RADIASOFT LLC — Department of Energy SBIR Phase I: 02a

Description

The advanced manufacturing and engineering sectors of the US economy require turnkey high- performance computing (HPC) software to reliably automate all aspects of the design, manufacture and operation of emerging hardware solutions to societal needs. Robotic systems are a key element of advanced manufacturing. An optimized 4π radiotherapy system is under development, based on a robotic radiation delivery platform, which will significantly improve patient outcomes and reduce the treatment planning and delivery times for hospitals. The associated treatment planning software, developed in the UCLA Department of Radiation Oncology, requires hardening for commercial use. General statement of how this problem is being addressed We will develop a compelling prototype suite of applications, suitable for attracting a large venture capital investment in Phase III, which will integrate all aspects of treatment planning and machine control: 1) import of magnetic resonance imaging (MRI) and computed tomography (CT) images; 2) automatic image registration, with the option for manual adjustments; 3) dose calculation and visualization; 4) dose optimization with template treatment plans and interactive modification; 5) machine learning from accepted plans with the goal of minimizing the time for modification and knowledge sharing; 6) interface to the control system, which includes modeling of the robotic delivery space and maximizing efficiency of the robotic arm movement around the patient. Our treatment planning suite will run on NVIDIA GPUs, including a state-of-the-art cloud-based GUI. What is to be done in Phase I? During Phase I, we will harden the UCLA software for calculating the 3D radiation dose resulting from a specified X-ray source, as follows: 1) develop an automated test suite; 2) refactor as necessary and port to the NVIDIA GPU; and 3) develop a prototype browser-based GUI suitable for cloud computing. We will also harden (in the same three ways) the UCLA software for dose optimization, based on a uniquely powerful algorithm which will be patented. We will test machine learning algorithms on image recognition for relevant MRI and CT images, also exploring the inclusion of GUI events that represent decisions made by the physician. Commercial Applications and Other Benefits RadiaBeam Technologies, LLC and the UCLA Department of Radiation Oncology are collaborating to develop an optimized 4π radiotherapy system based on a robotic radiation delivery platform, with an agreement to create a new VC funded company that will obtain FDA approval and sell the integrated hardware and software system to hospitals and radiation oncology centers around the world. RadiaSoft will license our software (some of it to be jointly owned with UCLA) to the new firm, and we also expect some after sales service and support contracts to be passed through to RadiaSoft.