig. 6A. —Digital image shows typical benign cluster in which both
shape of individual microcalcifications and cluster geometry indicate
benignancy. Patient was asymptomatic 56-year-old woman. Original digitized
radiograph shows cluster of microcalcifications.
|
|
View larger version (102K):
[in a new window]
|
Fig. 6B. —Digital image shows typical benign cluster in which both
shape of individual microcalcifications and cluster geometry indicate
benignancy. Patient was asymptomatic 56-year-old woman. Computer-processed
image shows rounded microcalcifications and triangulation lines that indicate
scattered cluster geometry.
|
|
View larger version (88K):
[in a new window]
|
Fig. 7A. —Digital image shows malignant cluster in which cluster
geometry indicates malignancy, but shape is more rounded than in typical
malignant cluster. Patient was asymptomatic 64-year-old woman. Original
digitized radiograph shows cluster.
|
|
View larger version (125K):
[in a new window]
|
Fig. 7B. —Digital image shows malignant cluster in which cluster
geometry indicates malignancy, but shape is more rounded than in typical
malignant cluster. Patient was asymptomatic 64-year-old woman.
Computer-processed image shows triangulation lines that indicate tight cluster
geometry.
|
|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
Copyright © 2004 by the American Roentgen Ray Society.
