Thresholds for visual detection of Young's modulus contrast in simulated ultrasound image movies

Citation
Nr. Miller et Jc. Bamber, Thresholds for visual detection of Young's modulus contrast in simulated ultrasound image movies, PHYS MED BI, 45(8), 2000, pp. 2057-2079
Citations number
36
Categorie Soggetti
Multidisciplinary
Journal title
PHYSICS IN MEDICINE AND BIOLOGY
ISSN journal
00319155 → ACNP
Volume
45
Issue
8
Year of publication
2000
Pages
2057 - 2079
Database
ISI
SICI code
0031-9155(200008)45:8<2057:TFVDOY>2.0.ZU;2-#
Abstract
Elasticity imaging (EI) is being developed to allow the evaluation of the m echanical properties of soft tissue, but these properties are already asses sed in routine ultrasound breast examination using a method that involves t he subjective interpretation of tissue motion seen in real-time B-mode imag e movies during palpation. We refer to this method as relative motion asses sment (RMA). The purpose of this study was to begin a process of learning a bout the usefulness and limitations of RMA relative to the emerging method of elasticity imaging. Perception experiments were performed to measure You ng's modulus contrast thresholds for positive contrast lesions under contro lled conditions that could subsequently be repeated to evaluate elasticity imaging for the same task. Observer ability to grade relative lesion contra st using RMA was also assessed. Simulated sequences of B-scans of tissue moving in response to an applied f orce were generated and used in a two-alternative forced-choice (2-AFC) exp eriment to measure contrast thresholds for the detection of disc-shaped ela stic lesions by RMA in the absence of ultrasound echo contrast. Results wer e obtained for four observers at a lesion area of about 77 speckle cells an d for live observers at lesion areas of about 42 and 139 speckle cells. Young's modulus contrast thresholds were found to decrease with increasing lesion size and were well within the range of contrast values that have bee n measured for breast tumours in vitro. It was also found that observers we re quite skilled at using RMA to grade the relative strain contrast of lesi ons. The nonlinear relationship between the object contrast (Young's modulu s contrast) and the image contrast (strain contrast) prevented observers fr om detecting very small lesions with 100% accuracy, no matter how high the object contrast. A preliminary comparison of the results for RMA with publi shed thresholds for elastography indicated that elastography is likely to o ffer great benefit in reducing modulus contrast thresholds, but further stu dy is required to confirm this.