Depth variation bias and interaction with gain setting in ultrasonic tissue characterization by integrated backscatter analysis.

Overview

abstract

Integrated backscatter signal (IBS) has been proposed as a tool to measure cardiac fibrosis. To overcome problems associated with machine settings and attenuation of the chest wall, IBS has been expressed in relation to posterior pericardium, as a variation across cardiac cycle, or both. Depth of the reflecting structure has never been considered as a source of variability. Accordingly, we studied the effect of structure depth on IBS and examined its interaction with gain setting. Backscatter signals were recorded from plastic phantoms containing identical structures set at increasing depth and in 1 healthy volunteer using silicone spacers to modify depth, on a wide range of gain settings. In the phantom, IBS signal linearly decreased with increasing depth and nonlinearly increased with increasing gain (all r(2) > 0.97). In the healthy volunteer, results from septum were very similar to the phantom experiment. Values of septal IBS were adjusted using multiple regression coefficients for gain and depth from the phantom experiment and resulted in a near-complete offset of effect of depth and gain on septal IBS (P = not significant for both gain and depth). These assumptions were also used to compare IBS analysis between hypertensive patients and patients with hypertrophic cardiomyopathy. Thus, depth and its relation with gain should be taken into account and might be almost fully predicted. Using appropriate regression modeling may allow analysis in optimal imaging conditions, tolerating between-patient comparisons even in limited diastolic frames.