Keywords:
Image registration, Computer Applications-3D, Ultrasound, Breast
Authors:
J. C. H. Chiu1, A. Choate2, J. Ecanow2, G. Spear3, S. Mondello 4, Y. Lei5, C. Segarceanu5; 1Evanston, IL/US, 2Evanston/US, 3Evanston /US, 4Messina/IT, 5Chicago/US
DOI:
10.1594/ecr2018/C-2153
Aims and objectives
Ultrasound is an inexpensive,
non-invasive and widely used imaging modality which has been demonstrated to be a valuable tool in the detection and monitoring of breast lesions and cancer,
especially when it is combined with screening mammography in women with dense breasts(1).
However,
the power of breast ultrasound can be limited by its inherent technical difficulties and inter-operator variability.
One study revealed that less than half of the lesions measuring 5 mm or smaller were identified by all experienced examiners (2).
The landmark study ACRIN 6666 has demonstrated that in approximately 20% of the 2662 subjects requiring follow-up imaging of probably benign breast lesions,
38.7% of the lesions cannot be seen in the follow-up examinations (3).
The success of finding the target lesion in follow-up visits relies upon the accuracy and reproducibility of positional coordinates recorded during the index examination.
Nonetheless,
manual positional annotations have been significantly limited due to operator’s dependency.
The distance to nipple of the target lesion was recorded only 13.1% of the time in one study (4) and over 60% of examinations did not comply with at least one ACR (American College of Radiology) recommendation in a different study (5).
Therefore,
automated mapping systems that can accurately,
reliably,
and rapidly localize and re-localize breast lesions would be of immense clinical benefit and are highly needed. In our study,
we aimed to determine the effects of patient position changes on the exam table,
and ultrasound probe position and orientation changes relative to the patient coronal plane,
on the positional coordinates of breast lesions using the BVN Model G-1000 (Metritrack Inc.,Hillside,
IL,USA),
which is a novel breast imaging and mapping system.
We also evaluated the ability of the said system to re-localize breast lesions by matching the ultrasound probe position relative to the nipple and body planes and the patient position on the exam table,
between the index image and an image obtained in subsequent examination.
Finally,
we estimated the time-saving benefits of automated annotation compared to the conventional manual annotation.