Skip to main content
2024 BMES Annual Meeting Main banner
Icon Legend
This session is not in your schedule.
This session is in your schedule. Click again to remove it.

Biomedical Imaging and Instrumentation

Biomedical Imaging and Instrumentation - Poster Session E

Poster R8 - HURST/VARIABLE ORIENTATION TRANSFORM ANALYSIS OF ANTERIOR CERVICAL DISCECTOMY AND FUSION SURGICAL SITE

Saturday, October 26, 2024
10:00 AM - 11:00 AM EST
Location: Exhibit Hall E, F & G

Introduction:  Anterior cervical discectomy and fusion (ACDF) is a surgical procedure for degenerative disc disease in which problematic discs are removed and replaced with an intervertebral spacer filled with fusion inducing graft material. The leading cause of ACDF revisions (~60% of cases) is pseudarthrosis or vertebral non-union1. This occurs when the two adjacent vertebral bodies do not fuse properly, forming a false joint. Medical imaging (x-ray, CT, and SPEC-CT) and patient reported pain are used to diagnose delayed healing; however, diagnosis most often relies on subjective assessment of radiolucency in x-rays. Further, patients often present as asymptomatic. Even the definition of a failed union, typically determined after 1-year post-surgery, can vary across health systems. The goal of this study was to develop a quantitative method that increases the diagnostic capability of x-rays, improving our ability to detect incomplete spinal fusion following ACDF with high accuracy.

Materials and

Methods:  Longitudinal x-rays were prospectively collected from patients (n=2) undergoing ACDF surgery at a single institution enrolled in an IRB approved study. Patients with a full panel of sagittal plane x-rays at days 1, 42 and 90 post-op were utilized to quantify differences in surgical site healing. The degree of union at the individual surgical site was assessed using two methods: the Hurst Orientation Transform (HOT) which reports the greatest absolute difference in pixel intensity along every given angle and Variation Orientation Transform (VOT) which reports the variance of the pixel intensity along any given angle2-5 . High HOT/VOT values are indicative of large disparities in pixel intensity in a given direction, reflecting the presence of both highly mineralized and poorly mineralized tissue. As such, high HOT/VOT values indicate incomplete union or tissue heterogeneity in the search path. In contrast, the presence of low HOT/VOT values indicates that tissue along the search path is more uniformly mineralized, indicative of boney fusion. The analysis was conducted by evaluating grayscale images with a pixel intensity range of 0 and 255 . The search region for both transforms was denoted as a circular patch with a radius of 17 pixels over a search length of 49 pixels to avoid interference of the implant (Figure 1). The values were mapped to their respective angles on a polar plot. To assess changes in the surgical site environment over time the average and the absolute maximum value of both the HOT and VOT data was calculated.

Results, Conclusions, and Discussions: Analysis of patient spines returned an hourglass pattern, with the primary axis of variation lying between 45°-135° (Figure 2, Table 1). The hourglass shape of the HOT/VOT polar plot is representative of bone location in relation to the origin within the intervertebral space. For example, values between ±45° and 135° have the highest reported values from sampling the vertebrae, while values between 0 and ±45 degrees have less variation from sampling mostly soft tissue and disc space. A notable drop in both the average and absolute maximum recorded value of HOT absolute difference and VOT variance across time points was observed. The drop in these reported values decreases from baseline until 42 days post-op, after which an increase was observed 90 days post-op (Figure 3).
Using the HOT/VOT methodologies, the degree of anisotropy at the surgical site is exaggerated, and thus can be more readily assessed for interpreting spinal fusion from x-rays. Bone resorption has been shown to occur between 2 to 3 months (60-90 days) post-surgery, and the trends in the HOT/VOT plots follow this pattern6 7. After this period of bone resorption, an increase of reported magnitudes is observed, which can potentially be attributed to the bridging of trabecular bone intruding into the search region and thus increasing the absolute difference between intervertebral space and the vertebral bodies. This is because soft tissue has a lower attenuation than mature trabecular bone. Thus, changes in the vertebral structures can be tracked using the bounding angles of the polar plots. Using the HOT/VOT method and its derivations can potentially provide a new approach for assessing surgical site integrity and tracking the degree of fusion in spinal procedures.
All evaluated surgical sites presented with a consistent hourglass shape conserved across all patients and time points, indicative of the original bone on either side of the surgical site and confirming consistent spatial assessment. The variation in the magnitude of the HOT/VOT plots enabled the objective interpretation of healing at the surgical site, with increasing pixel magnitudes indicating progression of healing. Further investigation with more clinical cases is required to validate these findings.

Acknowledgements (Optional):