Quantitative ultrasound imaging over the ischial tuberosity: An exploratory study to inform tissue health

Characterization of ischial tissue health using a standardized diagnostic ultrasound protocol capturing thickness and gray scale analysis has not been established.

This study evaluates inter-participant and inter-trial reliability of thickness and gray scale analysis of ultrasound images of tissues overlying the ischial tuberosity. It provides recommendations for the number of images required to minimize the standard error of measurement (SEM) and determines the number of images required for thickness, gray scale and contrast values that exceed an a-priori minimal detectable change (MDC) for repeated tissue assessment.

Brightness mode ultrasound images using a 12 MHz linear probe were collected on the dominant limb in the side lying position for ten healthy participants and partitioned into three regions of interest: skin, subcutaneous tissue and muscle. Thickness and gray scale measures of skin, muscle and subcutaneous tissue were calculated using a customized MATLAB program. Contrast of each region of interest was calculated using the Gray Scale Level Co-Occurrence Matrix. Generalizability theory was used to quantify indices of dependability and corresponding SEMs and MDCs with 90% Confidence Intervals.

Participants accounted for most of the total variance (75.56% to 94.78%). Coefficient of dependability (ϕ) for thickness, grey scale and contrast measures was greater than 0.80 when more than two images were averaged. In order to detect a MDC of 21% in thickness and echogenicity measures, at least three images are required, while at least 5 images are required for a MDC of 25% for contrast measures.

Obtaining reliable thickness, echogenicity and contrast measures of tissue overlying the ischial tuberosity can be achieved from two ultrasound images by a single therapist on an individual participant however three and five images are required to use a MDC of 21% for thickness measures and MDC of 25% for contrast measures respectively.