Click here to

Session: Therapy: Treatment Verification [Return to Session]

Impact of Delivery and Measurement Uncertainties On the Calculation of Pixel Sensitivity Maps for MV Portal Imagers

M Ahmed1*, H Nourzadeh2, J Siebers3, (1) Vanderbilt University Medical Center, Nashville, TN,(2) Thomas Jefferson University, Philadelphia, PA, (3) University of Virginia Health System, Charlottesville, VA


TH-IePD-TRACK 5-4 (Thursday, 7/29/2021) 3:00 PM - 3:30 PM [Eastern Time (GMT-4)]

Purpose: To describe possible error pathways during calculation of the Pixel Sensitivity Maps (PSMs) and to study the impact on the calculated PSMs with different approaches.

Methods: PSMs are used to normalize EPID's diode-pixels response to the incident radiation. Multiple approaches were developed to calculate EPID-specific PSMs. A shifting-based method uses multiple laterally shifted measurements of same flood field signals to find pixel-gain through utilizing the relative gains of pixels irradiated identically. A shiftless-based method, uses regression surface models to capture low-frequency features of the incident fluence and ascertain the device PSM by removal of these features from signals superimposing high-frequency PSM features with low-frequency fluence features. Here, we use Monte-Carlo calculated EPID signals with introduced errors to study the impact on the reconstructed PSMs. Errors introduced include misalignment of signals used during the PSM calculation, change of output between irradiations and inconsistent background removal from signals used in PSM generation algorithms.

Results: In an ideal setup the shifting-based and the shiftless methods reconstructed the original PSM with 99% and 95% accuracies, respectively. For the shifting method; a one pixel misalignment during the PSM calculation resulted in 90% and 100% of the pixels retained 99% and 90% of the original gain values, respectively. A 0.1% and 0.5% change in linac output between images resulted in 27% and 11% of the gains have 99% of the original values. A 0.1% and 0.5% pixel-to-pixel variance in the individual pixel response resulted in 1.7% and 32% of the pixel gains have 99% and 90% of the original values, respectively. Introduction of the aforementioned errors in the shiftless method worsen the agreement results by a max of ~4%.

Conclusion: The simulation conducted here shows that both methods are accurate in impeccable clinical setups. The shiftless PSM method performs better in sub-standard calibration scenarios.

Funding Support, Disclosures, and Conflict of Interest: This work is supported in part by Varian Medical Systems



    Not Applicable / None Entered.


    Not Applicable / None Entered.

    Contact Email