Objective: To determine the effect of noise-reducing innovation-precision imaging (PI)-on image quality and diagnostic efficacy in breast ultrasound.
Methods: The study, which assessed four levels of PI from zero to three, consisted of two parts: image quality assessment and diagnostic efficacy evaluation. For the first part, 247 sets of ultrasound images displayed at each PI level were evaluated by 6 experienced breast imaging observers, by rating image quality using visual grading analysis on a 1-4 scale. For the diagnostic efficacy part 51 breast lesions were displayed at each PI level and scored 1-6 to generate a receiver operating characteristic (ROC) curve. These images were evaluated by radiologists and sonographers. Analyses were performed using non-parametric Friedman and Wilcoxon signed rank tests and a multireader multicase methodology.
Results: Statistically, higher scores of image quality were observed with increased levels of PI than with the zero setting (p < 0.001). The ROC analysis did not demonstrate any significant change in diagnostic efficacy, with mean scores for all observers being 0.79, 0.80, 0.81 and 0.81 for settings zero, one, two and three, respectively.
Conclusion: This study suggested a perceived improvement in image quality with increasing levels of PI; however, no changes in diagnostic efficacy were noted. The importance of looking at the impact of new imaging technologies in a multifaceted way is emphasized.
Advances in knowledge: To our knowledge, this is the first article investigating the impact of the PI algorithm on ultrasound image quality and breast lesion characterization.