TY - JOUR
T1 - Palpation tomography - A new technique for modulus estimation in elastography
AU - Konofagou, E. E.
AU - Harrigan, T. P.
N1 - Funding Information:
This study was partially supported by the International Collaboration for Capitalizing on Cost-Effective and Life-Saving Commodities (i4C) that is funded through the Research Council of Norway's Global Health & Vaccination Programme (GLOBVAC Project #234608) and contract funding from the Public Health Agency of Canada. Neither funder had any role in study design, data collection, data analysis, data interpretation or writing of the report. S.J.H. is additionally funded by the Canadian Institutes of Health Research and the Ontario Government's Ministry of Research, Innovation and Science. S.R.V.K. is supported by an Ontario Graduate Scholarship. J.M.G. holds a Canada Research Chair in Health Knowledge Transfer and Uptake.
PY - 2003
Y1 - 2003
N2 - In Elastography, strain estimation has been shown to be far more reliable compared to the elastic properties obtained using reconstruction techniques. In this study, to make the method less sensitive to noise in the experimental data and inspired by the clinical practice of palpation (i.e., the use of sequential finger loading), we investigated the effect of using several different smaller quasi-static load cases (instead of a one-time load on the whole boundary), with the error indicator taken as the sum of the errors from each load case. This increased the ratio of measurements to the fitted parameters, which made the method less sensitive to random errors. To demonstrate this effect, we calculated displacements from a two-dimensional, quadrilateral, plane-strain, finite-element model of a 40-by-40 mm region containing a cylindrical inclusion (7 mm in diameter) three-times stiffer than the background. The ratio of nodal pressures was chosen to produce approximately 0.75% strain. Known amounts of random displacement errors were then added at a signal-to-noise ratio varying from 60 dB to 20 dB. Elastic modulus reconstructions using the noisy displacement results from a single, total-boundary, pressure load (as is typically applied in elastography) were compared to reconstructions using data from nine smaller-width loading cases, and the reconstructed modulus distributions were compared to the original model parameters. It was found that in the cases of 60 dB and 40 dB the multiple loading cases resulted in noise reduction in the modulus reconstruction by at least a two-fold compared to the single-loading case, at the expense of a 'shadowing' effect (i.e., erroneous modulus estimates) underneath the inclusion that could be eliminated by using larger loading areas for the individual loading cases. Finally, at 20 dB both the large single-load and combined, smaller five-load cases failed to accurately reconstruct the modulus of the inclusion; depicting thus a fundamental limit on the reconstruction method.
AB - In Elastography, strain estimation has been shown to be far more reliable compared to the elastic properties obtained using reconstruction techniques. In this study, to make the method less sensitive to noise in the experimental data and inspired by the clinical practice of palpation (i.e., the use of sequential finger loading), we investigated the effect of using several different smaller quasi-static load cases (instead of a one-time load on the whole boundary), with the error indicator taken as the sum of the errors from each load case. This increased the ratio of measurements to the fitted parameters, which made the method less sensitive to random errors. To demonstrate this effect, we calculated displacements from a two-dimensional, quadrilateral, plane-strain, finite-element model of a 40-by-40 mm region containing a cylindrical inclusion (7 mm in diameter) three-times stiffer than the background. The ratio of nodal pressures was chosen to produce approximately 0.75% strain. Known amounts of random displacement errors were then added at a signal-to-noise ratio varying from 60 dB to 20 dB. Elastic modulus reconstructions using the noisy displacement results from a single, total-boundary, pressure load (as is typically applied in elastography) were compared to reconstructions using data from nine smaller-width loading cases, and the reconstructed modulus distributions were compared to the original model parameters. It was found that in the cases of 60 dB and 40 dB the multiple loading cases resulted in noise reduction in the modulus reconstruction by at least a two-fold compared to the single-loading case, at the expense of a 'shadowing' effect (i.e., erroneous modulus estimates) underneath the inclusion that could be eliminated by using larger loading areas for the individual loading cases. Finally, at 20 dB both the large single-load and combined, smaller five-load cases failed to accurately reconstruct the modulus of the inclusion; depicting thus a fundamental limit on the reconstruction method.
UR - http://www.scopus.com/inward/record.url?scp=4143099375&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=4143099375&partnerID=8YFLogxK
M3 - Conference article
AN - SCOPUS:4143099375
SN - 1051-0117
VL - 1
SP - 652
EP - 655
JO - Proceedings of the IEEE Ultrasonics Symposium
JF - Proceedings of the IEEE Ultrasonics Symposium
T2 - 2003 IEEE Ultrasonics Symposium - Proceedings
Y2 - 5 October 2003 through 8 October 2003
ER -