Brachyverifier: An Automated System for Plan Quality Assurance in High-Dose-Rate Brachytherapy

Abstract

To improve the quality assurance (QA) process of high-dose-rate brachytherapy (HDR) by automating the physics plan check process. BrachyVerifier is an application written in Java for the automation of plan QA in HDR. The program extracts information from the record-and-verify (R&V) data stored in an ARIA database (Varian Medical Systems, Inc., Palo Alto, CA) and from portable document format (PDF) files of (i) the plan printout from the Oncentra Brachytherapy planning system (Nucletron, an Elekta company, Elekta AB, Stockholm, Sweden), (ii) the written directive, and (iii) the catheter-length measurement report. Checks are customized to the type of plan, which is detected from the plan printout. If more than one type of plan is compatible with the plan printout, the user is prompted to choose the plan type. Check speed was assessed through a review of the time elapsed between plan approval and final plan check in the 43 tandem-and-ring fractions that occurred between July 2013 and November 2013. Statistical significance (p < 0.05) was assessed using an unpaired Student T-test. Checks specific to a given type of plan have been developed for the following types: tandem and ring, tandem and ring + interstitial needles, tandem and ovoids, tandem and ovoids + interstitial needles, tandem + interstitial needles, interstitial needles only (gynecologic and non-gynecologic), straight line cylinder, graphical cylinder and surface applicator. Other types of plans are processed under an “other” category, with no type-specific checks. BrachyVerifier was written and tested to perform the following checks:-Secondary calculation: the point doses in the plan printout are compared to the dose resulting from a 1D TG43 calculation. Tolerance is 5%.-Channel assignments: the channel assigned to each catheter is compared to our clinical practice for a given plan type (e.g., tandem to channel 3). Match must be exact.-Catheter lengths: the catheter lengths in the plan printout are compared to either a) the measurement report or b) the lengths that are expected for a given plan type (e.g., 1500 for tandem). Match must be exact.-Written directive compliance: the fractionation scheme, number of catheters and step size indicated in the written directive are compared to the plan printout. Match must be exact.-R&V check: the patient name, fractionation scheme and plan properties (such as TRAK value) in the plan printout are compared to the R&V system. TRAK tolerance is 1%. Three documents are generated: a summary sheet of all the checks being performed and their results, the results of the secondary calculation and a worksheet summarizing the plan information to be checked during treatment. Since its introduction into clinical use in September 2013, BrachyVerifier has been used to check 51 HDR plans. Mean physics check time for tandem and ring cases was 1217 seconds with BrachyVerifier (n = 23) and 1403 seconds without (n = 20; p = 0.045). BrachyVerifier has standardized the physics checks of HDR plans and produced a significant reduction in plan check time. The integration of BrachyVerifier into our clinical practice automated the checks on channel assignments, catheter lengths and prescription, thus reducing the possibility of missing errors that occurred during planning. BrachyVerifier has not eliminated the need for plan checking by a physicist. Moreover, checks such as correct attachment of transfer tubes and other safety-related checks cannot be automated via software. Although the exact implementation of BrachyVerifier is specific to our clinical practices and equipment, the general architecture is portable from one clinic to another.