Effect of timing of initial scan and interval between scans on Rossavik growth model specification

The effect of the time of the initial scan and the interval between the two scans needed for Rossavik growth model specification was evaluated in 20 normally growing fetuses. Based on systematic and random prediction errors, determined by comparing predicted values to measured values obtained after 27 weeks, menstrual age, optimal results for all parameters studied were obtained with an initial scan at 18 weeks and an interval of 6 weeks between scans. For one-dimensional parameters--head circumference (HC), abdominal circumference (AC), and femur diaphysis length (FDL)--similar results were obtained with an initial scan at 15 weeks and an interval of 3 weeks, but systematic overestimations were seen with 9-week intervals for HC. Studies with three-dimensional parameters--head cube, abdominal cube, and estimated weight--indicated that the timing of the initial scan was not critical, although less variability was seen with the initial scan at 18 weeks. However, the use of 3-week and 9-week intervals (instead of 6 weeks) produced significant systematic errors that varied with the parameter studied. These results suggest that Rossavik growth model specification is possible under a variety of conditions, but both the timing of the initial scan and the interval between scans must considered for each parameter.