Click here to

Session: Imaging for Patient Setup and Target Alignment [Return to Session]

Evaluation of Different Holographic Visualization Techniques for Surface-Guided Patient Alignment Using Mixed Reality (MixR) and the Microsoft Hololens 2

A Jackson1*, P Johnson2, E Feldman2, A Martin-gomez3, (1) Mayo Clinic, Jacksonville, FL, (2) University of Florida Health Proton Therapy Institute, Jacksonville, FL, (3) Johns Hopkins University Whiting School Of Engineering, Baltimore, MD,

Presentations

TU-D1030-IePD-F1-4 (Tuesday, 7/12/2022) 10:30 AM - 11:00 AM [Eastern Time (GMT-4)]

Exhibit Hall | Forum 1

Purpose: To evaluate four different holographic visualization techniques for surface-guided phantom alignment using mixed reality (MixR) and the Microsoft Hololens 2.

Methods: A CT scan of an anthropomorphic RANDO female phantom was acquired and used to generate an external body contour. This contour was exported as a DICOM structure, converted to an OBJ format, and imported into the Unity game engine. Within Unity, four different holographic shaders were used to affect different MixR visualizations of the body surface including Fresnel-derivative, semitransparent, silhouette, and wireframe renderings. A MixR application was created to enable the registration of the holograms with a TrueBeam linear accelerator using marker-based and marker-less tracking. Nine users performed 2-3 trials in which they wore the HoloLens 2 to view the holographic reference surface while aligning the phantom accordingly. CBCT was used to quantify the offset from isocenter, and for each trial users indicated their confidence in the alignment. Mean absolute errors (MAE) of vertical, longitudinal, and lateral offsets were evaluated, along with user confidence, setup time, and hologram preference. Each user completed a questionnaire to determine a score using the system usability scale (SUS).

Results: MAE was lowest for the silhouette and Fresnel-derivative holograms at approximately 3 mm, though, a one-way ANOVA revealed no statistically significant difference between the shaders. Average setup time was approximately 4 minutes with faster setup favoring silhouette. Confidence score favored Fresnel-derivative. User preference was 4/9 for silhouette, 2/9 for Fresnel-derivative, 2/9 for wireframe, and 1/9 for semitransparent. The mean system usability score was 75.6, indicating “good” and “acceptable”.

Conclusion: More trials are planned to reveal a statistical difference between the accuracy of the holograms, but the existing results imply that user preference should be the guide. Future trials are planned to evaluate against current SGRT systems and also pilot the system clinically.

Keywords

Surface Matching, Registration, Patient Positioning

Taxonomy

TH- RT Interfraction Motion Management: Development (new technology and techniques)

Contact Email

Share: