Patient dose optimization for computed radiography using physical and observer-based measurements as image quality metrics

Abstract Radiation protection of patients undergoing diagnostic x-ray examinations requires practical evaluation of doses and image quality under clinical conditions. On this subject, optimization of the dose-image quality relationship plays key role in order to achieve this goal. In this study, patient dose optimization was implemented for a computed radiography (CR) system used in general x-ray examinations, considering physical and observer-based measurements as image quality metrics. An anthropomorphic phantom was used to simulate the patient under clinical conditions of chest and abdominal x-rays. Entrance surface doses (ESDs) were measured using a solid-state dose detector positioned at phantom entrance surface during simulated x-rays with different combinations of tube potential (kV) and tube current-time product (mAs), including the kV-mAs used clinically. Agfa's CR System with CR12-X digitizer and a set of 35x43cm cassettes and imaging plates (IP) were employed to capture digital images. Contrast-to-noise ratio (CNR) determined from different regions in the acquired images was used as physical measurement of image quality. Two experienced radiologists evaluated the images qualifying them in terms of acceptable noise. The relationship between calculated CNR and ESD measured for each exposure setting in association to the acceptable images by radiologists were employed as optimization strategies: maximize CNR for a constant dose, minimize dose for a constant CNR and, finally, maximize the figure of merit (FoM) that relates CNR and dose. Prior to the dose and image quality optimization, standardized exposure index (EI) from clinically accepted images and its associated deviation index (DI) were collected for one month. Large reductions in ESD (up to 65%) with clinical image quality assurance were achieved with optimization strategies. The results indicate that the physical parameters of digital image quality assessment validated by experts observation can be an efficient way in the practice of optimization.