Automated palpation for breast tissue discrimination based on viscoelastic biomechanical properties

• Published in: International journal for computer assisted radiology and surgery Vol. 10; no. 5; pp. 593 - 601
• Main Authors: Tsukune, Mariko, Kobayashi, Yo, Miyashita, Tomoyuki, Fujie, G. Masakatsu
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.05.2015
• Subjects: Animals; Biomechanical Phenomena; Breast - pathology; Breast Neoplasms - diagnosis; Breast Neoplasms - pathology; Computer Imaging; Computer Science; Elasticity; Female; Health Informatics; Humans; Imaging; Medicine; Medicine & Public Health; Models, Theoretical; Nonlinear Dynamics; Original Article; Palpation - methods; Pattern Recognition and Graphics; Radiology; Robotics; Surgery; Swine; Vision; Breast tumor diagnosis; Palpation; Dynamic viscoelastic test; Creep test; Nonlinear viscoelastic parameter
• ISSN: 1861-6410; 1861-6429
• DOI: 10.1007/s11548-014-1100-2

Purpose Accurate, noninvasive methods are sought for breast tumor detection and diagnosis. In particular, a need for noninvasive techniques that measure both the nonlinear elastic and viscoelastic properties of breast tissue has been identified. For diagnostic purposes, it is important to select a nonlinear viscoelastic model with a small number of parameters that highly correlate with histological structure. However, the combination of conventional viscoelastic models with nonlinear elastic models requires a large number of parameters. A nonlinear viscoelastic model of breast tissue based on a simple equation with few parameters was developed and tested. Methods The nonlinear viscoelastic properties of soft tissues in porcine breast were measured experimentally using fresh ex vivo samples. Robotic palpation was used for measurements employed in a finite element model. These measurements were used to calculate nonlinear viscoelastic parameters for fat, fibroglandular breast parenchyma and muscle. The ability of these parameters to distinguish the tissue types was evaluated in a two-step statistical analysis that included Holm’s pairwise t test. The discrimination error rate of a set of parameters was evaluated by the Mahalanobis distance. Results Ex vivo testing in porcine breast revealed significant differences in the nonlinear viscoelastic parameters among combinations of three tissue types. The discrimination error rate was low among all tested combinations of three tissue types. Conclusion Although tissue discrimination was not achieved using only a single nonlinear viscoelastic parameter, a set of four nonlinear viscoelastic parameters were able to reliably and accurately discriminate fat, breast fibroglandular tissue and muscle.