br arc based method is served well by
940
896
arc-based method is served well by selecting an appropriate
proposed approach is more realistic as it KPT 330 closely mimics the
941
897
width of the annular region. The prediction of the radial
radiologist's clinical practice of two views-based analysis of
942
898
distances of the lesion from nipple on CC and MLO views and
breast anomalies. The preliminary results after fusion of single
943
899
the width of annular search region are determined statically in
view features with two views information have demonstrated
944
900
Refs. [9,28] which are applicable only to the images in their
a slight improvement in the case-based diagnostic accuracy
945
901
respective studies. In this regard, we have selected the width of
and FPs/I close to the film-based scheme. The findings of
946
902
annular region dynamically based on the length of minor axis
proposed two view scheme are significantly useful compared
947
903
of both lesions involved in the process. The success of this task
to the traditional single view CAD systems. The proposed
948
904
is dependent on the accurate positioning of the lesion and its
algorithms are helpful to improve the outcomes and efficiency
949
905
centre during the segmentation phase.
of two views CAD system for assisting the radiologists during
950
906
During matching the lesions on ipsilateral views in two
their diagnostic decisions in clinical practice. However, the
951
907
view analysis, one lesion may be paired with two or more
results obtained can be further optimized in terms of case-
952
908
lesions at approximately the same distance in the annular
based sensitivity and specificity. The perfection and accuracy
953
909
search region. Obviously, the only distance metric is not
in automatic nipple detection can be improved further and we
954
910
sufficient to select the real TP–TP pair. We have selected four
are attempting to do the same to enhance the performance of
955
911
simple similarity measures based on our intuitive expectation
our proposed approach. The overall performance improve-
956
912
that they would provide useful information for matching the
ments in our proposed algorithms are encouraging to increase
957
913
corresponding lesions. The radial distance-based score com-
the potential usage of our two views CAD system as a ‘‘second
958
914
bined with a similarity score, as described in Section 4.3.3 to
reader’’ in breast cancer diagnosis.
959
Please cite this article in press as: Sapate S, et al. Breast cancer diagnosis using abnormalities on ipsilateral views of digital mammograms.
960 Conflict of interest
962 Q3 Uncited reference
Appendix A. Supplementary data
965 Supplementary data associated with this article can be
970 et al. Bench to bedside molecular functional imaging in
971 translational cancer medicine: to image or to imagine? Clin
974 descriptor based on sparse curvelet coefficients for false-
975 positive reduction in mammograms. J Healthcare Eng
978 and new developments in breast cancer diagnosis and
981 A review on automatic mammographic density and
985 analysis for risk assessment, detection, diagnosis, and
988 MA, et al. Computer-aided detection of breast masses on
989 mammograms: dual system approach with two-view
992 aided detection of breast cancer. Comput Med Imaging
994 [8] Goudarzi M, Maghooli K. Extraction of fuzzy rules at
995 different concept levels related to image features of
996 mammography for diagnosis of breast cancer. Biocybern
999 Improvement of computerized mass detection on
mammograms: fusion of two-view information. Med Phys
1001
cancer detection: a CADx approach. Biocybern Biomed Eng
1004
mass detection based on ipsilateral multiview
1007
variability in cancer detection during routine repeat
1010
(incident) mammographic screening in a study of two
1011
of breast lesions: comparison of digital breast
1013
tomosynthesis and ultrasonography. Korean J Radiol
1014
Kerlikowske K, Buist DSM, et al. Accuracy and outcomes of
1017
screening mammography in women with a personal
1018
history of early-stage breast cancer. J Am Med Assoc
1019
mass detection based on ipsilateral multiview
1022
A new near-term breast cancer risk prediction scheme
1025
based on the quantitative analysis of ipsilateral view
1026
mammograms. Comput Methods Programs Biomed
1027
Brenner R, Flowers CI, et al. Benefit of semiannual