Prenatal prediction of postnatal survival in fetuses with congenital diaphragmatic hernia using MRI: lung volume measurement, signal intensity ratio, and effect of experience. (19th March 2022)

• Record Type: Journal Article
• Authors: Dütemeyer, Vivien; Cordier, Anne-Gael; Cannie, Mieke M.; Bevilacqua, Elisa; Huynh, Van; Houfflin-Debarge, Véronique; Verpillat, Pauline; Olivier, Camille; Benachi, Alexandra; Jani, Jacques C.
• Is Part Of: Journal of maternal-fetal & neonatal medicine. Volume 35:Number 6(2022)
• Page Start: 1036
• Page End: 1044
• Publication Date: 2022-03-19
• Subjects: Congenital diaphragmatic hernia -- fetoscopic endoluminal tracheal occlusion -- fetal lung signal intensity ratios -- magnetic resonance imaging -- operator's experience; Obstetrics -- Periodicals; Perinatology -- Periodicals; Infants (Newborn) -- Diseases -- Periodicals; Neonatology -- Periodicals; 618.2
• DOI: 10.1080/14767058.2020.1740982
• Languages: English
• ISSNs: 1476-7058

Abstract

Abstract: Objective: To evaluate various signal intensity ratios in isolated congenital diaphragmatic hernia (CDH) and to compare their potential in predicting survival with that of the observed-to-expected (O/E) ratio of total fetal lung volume (TFLV) using magnetic resonance imaging (MRI) measurements. Our second objective was to evaluate the impact of operator's experience in comparing the prediction of postnatal survival by O/E-TFLV. Methods: In 75 conservatively managed CDH fetuses and in 50 who underwent fetoscopic endoluminal tracheal occlusion (FETO), the fetal lung-to-amniotic fluid, lung-to-liver, lung-to-muscle, lung-to-spinal fluid signal intensity ratios, respectively LAFSIR, LLSIR, LMSIR, and LSFSIR, were measured, as was O/E-TFLV. Receiver operating characteristic (ROC) curves were constructed and used to compare the various signal intensity ratios with O/E-TFLV in the prediction of postnatal survival. In 72 MRI lung volumes assessed by the referring radiologists in Paris and Lille and secondarily by our expert radiologist in Brussels (M.M.C.) using the same MRI examinations, ROC curves were constructed and used to compare the value of O/E-TFLV determined by the two centers in the prediction of postnatal survival. Results: In the total cohort of CDH fetuses, O/E-TFLV and LLSIR were predictive of postnatal survival whereas in the conservatively managed group O/E-TFLV, LLSIR, and LMSIR predicted postnatal survival. O/E-TFLV predicted postnatal survival far better than the signal intensity ratios: area under the ROC curve for prediction by O/E-TFLV in the total cohort was 0.866 ( p < .001; standard error = 0.031). The area under the ROC curve for prediction of postnatal survival using O/E-TFLV by MRI evaluated at the referral centers was 0.640 ( p = 102; standard error = 0.085), and with O/E-TFLV reevaluated by M.M.C., it was 0.872 ( p < .001; standard error = 0.061). Pairwise comparison showed a significant difference between the areas under the ROC curves (difference = 0.187, p = .012; standard error = 0.075). Conclusion: In fetuses with CDH with/without FETO, LLSIR was significantly correlated with the prediction of postnatal survival. However, measurement of O/E-TFLV was far better in predicting postnatal outcome. Operator experience in measurement of lung volumes using MRI seem to play a role in the predictive value of the technique.