TY - JOUR
T1 - Validation of three-dimensional models of in situ scapulae
AU - Bryce, Chris D.
AU - Pennypacker, Jason L.
AU - Kulkarni, Nikhil
AU - Paul, Emmanuel M.
AU - Hollenbeak, Christopher S.
AU - Mosher, Timothy
AU - Armstrong, April
N1 - Copyright:
Copyright 2009 Elsevier B.V., All rights reserved.
PY - 2008/9
Y1 - 2008/9
N2 - A principal challenge in creating accurate models of in situ scapulae is delineating bone from surrounding soft tissues. Computed tomography scans were obtained of both shoulders of 20 embalmed cadavers. Each shoulder was rescanned after repositioning of the cadavers to test for rescan reliability. After scans were complete, all scapulae were excised and stripped of all soft tissue. Thresholding, region growing, and manual processing were used to create computer-generated 3-dimensional (3D) models. Seven anatomic measurements were performed on each scapula and 3D model. Mean differences between corresponding measurements of specimen and model were small (<3 mm). Intraobserver and interobserver reliability for cadaveric measurements and rescan and interobserver reliability for model measurements were all excellent (R2 = 0.99). Patient positioning was not a significant source of error in obtaining measurements from 3D models. Results from this work verify that accurate and reproducible 3D models can be created from in situ scapulae by use of effective segmentation.
AB - A principal challenge in creating accurate models of in situ scapulae is delineating bone from surrounding soft tissues. Computed tomography scans were obtained of both shoulders of 20 embalmed cadavers. Each shoulder was rescanned after repositioning of the cadavers to test for rescan reliability. After scans were complete, all scapulae were excised and stripped of all soft tissue. Thresholding, region growing, and manual processing were used to create computer-generated 3-dimensional (3D) models. Seven anatomic measurements were performed on each scapula and 3D model. Mean differences between corresponding measurements of specimen and model were small (<3 mm). Intraobserver and interobserver reliability for cadaveric measurements and rescan and interobserver reliability for model measurements were all excellent (R2 = 0.99). Patient positioning was not a significant source of error in obtaining measurements from 3D models. Results from this work verify that accurate and reproducible 3D models can be created from in situ scapulae by use of effective segmentation.
UR - http://www.scopus.com/inward/record.url?scp=51249106880&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=51249106880&partnerID=8YFLogxK
U2 - 10.1016/j.jse.2008.01.141
DO - 10.1016/j.jse.2008.01.141
M3 - Article
C2 - 18490182
AN - SCOPUS:51249106880
SN - 1058-2746
VL - 17
SP - 825
EP - 832
JO - Journal of Shoulder and Elbow Surgery
JF - Journal of Shoulder and Elbow Surgery
IS - 5
ER -