Improving treatment plan evaluation with automation

Abstract

The goal of this work is to evaluate the effectiveness of Plan-Checker Tool (PCT) which was created to improve first-time plan quality, reduce patient delays, increase the efficiency of our electronic workflow, and standardize and automate the phys-ics plan review in the treatment planning system (TPS). PCT uses an application programming interface to check and compare data from the TPS and treatment management system (TMS). PCT includes a comprehensive checklist of automated and manual checks that are documented when performed by the user as part of a plan readiness check for treatment. Prior to and during PCT development, errors identified during the physics review and causes of patient treatment start delays were tracked to prioritize which checks should be automated. Nineteen of 33checklist items were automated, with data extracted with PCT. There was a 60% reduction in the number of patient delays in the six months after PCT release. PCT was suc-cessfully implemented for use on all external beam treatment plans in our clinic. While the number of errors found during the physics check did not decrease, automation of checks increased visibility of errors during the physics check, which led to decreased patient delays. The methods used here can be applied to any TMS and TPS that allows queries of the database.

Figure 1

Schematic of the PCT architecture which includes a framework upon which checkers are built. Data can be accessed from multiple sources.

Figure 2

Example of the GUI of PCT with a schematic of the hierarchy added. The highest level of the hierarchy, the site, can be selected at the top left (this version has only one site, “Default”). The next level under site are the stages. In this version of PCT, there are 5 stages which are displayed in blue and the top of the display. All stages contain unique checkers (e.g., “Check dataset name against standards”) that are listed under the stage title (e.g., “Stage 1: Prior to planning”).

Figure 3

An example of flagged, passing, report, and manual checks. In this example, “Verify origin is set correctly” is a manual check that was toggled to “M” once manually verified. The “Check consistency of RX” is flagged due to a mismatch between the planned and prescribed energy. The green checkmark for “Check dose limits in Reference Points” indicates that the session dose limits match the planned reference point dose per fraction and total dose. “Verify beam energy and modality” is a manual check that has not been toggled. The bolus check displays a report symbol because structure and tray bolus were not found in the plan.

Figure 4

Treatment plan review workflow with Plan‐Checker Tool (PCT).

Figure 5

Comparison of errors found in the physics check before and after release of PCT with checkers that became automated.

Figure 6

After release of PCT, a number of delays have been eliminated. One prescription delay occurred because PCT was not run on a simulation on‐set plan. Retraining was done to ensure all plans have a PCT report prior to treatment.