Prenatal prediction of survival in congenital diaphragmatic hernia: An audit of postnatal outcomes

A machine learning approach to predict mortality and neonatal persistent pulmonary hypertension in newborns with congenital diaphragmatic hernia. A retrospective observational cohort study

Congenital diaphragmatic hernia (CDH) has high morbidity and mortality rates. This study aimed to develop a machine learning (ML) algorithm to predict outcomes based on prenatal and early postnatal data. This retrospective observational cohort study involved infants with left-sided CDH, born from 2012 to 2020. We analyzed clinical and imaging data using three classification algorithms: XGBoost, Support Vector Machine, and K-Nearest Neighbors. Medical records of 165 pregnant women with CDH fetal diagnosis were reviewed. According to inclusion criteria, 50 infants with isolated left-sided CDH were enrolled. The mean o/eLHR was 37.32%, and the average gestational age at delivery was 36.5 weeks. Among these infants, 26 (52%) had severe persistent neonatal pulmonary hypertension (PPHN), while 24 (48%) had moderate or mild form; 37 survived (74%), and 13 did not (26%). The XGBoost model achieved 88% accuracy and 95% sensitivity for predicting mortality using ten features and 82% accuracy for PPHN severity with 14 features. The area under the ROC curve was 0.87 for mortality and 0.82 for PPHN severity.

CONCLUSION

ML models show promise in predicting CDH outcomes and supporting clinical decisions. Future research should focus on more extensive studies to refine these algorithms and improve care management.

CLINICAL TRIAL REGISTRATION

NCT04609163.

WHAT IS KNOWN

• Congenital diaphragmatic hernia (CDH) is a serious condition characterized by high morbidity and mortality rates, making it critical to predict neonatal outcomes for effective clinical management accurately. • Traditional prenatal diagnostic methods often struggle to predict complications such as Neonatal Persistent Pulmonary Hypertension (PPHN) in CDH, highlighting the need for innovative predictive approaches.

WHAT IS NEW

• Machine learning (ML) models, particularly XGBoost, have been shown to accurately forecast mortality and the severity of PPHN in infants with CDH based on prenatal and early postnatal clinical and imaging data. • ML-based predictive models can enhance prenatal counseling, optimize birth planning, and tailor postnatal care for patients with CDH, enabling real-time risk assessment and adaptive management strategies.

Congenital diaphragmatic hernia: automatic lung and liver MRI segmentation with nnU-Net, reproducibility of pyradiomics features, and a machine learning application for the classification of liver herniation

Prenatal assessment of lung size and liver position is essential to stratify congenital diaphragmatic hernia (CDH) fetuses in risk categories, guiding counseling, and patient management. Manual segmentation on fetal MRI provides a quantitative estimation of total lung volume and liver herniation. However, it is time-consuming and operator-dependent. In this study, we utilized a publicly available deep learning (DL) segmentation system (nnU-Net) to automatically contour CDH-affected fetal lungs and liver on MRI sections. Concordance between automatic and manual segmentation was assessed by calculating the Jaccard coefficient. Pyradiomics standard features were then extracted from both manually and automatically segmented regions. The reproducibility of features between the two groups was evaluated through the Wilcoxon rank-sum test and intraclass correlation coefficients (ICCs). We finally tested the reliability of the automatic-segmentation approach by building a ML classifier system for the prediction of liver herniation based on support vector machines (SVM) and trained on shape features computed both in the manual and nnU-Net-segmented organs. We compared the area under the classifier receiver operating characteristic curve (AUC) in the two cases. Pyradiomics features calculated in the manual ROIs were partly reproducible by the same features calculated in nnU-Net segmented ROIs and, when used in the ML procedure, to predict liver herniation (both AUC around 0.85).          Conclusion: Our results suggest that automatic MRI segmentation is feasible, with good reproducibility of pyradiomics features, and that a ML system for liver herniation prediction offers good reliability.          Trial registration: https://clinicaltrials.gov/ct2/show/NCT04609163?term=NCT04609163&draw=2&rank=1 ; Clinical Trial Identification no. NCT04609163. What is Known: • Magnetic resonance imaging (MRI) is crucial for prenatal congenital diaphragmatic hernia (CDH) assessment. It enables the quantification of the total lung volume and the extent of liver herniation, which are essential for stratifying the severity of CDH, guiding counseling, and patient management. • The manual segmentation of MRI scans is a time-consuming process that is heavily reliant upon the skill set of the operator. What is New: • MRI lung and liver automatic segmentation using the deep learning nnU-Net system is feasible, with good Jaccard coefficient values and satisfactory reproducibility of pyradiomics features compared to manual results. • A feasible ML system for predicting liver herniation could improve prenatal assessments and CDH patient management.

Perinatal Outcome in Congenital Diaphragmatic Hernia (CDH): A Single-Center Experience

Objective

To study the perinatal outcome in fetuses diagnosed with congenital diaphragmatic hernia (CDH).

Methods

Thirty-two pregnant women with antenatal diagnosis of CDH in fetus, who delivered between 2018 and 2021, were included in the study. Postnatally eventration of diaphragm was diagnosed in 3 neonates and were excluded.

Results

The median gestational age at diagnosis was 23 weeks (IQR: 216-261 weeks). The mean O/E LHR was 34.88 ± 9.03%, and the O/E LHR was significantly lower in fetuses who did not survive (40.81 ± 4.25 vs 31.26 ± 9.33; p = 0.0037). On ROC analysis, at a cutoff of ≤ 32.93, O/E LHR had a specificity of 100% with a sensitivity of 72.22% in predicting mortality. Cases with liver herniation were not significantly different between survivors versus non-survivors. The overall survival rate was 37.93%, and the leading cause of death was severe persistent pulmonary hypertension.

Conclusion

O/E LHR can predict mortality in neonates with antenatal diagnosis of CDH. The presence of pulmonary hypertension was the leading cause of death in these neonates.

In Utero Therapy for Congenital Diaphragmatic Hernia

Congenital diaphragmatic hernia is an anomaly that is often prenatally diagnosed and spans a wide spectrum of disease, with high morbidity and mortality associated with fetuses with severe defects. Congenital diaphragmatic hernia is thus an ideal target for fetal intervention. We review the literature on prenatal diagnosis, describe the history of fetal intervention for congenital diaphragmatic hernia, and discuss fetal endoscopic tracheal occlusion and the Tracheal Occlusion To Accelerate Lung growth trial results. Finally, we present preclinical studies for potential future directions.

Prenatal predictors of mortality in fetuses with congenital diaphragmatic hernia: a systematic review and meta-analysis

PURPOSE

This study aimed to evaluate prenatal predictors of mortality in fetuses with congenital diaphragmatic hernia (CDH).

METHODS

A systematic literature search was performed to identify relevant observational studies that evaluated the ability of lung-to-head ratio (LHR), observed-to-expected LHR (o/e-LHR), observed-to-expected total fetal lung volume (o/e-TFLV), lung-to-thorax transverse area ratio (L/T ratio), intrathoracic herniation of the liver and the stomach, and side of diaphragmatic hernia, using a threshold for the prediction of mortality in fetuses with CDH. Study quality was assessed using the QUADAS-2 tool. Hierarchical summary receiver operating characteristic curves were constructed.

RESULTS

A total of 50 articles were included in this meta-analysis. The QUADAS-2 tool identified a high risk of bias in more than one domain scored in all parameters. Among those parameters, the diagnostic odds ratio of mortality with o/e-LHR < 25%, o/e-TFLV < 25%, and L/T ratio < 0.08 were 11.98 [95% confidence interval (CI) 4.65-30.89], 11.14 (95% CI 5.19-23.89), and 10.28 (95% CI 3.38-31.31), respectively. The predictive values for mortality were similar between the presence of liver herniation and retrocardiac fetal stomach position.

CONCLUSIONS

This systematic review suggests that o/e-LHR, o/e-TFLV, and L/T ratio are equally good predictors of neonatal mortality in fetuses with isolated CDH.

Motion corrected fetal body MRI provides reliable 3D lung volumes in normal and abnormal fetuses

Objectives

To calculate 3D‐segmented total lung volume (TLV) in fetuses with thoracic anomalies using deformable slice‐to‐volume registration (DSVR) with comparison to 2D‐manual segmentation. To establish a normogram of TLV calculated by DSVR in healthy control fetuses.

Methods

A pilot study at a single regional fetal medicine referral centre included 16 magnetic resonance imaging (MRI) datasets of fetuses (22–32 weeks gestational age). Diagnosis was CDH (n = 6), CPAM (n = 2), and healthy controls (n = 8). Deformable slice‐to‐volume registration was used for reconstruction of 3D isotropic (0.85 mm) volumes of the fetal body followed by semi‐automated lung segmentation. 3D TLV were compared to traditional 2D‐based volumetry. Abnormal cases referenced to a normogram produced from 100 normal fetuses whose TLV was calculated by DSVR only.

Results

Deformable slice‐to‐volume registration‐derived TLV values have high correlation with the 2D‐based measurements but with a consistently lower volume; bias −1.44 cm³ [95% limits: −2.6 to −0.3] with improved resolution to exclude hilar structures even in cases of motion corruption or very low lung volumes.

Conclusions

Deformable slice‐to‐volume registration for fetal lung MRI aids analysis of motion corrupted scans and does not suffer from the interpolation error inherent to 2D‐segmentation. It increases information content of acquired data in terms of visualising organs in 3D space and quantification of volumes, which may improve counselling and surgical planning.

What's already known about this topic?

Congenital diaphragmatic hernia (CDH) and congenital lung lesions (CLL) are prognosticated with ultrasound‐based measurements of the fetal lung in a single dimension; however true volumes may provide greater sensitivity for high risk cases. Current use of magnetic resonance imaging (MRI) to calculate fetal lung volumes is limited as two‐dimensional segmentation is labour intensive and risks interpolation and motion‐corruption errors.

What does this study add?

Three‐dimensional lung volumes can be computed from deformable slice‐to‐volume registration (DSVR) 3D reconstructions and highly correlate with traditional 2D‐derived volumes. DSVR‐derived volumes, however, should be more reliable owing to higher resolution and semi‐automated calculations that do not rely on interpolation between slices on motion‐corrupted stacks.

The role of magnetic resonance imaging in the diagnosis and prognostic evaluation of fetuses with congenital diaphragmatic hernia

In recent years, magnetic resonance imaging (MRI) has largely increased our knowledge and predictive accuracy of congenital diaphragmatic hernia (CDH) in the fetus. Thanks to its technical advantages, better anatomical definition, and superiority in fetal lung volume estimation, fetal MRI has been demonstrated to be superior to 2D and 3D ultrasound alone in CDH diagnosis and outcome prediction. This is of crucial importance for prenatal counseling, risk stratification, and decision-making approach. Furthermore, several quantitative and qualitative parameters can be evaluated simultaneously, which have been associated with survival, postnatal course severity, and long-term morbidity.

CONCLUSION

Fetal MRI will further strengthen its role in the near future, but it is necessary to reach a consensus on indications, methodology, and data interpretation. In addition, it is required data integration from different imaging modalities and clinical courses, especially for predicting postnatal pulmonary hypertension. This would lead to a comprehensive prognostic assessment.

WHAT IS KNOWN

• MRI plays a key role in evaluating the fetal lung in patients with CDH. • Prognostic assessment of CDH is challenging, and advanced imaging is crucial for a complete prenatal assessment and counseling.

WHAT IS NEW

• Fetal MRI has strengthened its role over ultrasound due to its technical advantages, better anatomical definition, superior fetal lung volume estimation, and outcome prediction. • Imaging and clinical data integration is the most desirable strategy and may provide new MRI applications and future research opportunities.

Fetal MRI mediastinal shift angle and respiratory and cardiovascular pharmacological support in newborns with congenital diaphragmatic hernia

In newborns with congenital diaphragmatic hernia (CDH), the mediastinal shift caused by the herniated organs negatively affects lung development. Assessment of the fetal magnetic resonance imaging (MRI) mediastinal shift angle (MSA) was shown to have an inverse correlation with the total fetal lung volume (TFLV), being associated with neonatal survival. However, a possible association with postnatal morbidity has never been investigated. We hypothesize that the degree of the mediastinal shift could be associated with higher respiratory and cardiocirculatory impairment, requiring intensive treatments and extended hospitalization in survivors. We retrospectively consider a cohort of isolated, left-sided CDH, for whom we calculated the MSA and the observed/expected (O/E) TFLV at fetal MRI. We performed a data collection regarding inotropic or vasoactive support, treatment with pulmonary vasodilators, mechanical ventilation, and length of stay. General linear models were performed. The MSA and O/E TFLV were inversely correlated (Pearson's coefficient - 0.65, p < 0.001), and deceased patients showed higher MSA values then survivors (p = 0.011). Among survivors, an increase in MSA was associated with longer pharmacological treatments (dobutamine: p = 0.016; dopamine: p = 0.049; hydrocortisone: p = 0.003; nitric oxide: p = 0.002; sildenafil: p = 0.039; milrinone: p = 0.039; oxygen: p = 0.066), and mechanical ventilation (p = 0.005), with an increasing trend in the length of hospitalization (p = 0.089).Conclusion: The MSA indirectly reflects lung hypoplasia and is associated with a higher neonatal intensity of cares. However, further studies are needed to consolidate the results.Trial registration: The study is an exploratory post hoc analysis of the registered NeoAPACHE protocol at ClinicalTrials.gov with the identifier NCT04396028. What is Known: • In congenital diaphragmatic hernia, the lung size, liver position, and defect side are the most common prenatal prognostic parameters used in clinical practice for morbidity and mortality prediction. • Lung hypoplasia, strictly associated with lung size, is estimated by observed/expected lung to head ratio and observed/expected total fetal lung volume with prenatal ultrasound and fetal magnetic resonance imaging, respectively. What is New: • A new, faster, more straightforward, and less operator-dependent tool to assess CDH severity could be the mediastinal shift angle calculation with fetal magnetic resonance imaging. • Postnatal clinical severity, considered as a postnatal cardiovascular and respiratory impairment that indirectly reflects lung hypoplasia, is associated with an increased mediastinal shift angle calculation.

Treatment of pulmonary hypertension during initial hospitalization in a multicenter cohort of infants with congenital diaphragmatic hernia (CDH)

OBJECTIVE

Describe inpatient pulmonary hypertension (PH) treatment and factors associated with therapy at discharge in a multicenter cohort of infants with CDH.

METHODS

Six years linked records from Children's Hospitals Neonatal Database and Pediatric Health Information System were used to describe associations between prenatal/perinatal factors, clinical outcomes, echocardiographic findings and PH medications (PHM), during hospitalization and at discharge.

RESULTS

Of 1106 CDH infants from 23 centers, 62.8% of infants received PHM, and 11.6% of survivors were discharged on PHM. Survivors discharged on PHM more frequently had intrathoracic liver, small for gestational age, and low 5 min APGARs compared with those discharged without PHM (p < 0.0001). Nearly one-third of infants discharged without PHM had PH on last inpatient echo.

CONCLUSIONS

PH medication use is common in CDH. Identification of infants at risk for persistent PH may impact ongoing management. Post-discharge follow-up of all CDH infants with echocardiographic evidence of PH is warranted.

Risk factors for mortality in infants with congenital diaphragmatic hernia: a single center experience

Background

Despite current progress in research of congenital diaphragmatic hernia, its management remains challenging, requiring an interdisciplinary team for optimal treatment.

Objective

Aim of the present study was to evaluate potential risk factors for mortality of infants with congenital diaphragmatic hernia.

Methods

A single-center chart review of all patients treated with congenital diaphragmatic hernia over a period of 16 years, at the Medical University of Vienna, was performed. A comparison of medical parameters between survivors and non-survivors, as well as to published literature was conducted.

Results

During the observational period 66 patients were diagnosed with congenital diaphragmatic hernia. Overall survival was 84.6%. Left-sided hernia occurred in 51 patients (78.5%) with a mortality of 7.8%. In comparison, right-sided hernia occurred less frequently (n = 12) but showed a higher mortality (33.3%, p = 0.000). Critically instable patients were provided with venoarterial extracorporeal membrane oxygenation (ECMO, 32.3%, n = 21). Survival rate among these patients was 66.7%. Right-sided hernia, treatment with inhaled nitric oxide (iNO) over 15 days and the use of ECMO over 10 days were significant risk factors for mortality.

Conclusion

The survival rate in this cohort is comparable to the current literature. Parameters such as the side of the diaphragmatic defect, duration of ECMO and inhaled nitric oxide were assessed as mortality risk factors. This analysis of patients with congenital diaphragmatic hernia enhances understanding of risk factors for mortality, helping to improve management and enabling further evaluation in prospective clinical trials.

Discordant prenatal ultrasound and fetal MRI in CDH: wherein lies the truth?

PURPOSE

Prenatal risk assessment of congenital diaphragmatic hernia (CDH) relies on prenatal ultrasound (U/S) and fetal magnetic resonance imaging (MRI). When the modalities differ in prognosis, it is unclear which is more reliable.

METHODS

Retrospective chart review identified cases of prenatally diagnosed CDH from 4/2010-6/2018 meeting inclusion criteria. Demographic, radiologic, and postnatal outcomes data were collected. Ultrasound- versus MRI-based prognosis (mild, moderate, and severe) was compared with clinical outcomes. Kappa measures compared congruency in disease severity scaling between imaging modalities, while logistic regression and receiver operating characteristics curves compared the ability of each modality to predict outcomes.

RESULTS

Forty-two patients met criteria. Both U/S- and MRI-based prognosis categories differentiated for survival. MRI categories differentiated for ECMO use, surgical repair, and defect type. O/e TFLV better discriminated for survivors and defect type than o/e LHR. Seventeen (40.5%) had discordant prenatal prognostic categories. In 13/17 (76.5%), o/e TFLV predicted higher severity when compared to o/e LHR, but sample size was insufficient to compare accuracy in cases of discordance.

CONCLUSIONS

Clinical outcomes suggest fetal MRI may more accurately predict severe pulmonary hypoplasia compared to prenatal ultrasound. Our analysis suggests fetal MRI is a valuable adjunct in the prenatal evaluation of CDH.

LEVEL OF EVIDENCE

Level III.

TYPE OF STUDY

Retrospective Review.

Calculating Observed-to-Expected Total Fetal Lung Volume in CDH Fetuses in Twin Gestation: Is There a Better Way?

BACKGROUND

Congenital diaphragmatic hernia (CDH) is a potentially lethal birth defect, and identifying prenatal predictors of outcome is important. Observed-to-expected total fetal lung volume (o/e TFLV) has been shown to be a predictor of severity and useful in risk stratification but is variable due to different TFLV formulas.

OBJECTIVES

To calculate o/e TFLV for CDH patients part of a twin gestation using the unaffected sibling as an internal control and comparing these values to those calculated using published formulas for TFLV.

METHODS

Seven twin gestations with one fetus affected by CDH were identified between 2006 and 2017. The lung volume for each twin was calculated using magnetic resonance imaging (MRI), and o/e TFLV was calculated using the unaffected twin's TFLV. This percentage was then compared to the o/e TFLV calculated using published formulas.

RESULTS

Lung volumes in the unaffected twins were within normal ranges at the lower end of the spectrum. No single TFLV formula was found to correlate perfectly. Intraclass correlation coefficient estimate was most consistent for o/e TFLV calculated with the Meyers formula and supported by Bland-Altman plots.

CONCLUSIONS

O/e TFLV measured in CDH/non-CDH twin gestations using the unaffected sibling demonstrated agreement with o/e TFLV calculated using the Meyers formula. We urge the fetal community to standardize the method, use, and interpretation of fetal MRI in the prenatal evaluation of CDH.