Open Collections

Image-based enhancement and tracking of an epidural needle in ultrasound using time-series analysis

University of British Columbia

Accurate placement of the needle to the target spot is crucial to the safety, efficiency and success of any ultrasound (US) guided procedures, e.g. epidurals. Real-time single-operator US guidance of epidurals is currently impractical with conventional 2D US and a standard needle. A novel 3D US imaging technology (3DUS+Epiguide) has been developed by our group to provide such capability, which comprises thick-slice rendering and a custom needle guide, Epiguide. This system aims to facilitate a single-operator real-time midline epidural needle insertion by visualizing the needle progression toward the epidural space. The visibility of the needle in US-guided procedures is, however, limited by dispersion of the needle's echoes away from the transducer. This makes needle visibility an ongoing challenge in US-guided procedures, e.g. epidurals. The aim of this thesis is to provide a software-based clinically-suitable solution to enhance needle visibility in US. In particular, we are interested in difficult cases, where the needle is invisible (i.e. minimal visibility) in an US image. To this end, we have developed frameworks to extract signatures from the needle using time-series analysis. We demonstrate that nearly invisible changes in motion dynamics of the needle can be revealed through spatio-temporal processing of the standard US image sequences. The extracted needle features are used to detect, track and localize a hand-held needle in a machine learning-based framework. Clinical, animal and phantom studies are designed to evaluate: 3DUS+Epiguide's capability to identify the needle puncture site for a midline insertion in the lumbar spine, and the capability of the proposed detection frameworks in localizing a needle within clinical acceptance. Methods are evaluated on the data from studies conducted at BC Women's Hospital and Jack Bell Research Facility, and are compared to the gold standard (GS). Results demonstrate that compared to the state-of-the-art needle detection methods using a finely-tuned Hough Transform, with 14% success rate, our proposed method is 100% successful in detecting the trajectory within the GS's close proximity (success: angular error < 10°). The proposed method is especially suitable for barely-visible needles, for which appearance-based enhancement methods fail.

Text

2017-08-31

Attribution-NonCommercial-NoDerivatives 4.0 International

http://hdl.handle.net/2429/62922

Electrical and Computer Engineering

University of British Columbia

UBCV

http://creativecommons.org/licenses/by-nc-nd/4.0/