Jump to ContentJump to Main Navigation
Show Summary Details
More options …

Current Directions in Biomedical Engineering

Joint Journal of the German Society for Biomedical Engineering in VDE and the Austrian and Swiss Societies for Biomedical Engineering

Editor-in-Chief: Dössel, Olaf

Editorial Board: Augat, Peter / Buzug, Thorsten M. / Haueisen, Jens / Jockenhoevel, Stefan / Knaup-Gregori, Petra / Kraft, Marc / Lenarz, Thomas / Leonhardt, Steffen / Malberg, Hagen / Penzel, Thomas / Plank, Gernot / Radermacher, Klaus M. / Schkommodau, Erik / Stieglitz, Thomas / Urban, Gerald A.

CiteScore 2018: 0.47

Source Normalized Impact per Paper (SNIP) 2018: 0.377

Open Access
See all formats and pricing
More options …

Performance behavior of prediction filters for respiratory motion compensation in radiotherapy

Alexander Jöhl
  • Corresponding author
  • Product Development Group Zurich, Department of Mechanical and Process Engineering, ETH Zurich, CLA G19.2 Tannenstrasse 3 8092 Zurich, Switzerland
  • Email
  • Other articles by this author:
  • De Gruyter OnlineGoogle Scholar
/ Yannick Berdou / Matthias Guckenberger / Stephan Klöck / Mirko Meboldt / Melanie Zeilinger
  • Institute for Dynamic Systems and Control, Department of Mechanical and Process Engineering, ETH Zurich, Switzerland
  • Email
  • Other articles by this author:
  • De Gruyter OnlineGoogle Scholar
/ Stephanie Tanadini-Lang / Marianne Schmid Daners
Published Online: 2017-09-07 | DOI: https://doi.org/10.1515/cdbme-2017-0090


Introduction: In radiotherapy, tumors may move due to the patient’s respiration, which decreases treatment accuracy. Some motion mitigation methods require measuring the tumor position during treatment. Current available sensors often suffer from time delays, which degrade the motion mitigation performance. However, the tumor motion is often periodic and continuous, which allows predicting the motion ahead. Method and Materials: A couch tracking system was simulated in MATLAB and five prediction filters selected from literature were implemented and tested on 51 respiration signals (median length: 103 s). The five filters were the linear filter (LF), the local regression (LOESS), the neural network (NN), the support vector regression (SVR), and the wavelet least mean squares (wLMS). The time delay to compensate was 320 ms. The normalized root mean square error (nRMSE) was calculated for all prediction filters and respiration signals. The correlation coefficients between the nRMSE of the prediction filters were computed. Results: The prediction filters were grouped into a low and a high nRMSE group. The low nRMSE group consisted of the LF, the NN, and the wLMS with a median nRMSE of 0.14, 0.15, and 0.14, respectively. The high nRMSE group consisted of the LOESS and the SVR with both a median nRMSE of 0.34. The correlations between the low nRMSE filters were above 0.87 and between the high nRMSE filters it was 0.64. Conclusion: The low nRMSE prediction filters not only have similar median nRMSEs but also similar nRMSEs for the same respiration signals as the high correlation shows. Therefore, good prediction filters perform similarly for identical respiration patterns, which might indicate a minimally achievable nRMSE for a given respiration pattern.

Keywords: radiotherapy; tumor motion; respiration; prediction filter

About the article

Published Online: 2017-09-07

Citation Information: Current Directions in Biomedical Engineering, Volume 3, Issue 2, Pages 429–432, ISSN (Online) 2364-5504, DOI: https://doi.org/10.1515/cdbme-2017-0090.

Export Citation

©2017 Alexander Jöhl et al., published by De Gruyter. This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 License. BY-NC-ND 4.0

Comments (0)

Please log in or register to comment.
Log in