Samantha St Pierre1, Jenifer Siegelman2, Nancy A Obuchowski3, Xiaonan Ma1, David Paik1, Andrew J Buckler4. 1. Elucid Bioimaging Inc., 225 Main Street, Wenham, MA 01984. 2. Brigham and Women's Hospital, Boston, Massachusetts. 3. Quantitative Health Sciences, Cleveland Clinic Foundation, Cleveland, Ohio. 4. Elucid Bioimaging Inc., 225 Main Street, Wenham, MA 01984. Electronic address: andrew.buckler@elucidbio.com.
Abstract
RATIONALE AND OBJECTIVES: The purpose of this study was to characterize analytic performance of software-aided arterial vessel structure measurements across a range of scanner settings for computed tomography angiography where ground truth is known. We characterized performance for measurands that may be efficiently measured for clinical cases without use of software, as well as those that may be done manually but which is generally not done due to the effort level required unless software is employed. MATERIALS AND METHODS: Four measurands (lumen area, stenosis, wall area, wall thickness) were evaluated using tissue-mimicking phantoms to estimate bias, heteroscedasticity, and limits of quantitation both pooled across scanner settings and individually for eight different settings. Reproducibility across scanner settings was also estimated. RESULTS: Measurements of lumen area have a near constant bias of +1.3 mm for measurements ranging from 3 mm2 to 40 mm2; stenosis bias is +7% across a 30%-70% range; wall area bias is +14% across a 50-450 mm2 range; and wall thickness bias is +1.2 mm across a 3-9 mm range. All measurements possess properties that make them suitable for measuring longitudinal change. Lumen area demonstrates the most sensitivity to scanner settings (bias from as low as +.1 mm to as high as +2.7 mm); wall thickness demonstrates negligible sensitivity. CONCLUSIONS: Variability across scanner settings for lumen measurands was generally higher than bias for a given setting. The converse was true for the wall measurands, where variability due to scanner settings was very low. Both bias and variability due to scanner settings of vessel structure were within clinically useful levels.
RATIONALE AND OBJECTIVES: The purpose of this study was to characterize analytic performance of software-aided arterial vessel structure measurements across a range of scanner settings for computed tomography angiography where ground truth is known. We characterized performance for measurands that may be efficiently measured for clinical cases without use of software, as well as those that may be done manually but which is generally not done due to the effort level required unless software is employed. MATERIALS AND METHODS: Four measurands (lumen area, stenosis, wall area, wall thickness) were evaluated using tissue-mimicking phantoms to estimate bias, heteroscedasticity, and limits of quantitation both pooled across scanner settings and individually for eight different settings. Reproducibility across scanner settings was also estimated. RESULTS: Measurements of lumen area have a near constant bias of +1.3 mm for measurements ranging from 3 mm2 to 40 mm2; stenosis bias is +7% across a 30%-70% range; wall area bias is +14% across a 50-450 mm2 range; and wall thickness bias is +1.2 mm across a 3-9 mm range. All measurements possess properties that make them suitable for measuring longitudinal change. Lumen area demonstrates the most sensitivity to scanner settings (bias from as low as +.1 mm to as high as +2.7 mm); wall thickness demonstrates negligible sensitivity. CONCLUSIONS: Variability across scanner settings for lumen measurands was generally higher than bias for a given setting. The converse was true for the wall measurands, where variability due to scanner settings was very low. Both bias and variability due to scanner settings of vessel structure were within clinically useful levels.
Authors: Tobias Saam; William S Kerwin; Baocheng Chu; Jianming Cai; Annette Kampschulte; Thomas S Hatsukami; Xue-Qiao Zhao; Nayak L Polissar; Blazej Neradilek; Vasily L Yarnykh; Kelly Flemming; John Huston; William Insull; Joel D Morrisett; Scott D Rand; Kevin J DeMarco; Chun Yuan Journal: J Cardiovasc Magn Reson Date: 2005 Impact factor: 5.364
Authors: Stefan B Puchner; Ting Liu; Thomas Mayrhofer; Quynh A Truong; Hang Lee; Jerome L Fleg; John T Nagurney; James E Udelson; Udo Hoffmann; Maros Ferencik Journal: J Am Coll Cardiol Date: 2014-08-19 Impact factor: 24.094
Authors: Anna E H Zavodni; Bruce A Wasserman; Robyn L McClelland; Antoinette S Gomes; Aaron R Folsom; Joseph F Polak; João A C Lima; David A Bluemke Journal: Radiology Date: 2014-03-04 Impact factor: 11.105
Authors: Larry G Kessler; Huiman X Barnhart; Andrew J Buckler; Kingshuk Roy Choudhury; Marina V Kondratovich; Alicia Toledano; Alexander R Guimaraes; Ross Filice; Zheng Zhang; Daniel C Sullivan Journal: Stat Methods Med Res Date: 2014-06-11 Impact factor: 3.021
Authors: David L Raunig; Lisa M McShane; Gene Pennello; Constantine Gatsonis; Paul L Carson; James T Voyvodic; Richard L Wahl; Brenda F Kurland; Adam J Schwarz; Mithat Gönen; Gudrun Zahlmann; Marina V Kondratovich; Kevin O'Donnell; Nicholas Petrick; Patricia E Cole; Brian Garra; Daniel C Sullivan Journal: Stat Methods Med Res Date: 2014-06-11 Impact factor: 3.021
Authors: Nancy A Obuchowski; Anthony P Reeves; Erich P Huang; Xiao-Feng Wang; Andrew J Buckler; Hyun J Grace Kim; Huiman X Barnhart; Edward F Jackson; Maryellen L Giger; Gene Pennello; Alicia Y Toledano; Jayashree Kalpathy-Cramer; Tatiyana V Apanasovich; Paul E Kinahan; Kyle J Myers; Dmitry B Goldgof; Daniel P Barboriak; Robert J Gillies; Lawrence H Schwartz; Daniel C Sullivan Journal: Stat Methods Med Res Date: 2014-06-11 Impact factor: 3.021
Authors: Matthew T Chrencik; Amir A Khan; Lauren Luther; Laila Anthony; John Yokemick; Jigar Patel; John D Sorkin; Siddhartha Sikdar; Brajesh K Lal Journal: J Vasc Surg Date: 2019-03-06 Impact factor: 4.268