Transitioning from subjective to objective testing of a digital detector

Abstract

Background: Objective testing methods for digital detectors not only eliminate the inherent human errors and inconsistencies associated with subjective testing methods, but also provide a more detailed assessment of digital detectors. At the Major National Hospital, digital detector testing currently relies heavily on subjective tests. Thus, there is a clear need for the development of an objective quality assurance (QA) methodology for digital detector testing. Objectives: To develop the best objective methodology and standard operating procedure (SOP) for the signal transfer property (STP), image uniformity, dark noise, modulation transfer function (MTF), normalised noise power spectrum (NNPS), detective quantum efficiency (DQE) and low contrast-detail detectability tests. Methodology: A draft SOP outlining multiple methods for each test was initially developed and evaluated on a table detector. The methods chosen were based on quantifiable and reproducible image quality metrics, and required the use of multiple phantoms, such as the CDRAD and TO MTF phantom, and several software tools, such as ImageJ, COQ and DRIQ. Based on the results obtained, a finalised methodology was established, repeated for reproducibility, and then applied to a wall and a wireless detector to ensure its broader applicability. Results: All STP plots yielded R 2 values above 0.99 while coefficient of variation values for the image uniformity were all below 4%. The maximum percentage difference for the dark noise was 1.1%. MTF plots from the MTF Analyser and DRIQ (both using 30 mm ROIs) showed good agreement, whereas those from COQ and DRIQ (IEC ROIs) were computed incorrectly. NNPS plots from COQ and DRIQ showed no significant differences, though DQE plots from COQ were found to be overestimated. Finally, PMMA thickness, mAs, imaging protocol, and image format were found to affect the contrast-detail curves and IQFINV values. Conclusions and Recommendations: An efficient, reproducible and reliable QA methodology for digital detector testing was developed. Such methodology was validated on three different detectors as part of a single digital radiography system. For future research, it was recommended that the developed methodology should also be validated on other systems.