Morphometric Analysis of the Human Lower Lumbar Intervertebral Discs and Vertebral Endplates: Experimental Approach and Regression Models

Abstract

Low back pain (LBP) has been a major socioeconomic problem to the modern society for decades. In industry, one of the most challenging issues in occupational ergonomics and health practices has been the reliable and accurate estimation of risks of work-related musculoskeletal disorders (WMSDs), particularly work-related low back pain (WRLBP) and injuries which represent a large portion of all Workers' Compensation (WC) cost. To date, ergonomics evaluation measures developed to pinpoint jobs with elevated risks of WRLBP primarily rely on biomechanical models of the musculoskeletal structures of the human spine to estimate the internal response in terms of muscle induced compressive forces and to characterize the risk associated with the postures and forceful motions. However, morphometric characteristics of the human spine has not yet been thoroughly investigated and incorporated in the development of biomechanical models. In particular, the size of the load-bearing surface (cross-sectional area) of lumbar motion segments has been lacking in the literature, despite the fact that there is strong correlation between the cross-sectional area (CSA) and the ultimate compressive strength. Morphometric data regarding the human spine have been obtained with either direct measurements on cadaveric specimens or using medical imaging techniques, which require strict measurement protocol and incur high cost. In industry, occupational safety and health practitioners would prefer a more cost-effective means to obtain these morphometric data to improve the ergonomic evaluations and risk estimation of WRLBP.

The objective of this study was 1) to develop standardized protocol using magnetic resonance (MR) scans to measure the cross-sectional areas (CSAs) of the lower lumbar intervertebral discs and vertebral endplates, and 2) to develop regression models to predict these CSAs with different hierarchy of model complexity and predictor selection criteria. MR scans were 1) collected from a medical database and 2) performed in a research institute (Auburn University MRI Research Center). MR scans were analyzed using imaging processing software package with a research protocol developed in this dissertation. The protocol standardized the definitions of each geometric dimension and the measurement techniques and achieved excellent measurement reliability.

This dissertation provides comprehensive morphometric data regarding both the linear and planar aspects of the lower lumbar intervertebral discs (IVDs) and vertebral endplates (EPs), which has been lacking in the literature. Results of this dissertation also indicate that it is feasible to perform satisfactory predictions of the CSAs of the human lower lumbar IVDs and EPs using subject variables (characteristics and anthropometric measures). Results of this dissertation also suggest that the discrepancy in historical geometric data regarding the human lower lumbar may not only be attributed to gender alone but also related to other anthropometric measures. In addition, it is also evident that superior model performance can be achieved when certain anthropometric measures, such as the dimensions of ankle and elbow joints, are included as predictors in the prediction equations.