OA038] Does automation reduce the number of errors in quality control of treatment plans for external beam radiotherapy?

Abstract

Purpose Any single treatment plan error should be detected and corrected prior to treatment either during a check procedure or by built-in safety features of the treatment planning (TPS) or record and verify systems. However, as delivery techniques have become increasingly complex the number of possible errors in a plan has increased dramatically. It is therefore desirable to automate as many check procedures as possible to eliminate manual errors. In this work we investigate the effect on error rates of introducing automation in quality control of patient treatment plans. Methods Dose constraints, fractionation and best practice guidelines for all treatment schemes in our clinic taken from relevant guidelines (institutional, national, international or clinical trial) were collected in a database. A TPS script was written to generate a report comparing plan information with reference values from the database as pass/fail criteria. To determine if automation reduces the number of errors compared to manual quality control, 322 consecutive plans approved for treatment with manual quality control between September 1st and October 1st 2017 were retrospectively subjected to automated quality control with the script. All errors were recorded and severity was scored using the recommendations from the AAPM TG-100 report. Results 320 errors were detected in 10,243 individual checks (3.1%). Three errors were found to have had impact on target dose, ranging from 0.5% (severity 5) to 7% (severity 7), while another 18 could have caused either geographic or dosimetric impact (severity 5+). The remaining 299 errors were either purely clerical or could at worst cause minor inconvenience to staff, severity score 1–2. Conclusions Automation of treatment plan quality control reduces error rates and increases adherence to guidelines compared to a purely manual workflow.