Non-contact quantitative imaging of limbs, bone and tissue, using ultrasound tomographic techniques
Author(s)Fincke, Jonathan Randall
Massachusetts Institute of Technology. Department of Mechanical Engineering.
Brian W. Anthony.
MetadataShow full item record
Non-contact and quantitative ultrasound images of bone and soft tissue are produced from original algorithms applied to observational data sets collected on two custom built ultrasound imaging systems for limb imaging. The images are quantitative in that the distribution of sound speeds and dimensionally accurate geometry of tissue structures are reconstructed. The first imaging system is based on laser generation and detection of ultrasound (LUS) and the second is a water immersion ultrasound tomography (UST) system. The LUS and UST systems and algorithms, in aggregate, can be used to generate large volume, quantitative 2D and traditional 2D ultrasound images. Existing medical ultrasound systems are unable to acquire or generate large volume, quantitative and clinically useful bone images. A medical ultrasound system with these capabilities would have significant clinical value. LUS and UST systems could improve the quality, cost and safety of osteoporosis diagnosis and tracking, prosthetic fitting, bone fracture detection and tracking, intraoperative imaging and volumetric imaging in intensive care units. The algorithms and systems established for this thesis contribute broadly to non-contact ultrasound imaging: LUS for medical imaging, and UST for bone and soft tissue imaging and quantification. Non-contact techniques are clinically valuable because they can deliver operator independent image quality and large volume imagery without making contact with the body and distorting the tissue. Quantitative ultrasound imaging techniques are clinically useful because they provide intrinsic information about tissue mechanical properties, such as stiffness and density. Successful bone and soft tissue quantification using UST techniques could yield an entirely new and radiation free means of assessing bone and soft tissue strength and health. The experiments completed with the LUS system demonstrate its capability to generate images without contacting or treating the skin surface. Further, soft tissue (weak reflector), as well as bone (strong reflectors) are resolved at skin safe optical exposures. To enhance the LUS system performance, the optical wavelength of the generation laser is studied and optimized to deliver the largest acoustic source possible while also meeting optical exposure thresholds for skin. Additionally, commercially available laser vibrometer technology optimized for detecting vibrations on rough surfaces, such as skin, is identified and tested. Three original algorithms yield images of bone and soft tissue geometry and sound speed when applied to experimental data from the UST device. The first algorithm is a backscatter/reflection adaptive imaging technique that enables high resolution, SNR and volumetric imagery of bone and soft tissue to be formed from a single, mechanically scanned ultrasound transducer. The second algorithm uses a travel-time sound speed inversion technique that estimates water, soft tissue and bone sound speed to within 10% of ground truth estimates. The performance of this algorithm is validated on multiple samples and a simulated data set. The third algorithm is a full waveform inversion (FWI) algorithm regularized with the level set technique to enable quantification of bone properties. This algorithm is validated on an animal tissue sample and simulated data sets. The FWI technique resolves the soft tissue spatial sound speed distribution with half to one-quarter wavelength (1 - 0.5 mm) resolution and average sound speed values are within 10% of ground truth measurements.
Thesis: Ph. D., Massachusetts Institute of Technology, Department of Mechanical Engineering, 2018.Cataloged from PDF version of thesis.Includes bibliographical references (pages 123-131).
DepartmentMassachusetts Institute of Technology. Department of Mechanical Engineering.
Massachusetts Institute of Technology