Encrypted login | home

Program Information

QA Methodology for CBCT Flat Panel Calibration

Q Han

Q Han1*, J Park1 , G Yan1 , P Bassett1 , S Lee2 , S Samant1 , (1) University of Florida,Gainesville, FL, (2) University Hospitals Cleveland Medical Center, Cleveland, OH


SU-I-GPD-J-48 (Sunday, July 30, 2017) 3:00 PM - 6:00 PM Room: Exhibit Hall

Purpose: To provide practical and fast QA tools for evaluation of individual steps in the CBCT flat panel calibration chain as an alternative to the CATPHAN imaging QA carried out at the end of the calibration sequence.

Methods: CBCT flat panel calibration consists of implementing a vendor supplied “bad pixel” utility, based on a reference image acquired during panel installation, to interpolate across bad pixels; and measuring a pixel response map during kV irradiation using a single level gain (SLG) or multilevel gain (MLG) calibration. Following this, a CATPHAN phantom is imaged and assessed for image quality. To improve the efficiency of this procedure we have developed QA tools to identify bad pixels, measure the modulation transfer function (MTF), identify inter-panel variations in pixel intensity uniformity across the 9 subpanels of the flat panel, and measure the 1D noise power spectrum (NPS). Bad pixels are identified using two-level Harr wavelet transform. The MTF is measured using a bar pattern from single kV exposure. Inter-subpanel uniformity is measured by comparing mean pixel intensity and standard deviation for each subpanel. NPS is measured using an in-air open field image. All the images were acquired on Elekta XVI CBCT.

Results: MTF measurements were accurately measured with the bar pattern, and remained stable since the installation of CBCT panel. Bad pixels were correctly identified by the wavelet transform based on a visual assessment and typically were addressed on annual basis using the bad pixel utility. NPS and inter-subpanel uniformity were used to assess the goodness of the SLG or MLG calibrations, and typically done every 6 months.

Conclusion: The proposed methodology provides the medical physicist or linac engineer with fast and robust tools for an end-to-end assessment of the reliability of a calibration without having to rely on an end-of-chain CATPHAN imaging measurement.

Contact Email: