Radiation therapy plan checks in a paperless clinic

Abstract

Traditional quality assurance checks of a patient's radiation therapy plan involve printing out treatment parameters from the treatment planning system and the "record and verify" (R&V) system and visually checking the information for one-to-one correspondence. In a paperless environment, one can automate this process through independent software that can read the treatment planning data directly and compare it against the parameters in the R&V system's database. In addition to verifying the data integrity, it is necessary to check the logical consistency of the data and the accuracy of various calculations. The results are then imported into the patient's electronic medical record. Appropriate workflows must be developed to ensure that no steps of the QA process are missed. This paper describes our electronic QA system (EQS), consisting of in-house software and workflows. The EQS covers 3D conformal and intensity modulated radiation therapy, electrons, stereotactic radiosurgery, total body irradiation, and clinical set ups with and without virtual simulation. The planning systems handled by our EQS are ADAC Pinnacle and Varian FASTPLAN, while the R&V systems are LANTIS and VARIS. The improvement in our plan check process over the paperless system is described in terms of the types of detected errors. The potential problems with the implementation and use of the EQS, as well as workarounds for data that are not easily accessible through electronic means, are described.

Figure 1

Screen shot of LEX. The numbered areas correspond to the following descriptions: (1) Prescription and beam data are read directly from Pinnacle's data file; (2) The transfer software looks for conditions that do not follow our clinical practice; (3) Pinnacle has no provision for course or prescription number; the operator can select a course number and prescription number in order to create a unique ID for each beam; (4) Pinnacle has no fields for table position; the operator can edit the default values determined by LEX from the isocenter; (5) The Siemens ONCOR linear accelerators in our clinic have matched beam energies, so the operator must select a machine from the list; (6) The operator can choose to have the beams gated, re‐order the segments in multi‐segment beams to minimize the delivery time, and/or convert multi‐segment beams into a series of individual beams to handle gated IMRT; (7) Pinnacle does not provide tolerance tables, so the operator must choose the proper one from the drop down list; (8) The button to create the RTPLINK file is disabled until all errors are cleared; this prevents the operator from sending incomplete data; (9) By checking the QA box, beams will be created that have a unique “QA” field ID; (10) These buttons control MLC options such as adding a 2 mm gap between closed leaf pairs that are blocked by the collimators; (11) Unresolved errors are displayed in this box. Once all the errors have been cleared, the button to create the transfer file is enabled.

Figure 2

Screen shot of the RTP‐Filter. The numbered areas correspond to the following descriptions: (1) The prescriptions, beams and segments can be selected by the operator; (2) The position parameters (gantry, collimator, couch, MLC leaves) of each segment can be viewed in the panels in this section; (3) More information about the plan, the selected prescription, field and MLC, the dose, beam IDs, interlocks and intensity maps can be reviewed by selecting the appropriate tabs; (4) Pinnacle plan to LANTIS field comparisons are done when the user performs a dose calculation using the LANTIS data; calculation parameters for flash and effective depths can also be modified in this tab.

Figure 3

Sample MU check report from LANTIS. (Note: patient identifiers and physicist's signatures were removed). The calculation uses the field parameters from LANTIS, and the patient SSD and depth information from Pinnacle as a starting point. The depths are modified by the physicist to account for the 3 dimensional nature of the heterogeneities and patient contours. All dosimetric parameters were taken from an electronic version of the dosimetry calculation book. The parameters in the report are effective parameters, accounting for the variations among the segments for each beam.

Figure 4

Sample IMRT QA report from LANTIS. (Note: patient identifiers and physicist's signatures were removed.) The QA method in this report is the 2D calculation using a ray trace through a solid water phantom. All SSDs and depths are calculated from intersections of each ray line with the phantom surface planes. This is done for each point on the $2\text{mm}\times 2\text{mm}$ grid on the coronal plane through isocenter. The resulting calculated “virtual film” (middle left) is compared against the corresponding $4\text{mm}\times 4\text{mm}$ Pinnacle calculation (top left) in order to determine the error (bottom left), distance to agreement (top right), and decision (middle right) maps. The calculated data were interpolated onto a common $1\text{mm}\times 1\text{mm}$ grid.

Figure 5

An overview of the electronic plan check workflow. The workflow starts from the TPS and ends with the treatment. $\text{Rectangles}=\text{processes or software}$, $\text{diamonds}=\text{decisions}$, $\text{trapezoids}=\text{manual input}$, $\text{rectangles with curved bottom}=\text{files}$. The boxes are color coded by function: $\text{green}=\text{dosimetrists}$, $\text{blue}=\text{physicists}$, $\text{yellow}=\text{therapists}$. The abbreviations are as follows: $\text{TPS}=\text{treatment planning system}\left(\text{Pinnacle}\right)$, $\text{LEX}=\text{in‐house plan transfer software}$, $\text{RTP‐C}=\text{RTP‐Connect file}$, $\text{R\&V}=\text{Record and Verify}\left(\text{LANTIS}\right)$, $\text{Rx}=\text{prescription}$, $\text{IE}=\text{initial export}$, $\text{Filter}=\text{RTP‐Filter}\left(\text{an in‐house application for plan checking}\right)$, $\text{LC}=\text{logically consistent}$, $\text{PM}=\text{Plan matches R\&V}$, $\text{PQ}=\text{Plan quality is acceptable}$, $\text{SC}=\text{patient setup consistent with the plan}$, $\text{FE}=\text{final export}$.