Geomagic wrap 2021

IVD deformations were assessed via segmentation of MR images as described previously^{16 (link)}. Briefly, a single investigator manually segmented the outer surfaces of the L1-L2 – L5-S1 IVDs in both the pre- and post-exercise anatomic scans. A musculoskeletal radiologist with more than 30 years of experience trained the investigator and reviewed all segmentations. During segmentation, the investigator was blinded to subject descriptive characteristics and demographics, including sex, BMI, NP T1rho relaxation times and degeneration status (Pfirrmann grade). These segmentations were then used to generate 3D surface models of the IVDs using solid modeling software (Rhinoceros 3D 4.0, Robert McNeel and Associates, Seattle, WA) and Geomagic Wrap 2021 (3D Systems, Cary, NC)^{16 (link)} (Figure 1 A–D). IVD height was calculated by isolating the superior and inferior surfaces of the IVD and subsequently estimated by calculating the mean distance between the two surfaces. After evaluating both pre- and post-exercise models, IVD deformation (%) was calculated and defined as the percent change in IVD height from pre- to post-exercise scans^{16 (link)}. Using these methods, IVD height calculations have been shown to be repeatable to within 1% across repeated segmentations^{16 (link)}. Additionally, prior work in our lab found comparable repeatability in disc heights between different days^{17 (link)}.