Minimal mask immobilization with optical surface guidance for head and neck radiotherapy

Abstract Purpose Full face and neck thermoplastic masks provide standard‐of‐care immobilization for patients receiving H&N IMRT. However, these masks are uncomfortable and increase skin dose. The purpose of this pilot trial was to investigate the feasibility and setup accuracy of minimal face and neck mask immobilization with optical surface guidance. Methods Twenty patients enrolled onto this IRB‐approved protocol. Patients were immobilized with masks securing only forehead and chin. Shoulder movement was restricted by either moldable cushion or hand held strap retractors. Positional information, including isocenter location and CT skin contours, were imported to a commercial surface image guidance system. Patients typically received standard‐of‐care IMRT to 60–70 Gy in 30–33 fractions. Patients were first set up to surface markings with optical image guidance referenced to regions of interest (ROIs) on simulation CT images. Positioning was confirmed by in‐room CBCT. Following six‐dimensional robotic couch correction, a new optical real‐time surface image was acquired to track intrafraction motion and to serve as a reference surface for setup at the next treatment fraction. Therapists manually recorded total treatment time as well as couch shifts based on kV imaging. Intrafractional ROI motion tracking was automatically recorded by the optical image guidance system. Patient comfort was assessed by self‐administered surveys. Results Setup error was measured as six‐dimensional shifts (vertical/longitudinal/lateral/rotation/pitch/roll). Mean error values were −0.51 ± 2.42 mm, −0.49 ± 3.30 mm, 0.23 ± 2.58 mm, −0.15 ± 1.01o, −0.02 ± 1.19o, and 0.06 ± 1.08o, respectively. Average treatment time was 21.6 ± 8.4 mins). Subjective comfort during surface‐guided treatment was confirmed on patient surveys. Conclusion These pilot results confirm feasibility of minimal mask immobilization combined with commercially available optical image guidance. Patient acceptance of minimal mask immobilization has been encouraging. Follow‐up validation, with direct comparison to standard mask immobilization, appears warranted.


| INTRODUCTION
Patient immobilization is critical for safe, reproducible delivery of H&N radiotherapy. Thermoplastic masks routinely provide this immobilization. Many patients find masks constrictive and stressful. The density of thermoplastic is water equivalent and can create a skin bolus effect which intensifies skin reactions to treatment. 1 Minimal open face masks may make treatment less uncomfortable and toxic for patients.
Three-dimensional optical surface imaging can effectively monitor setup for surface-guided radiation therapy (SGRT). 2 SGRT is noninvasive and does not expose patients to ionizing radiation. Realtime surface capture can be registered with a baseline reference surface, such as skin contours rendered from a CT or optical images taken at the time of treatment simulation. Displacement errors can be displayed in real time as six-dimensional deltas to guide therapists during daily setup. Unlike online kV/MV imaging, SGRT provides continuous motion tracking during treatment. This has been leveraged to confirm breast setup accuracy during breath hold 3,4 and to track stereotactic treatment to cranial 5,6 and thoracic 7 sites.
Several small series have quantified setup reproducibility of open face mask prototypes; however, these prototypes typically employed small openings limited to central face. 8,9 In contrast, we wished to more significantly reduce mask coverage to selectively immobilize only fulcrums of movement at the chin and forehead, and to validate optical surface guidance of these minimal mask setups via in-room CBCT reference imaging.

2.A | SGRT platform and procedures
We used a commercial SGRT platform (AlignRT, Vision RT Ltd., London, UK) in all cases. The primary components of the system are three ceiling-mounted optical camera units capable of capturing three-dimensional real-time surface data from the patient on the treatment couch. 2 The cameras capture surface patterns projected onto patients, permitting in silico three-dimensional surface renderings. We used an elevated grid phantom provided by the manufacturer to calibrate isocenter localization. Daily grid QA was performed with a threshold of 1 mm. LINAC imaging center coincidence was cross-checked via an isocube phantom monthly. 10 We set 1 mm/0.5°a s a threshold to apply isocenter calibration correction to AlignRT.

2.B | Study cohort
This study was approved by our institutional IRB with a target enrollment of 20 patients. Patients receiving curative head and neck radiotherapy requiring an extended thermoplastic mask to cover shoulders were eligible for enrollment. Patients in the study cohort are listed in Table 1. All patients received standard-of-care IMRT to 60-70 Gy in 30-33 daily fraction, except for one patient (#17) who received hypofractionated SBRT for supraglottic laryngeal cancer on protocol. Mean age was 60.0 AE 8.9 yr.

2.C | Clinical workflow
We modified commercial thermoplastic masks (Qfix, Avondale, Pennsylvania, USA, model RT-1876KSDGLF) to immobilize only forehead and chin. The original mask has a precut 5 9 9 cm mid-face opening At each treatment, therapists recorded total time from patient entry to exit from the treatment room. Beam gating was manually controlled by therapists. "How securely did the mask keep you in one place?" The third item:

2.D | Patient comfort survey
"How confident are you that you will be able to tolerate this mask every day during treatment?" The fourth item: "How satisfied are you with this overall experience?" Patient survey data were tabulated and reported as raw values.

2.E | Data analysis
Setup accuracy based on SGRT was compared against CBCT. Group mean and standard deviation were calculated for all treatment fractions from all patients. Systematic setup error(∑) and the random error (r) were also calculated. 11 Systematic error is the standard deviation of the individual patient means from his/her entire treatment fractions. Random error is the root mean square of the individual patient standard deviation from treatment fractions. These two measures are ingredients of popular margin recipe from Van Herk. 12 Similarly these metrics were also performed on the intrafraction motion collected by AlignRT.

3.A | Position verification
A total of 591 CBCTs were obtained for reference to SGRT. Average couch shifts following CBCT verification of SGRT-guided setup are listed in Table 2. Also shown are systematic and random errors based on the two shoulder restriction methods we used, as well as for the cohort as a whole. Average shifts and errors were smaller with molded shoulder cushions vs shoulder retractors. Overall systematic error on translational shifts was small (<1.4 mm) and random

3.B | Intrafraction motion
Similar analysis was performed for intrafraction motion data. A total of 596 treatments were analyzed. As shown in Table 2

3.D | Patient comfort survey
A total of 19 of 20 surveys were returned. We specifically desired patients to provide stand-alone assessments of the minimal masks; no study patient tried a regular fully closed mask for comparison.
Out of a maximum score of 6, average score for mask comfort was   (Table 2) with surface guidance.
Therefore, the need for angle corrections with an expensive six-dimensional robotic couch is limited. Treatment time with our minimal mask platform is already comparable to standard closed mask treatment time.
Intrafractional motion in our study was also small. We used a composite ROI which included face and neck to cover the entire treatment area. Six-dimensional intrafractional motion tracking with this ROI was comparable to previous studies. 15  Another component of our study included assessment of shoulder immobilization methods. We compared two shoulder restriction methods and found that a moldable cushion provides better setup than shoulder stirrups. With the moldable cushion, patients are able to rest their arms and elbows at anchoring points, which makes shoulder and neck position more reproducible.
Interestingly, we found that translational and rotational errors spiked during the final week of treatment (Fig. 2)  Although set up accuracy was comparable to that of standard coverage masks for patients with similar treatment anatomy, 8,13-18 surface guidance alone is not enough to ensure reproducible daily setup given the large translational random setup errors. Therefore, onboard imaging is recommended to confirm setup. Considering small rotational setup errors, planar KVs may be sufficient.
Finally, although patient-reported comfort appeared to be high, we did not conduct a direct comparison between minimal vs standard mask comfort in the study patients. This was intentional, since we wished to minimize patient bias toward higher comfort scores for the minimal mask after trying a standard mask on. Future studies may directly compare patient comfort between both systems.

| CONCLUSION
We present a minimal mask immobilization solution with a stream-

CONF LICT OF I NTEREST
This study was supported by Vision RT, Ltd. (London, UK).