Towards Ultrasound-guided Spinal Fusion Surgery

Each year, around 1.3 million cases of spine surgery are performed in the United States alone, 53 % of which could be categorized as spinal fusion surgeries (McWilliams 2011; Solomon 2010), necessitating the insertion of screw implants into small bones in the spinal vertebrae, called pedicles (see Figs. 1.1 and 1.3). Pedicle screw placement is a complex 3-D navigation problem that involves risk to the patient. Of major concern is the fact that improperly inserted screws could place the surrounding vital structures at risk. This could lead to lawsuits for hospitals and the financial burden of revision surgeries for healthcare. Statistics show a 6.8 % readmission rate within 30 days of this procedure in USA hospitals (Weiss et al. 2013).

Many studies have demonstrated the need for accuracy in spinal fusion procedures as well as for improvements in education and training of residents. This chapter provides a review of current approaches for the training in orthopedic screw insertions for spinal fusion surgery together with a detailed review of the various approaches that can be used to provide the surgeon with navigational guidance for screw insertion as a means of improving the surgical outcome.

It is well known in the field that variations in the density and microstructure of trabecular bone can have considerable impact on the resulting ultrasound images (Chaffaí et al. 2002; Haïat et al. 2007; Wear 2008). For example, Wear (2000) reported average backscatter coefficient to be 50 % higher in the medial-lateral (ML) direction than in the anterior-posterior (AP) direction (human calcaneus, 500 kHz). Our interest in the trabecular alignment within bone stems from the need to better understand the manner in which it can affect ultrasound propagation, particularly in pedicles. This chapter describes our method for quantitatively determining the directions of the trabeculae within pedicle trabecular bone. It uses micro-CT images and is based on the special properties of the Gabor filter that have been exploited for this purpose. Extensive testing of the method was accomplished using 2D simulated structures and 3D printed bone models. The method has been applied to six human pedicle samples in two orthogonal planes, with results that provide good evidence that the method is well suited for estimating the directionality distribution within smaller bones. In order to gain further understanding of this method and its results, readers are encouraged to refer to Gdyczynski et al. (2014).

The aim of this chapter is to describe the design, fabrication, and characteristics of three different 2 MHz ultrasound transducers incorporated in a 3.6 mm diameter (11 Fr) probe designed for pedicle screw guidance using ultrasound imaging. In order to compare the imaging performance of these transducers, various acoustic parameters (sensitivity, bandwidth, and center frequency) were obtained experimentally and finally, B-mode image acquisition was done on two human pedicle samples (U of T Research Ethics Approval Protocol 29799). The work reported in this chapter represents the first stage in our ultimate goal of developing a 32-element radial-phased array capable of generating a cross-sectional image of the pedicle from within.

In the previous chapter, we demonstrated that the use of low-frequency ultrasound results in adequate penetration depth in bone for the detection of cortical bone boundaries in pedicle bones. Finding the optimal acoustic design and material that would be best suited for imaging within the bone was also studied. Although all the experiments were performed with a single element transducer, we believe that to make the image guidance technique more acceptable to spine surgeons, it is necessary to use a method that avoids the need for mechanical transducer rotation. To achieve this, we have designed and fabricated a 32-element cylindrical imaging array that is intended to be used with phased focusing at a center frequency of around 2 MHz using an Ultrasonix^™ imaging system. This platform is an open-source ultrasound system that allows a variety of ultrasound transducers to be connected and controlled by the user.

This chapter concerns the experimental evaluations of our custom-built array when connected to and controlled by a SonixTablet system (Ultrasonix™ Analogic Ultrasound, Richmond, BC). The purpose of these evaluations was to demonstrate functionality under relatively ideal circumstances, representing a first step toward bone imaging with a low-frequency radial array. Following the fabrication of the radial array prototype, software was then developed to interface the prototype with an Ultrasonix™ system. This platform is one of the few that allows access to the data acquisition toolbox. Its research platform allows the user to store, modify, and conduct signal processing on the raw radio frequency (RF) data obtained by each element of the array.

The work described in this book is based on the results of prior studies in our laboratory. Such studies suggested an optimal range of ultrasound frequency (1–3 MHz) for bone sonography, for the purpose of providing a guidance system for pedicle screw implant insertion during spinal fusion surgery.