This work evaluated, on a 3D mobile phantom, the accuracy and sensitivity of two motion management systems, known as OSMS and HDMM, that have been utilized for frameless stereotactic treatments. Overall, the results showed both systems can achieve sub-millimeter accuracy in measuring linear motion. However, the positioning accuracy in the Y-direction (S-I) for OSMS forehead monitoring (0.8 ± 0.5 mm) was poorer than OSMS nose monitoring (0.3 ± 0.2 mm, p = 0.01*) and HDMM monitoring (0.1 ± 0.2 mm, p < 0.001*). Both systems were also able to detect sudden changes of the motion with minimal time delay. If motion speed was beyond 20 mm/s along Y-direction (S-I), OSMS was relatively insensitive in detecting the motion.
Wiant et. al reported that if the phantom moves in parallel to the sight line of the cameras, OSMS has decreased accuracy in detecting motion . Our study confirmed this finding that the accuracy and sensitivity decreased along the direction of longitudinal motion. A similar trend was also observed for HDMM. This trend is probably due to the same reason, as the camera was attached at the foot of the couch, making the reflection signal relatively harder to detect for the longitudinal direction compared to lateral or vertical shifts.
A larger measurement delay was also observed with higher motion speed especially for OSMS due to its lower frame rate. Stereotactic radiosurgery (SRS) has become a popular tool to treat intracranial brain metastases due to convenience for the patients, durable local control and the possibility of reduced cognitive impairment versus whole brain radiotherapy [15–17]. Thomas et. al reported that mean delivery time was 1.7 min per target for linac-based SRS using flattening-free beam (FFF) and an average 31.6 min delivery per target for GK-SRS . With faster delivery but less frequent monitoring, the radiation delivery may be misaligned due to involuntary patient shifting. For example, OSMS system generates a 3D surface at a rate of 5 fps while the frame rate for HDMM is 20 fps. If the target moves out of position during one frame acquisition, a 0.2% dose may be missed for the linac-based FFF SRS treatment with OSMS monitoring, compared to 0.003% with HDMM monitoring for GK treatment, indicating that the latter is more forgiving and less susceptible to patient motion at the expense of longer treatment times. The SRS with frame fixation has minimal reported shifts during the whole course of treatment . However, mask fixation demonstrated significant higher variability and overall errors than frame fixation . The reasons include conformity of the mask to patient’s face, amount of pressure applied by the mask on the skin, and deformations in mask assembly. Despite the monitoring systems allows for real-time tracking, the frequent patient motion should still be minimized.
In the present study, two reference ROIs, the forehead and nose areas, were chosen and investigated. We demonstrated that OSMS was less sensitive to motion fluctuations of smooth/flat surfaces such as the forehead. This may be caused by the longer integration time and complexity to solve smooth surface matching with current registration algorithms in OSMS. Therefore, it is important to select the appropriate reference ROI for OSMS monitoring. Yet, there also exist issues for with single marker-based monitoring system. The question of how intra-cranial targets moves in relationship to the nose is not resolved yet. Prior studies suggested the intracranial displacement, although not guaranteed in all cases, is typically less than nose displacement [21, 22]. This sensitive monitoring may increase treatment gating events and hinder treatment delivery with decreased patient comfort.
External beam machine with OSMS capability is typically coupled with a 6 degree of freedom (6-DOF) couch. OSMS system can report 6 degrees of change as lateral (x-), longitudinal (y-), vertical (z-) translations, rotational yaw, pitch and roll. However, the Gamma Knife Icon only has three degrees of freedom couch and HDMM systems reports three translational positioning shifts only. We designed this phantom used in this study for translational shifts but not rotational movements, and only assessed linear motion monitoring. In the future, a more sophisticated phantom which allows rotational changes and the capability in recognizing those changes should be evaluated.
It is important to stress, at the present time, mask-based SRS patient treatment in conjunction with motion tracking and image-guidance is still a relatively new paradigm for most users, especially in GK community. As such there has been limited data to report or to compare the performance of these real-time motion management systems. In this study, we have evaluated two real-time motion management and monitoring systems used for frameless stereotactic treatment. We demonstrated that both OSMS and HDMM are efficient tools to improve the accuracy for verifying and complementing patient positioning in stereotactic treatment. However, performance variations were observed along different directions, as well as in the selection of reference images. Caution is needed when using real-time monitoring system for frameless SRS/SRT treatment and proper evaluation of the system prior to clinical use should be conducted.