Basic Original ReportClinical log data analysis for assessing the accuracy of the CyberKnife fiducial-free lung tumor tracking system
Introduction
Stereotactic body radiation therapy (SBRT) is considered a treatment option for stage I non-small cell lung cancer patients.[1], [2] In SBRT, some techniques for the management of respiratory tumor motion are applied to decrease the irradiated volume of healthy lung tissues.[3], [4], [5], [6] Real-time tumor tracking is a technique in which the target location is identified by using a correlation model that incorporates the respiratory movements and the target, which then allows for continuous irradiation that adjusts with respiratory movements.[7], [8] The tracking accuracy depends on the models used in each system[9], [10], [11]; therefore, assessment of tracking uncertainties is required to determine the proper margins in treatment plans.
The CyberKnife Synchrony Respiratory Tracking System (Accuray Incorporated, Sunnyvale, CA) can perform real-time tracking delivery of SBRT to the lung.[5], [7], [12] The system has 2 orthogonal diagnostic x-ray sources on the ceiling combined with flat-panel detectors under the floor to image the patient’s internal targets and correct the beam positions during treatment delivery, thereby minimizing any errors caused by intrafractional patient motion. The respiratory motion of the patient is obtained by a camera array mounted on the ceiling as well as light-emitting diode (LED) markers attached to the patient. For real-time tracking, a correlation model that relates the external LED marker positions to internal target positions is generated before each treatment. The system has also a prediction model that calculates the target’s future position to compensate for the mechanical time delay.[13], [14] Currently, 2 real-time tracking methods are available for the CyberKnife Synchrony system: fiducial-based and fiducial-free tracking methods. Whereas fiducial-based tracking requires fiducial markers implanted near or inside the tumor, fiducial-free tracking allows for direct tumor tracking without the need for implanted fiducials. This avoids the complications associated with the use of fiducial markers (eg, pneumothorax, delay of treatment, irregular tracking from fiducial migration and placement). Moreover, the errors caused by interfractional geometry changes, such as those that occur during fiducial-based tracking, can be minimized because of direct tumor tracking. The fiducial-free tracking algorithm segments a tumor tracking volume (TTV) into small regions and searches the digitally reconstructed radiographs (DRRs) and live radiography images for the intensity pattern most similar to that of the tumor, thereby localizing the tumor using image registration. There are 2 fiducial-free tracking systems: Xsight Lung Tracking (XLT) and 1-View tracking.[15], [16] The XLT system is available for tumors that can clearly be recognized by 2 x-ray cameras, and tracks the 3-dimensional position of the target. The 1-View tracking system is used for tumors that are visible on only 1 camera, and tracks each tumor in the plane of motion visible to that camera.
The CyberKnife Synchrony systems generate log files that contain tracking information for each treatment; this allows for the assessment of tracking discrepancies. Previous studies have revealed uncertainties in the correlation and prediction models for fiducial-based tracking by analyzing clinical log data or phantom experiments.[13], [14], [17], [18] No studies that closely investigate the accuracy for the fiducial-free tracking systems (XLT and 1-View tracking) have been performed, however. In contrast to fiducial markers such as gold spheres, the tracking lung tumor differs in shape, volume, and density per patient; thus, it is unclear if fiducial-free tracking can provide the same tracking accuracy as a fiducial-based tracking system.
In this study, we retrospectively analyzed the Synchrony log files of patients with lung tumors who underwent treatment using the XLT or 1-View tracking system at our institution and evaluated the clinical correlation and prediction errors. Furthermore, we investigated the tracking tumor-related parameters for these errors.
Section snippets
Patients
Data from all 42 patients with lung tumors (28 with primary lung cancers and 14 with metastatic tumors) who underwent SBRT using the CyberKnife (VSI, version 9.6.0) XLT or 1-View tracking system without implanted fiducials at our institution between March 2015 and October 2016 were retrospectively analyzed. The patients were treated with a prescription dose of 36 to 66 Gy in 4 to 12 fractions, depending on each patient’s attributes. Before devising a treatment plan, simulations were performed
Tracking tumor characteristics
As shown in Table 2, the overall mean tumor motion amplitude was 6.7 ± 6.8 mm, 1.4 ± 1.0 mm, 3.5 ± 2.0 mm, and 8.2 ± 6.5 mm in the SI, LR, AP, and radial directions, respectively. The amplitude in the SI direction was significantly larger for lower lobe tumors than for upper or middle lobe tumors (9.0 ± 8.7 mm vs 4.6 ± 3.3 mm, respectively, P = 0.03). The mean tumor tracking volume was 11.7 ± 15.3 cm3 and the mean tumor density was 26.2 ± 13.8 Hounsfield unit. There were no significant
Discussion
We successfully described the correlation and prediction errors of the CyberKnife fiducial-free tumor tracking systems via analyzing clinical log files. To our knowledge, no studies have previously investigated clinical model uncertainties in the XLT or 1-View tracking systems; hence, our results ought to provide information regarding proper margins when devising treatment plans for fiducial-free tumor tracking methods. Our data revealed smaller magnitudes of errors than those of previous
Conclusions
Upon analyzing the correlation and prediction model errors in CyberKnife fiducial-free lung tumor tracking systems using clinical log data, we found that the overall mean correlation error was 0.95 ± 0.43 mm and prediction error was 0.14 ± 0.11 mm in the radial directions. The tumor motion amplitude may be one of the factors that affect these model errors; therefore, the errors should be carefully verified to determine the margins for tumors with large motion amplitudes as accurately as
References (22)
- et al.
Stereotactic ablative radiotherapy versus lobectomy for operable stage I non-small-cell lung cancer: A pooled analysis of two randomised trials
Lancet Oncol
(2015) - et al.
The deep inspiration breath-hold technique in the treatment of inoperable non-small-cell lung cancer
Int J Radiat Oncol Biol Phys
(2000) - et al.
Benefit of respiration-gated stereotactic radiotherapy for stage I lung cancer: An analysis of 4DCT datasets
Int J Radiat Oncol Biol Phys
(2005) - et al.
Geometric accuracy of a real-time target tracking system with dynamic multileaf collimator tracking system
Int J Radiat Oncol Biol Phys
(2006) - et al.
Synchrony--cyberknife respiratory compensation technology
Med Dosim
(2008) - et al.
Development of a four-dimensional image-guided radiotherapy system with a gimbaled X-ray head
Int J Radiat Oncol Biol Phys
(2006) - et al.
Accuracy of robotic radiosurgical liver treatment throughout the respiratory cycle
Int J Radiat Oncol Biol Phys
(2015) - et al.
Clinical accuracy of the respiratory tumor tracking system of the cyberknife: assessment by analysis of log files
Int J Radiat Oncol Biol Phys
(2009) - et al.
Verification of accuracy of CyberKnife tumor-tracking radiation therapy using patient-specific lung phantoms
Int J Radiat Oncol Biol Phys
(2015) - et al.
Quantifying rigid and nonrigid motion of liver tumors during stereotactic body radiation therapy
Int J Radiat Oncol Biol Phys
(2014)
Predictive parameters of CyberKnife fiducial-less (XSight Lung) applicability for treatment of early non-small cell lung cancer: a single-center experience
Int J Radiat Oncol Biol Phys
Cited by (22)
Quantification of Intrafraction and Interfraction Tumor Motion Amplitude and Prediction Error for Different Liver Tumor Trajectories in Cyberknife Synchrony Tracking
2021, International Journal of Radiation Oncology Biology PhysicsCitation Excerpt :The key issue in real-time tumor tracking is that the target can be accurately and simultaneously predicted by the synchronized motion tracking system of the Cyberknife. The research of Descovich,21 Hoogeman,22 Nakayama,23,24 Nioutsikou,25 Pepin,26 and Seppenwoolde27 has provided detailed insight into how Synchrony works, and we will not repeat it in this article. The prediction model and its errors (not only intrafraction error but also interfraction error) are essential to improve the precision of the tracking system.
Localization accuracy of robotic radiosurgery in 1-view tracking
2019, Physica MedicaCitation Excerpt :In April 2018 1-view tracking was mentioned in two different and extensive works [22,23]. In particular, Nakayama et al. [22] asserted that there are no significant differences in correlation and prediction errors between XLTS and 1-view tracking, while Ricotti et al. [23] estimated an overall geometric accuracy for each fiducial-free modality using retrospectively 2-view log data and 4D-CT acquisitions. We conducted a retrospective study addressed to define the ITV-to-PTV margin necessary to compensate the localization uncertainties arising in the management of the case 1-view compared with the reference 2-view.
Retrospective assessment of a single fiducial marker tracking regimen with robotic stereotactic body radiation therapy for liver tumours
2019, Reports of Practical Oncology and RadiotherapyCitation Excerpt :Two radiation technologists carefully monitored patient movement and respiratory motion during treatment. Data in ModelPoint.log, which is the log data file produced by the Synchrony system, was obtained after each treatment fraction in each patient to analyse the correlation model error, as described previously.13 The data in this file were recorded at the time of each image acquisition during treatment.
Evaluation of a New Method for CyberKnife Treatment for Central Lung and Mediastinal Tumors by Tracheobronchial Tracking
2024, Technology in Cancer Research and TreatmentDevelopment of a prediction model for target positioning by using diaphragm waveforms extracted from CBCT projection images
2023, Journal of Applied Clinical Medical Physics
Conflicts of interest: None.
Sources of support: This work was partially supported by JSPS KAKENHI Grant Number 16K10395.