Click here to


Are you sure ?

Yes, do it No, cancel

IMRT QA Workflow Automation Using a Custom Script

T Wolken*, R Sheu, Mount Sinai Medical Center, New York, NY


(Sunday, 7/12/2020)   [Eastern Time (GMT-4)]

Room: AAPM ePoster Library

Purpose: To improve efficiency by automating the IMRT QA workflow

Methods: Our existing workflow uses Varian’s® Portal Dosimetry (PD) software to analyze IMRT QA and generate a report document. For hospitals/clinics who use Mosaiq® as their record and verify system, this involves manual steps such as importing and linking images. An ESAPI program was created to export the predicted images to a python script that would read in both the predicted and measured images, and perform gamma analysis. The python script used the open source library “PyMedPhys” which contains a built in gamma analysis function. The python script was benchmarked against PD by using both a contrived dose distribution and 20 clinical plans. Gamma passing rate, maximum gamma, average gamma, and gamma histograms were used to understand how PD performed its gamma analysis. For example, we determined which dose distribution was treated as the reference in PD. The python script parameters were updated to match PD.

Results: Average time saved was 2 minutes and 36 seconds per plan. A hospital/clinic with 20 IMRT QAs per week will save approximately 43 hours over the span of 1 year with this script. The maximum percent difference in gamma passing rates between the python script and PD across all tests was less than 1.2%. The average percent difference of maximum and average gamma values was 1.3% and 1.2% respectively.

Conclusion: Certain parameters can have a large effect on gamma passing rate such as the choice of using the measured or predicted image as the reference image and how the maximum dose is chosen for normalization. A deep understanding of how PD conducts gamma analysis allowed for the creation of an automated python script that performs nearly identically to PD. The python script benchmarked in this report is ready for clinical use.


Not Applicable / None Entered.


Not Applicable / None Entered.

Contact Email