Methodology based on genetic heuristics for in-vivo characterizing the patient-specific biomechanical behavior of the breast tissues. Issue 21 (30th November 2015)

• Record Type: Journal Article
• Authors: Lago, M.A.; Rupérez, M.J.; Martínez-Martínez, F.; Martínez-Sanchis, S.; Bakic, P.R.; Monserrat, C.
• Is Part Of: Expert systems with applications. Volume 42:Issue 21(2015)
• Page Start: 7942
• Page End: 7950
• Publication Date: 2015-11-30
• Subjects: Genetic heuristics -- In-vivo tissue characterization -- Breast biomechanical modeling -- Parameter estimation; Expert systems (Computer science) -- Periodicals; Systèmes experts (Informatique) -- Périodiques; Electronic journals; 006.33
• DOI: 10.1016/j.eswa.2015.05.058
• Languages: English
• ISSNs: 0957-4174

Abstract

Highlights: Genetic heuristics were used to in-vivo model the breast biomechanical behavior. An iterative search was used to optimize the elastic constants of the model. Phantom images of the breast were used to compare the internal tissue distribution. The accuracy of the estimated model was achieved with an error of less than 10%. Abstract: This paper presents a novel methodology to in-vivo estimate the elastic constants of a constitutive model proposed to characterize the mechanical behavior of the breast tissues. An iterative search algorithm based on genetic heuristics was constructed to in-vivo estimate these parameters using only medical images, thus avoiding invasive measurements of the mechanical response of the breast tissues. For the first time, a combination of overlap and distance coefficients were used for the evaluation of the similarity between a deformed MRI of the breast and a simulation of that deformation. The methodology was validated using breast software phantoms for virtual clinical trials, compressed to mimic MRI-guided biopsies. The biomechanical model chosen to characterize the breast tissues was an anisotropic neo-Hookean hyperelastic model. Results from this analysis showed that the algorithm is able to find the elastic constants of the constitutive equations of the proposed model with a mean relative error of about 10%. Furthermore, the overlap between the reference deformation and the simulated deformation was of around 95% showing the good performance of the proposed methodology. This methodology can be easily extended to characterize the real biomechanical behavior of the breast tissues, which means a great novelty in the field of the simulation of the breast behavior for applications such as surgical planing, surgical guidance or cancer diagnosis. This reveals the impact and relevance of the presented work.