Validation of the Calypso Surface Beacon Transponder

Maxwell Belanger, Ziad Saleh, Tom Volpe, R Margiasso, Xiang Li, Maria F. Chan, Xiaofeng Zhu, Xiaoli TangJournal of Applied Clinical Medical Physics 2016 · DOI

TL;DR

This paper validates the new L-shaped Calypso Surface Beacon transponder for real-time tracking in radiotherapy. The study demonstrates that the system achieves submillimeter precision and accuracy across various conditions, including different breathing rates and orientations, making it a stable and reproducible tool for patient motion management.

What Problem Does This Paper Address?

Precise patient positioning and real-time tracking of internal anatomy are crucial for modern radiation therapy, especially for targets that move with respiration. The Calypso system, which uses electromagnetic transponders, has been used clinically for this purpose. Recently, Varian Medical Systems released an L-shaped Surface Beacon transponder designed for external placement, offering a potentially less invasive tracking solution. However, before widespread clinical adoption, its performance needed rigorous validation. This paper addresses the need to evaluate the stability, reproducibility, sensitivity to orientation and breathing rate, and overall accuracy of this new Surface Beacon transponder in tracking simulated and recorded respiratory motion.

Methods

The researchers employed a Quasar respiratory phantom to simulate breathing motion. This phantom can generate both sinusoidal waveforms and reproduce recorded patient breathing patterns with high positional accuracy, serving as a ground truth. The Calypso Surface Beacon transponder was placed on the phantom's linear translation arm.

The validation involved several key experiments:

  • Stability and Drift Analysis: Sinusoidal waveforms were generated, and the transponder's position was tracked over 200 cycles. Linear least-squares fitting was used to assess any systematic drift in position over time.
  • Reproducibility: The Calypso system was re-setup multiple times, and the transponder's motion was tracked for 100 cycles in each run. This assessed the impact of electromagnetic array setup variations on accuracy.
  • Beacon Orientation Sensitivity: The transponder was tested at various angular offsets (0°, 5°, 10°, 15°) relative to the phantom's motion to determine if orientation affected tracking accuracy.
  • Rate Dependence: The phantom generated sinusoidal motion at different cycle rates (0.2 Hz, 0.33 Hz, and 1 Hz) to evaluate how breathing frequency impacts tracking.
  • Respiratory Waveform Tracking Accuracy: The system was tested using recorded breathing waveforms from volunteers and patients, as well as preprogrammed breathing sets. This assessed the system's ability to track complex, non-sinusoidal motion.

Datasets

The study utilized the following:

  • Simulated Breathing Waveforms: Sinusoidal waveforms generated by the Quasar respiratory phantom.
  • Recorded Breathing Waveforms: Data sourced from recordings of volunteers (6 of 37) and patients (13 of 37) using the Real-Time Position Management (RPM) system.
  • Preprogrammed Breathing Sets: 18 of 37 datasets were preprogrammed breathing sets from the Calypso phantom.

The Quasar phantom's on-board optical encoder provided the ground truth position measurements, offering a higher resolution than the Calypso system itself.

Key Insights and Ideas

  • Submillimeter Precision is Achievable: The Surface Beacon consistently delivered tracking with submillimeter resolution, which is critical for clinical radiotherapy.
  • Drift is Clinically Insignificant: While minor drift was detected, its velocity was on the order of nanometers per second, deemed too small to impact clinical treatments.
  • Setup Error is a Key Factor: Reproducibility tests highlighted that electromagnetic array setup variations contribute significantly to overall system precision. Careful setup is therefore paramount.
  • Orientation Matters, But Not Linearly: While beacon orientation didn't show a linear increase in standard deviation (SD) within the tested range, there was a systematic error observed, suggesting potential issues with angle measurement accuracy.
  • High Breathing Rates Pose a Challenge: While the system performed well at typical breathing rates, a 1 Hz cycle rate (higher than most patient rates) showed a noticeable decrease in resolution and amplitude shifts, indicating a potential limitation at extreme rates.
  • Real-World Breathing is Tracked Well: The system demonstrated robust tracking of complex, recorded breathing waveforms, with a standard deviation of 0.233 mm.

Key Results

  • Stability and Drift: Mean amplitudes were 20.01 ± 0.04 mm (phase 0%) and -19.65 ± 0.08 mm (phase 50%). Drift velocity was minimal (e.g., 5.1 × 10⁻⁵ mm/s at phase 0%).
  • Reproducibility: Combined SDs (setup + intrarun error) were 0.07 mm (phase 0%) and 0.11 mm (phase 50%). Setup error had a greater impact on phase 0% location.
  • Beacon Orientation: SD ranged from 0.032 mm to 0.039 mm (phase 0%) and 0.084 mm to 0.096 mm (phase 50%). Trends were observed but deemed suspect due to high reduced chi-squared values.
  • Rate Dependence: At 0.2 Hz and 0.33 Hz, SD and displacement amplitudes were stable. At 1 Hz, SD increased significantly (1.6x at phase 0%, 5.4x at phase 50%), and amplitudes shifted by ~0.7 mm.
  • Respiratory Waveform Tracking: SD was 0.233 mm with an approximately normal distribution over 134 minutes of tracking.

Limitations and Future Work

  • Angle Measurement Discrepancy: The study noted that the Calypso's measured Beacon angle did not always agree with the calibrated couch angle, and this disagreement was not systematic. This suggests potential issues with the Calypso's angle measurement itself.
  • Limited 3D Motion Analysis: The experimental design primarily focused on anterior-posterior motion and did not fully explore three-dimensional motion or rotation during motion. While the authors don't anticipate significant changes, this remains an area for future investigation.
  • High Rate Performance: The decrease in resolution at 1 Hz, while not typical for patient breathing, suggests that extreme respiratory rates might impact accuracy. Further investigation into the source of this shift (e.g., phantom drive motor interference) is warranted.
  • Setup Tolerance: The authors recommend that therapists optimize electromagnetic array positioning beyond manufacturer tolerances to further improve accuracy.

Future work could involve more comprehensive 3D motion tracking studies and further investigation into the factors affecting performance at higher respiratory rates.

Summary updated

Abstract

Calypso L-shaped Surface Beacon transponder has recently become available for clinical applications. We herein conduct studies to validate the Surface Beacon transponder in terms of stability, reproducibility, orientation sensitivity, cycle rate dependence, and respiratory waveform tracking accuracy. The Surface Beacon was placed on a Quasar respiratory phantom and positioned at the isocenter with its two arms aligned with the lasers. Breathing waveforms were simulated, and the motion of the transponder was tracked. Stability and drift analysis: sinusoidal waveforms (200 cycles) were produced, and the amplitudes of phases 0% (inhale) and 50% (exhale) were recorded at each breathing cycle. The mean and standard deviation (SD) of the amplitudes were calculated. Linear least-squares fitting was performed to access the possible amplitude drift over the breathing cycles. Reproducibility: similar setting to stability and drift analysis, and the phantom generated 100 cycles of the sinusoidal waveform per run. The Calypso system's was re-setup for each run. Recorded amplitude and SD of 0% and 50% phase were compared between runs to assess contribution of Calypso electromagnetic array setup variation. Beacon orientation sensitivity: the Calypso tracks sinusoidal phantom motion with a defined angular offset of the beacon to assess its effect on SD and peak-to-peak amplitude. Rate dependence: sinusoidal motion was generated at cycle rates of 1 Hz, .33 Hz, and .2 Hz. Peak-to-peak displacement and SDs were assessed. Respiratory waveform tracking accuracy: the phantom reproduced recorded breathing cycles (by volunteers and patients) were tracked by the Calypso system. Deviation in tracking position from produced waveform was used to calculate SD throughout entire breathing cycle. Stability and drift analysis: Mean amplitude ± SD of phase 0% or 50% were 20.01 ± 0.04 mm and -19.65 ± 0.08 mm, respectively. No clinically significant drift was detected with drift measured as 5.1 × 10-5 mm/s at phase 0% and -6.0 × 10-5 mm/s at phase 50%. Reproducibility: The SD of the setup was 0.06 mm and 0.02 mm for phases 0% and 50%, respectively. The combined SDs, including both setup and intrarun error of all runs at phases 0% and 50%, were 0.07mm and 0.11 mm, respectively. Beacon orientation: SD ranged from 0.032mm to 0.039 mm at phase 0% and from 0.084 mm to 0.096 mm at phase 50%. The SD was found not to vary linearly with Beacon angle in the range of 0° and 15°. A positive systematic error was observed with amplitude 0.07 mm/degree at phase 0% and 0.05 mm/degree at phase 50%. Rate dependence: SD and displacement amplitudes did not vary significantly between 0.2 Hz and 0.33 Hz. At 1 Hz, both 0% and 50% amplitude measurements shifted up appreciably, by 0.72 mm and 0.78mm, respectively. As compared with the 0.33 Hz data, SD at phase 0% was 1.6 times higher and 5.4 times higher at phase 50%. Respiratory waveform tracking accuracy: SD of 0.233 mm with approximately normal distribution in over 134 min of tracking (201468 data points). The Surface Beacon transponder appears to be stable, accurate, and reproducible. Submillimeter resolution is achieved throughout breathing and sinusoidal waveforms.
BibTeX
@article{belanger2016validation,
  author = {Maxwell Belanger and Ziad Saleh and Tom Volpe and R Margiasso and Xiang Li and Maria F. Chan and Xiaofeng Zhu and Xiaoli Tang},
  title = {{Validation of the Calypso Surface Beacon Transponder}},
  journal = {Journal of Applied Clinical Medical Physics},
  year = {2016},
  doi = {10.1120/jacmp.v17i4.6152},
}

Loading related content...

Figure