Background Artificial intelligence (AI) is a promising field in the neonatal field.  We focused on lung ultrasound (LUS), a useful tool for the neonatologist. Our aim was to train a neural network to create a model able to interpret LUS. Methods Our multicentric, prospective study included newborns with gestational age (GA) ≥ 33+0 weeks with early tachypnea/dyspnea/oxygen requirements. For each baby, three LUS were performed: within 3 hours of life (T0), at 4–6 hours of life (T1) and in the absence of respiratory support (T2). Each scan was processed to extract ROI used to train a neural network to classify it according to the LUS score. We assessed sensitivity, specificity, positive and negative predictive value of the AI model’s scores in predicting the need for respiratory assistance with nasal Continuous Positive Airway Pressure (nCPAP) and for surfactant, compared to the “classical” scores. Results We enrolled 62 newborns (GA=36±2 weeks). In the prediction of the need for CPAP, we found a cut-off of 6 (at T0) and 5 (at T1) for both the classical nLUS and AI score. In the prediction of surfactant therapy we found a cut-off of 9 for both scores at T0, at T1 the nLUS cut-off was 6, while the AI’s one was 5. Classification accuracy was good both at the image and classes level. Conclusions This is, to our knowledge, the first attempt to use an AI model to interpret early neonatal LUS and can be extremely useful for neonatologist in the clinical setting.

Background: MYCN amplification represents a powerful prognostic factor in neuroblastoma (NB) and may occasionally account for intratumoral heterogeneity. Radiomics is an emerging field of advanced image analysis that aims to extract a large number of quantitative features from standard radiological images, providing valuable clinical information Procedure: In this retrospective study, we aimed to create a radiogenomics model by correlating computed tomography (CT) radiomics analysis with MYCN status and overall survival (OS). NB lesions were segmented on pre-therapy CT scans and radiomics features subsequently extracted using a dedicated library. Dimensionality reduction/features selection approaches were then used for features procession and logistic regression models have been developed for the considered outcome. Results: Seventy-eight patients were included in this study, 24 presented MYCN amplification. In total, 232 radiomics features were extracted. Eight features were selected through Boruta algorithm and 2 features were lastly chosen through Pearson correlation analysis: mean of voxel intensity histogram (p=0.0082) and zone size non-uniformity (p=0.038). Five-times repeated 3-fold cross-validation logistic regression models yielded an Area Under the Curve (AUC) value of 0.879 on the training and 0.865 on the testing set for MYCN. No statistical significant difference has been observed comparing radiomics predicted and actual OS data. Conclusions: CT based radiomics is able to predict MYCN amplification status and OS in NB, paving the way to the in depth analysis of imaging based biomarkers that could enhance outcomes prediction.