Dose reduction in digital chest radiography and perceived image quality.

The aim of the study was to investigate whether radiologists can rank the image quality of digital radiographs with different doses; a preliminary study investigated whether reduced dose images provide sufficient diagnostic quality. Raw data of 40 chest radiographs (posteroanterior (PA) and lateral) obtained with a full-field slot-scan charge-coupled device system in 20 patients with chest pathology were used. Noise was added to simulate reduced dose levels to 50%, 25% and 12%. Four observers ranked the quality of the corresponding images and judged the diagnostic quality. Linear regression analysis was performed. Differences were found in image quality at the different dose levels for both PA (p</=0.0001) and lateral images (p<0.002). The quality was graded sufficient in 159/160 100% dose-level observations and in 159/160 50% dose-level observations. At lower doses, substantially fewer images were graded as having sufficient quality: for PA images, 69/80 for the 25% dose and 63/80 for the 12% dose; for lateral images, 74/80 for the 25% dose and 63/80 for the 12% dose images. In conclusion, observers recognized the reduced quality of lower dose radiographs, although 50% dose imaging was regarded as equal to the 100% dose in having sufficient quality to answer clinical questions. Preliminary findings suggest that a 50% dose reduction seems feasible in a variety of chest pathologies, whereas further dose reduction reduces the diagnostic quality.