3685 - Feasibility of Low Latency kV Projection Steaming for Position Verificationand Intrafraction Motion Monitoring in Ultrahypofractionated Radiotherapy

Purpose/Objective(s): Ultrahypofractionated treatment regimens, including single fraction SBRT, stereotactic arrhythmia radioablation and PULSAR, are quickly emerging. Position verification and intrafraction motion monitoring is critical for these procedures. However, many patients have artifact producing implanted devices that can challenge accurate position verification on cone beam CT (CBCT) images. Several commercial treatment delivery platforms support acquisition of real-time 2D kV projections, which often demonstrate reduced artifacts. However, flexible and dynamic display of motion monitoring structures on 2D projection images can be challenging, as these platforms may not support overlay of target and organs at risk (OAR) contours on the real-time projection steam. We demonstrate here the feasibility of using alternative motion monitoring systems for position verification and intrafraction motion monitoring for ultrahypofractionated radiation therapy.

Materials/Methods: A computer workstation running a research version of software with a real-time API was installed on the local machine network of a Versa HD linac from a precision radiation medicine company, enabling gigabit connectivity to XVI R5.0. An XVI projection streaming client (PSC) was also installed on this workstation. The PSC receives raw images and metadata from XVI but does not permit overlay of target or OAR contours. Custom Python code was developed to receive images and metadata from the PSC, convert images to DICOM, and transmit to the commercially available software real-time API using HTTP requests. For testing, a motion phantom was placed on the linac couch and a 3D kV CBCT was acquired and exported to commercially available software. Two sample motion monitoring contours were drawn on the phantom. XVI was then configured to stream real-time static gantry (MotionView) and rotating gantry (VolumeView) projections to commercially available software to simulate position verification and intrafraction motion monitoring, respectively.

Results: Real-time kV projection images of the motion phantom were successfully streamed to the commercially available software real-time API for both static and rotating gantry acquisitions. System latency was estimated to be approximately 280–380 ms, depending on XVI configuration. Streaming projections to software enabled users to dynamically toggle display of contours to overlay on the projection stream for motion monitoring.

Conclusion: We developed a low latency projecting streaming approach using alternative motion monitoring systems to facilitate position verification and intrafraction motion monitoring in ultrahypofractionated radiotherapy.