Session: 06-14-01: Biotechnology and General Applications

Paper Number: 110141

110141 - Statistical Shape Modelling of the Lumbar Spine With Reference to Gender and Principal Component Analysis 

Biomechanical evaluation of the spine relies on the anatomical features of the vertebrae. Understanding the morphology of vertebrae helps in planning of treatments for chronic conditions such as lower back pain (LBP). Also, assessment of lumbar vertebrae is important for the success of  many medical procedures related to lower back pain, including but not limited to radiofrequency ablation (RFA) and pedicle screw insertion. The anatomy of the spine is not uniform in the normal population and lumbar vertebrae have variations in their shape between individuals and between each gender. The anatomy of lumbar vertebrae along with gender differences is studied in this research using statistical shape modelling (SSM). This modelling method utilizes three-dimensional representation of the of the lumbar spine. Shapeworks software was used to analyze these models. The study used publicly available CT images and lumbar three-dimensional segmentations from Vertebrae Segmentation challenge (VerSe) with a total of 46 patients were selected from the VerSe CT images. The following vertebrae were chosen for analysis: L3, L4, and L5 vertebrae. Using Slicer software, the 3D segmentations from the VerSe samples were exported as separate objects for statistical analysis. The method that Shapeworks software implemented  to create statistical shape models was population-specific anatomical mapping. This method doesn’t use depend on landmarks or templates for statistical assessment. The sample included equal size of patients from each gender (23 individuals per group) to increase accuracy of comparison. Particle-based shape modelling was used to create more than one statistical shape model: the overall mean, the females mean, and the males mean statistical shape models. Moreover, the differences between females and males were illustrated using difference color maps to better delineate the areas with variation between genders. Also, areas with significant differences, according to p-values, between females and males were shown on another SSM model. It was found that the differences between females and males were mostly in the posterior element of the vertebrae. The location of shape differences was not consistent among the lumbar vertebrae. The study also examined lumbar vertebrae using principal component analysis (PCA) to investigate the modes of variation which helps in distinguishing how the lumbar vertebral shape varies among the individuals. The principal component analysis showed that the overall mean for the lumbar vertebrae differ in the scale along with other discrete areas of shape variation such as posterior elements. For example, within 2 standard deviations, the first PC (34.4 % explained variance) had most of the variations in the posterior elements and the posterior area of vertebral body in L1. For L2, the variation is in the posterior part of the vertebral body. Meanwhile for L3, the variations are in the posterior elements and the anterior part of the vertebral body. For the second PC (16.9 % explained variance), the changes are mostly in L5 covering most of the overall surface except the spinous process. The other vertebrae, L3 and L4, didn’t have as much variations as L5.  Therefore, the study shows the existence of sexual dimorphism in lumbar vertebrae of the spine along with the variations in PCA for the lumbar vertebra. This finding enables better planning of medical procedures using statistical shape modelling depending on different shapes for lumbar vertebrae and depending on the gender of the patient which are essential for the success of the treatment modalities.

Presenting Author: Faris Almalki The Pennsylvania State University

Presenting Author Biography: Faris Almalki is a PhD candidate in Mechanical Engineering at The Pennsylvania State University. He holds Bachelor’s and Master’s degrees in Mechanical Engineering from King Abdulaziz University and Penn State, respectively. Faris is the recipient of Saudi Arabian Cultural Mission (SACM) scholarship from the Saudi government in order to pursue his Ph.D. study in the United States. Faris will be a faculty member in University of Jeddah in Saudi Arabia once he finishes his Ph.D. degree. His PhD research focuses on the effect of radiofrequency ablation on multifidus muscle and lumbar spine. Faris has three years of work experience in steelmaking industry before joining academia. He also has one year of experience in teaching mechanical engineering drawings. He presented at SB3C 2022 about alteration of the biomechanics of the lumbar spine after multifidus denervation due to radiofrequency ablation. Also, he has two publications under review. One journal paper addresses how joint angles and forces are affected by altering the biomechanics of spine caused by radiofrequency ablation. The other journal paper is about machine learning modelling in additive manufacturing with respect to Ti-6Al-4V.

Authors:

Faris A. Almalki Penn State University
Daniel H. Cortes Penn State University