Reliability and validity of splenic volume measurement by 3‐D ultrasound

Abstract

We prospectively evaluated the reliability and validity of splenic volume with 3-D ultrasound measurement and clarified its clinical usefulness. Thirty healthy volunteers and 30 patients with cirrhosis were included in this study. All 3-D ultrasound examinations of splenic volumes were performed twice by two experienced sonographers with transabdominal ultrasound using virtual organ computer-aided analysis (VOCAL). Reliability was confirmed among all subjects by evaluating within-observer repeatability and between-observer reproducibility using intraclass correlation coefficients (ICC) and Bland-Altman plots. Overall between-instrument agreement of the measurements and computed tomography (CT) volumetry among cirrhotic patients were performed to determine validity. For all 240 examinations, 3-D ultrasound visualization and measurement of the spleen volume was possible. Mean spleen volume was 104.0 mL for the volunteers and 283.5 mL for the cirrhotic patients. The repeatability was high, with ICC (95% confidence interval) of 0.996 (0.993-0.997) for observer A and 0.997 (0.994-0.998) for observer B. Moreover, the interobserver ICC was 0.996, indicating high reproducibility. Despite the difference in volume between the volunteers and cirrhotic patients, sensitivity analyses indicated consistent results for both groups. Regarding the validity of the 3-D ultrasound measurement, it also showed moderate to high agreement with CT volumetry, with mean ICC of 0.922 and 0.924 for observers A and B, respectively. The reliability and validity results from the Bland-Altman plots were similar to those from the ICC, with limits of agreement consistently narrow from a clinically practical view. 3-D ultrasound measurements using VOCAL are valid and reliable in spleen volume examinations.