Reconstructed measurements produced from ultrasound images using a computerized system

Objective The aim of this study was to evaluate the feasibility and accuracy of reconstructed measurements produced from video-printed ultrasound images. Design Reproduction of reconstructed measurements using designated software and comparison of the results with regular measurements derived from the same images. Subjects Seventy regular sonographic examinations of biparietal diameter (BPD) and nuchal translucency (NT) thickness. Methods The study included 35 conventional-size (BPD) and 35 small-size (NT) measurements. Each image was measured twice during a dynamic ultrasound examination (regular measurements) and twice using a computerized system (reconstructed measurements). The latter comprised three steps: (i) computerization (computer-generated images by scanning); (ii) calibration (using the scale located alongside the image); (iii) measurement (placing the calipers at the desired points). Results All images were successfully scanned, and the computer-generated images were of sufficient quality for proper measurement. There was no difference between mean values of regular and reconstructed BPD measurements, 63.9 mm (3.7 mm, SE) and 63.6 mm (3.8 mm, SE), respectively. Similarly, mean values of regular NT measurements were similar to the reconstructed ones, 1.48 mm (0.09 mm, SE) and 1.49 mm (0.09 mm, SE), respectively. Repeated regular BPD and NT measurements exhibited repeatability coefficients of 1.6 mm and 0.4 mm, respectively. These values were smaller than those obtained by repeated reconstructed measurements of 2.1 mm and 0.52 mm, respectively. Conclusions Reconstructed measurements of ultrasound video-printed images are feasible, require modest facilities and exhibit more than reasonable accuracy. This option may contribute to medical research, audit, quality control and training, as well as to medico-legal issues.