Related Experiment Video
Updated: Apr 17, 2026
13:35
Segmentation and Linear Measurement for Body Composition Analysis using Slice-O-Matic and Horos
Published on: March 21, 2021
12.2K
Variability of manual lumbar spine segmentation
Daniel J Cook1, David A Gladowski1, Heather N Acuff1
1Department of Neurosurgery, Allegheny General Hospital, Pittsburgh, PA.
International Journal of Spine Surgery
|February 20, 2015
Summary
Manual segmentation of 3D spine models is highly accurate and precise. This study quantifies segmentation error, providing crucial data for future biomechanical research using these detailed spinal models.
Area of Science:
- Biomechanics
- Spinal Research
- Medical Imaging
Background:
- Three-dimensional (3D) models of the spine are vital for biomechanical research, utilizing kinematic data from testing.
- Model development faces segmentation errors due to spinal complexity and pathology.
- Previous characterization of this segmentation error was lacking.
Purpose of the Study:
- To characterize segmentation error in specimen-specific 3D lumbar spine models.
- To evaluate the accuracy and precision of manual segmentation techniques.
Main Methods:
- Eight cadaveric lumbar spines underwent computed tomography (CT) scanning.
- Individual vertebrae were scanned post-disarticulation and soft-tissue removal.
- Multiple operators performed manual segmentation on full specimens and individual vertebrae, with subsequent analysis of model registration and point-distance distributions.
Main Results:
- Manual segmentation demonstrated high accuracy with root-mean-square errors below 0.39 mm.
- Intrauser and interuser precision were excellent, with errors below 0.33 mm and 0.35 mm, respectively.
- The 95th percentile of distances remained below 0.75 mm across all accuracy and precision analyses.
Conclusions:
- Specimen-specific 3D lumbar spine models derived from manual segmentation are highly accurate.
- A high degree of intrauser and interuser precision is achievable with these methods.
- Quantified error magnitudes are essential for designing and interpreting future studies employing manual segmentation for lumbar spine modeling.
