A maChine and deep Learning Approach to predict pulmoNary hyperteNsIon in newbornS with congenital diaphragmatic Hernia (CLANNISH): Protocol for a retrospective study

• DOI: 10.1371/journal.pone.0259724
• Publication Date: 2021-11-09T18:49:43Z
• Authors: Ilaria Amodeo; Giorgio De Nunzio; Genny Raffaeli; Irene Borzani; Alice Griggio; Luana Conte; Francesco Macchini; Valentina Condò; Nicola Persico; Isabella Fabietti; Stefano Ghirardello; Maria Pierro; Benedetta Tafuri; Giuseppe Como; Donato Cascio; Mariarosa Colnaghi; Fabio Mosca; Giacomo Cavallaro
• Keywords: Fetoscopy

Introduction Outcome predictions of patients with congenital diaphragmatic hernia (CDH) still have some limitations in the prenatal estimate postnatal pulmonary hypertension (PH). We propose applying Machine Learning (ML), and Deep (DL) approaches to fetuses newborns CDH develop forecasting models epoch, based on integrated analysis clinical data, provide neonatal PH as first outcome and, possibly: favorable response fetal endoscopic tracheal occlusion (FETO), need for Extracorporeal Membrane Oxygenation (ECMO), survival ECMO, death. Moreover, we plan produce a (semi)automatic fetus lung segmentation system Magnetic Resonance Imaging (MRI), which will be useful during project implementation but also an important tool itself standardize volume measures fetuses. Methods analytics Patients isolated from singleton pregnancies enrolled, whose checks were performed at Fetal Surgery Unit Fondazione IRCCS Ca’ Granda Ospedale Maggiore Policlinico (Milan, Italy) 30 th week gestation. A retrospective data collection radiological variables newborns’ mothers’ records eligible born between 01/01/2012 31/12/2020. The native sequences magnetic resonance imaging (MRI) collected. Data different sources analyzed using ML DL, algorithms developed each outcome. augmentation dimensionality reduction (feature selection extraction) employed increase sample size avoid overfitting. software automatic MRI DL 3D U-NET approach developed. Ethics dissemination This study received approval local ethics committee (Milan Area 2, Italy). development predictive outcomes key contribution disease prediction, early targeted interventions, personalized management, overall improvement care quality, resource allocation, healthcare, family savings. Our findings validated future prospective multicenter cohort study. Registration was registered ClinicalTrials.gov identifier NCT04609163.