SBIR Phase II: Cloud-based Automated Dose Accumulation for Online Adaptive Radiotherapy

The broader impact/commercial potential of this Small Business Innovation Research (SBIR) Phase II project is helping Community Cancer Centers serving the estimated 65% of patients receiving radiation therapies. Cancer is still the second leading cause of death in the United States. Currently accomplished only by significant investment in new equipment, successful therapy hinges upon the ability to delineate and adapt to a precise tumor location "on the day of" radiation delivery. Community Centers cannot afford to replace existing machines with more accurate and expensive modalities. Especially in rural communities with limited resources, few people diagnosed with cancer can afford to travel and incur the cost of staying 4-6 weeks in a local hotel to receive daily treatments from state of the art technologies located only in the wealthy institutions. Using a novel cloud-based architecture and proprietary software, this Phase II award will allow remote clinics, and any hospital, to treat patients with highly improved real-time accuracy. By enabling their existing radiotherapy machines to deliver an adaptive dose to the current location of the tumor, radiation oncologists can significantly improve treatment efficacy, a patient's quality of life, reduce patient re-hospitalizations, and reduce the cost of therapy. This Small Business Innovation Research (SBIR) Phase II project addresses the unmet need to see daily changes in anatomy throughout the course of cancer radiation treatment and effectively target only the tumor. Today, a physician designs a treatment plan and dose prescription based on images that are captured days or even weeks prior to initiation of radiation therapy, creating a margin of error around the tumor that encompasses healthy tissues. Unaccounted anatomical changes are quite frequent, leading to a diminished quality of treatment delivery, inferior outcomes, and a decreased quality of life. The phase I project investigated and developed a treatment-equipment agnostic approach to adapt the treatment to account for these anatomical changes. Phase II will focus on integrating artificial intelligence and deep neural analytics to predict two critical parameters for the treatment adaptivity: (a) treatment prognostics and (b) registration error quantification. Leveraging the company's cloud based, scalable, GPU computational framework, the project will develop and integrate patient-specific biomechanical models to automatically validate results and accurately predict future trends in patient-specific anatomical changes. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.