Learning curve in measurement of fetal frontomaxillary facial angle at 11-13 weeks of gestation.

OBJECTIVE To determine the number of ultrasound examinations required to train sonographers to accurately measure the fetal frontomaxillary facial (FMF) angle at 11-13 weeks of gestation. METHODS Eight sonographers accredited for nuchal translucency thickness (NT) measurement (and with different levels of experience) were trained to measure the fetal FMF angle using specially acquired three-dimensional (3D) volumes. Training was provided in cycles, and each cycle consisted of a training period on 20 randomly selected cases followed by an examination using 10 randomly selected cases. During training, the sonographer was informed of the true FMF angle value after each FMF angle measurement on a case-by-case basis. During examination, the difference between the measured and the true values of the FMF angle (i.e. the delta angle) was calculated. A measurement was considered accurate if the delta angle was less than 5 degrees . The sonographer was considered to be competent and the training finished if all 10 examination cases satisfied this criterion. Otherwise, the sonographer would undergo further cycles of training-examination, until he/she became competent. RESULTS The number of training cases required for a sonographer to become competent was 40 for two sonographers, 60 for one, 80 for one, 100 for two, 120 for one and 140 for one, with a median of 90. The median number of failed cases reduced from 2.5 (out of 10) at the first cycle to 0 by the 7(th) cycle. As training cycles increased, the mean angle deviation and measurement time required both reduced significantly. The average delta angle of the passing examination cycle was 2.06 +/- 1.40 degrees . The number of training cases required to become competent in FMF angle measurement was 40 for the two most experienced trainees and 80, 120 and 140 for the three least experienced ones. CONCLUSIONS We have demonstrated that competence in FMF angle measurement was achieved after a median number of 90 cases, with a range of up to 140. The number required was substantially lower, at 40 cases, among those with extensive experience of NT measurement.