Hard-copy versus soft-copy with and without simple image manipulation for detection of pulmonary nodules and masses

Purpose: To compare interpretation performance on soft-copy presentations, with and without simple image manipulation, and on unmodified hard-copy pre sentations with regard to detection of pulmonary nodules and masses. Material and Methods: Fifty chest digital radiograph combinations of patien ts with a total of 60 nodules, 32 of which were 2.0 cm in diameter, were se lected for the study. Three readers evaluated three separate image formats: unmodified hard- and soft-copies, and soft-copies with simple image manipu lation of lung and mediastinum window settings, and zooming. The screen dis play was 1600x1200 pixels with 8 bits/pixel. Results: The sensitivity, accuracy, detectability, and Az value of the soft -copy systems were clearly inferior to hard-copy evaluation. The mean Az va lues were 0.921 for unmodified hard-copy, 0.820 for image-manipulated soft- copy, and 0.781 for unmodified soft-copy. Conclusion: Soft-copy interpretations were not as sensitive in detecting pu lmonary nodules and masses as hard-copy evaluation.