Amoon Jamzad1, Seyed Kamaledin Setarehdan1. 1. Control and Intelligent Processing Center of Excellence, School of Electrical and Computer Engineering, College of Engineering, University of Tehran, Tehran, Iran.
Abstract
OBJECTIVES: The physical structures of renal stones are highly correlated with their breakability. Noninvasive estimation of stone roughness will be beneficial for management. The intensity of the twinkling artifact appearing at the site of renal stones on Doppler ultrasound imaging is also influenced by the stone's roughness level. This article proposes a quantitative method for roughness prediction of ex vivo renal stones based on a twinkling analysis of their color Doppler images. METHODS: Twenty surgically removed renal stones were first spatially modeled by an optical method, and 12 standard roughness measures were extracted from them. Stones were then embedded in an agar-based phantom and Doppler imaged with a calibrated ultrasound system. The images were preprocessed, and 11 twinkling intensities were measured numerically. The twinkling data along with the roughness labels were then analyzed by multiple linear regressions, and finally, a linear roughness predictor was trained for renal stones. RESULTS: The core height measure of roughness had the best linear fit to the twinkling data among other roughness parameters. The results of the multiple linear regression analysis indicated a strong linear relationship between twinkling data and stones' roughness, with an R2 value of 83.29% and high statistical significance of F(11,868) = 393.36 and P < .001. CONCLUSIONS: It was possible to predict the core roughness of renal stones using the proposed method and the twinkling artifact data acquired from the color Doppler images ex vivo.
OBJECTIVES: The physical structures of renal stones are highly correlated with their breakability. Noninvasive estimation of stone roughness will be beneficial for management. The intensity of the twinkling artifact appearing at the site of renal stones on Doppler ultrasound imaging is also influenced by the stone's roughness level. This article proposes a quantitative method for roughness prediction of ex vivo renal stones based on a twinkling analysis of their color Doppler images. METHODS: Twenty surgically removed renal stones were first spatially modeled by an optical method, and 12 standard roughness measures were extracted from them. Stones were then embedded in an agar-based phantom and Doppler imaged with a calibrated ultrasound system. The images were preprocessed, and 11 twinkling intensities were measured numerically. The twinkling data along with the roughness labels were then analyzed by multiple linear regressions, and finally, a linear roughness predictor was trained for renal stones. RESULTS: The core height measure of roughness had the best linear fit to the twinkling data among other roughness parameters. The results of the multiple linear regression analysis indicated a strong linear relationship between twinkling data and stones' roughness, with an R2 value of 83.29% and high statistical significance of F(11,868) = 393.36 and P < .001. CONCLUSIONS: It was possible to predict the core roughness of renal stones using the proposed method and the twinkling artifact data acquired from the color Doppler images ex vivo.
Authors: Christine U Lee; Matthew W Urban; A Lee Miller; Susheil Uthamaraj; James W Jakub; Gina K Hesley; Benjamin G Wood; Nathan J Brinkman; James L Herrick; Nicholas B Larson; Michael J Yaszemski; James F Greenleaf Journal: Eur Radiol Exp Date: 2022-06-17