Application of a simple phantom in assessing the effects of dose reduction on image quality in chest radiography

PurposeFirstly, to evaluate a commercial chest phantom incorporating a quasi anthropomorphic insert by comparing exposure measurements on the phantom with those of actual patients and, secondly, to assess the value of the phantom for image quality and dose optimisation.MethodsIn the first part of the study entrance surface doses (ESD), Beam transmission (BT), and optical density (OD) were obtained for 77 chest radiography patients and compared with measurements made from exposures of the phantom using the respective patient exposure factors from chest examination. Differences were assessed with a student t-test, while the Pearson's linear correlation coefficient was used to test for any linear relationship. The second part assessed the applicability of the phantom to image quality studies by investigating the effect, on the clarity and detectability of lung lesions made from gelatine, of reducing patient dose below current dose levels. Clarity of linear objects of different dimensions was also studied. Lesion detectability and clarity was assessed by four observers. The possibility of extending dose reduction below current dose levels (Dref) was assessed from comparison of doses that produced statistically significant differences in image quality from Dref.ResultsResults show that, with the exception of entrance doses and beam transmission through the diaphragm (P > 0.05), differences in OD and beam transmission between patients and phantom were statistically significant (P < 0.05). A Pearson's correlation test showed good positive correlation in ESD (r = 0.9) and beam transmission (r = 0.8) for all regions and a weaker positive correlation (r = 0.3) for OD in all areas. In both cases the correlation was statistically significant (P < 0.05). Extending dose reduction by between 41 and 65% below Dref produced significant changes in both clarity and detectability.ConclusionWithin limits posed by the observed differences, the phantom can be applied to image quality studies in diagnostic radiology.