Ruchi Deshpande MS, Presenter: Nothing to Disclose

Anh Hong Tu Le PhD, Abstract Co-Author: Nothing to Disclose

John J. Demarco PhD, Abstract Co-Author: Nothing to Disclose

Daniel A. Low PhD, Abstract Co-Author: Scientific Advisory Board, ViewRay, Inc

Patrick Kupelian MD, Abstract Co-Author: Consultant, ViewRay, Inc Consultant, Accuray Incorporated Speakers Bureau, Siemens AG Research Grant, Varian Medical Systems, Inc License agreement, VisionTree Software, Inc

Brent Julius Liu PhD, Abstract Co-Author: Nothing to Disclose

Radiation therapy treatment plans are determined by patient anatomy, which often limits the dose to the tumor and the degree of protection to surrounding organs-at-risk. Targeting the tumor sometimes holds priority over limiting damage to normal tissue, leading to radiation toxicity. Since the dose distribution and treatment plan are determined by the patient’s anatomy, it is possible that different patterns and combinations of anatomical features, plan parameters and dose characteristics lead to specific radiation toxicity outcomes. Our decision support module uses cloud computing for discovering and utilizing these patterns, thereby obviating the need to download and install software, by providing Software as a Service (SaaS). This follows current trends in Radiation Oncology departments, which are trying to move away from traditional in-house stand-alone workstations, towards a client-server architecture.

We have collected 80 treatment-planning data sets of patients who have undergone radiation therapy for prostate cancer. This data includes CT slices, DICOM RT Dose, Structure Set, Plan as well as quantified radiation toxicity outcomes. We are using this data to test our algorithms and evaluate the workflow of the system’s ability to predict toxicity outcomes in a cloud-computing environment.

We have created a knowledge base by quantifying anatomy and radiation toxicity outcomes of retrospective patients. This can be used to predict the radiation toxicity of future patients, or to search for treatment plans of previous patients with similar anatomy in order to optimize treatment for new patients. Our decision support tools are embedded in a cloud-based web application that features several presentation and visualization tools for analyzing treatment data.

We have developed a web application that utilizes cloud computing and quantifies patient anatomy using imaging studies, in order to categorize radiation toxicity risks associated with external beam radiation therapy. The methods and results of this work can also be applied to other computationally intensive post processing workflows in radiology.

Deshpande, R, Le, A, Demarco, J, Low, D, Kupelian, P, Liu, B, Cloud Computation of Anatomical Features from Imaging Studies to Discover Radiation Toxicity Trends Using a DICOM-based Decision Support System.  Radiological Society of North America 2013 Scientific Assembly and Annual Meeting, December 1 - December 6, 2013 ,Chicago IL. http://archive.rsna.org/2013/13026281.html