Automatic Localization of Target Vertebrae in Spine Surgery: Clinical Evaluation of the LevelCheck Registration Algorithm

Sheng Fu L Lo, Yoshito Otake, Varun Puvanesarajah, Adam S. Wang, Ali Uneri, Tharindu De Silva, Sebastian Vogt, Gerhard Kleinszig, Benjamin D. Elder, C. Rory Goodwin, Thomas A. Kosztowski, Jason A. Liauw, Mari Groves, Ali Bydon, Daniel M. Sciubba, Timothy F. Witham, Jean Paul Wolinsky, Nafi Aygun, Ziya L. Gokaslan, Jeffrey H. Siewerdsen

Research output: Contribution to journalArticle

Abstract

Study Design. A 3-dimensional-2-dimensional (3D-2D) image registration algorithm, "LevelCheck," was used to automatically label vertebrae in intraoperative mobile radiographs obtained during spine surgery. Accuracy, computation time, and potential failure modes were evaluated in a retrospective study of 20 patients. Objective. To measure the performance of the LevelCheck algorithm using clinical images acquired during spine surgery. Summary of Background Data. In spine surgery, the potential for wrong level surgery is significant due to the difficulty of localizing target vertebrae based solely on visual impression, palpation, and fluoroscopy. To remedy this difficulty and reduce the risk of wrong-level surgery, our team introduced a program (dubbed LevelCheck) to automatically localize target vertebrae in mobile radiographs using robust 3D-2D image registration to preoperative computed tomographic (CT) scan. Methods. Twenty consecutive patients undergoing thoracolumbar spine surgery, for whom both a preoperative CT scan and an intraoperative mobile radiograph were available, were retrospectively analyzed. A board-certified neuroradiologist determined the "true" vertebra levels in each radiograph. Registration of the preoperative CT scan to the intraoperative radiograph was calculated via LevelCheck, and projection distance errors were analyzed. Five hundred random initializations were performed for each patient, and algorithm settings (viz, the number of robust multistarts, ranging 50-200) were varied to evaluate the trade-off between registration error and computation time. Failure mode analysis was performed by individually analyzing unsuccessful registrations (>5 mm distance error) observed with 50 multistarts. Results. At 200 robust multistarts (computation time of ∼26 s), the registration accuracy was 100% across all 10,000 trials. As the number of multistarts (and computation time) decreased, the registration remained fairly robust, down to 99.3% registration accuracy at 50 multistarts (computation time ∼7 s). Conclusion. The LevelCheck algorithm correctly identified target vertebrae in intraoperative mobile radiographs of the thoracolumbar spine, demonstrating acceptable computation time, compatibility with routinely obtained preoperative CT scans, and warranting investigation in prospective studies.

Original languageEnglish (US)
Pages (from-to)E476-E483
JournalSpine
Volume40
Issue number8
DOIs
StatePublished - Apr 15 2015

Fingerprint

Spine
Palpation
Fluoroscopy
Retrospective Studies
Prospective Studies

Keywords

  • 3D-2D registration
  • anatomical deformation
  • automatic
  • intraoperative
  • LevelCheck
  • localization
  • target vertebrae
  • thoracolumbar
  • vertebral labeling
  • wrong-site surgery

ASJC Scopus subject areas

  • Clinical Neurology
  • Orthopedics and Sports Medicine

Cite this

Automatic Localization of Target Vertebrae in Spine Surgery : Clinical Evaluation of the LevelCheck Registration Algorithm. / Lo, Sheng Fu L; Otake, Yoshito; Puvanesarajah, Varun; Wang, Adam S.; Uneri, Ali; De Silva, Tharindu; Vogt, Sebastian; Kleinszig, Gerhard; Elder, Benjamin D.; Goodwin, C. Rory; Kosztowski, Thomas A.; Liauw, Jason A.; Groves, Mari; Bydon, Ali; Sciubba, Daniel M.; Witham, Timothy F.; Wolinsky, Jean Paul; Aygun, Nafi; Gokaslan, Ziya L.; Siewerdsen, Jeffrey H.

In: Spine, Vol. 40, No. 8, 15.04.2015, p. E476-E483.

Research output: Contribution to journalArticle

Lo, SFL, Otake, Y, Puvanesarajah, V, Wang, AS, Uneri, A, De Silva, T, Vogt, S, Kleinszig, G, Elder, BD, Goodwin, CR, Kosztowski, TA, Liauw, JA, Groves, M, Bydon, A, Sciubba, DM, Witham, TF, Wolinsky, JP, Aygun, N, Gokaslan, ZL & Siewerdsen, JH 2015, 'Automatic Localization of Target Vertebrae in Spine Surgery: Clinical Evaluation of the LevelCheck Registration Algorithm', Spine, vol. 40, no. 8, pp. E476-E483. https://doi.org/10.1097/BRS.0000000000000814
Lo, Sheng Fu L ; Otake, Yoshito ; Puvanesarajah, Varun ; Wang, Adam S. ; Uneri, Ali ; De Silva, Tharindu ; Vogt, Sebastian ; Kleinszig, Gerhard ; Elder, Benjamin D. ; Goodwin, C. Rory ; Kosztowski, Thomas A. ; Liauw, Jason A. ; Groves, Mari ; Bydon, Ali ; Sciubba, Daniel M. ; Witham, Timothy F. ; Wolinsky, Jean Paul ; Aygun, Nafi ; Gokaslan, Ziya L. ; Siewerdsen, Jeffrey H. / Automatic Localization of Target Vertebrae in Spine Surgery : Clinical Evaluation of the LevelCheck Registration Algorithm. In: Spine. 2015 ; Vol. 40, No. 8. pp. E476-E483.
@article{ad3b6ad301364886a53bce602564a211,
title = "Automatic Localization of Target Vertebrae in Spine Surgery: Clinical Evaluation of the LevelCheck Registration Algorithm",
abstract = "Study Design. A 3-dimensional-2-dimensional (3D-2D) image registration algorithm, {"}LevelCheck,{"} was used to automatically label vertebrae in intraoperative mobile radiographs obtained during spine surgery. Accuracy, computation time, and potential failure modes were evaluated in a retrospective study of 20 patients. Objective. To measure the performance of the LevelCheck algorithm using clinical images acquired during spine surgery. Summary of Background Data. In spine surgery, the potential for wrong level surgery is significant due to the difficulty of localizing target vertebrae based solely on visual impression, palpation, and fluoroscopy. To remedy this difficulty and reduce the risk of wrong-level surgery, our team introduced a program (dubbed LevelCheck) to automatically localize target vertebrae in mobile radiographs using robust 3D-2D image registration to preoperative computed tomographic (CT) scan. Methods. Twenty consecutive patients undergoing thoracolumbar spine surgery, for whom both a preoperative CT scan and an intraoperative mobile radiograph were available, were retrospectively analyzed. A board-certified neuroradiologist determined the {"}true{"} vertebra levels in each radiograph. Registration of the preoperative CT scan to the intraoperative radiograph was calculated via LevelCheck, and projection distance errors were analyzed. Five hundred random initializations were performed for each patient, and algorithm settings (viz, the number of robust multistarts, ranging 50-200) were varied to evaluate the trade-off between registration error and computation time. Failure mode analysis was performed by individually analyzing unsuccessful registrations (>5 mm distance error) observed with 50 multistarts. Results. At 200 robust multistarts (computation time of ∼26 s), the registration accuracy was 100{\%} across all 10,000 trials. As the number of multistarts (and computation time) decreased, the registration remained fairly robust, down to 99.3{\%} registration accuracy at 50 multistarts (computation time ∼7 s). Conclusion. The LevelCheck algorithm correctly identified target vertebrae in intraoperative mobile radiographs of the thoracolumbar spine, demonstrating acceptable computation time, compatibility with routinely obtained preoperative CT scans, and warranting investigation in prospective studies.",
keywords = "3D-2D registration, anatomical deformation, automatic, intraoperative, LevelCheck, localization, target vertebrae, thoracolumbar, vertebral labeling, wrong-site surgery",
author = "Lo, {Sheng Fu L} and Yoshito Otake and Varun Puvanesarajah and Wang, {Adam S.} and Ali Uneri and {De Silva}, Tharindu and Sebastian Vogt and Gerhard Kleinszig and Elder, {Benjamin D.} and Goodwin, {C. Rory} and Kosztowski, {Thomas A.} and Liauw, {Jason A.} and Mari Groves and Ali Bydon and Sciubba, {Daniel M.} and Witham, {Timothy F.} and Wolinsky, {Jean Paul} and Nafi Aygun and Gokaslan, {Ziya L.} and Siewerdsen, {Jeffrey H.}",
year = "2015",
month = "4",
day = "15",
doi = "10.1097/BRS.0000000000000814",
language = "English (US)",
volume = "40",
pages = "E476--E483",
journal = "Spine",
issn = "0362-2436",
publisher = "Lippincott Williams and Wilkins",
number = "8",

}

TY - JOUR

T1 - Automatic Localization of Target Vertebrae in Spine Surgery

T2 - Clinical Evaluation of the LevelCheck Registration Algorithm

AU - Lo, Sheng Fu L

AU - Otake, Yoshito

AU - Puvanesarajah, Varun

AU - Wang, Adam S.

AU - Uneri, Ali

AU - De Silva, Tharindu

AU - Vogt, Sebastian

AU - Kleinszig, Gerhard

AU - Elder, Benjamin D.

AU - Goodwin, C. Rory

AU - Kosztowski, Thomas A.

AU - Liauw, Jason A.

AU - Groves, Mari

AU - Bydon, Ali

AU - Sciubba, Daniel M.

AU - Witham, Timothy F.

AU - Wolinsky, Jean Paul

AU - Aygun, Nafi

AU - Gokaslan, Ziya L.

AU - Siewerdsen, Jeffrey H.

PY - 2015/4/15

Y1 - 2015/4/15

N2 - Study Design. A 3-dimensional-2-dimensional (3D-2D) image registration algorithm, "LevelCheck," was used to automatically label vertebrae in intraoperative mobile radiographs obtained during spine surgery. Accuracy, computation time, and potential failure modes were evaluated in a retrospective study of 20 patients. Objective. To measure the performance of the LevelCheck algorithm using clinical images acquired during spine surgery. Summary of Background Data. In spine surgery, the potential for wrong level surgery is significant due to the difficulty of localizing target vertebrae based solely on visual impression, palpation, and fluoroscopy. To remedy this difficulty and reduce the risk of wrong-level surgery, our team introduced a program (dubbed LevelCheck) to automatically localize target vertebrae in mobile radiographs using robust 3D-2D image registration to preoperative computed tomographic (CT) scan. Methods. Twenty consecutive patients undergoing thoracolumbar spine surgery, for whom both a preoperative CT scan and an intraoperative mobile radiograph were available, were retrospectively analyzed. A board-certified neuroradiologist determined the "true" vertebra levels in each radiograph. Registration of the preoperative CT scan to the intraoperative radiograph was calculated via LevelCheck, and projection distance errors were analyzed. Five hundred random initializations were performed for each patient, and algorithm settings (viz, the number of robust multistarts, ranging 50-200) were varied to evaluate the trade-off between registration error and computation time. Failure mode analysis was performed by individually analyzing unsuccessful registrations (>5 mm distance error) observed with 50 multistarts. Results. At 200 robust multistarts (computation time of ∼26 s), the registration accuracy was 100% across all 10,000 trials. As the number of multistarts (and computation time) decreased, the registration remained fairly robust, down to 99.3% registration accuracy at 50 multistarts (computation time ∼7 s). Conclusion. The LevelCheck algorithm correctly identified target vertebrae in intraoperative mobile radiographs of the thoracolumbar spine, demonstrating acceptable computation time, compatibility with routinely obtained preoperative CT scans, and warranting investigation in prospective studies.

AB - Study Design. A 3-dimensional-2-dimensional (3D-2D) image registration algorithm, "LevelCheck," was used to automatically label vertebrae in intraoperative mobile radiographs obtained during spine surgery. Accuracy, computation time, and potential failure modes were evaluated in a retrospective study of 20 patients. Objective. To measure the performance of the LevelCheck algorithm using clinical images acquired during spine surgery. Summary of Background Data. In spine surgery, the potential for wrong level surgery is significant due to the difficulty of localizing target vertebrae based solely on visual impression, palpation, and fluoroscopy. To remedy this difficulty and reduce the risk of wrong-level surgery, our team introduced a program (dubbed LevelCheck) to automatically localize target vertebrae in mobile radiographs using robust 3D-2D image registration to preoperative computed tomographic (CT) scan. Methods. Twenty consecutive patients undergoing thoracolumbar spine surgery, for whom both a preoperative CT scan and an intraoperative mobile radiograph were available, were retrospectively analyzed. A board-certified neuroradiologist determined the "true" vertebra levels in each radiograph. Registration of the preoperative CT scan to the intraoperative radiograph was calculated via LevelCheck, and projection distance errors were analyzed. Five hundred random initializations were performed for each patient, and algorithm settings (viz, the number of robust multistarts, ranging 50-200) were varied to evaluate the trade-off between registration error and computation time. Failure mode analysis was performed by individually analyzing unsuccessful registrations (>5 mm distance error) observed with 50 multistarts. Results. At 200 robust multistarts (computation time of ∼26 s), the registration accuracy was 100% across all 10,000 trials. As the number of multistarts (and computation time) decreased, the registration remained fairly robust, down to 99.3% registration accuracy at 50 multistarts (computation time ∼7 s). Conclusion. The LevelCheck algorithm correctly identified target vertebrae in intraoperative mobile radiographs of the thoracolumbar spine, demonstrating acceptable computation time, compatibility with routinely obtained preoperative CT scans, and warranting investigation in prospective studies.

KW - 3D-2D registration

KW - anatomical deformation

KW - automatic

KW - intraoperative

KW - LevelCheck

KW - localization

KW - target vertebrae

KW - thoracolumbar

KW - vertebral labeling

KW - wrong-site surgery

UR - http://www.scopus.com/inward/record.url?scp=84958111519&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84958111519&partnerID=8YFLogxK

U2 - 10.1097/BRS.0000000000000814

DO - 10.1097/BRS.0000000000000814

M3 - Article

C2 - 25646750

AN - SCOPUS:84958111519

VL - 40

SP - E476-E483

JO - Spine

JF - Spine

SN - 0362-2436

IS - 8

ER -