Image quality and diagnostic performance of silicone-specific breast MRI

Purpose: To compare the image quality of three techniques and diagnostic performance in detecting implant rupture. Materials and Methods: The study included 161 implants for the evaluation of image quality, composed of water-saturated short TI inversion recovery (herein called "water-sat STIR"), three-point Dixon techniques (herein called "Dixon"), and short TI inversion recovery fast spin-echo with iterative decomposition of silicone and water using least-squares approximation (herein called "STIR IDEAL") and included 41 implants for the evaluation of diagnostic performance in detecting rupture, composed of water-sat STIR and STIR IDEAL.Six image quality categories were evaluated and three classifications were used: normal implant, possible rupture, and definite rupture. Results: Statistically significant differences were noted for the image quality categories (p. <. 0.001). STIR IDEAL was superior or equal to water-sat STIR in all image quality categories except artifact effects and superior to Dixon in all categories. Water-sat STIR performed the poorest for water suppression uniformity.The sensitivity and specificity in detecting implant rupture of STIR-IDEAL were 81.8 % and 77.8 % and the difference between two techniques was not statistically significant. Conclusion: STIR-IDEAL is a useful silicone-specific imaging technique demonstrating more robust water suppression and equivalent diagnostic accuracy for detecting implant rupture, than water-sat STIR, at the cost of longer scan time and an increase in minor motion artifacts.