Abstract
Background/Aim: This study assessed the positional detection accuracy of the Catalyst+ HD system for intracranial stereotactic irradiation (STI) under clinically relevant conditions, including variations in head posture, isocenter position, and couch angle.
Materials and Methods: An anthropomorphic head phantom was used to simulate three head postures, chin-up, neutral, and chin-down, each stabilized with a corresponding thermoplastic mask. Seven isocenter positions were defined: one central position and six offset positions, each 5 cm away in a cardinal direction. Treatment plans incorporated multiple couch angles (0°, 30°, 45°, 60°, and 90°). The Catalyst+ HD system’s accuracy was evaluated by comparing its detected displacements to predefined shifts applied using a HexaPOD evo RT system. Translational shifts of ±3 mm and rotational shifts of ±2° were introduced. Statistical analysis was conducted using the Wilcoxon signed-rank test.
Results: Under standard conditions (neutral posture, central isocenter, and 0° couch angle), the system demonstrated submillimeter accuracy (mean translational error: 0.08 mm; mean rotational error: 0.12°). Detection errors were significantly larger in the chin-up posture compared to the neutral posture (p=0.028). Similarly, a superior isocenter position resulted in considerably larger errors (p=0.026). A couch rotation of 30° led to a significant increase in error, whereas other couch angles maintained high precision.
Conclusion: The Catalyst+ HD system exhibits high accuracy for intracranial STI under most tested conditions. However, to optimize performance and accuracy, configurations involving a chin-up posture or a superior isocenter position should be avoided.
- Surface-guided radiotherapy
- brain metastasis
- stereotactic radiotherapy
- intrafraction motion
- patient positioning
Introduction
Brain metastases are the most common intracranial malignancy in adults, affecting 20% to 40% of patients with cancer (1). In linear accelerator (linac)-based intracranial stereotactic irradiation (STI), delivering high doses in a single or few fractions demands precise treatment planning with minimal planning target volume (PTV) margins (2). This approach requires strict control of both setup errors and patient motion. While image-guided radiotherapy (IGRT) effectively corrects interfraction setup errors (3), the use of non-coplanar beams in STI limits the feasibility of cone-beam computed tomography (CBCT), making real-time motion management a critical challenge.
To address this limitation, the ExacTrac X-ray system enables image acquisition during treatment using floor- and ceiling-mounted X-ray devices (4). However, its limited field of view (FOV) and radiation exposure remain notable drawbacks. Alternatively, the Catalyst system offers a non-invasive, radiation-free tracking solution through visible light projection and surface imaging (5). Using a non-rigid algorithm, which calculates isocenter displacement from skin surface detection, positional accuracy comparable to CBCT for daily patient setup can be achieved (6). Furthermore, studies have shown that the Catalyst provides accuracy and reproducibility comparable to the ExacTrac X-ray at various couch angles simulating intracranial STI, highlighting its potential for position verification in non-coplanar treatments (7).
However, previous studies have not thoroughly investigated the impact of variations in isocenter position and patient posture on the measurement accuracy of these systems. In clinical practice, brain metastases can arise at diverse intracranial locations, leading to variations in isocenter positions among patients (8). Additionally, for patients with cervical spine disorders, standard head fixation positions can cause pain and discomfort, potentially leading to movement during irradiation. In such cases, adjusting the head position to a chin-up or chin-down posture can be an effective strategy. Furthermore, to the best of our knowledge, no studies have specifically examined these factors in the brain SRS-specific mode of the Catalyst+ HD (version 6.5.0, C-RAD AB, Uppsala, Sweden). Therefore, a comprehensive understanding of how these variables affect the accuracy of Catalyst+ HD is crucial for establishing appropriate clinical tolerances and ensuring safe, effective treatment delivery.
In this study, we evaluated the impact of isocenter position and postural variations, along with couch rotation angles, on the measurement accuracy of the catalyst system using an anthropomorphic phantom under intracranial STI conditions.
Materials and Methods
Treatment planning. As shown in Figure 1A, a PH-3 head phantom (Kyoto Kagaku, Kyoto, Japan) was positioned using two distinct head support systems to simulate three different head postures: chin-up, neutral, and chin-down. Individual thermoplastic immobilization masks were fabricated for each posture. Treatment planning CT scans for each configuration were acquired using a Siemens SOMATOM go.Sim scanner (Siemens Healthineers, Erlangen, Germany). The brain structure was contoured using the Monaco treatment planning system (version 6.1.4.0, Elekta, Stockholm, Sweden), and the isocenter was defined at the center of the brain, as shown in Figure 1B.
Experimental setup of the study. (A) Representative computed tomography images showing three different postures. (B) Six different isocenter positions (C: Center; A: Anterior; P: Posterior; S: Superior; I: Inferior; L: Left; R: Right).
Five additional isocenter positions were established by shifting 5 cm from the central position in each direction: anterior, posterior, superior, inferior, left, and right. For the neutral head position, treatment plans were generated incorporating various couch angles (0°, 30°, 45°, 60°, and 90°) to simulate non-coplanar beam arrangements.
Phantom setup. The positional detection accuracy of the Catalyst+ HD system was systematically evaluated. Initially, treatment planning CT images were registered with the Catalyst+ HD system’s live images to establish reference images. Subsequently, the HexaPOD evo RT system, integrated with a Synergy linac (Elekta), was used to apply precise known displacements. Translational shifts were introduced in 1 mm increments over a range of ±3 mm in the superior-inferior (S-I), left-right (L-R), and anterior-posterior (A-P) directions. Rotational adjustments were applied in 1° increments over a range of ±2° for roll, pitch, and yaw. Detection accuracy was assessed by calculating the differences between the displacements measured by the Catalyst+ HD system and the corresponding input displacement values at each position.
Experimental conditions. All measurements were initially conducted under standard conditions, defined as a neutral head position, an isocenter at the center of the phantom brain, and a couch angle of 0°. The evaluation then encompassed three primary variables: 1) head posture, with additional measurements taken in chin-up and chin-down positions while maintaining a fixed 0° couch angle and the “Center” isocenter; 2) isocenter position, with measurements performed at six additional locations under a neutral head position and 0° couch angle; and 3) couch rotation, with measurements obtained at five different couch angles while maintaining a neutral head position and the “Center” isocenter.
Detection errors and statistical analysis. The detection error was calculated as the root-sum-square difference between the displacement commanded by the HexaPOD evo RT system and the movement detected by the catalyst system. At nonzero couch angles, the Catalyst coordinate system remains fixed, while the HexaPOD coordinate system rotates with the couch. Consequently, all directions, except the vertical and pitch axes, require three-dimensional coordinate transformations using trigonometric functions. For instance, the detection error in the lateral direction (DElat.) is calculated using the following equation:
(1)
where and Ddetec. is the displacement detected using the catalyst system. Furthermore, the theoretical displacement value, Dtheor. for each axis was calculated as follows:
(2)
(3)
where Dcmd. represents the displacement commanded through the HexaPOD evo RT system, θ is the couch angle. Additionally, for other directions, the theoretical displacement values were compared with the detected values by evaluating the coordinate axis rotation and the displacements applied by the HexaPOD evo RT system in three dimensions. This comprehensive approach ensured accurate error quantification across all spatial dimensions.
Statistical analysis was performed using the Wilcoxon signed-rank test to compare detection errors under standard conditions with those observed under varying postures, isocenter positions, and couch angles. All statistical analyses were conducted using SPSS Statistics software version 27 (IBM Corp., Armonk, NY, USA), with statistical significance set at p<0.05.
Results
Standard conditions. Under standard conditions (neutral head position, isocenter at “Center”, and a couch angle of 0°), the mean±standard deviation of translational detection errors were 0.07±0.05 mm laterally, 0.08±0.04 mm longitudinally, and 0.08±0.04 mm vertically. Rotational errors were 0.13±0.05° for pitch, 0.13±0.10° for roll, and 0.10±0.08° for yaw.
Posture variations. Figure 2 illustrates the detection errors across different head postures, while Table I summarizes the mean translational and rotational errors. Among all six axes, the largest mean detection errors occurred in the chin-up position, measuring 0.17 mm in the vertical direction and 0.28° in yaw rotation. Statistical analysis of combined errors across all axes showed that detection errors in the chin-up position were significantly higher than in the baseline position (p=0.028), whereas no significant differences were observed in the chin-down position.
Detection errors across variations in head posture, illustrating (A) translational and (B) rotational measurements.
Summary of detection errors under different experimental conditions.
Couch rotation angles. Figure 3 displays detection errors across different couch angles, while Table I summarizes the mean errors for translational and rotational movements. The highest mean detection errors occurred at a 90° couch angle, with 0.18 mm in the lateral direction for translational movement, and at a 30° couch angle, with 0.14° in the pitch direction for rotational movement. Statistical analysis of errors across all six axes showed that detection errors at a 30° couch rotation were significantly higher than those under standard conditions (0°) (p<0.045). However, no significant differences were observed at other couch angles.
Detection errors across variations in couch angle, illustrating (A) translational and (B) rotational measurements.
Isocenter positions. Figure 4 illustrates detection errors across different isocenter positions, while Table I presents the mean errors for translational and rotational movements. The highest mean detection errors were observed at the superior position, with 0.13 mm for translational movement and 0.16° for rotational movement. Statistical analysis across all six axes showed that detection errors at the superior position were significantly higher than those at the baseline position (p=0.026), while no significant differences were found at other isocenter positions.
Detection errors across variations in isocenter position, illustrating (A) translational and (B) rotational measurements (C: Center; A: Anterior; P: Posterior; S: Superior; I: Inferior; L: Left; R: Right).
Discussion
This study evaluated the positional detection accuracy of the Catalyst+ HD system for intracranial stereotactic irradiation using an anthropomorphic phantom. In addition to assessing detection errors under varying couch angles – similar to a previous study (9), we examined how head posture variations (chin-up, chin-down) and isocenter position changes affect accuracy.
Under standard conditions (neutral head position, central isocenter, and 0° couch angle), the system demonstrated submillimeter accuracy, with mean errors of 0.08 mm in translation and 0.12° in rotation. These results align with those of the conventional Catalyst HD system and meet precision requirements for intracranial STI (10-12).
Head posture significantly influenced detection accuracy. The chin-up position produced larger errors than the neutral position (translational: 0.11±0.08 mm, rotational: 0.18±0.11°, p<0.028). Previous studies with the AlignRT system also reported greater variability in the chin-up position, likely due to blind spots caused by a lack of cranial camera coverage (13). In contrast, the chin-down position showed no significant differences, as FOV limitations were not an issue. These findings offer valuable guidance for adapting patient positioning, particularly for those with cervical spine disorders or other conditions that prevent maintaining a neutral posture.
With respect to couch angle variations, detection errors remained below 0.3 mm for translational directions and below 0.3° for rotational directions across all measurements. Although errors were significantly larger at a 30° couch angle, no significant differences were observed at other angles. This seemingly contradictory result may involve multiple contributing factors. First, at the specific angle of 30°, the relationship between the camera’s FOV and the relative position of the phantom surface may have changed, affecting the surface recognition accuracy of the system’s algorithm. Second, this angle may have created a configuration where partial occlusion by the gantry or treatment room structures was more likely to occur. Furthermore, with the p-value being close to the significance threshold, we cannot exclude the possibility that this observation was influenced by statistical variation. As reported in previous research (6, 9), while the Catalyst system generally demonstrates high precision in non-coplanar beam arrangements, its accuracy may fluctuate under specific angular conditions due to characteristics of the detection algorithm.
Regarding the effect of isocenter position variations, significantly larger errors were identified at the superior position (p=0.026). While previous studies have reported that errors increase with depth from the region of interest (ROI) (9), our results showed significant differences only at the superior position. This suggests that three-dimensional distance from the ROI, rather than depth alone, may influence accuracy. This finding has important implications for optimizing ROI settings and isocenter positioning in the treatment planning of metastatic brain tumors occurring at various intracranial locations. Overall, these results demonstrate that the Catalyst+ HD system achieves submillimeter accuracy across various clinical scenarios, supporting high-precision intracranial stereotactic irradiation. Although there are no published papers analyzing data from metastatic brain tumor patients using this system, it potentially offers comparable accuracy to the AlignRT system, which has been validated with actual patient data (14, 15). Accurate monitoring of patient movement during treatment is particularly crucial in stereotactic radiotherapy, as intrafraction motion errors can significantly impact organs at risk (16). This system has strong potential to meet the precision requirements for surface-guided radiotherapy (17).
Study limitations. First, although the anthropomorphic head phantom simulates human anatomical structures, it cannot fully reproduce the complexities of patient-specific factors such as skin tone and facial expressions (18, 19). Second, while this study focused on detection errors, it did not evaluate their clinical impact on treatment outcomes or dose distribution. Finally, because the configuration and positioning of catalyst systems may vary between institutions, this study was limited to verification at a single institution.
In the future, conducting multi-institutional validation studies based on the methodology established in this research will represent an important direction. Specifically, distributing standardized anthropomorphic head phantoms and detailed measurement protocols to multiple institutions would allow for a systematic evaluation of how different installation conditions of the catalyst system, such as camera positioning, room lighting, and treatment room layout, affect measurement accuracy across diverse clinical settings.
Conclusion
This study verified the accuracy of the Catalyst+ HD system for detecting intracranial STI under various clinical conditions. To ensure optimal accuracy, it is recommended to avoid the chin-up posture and isocenter positions that are distant from the camera’s FOV. Incorporating these findings into treatment planning can facilitate the clinical implementation of high-precision intracranial STI using non-invasive surface-guided technology.
Acknowledgements
The Authors would like to thank Editage (www.editage.jp) for English language editing.
Footnotes
Authors’ Contributions
T.S. collected the data, performed data analysis, and drafted the manuscript. S.O. contributed to the study design, collected data, and provided overall supervision. Y.I. contributed to data collection and provided the experimental facilities. Y.Y. supported with the experimental procedures. Y.N., T.O., T.H., M.M., A.K., H.Y., and K.N. contributed to the data analysis and interpretation and critically revised the manuscript.
Conflicts of Interest
Tenyoh Suzuki, Shingo Ohira, and Keiichi Nakagawa are affiliated with a department that receives unrestricted funding from Elekta AB. However, the funder had no role in this study.
Funding
None.
- Received March 21, 2025.
- Revision received April 8, 2025.
- Accepted April 9, 2025.
- Copyright © 2025 The Author(s). Published by the International Institute of Anticancer Research.
This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY-NC-ND) 4.0 international license (https://creativecommons.org/licenses/by-nc-nd/4.0).
References
In this issue
Jump to section
Related Articles
Cited By...
- No citing articles found.