A Computer Assisted Treatment Events Recognition System (CATERS) For Radiation Therapy
J Xia1*, C Mart1, J Bayouth1,2, (1) University Of Iowa, Iowa City, IA, (2) University of Wisconsin, Madison, WIWE-G-108-1 Wednesday 4:30PM - 6:00PM Room: 108
Purpose: To develop an automated treatment event detection system to safeguard radiation therapy patients by analyzing electronic treatment records and reporting treatment events.
Methods: CATERS is to make the treatment monitoring process more efficient by offloading repetitive tasks to the computer. Treatment events are defined as possible deviations from physician's intention, such as inconsistent or wrong treatment parameters. CATERS consists of a user interface, the CATERS core, and a data interface. CATERS core is responsible for all logic analysis and treatment event recognition. Through the user interface, users can access CATERS from either mobile devices or desktop computers. The data interface provides methods for retrieving electronic treatment records. A five months study was conducted to compare CATERS with manual chart checks.
Results: From July 2012 to November 2012, 491 patients were treated at our clinics for total of 7692 fractions. From these fractions, 553 events were detected by CATERS, none of the events has significant dosimetric impacts on patient treatments. On average, it takes about one second to check one patient. During the trial, CATERS was able to efficiently detect all events that were discovered with a manual chart check when used side-by-side. The top three types of events were couch position override (3.2%), cone beam CT frequency (1.85%), and significant couch position deviation (1.31%). CATERS is clinically deployed and routinely used by all the physicists in our clinics.
Conclusion: CATERS offers an effective method to detect and report treatment events. With CATERS, radiation treatments can be monitored and treatment events can be reported on a daily or even an hourly basis. Caters has the potential to improve treatment check efficacy and minimize the possibility of a misadministration.