The relation between different system gain settings and the accuracy of densitometric assessment of echogenicity.

Abstract

Two mathematical transformations of the measured density of a sonographic image of the fetal liver were used in 19 patients at nine different system gain settings (G) (optimal G number [dB] and optimal G number +/- 2, +/- 4, +/- 8, +/- 12). A significant correlation was found between the film-corrected density (Dcor) and gain interval from the optimal gain (Dcor = 1.39-0.052 [G interval], R = -0.81, P < 0.00001, 95% confidence limits = -0.052 +/- 0.052). After standardization of the values to the arbitrary G number of 49 (Dcor 49), a gain-independent convergence of the values was found. In 79% of the 19 cases the accuracy of the method was very high, as estimated from a small coefficient of variation. No significant difference was found in comparing the mean standardized densities derived from paired measurements (i.e., optimal G number +/- 2, +/- 4, +/- 8, and +/- 12). The data support background assumptions for the quantitative assessment or echogenicity by transmission densitometry and provide highly confident accuracy estimates.