Reproducibility of fetal lung-to-head ratio in left diaphragmatic hernia across the North American Fetal Therapy Network (NAFTNet).

Abstract

To determine the antenatal sonographic lung area measurement method in left congenital diaphragmatic hernia (CDH) with the highest interrater agreement among North American Fetal Therapy Network (NAFTNet) centers within and outside the fetoscopic tracheal occlusion (FETO) consortium and in comparison with a European "expert" reviewer (ER). Nineteen members from nine FETO consortium centers and 29 reviewers from 17 non-FETO centers reviewed ultrasound clips of the chest from 13 fetuses with isolated left CDH and were asked to select a static plane for lung area measurement using anteroposterior (AP), longest, and trace methods. Interrater agreement in lung area measurements was determined using intraclass correlation coefficient (ICC). Bland-Altman analysis was used to evaluate mean difference (bias) between NAFTNet reviewers and ER. Among FETO centers, agreement was highest using trace (ICC 0.94; 95% CI, 0.83-0.98), followed by longest (ICC 0.89; 95% CI, 0.75-0.97) and lowest for A-P (ICC 0.83; 95% CI, 0.67-0.94). Similar trends were noted in non-FETO centers. When compared with ER, bias was lowest for trace: 14±38mm2 and 19±36mm2 for FETO and non-FETO centers, respectively. The trace method demonstrated the highest interrater agreement and lowest bias for lung area estimation in left CDH across NAFTNet.