Bronchopulmonary Dysplasia Predicted by Developing a Machine Learning Model of Genetic and Clinical Information

Improving early diagnosis of bronchopulmonary dysplasia

INTRODUCTION

Bronchopulmonary disease (BPD) is associated with long-term neurodevelopmental and cardiorespiratory complications, often requiring significant use of resources. To reduce this healthcare burden, it is essential that those at high risk of BPD are identified early so that strategies are introduced to prevent disease progression. Our aim was to discuss potential methods for improving early diagnosis in the first week after birth.

AREAS COVERED

A narrative review was undertaken. The search strategy involved searching PubMed, Embase and Cochrane databases from 1967 to 2024. The results of potential biomarkers and imaging modes are discussed. Furthermore, the value of scoring systems is explored.

EXPERT OPINION

BPD occurs as a result of disruption to pulmonary vascular and alveolar development, thus abnormal levels of factors regulating those processes are promising avenues to explore with regard to early detection of high-risk infants. Data from twin studies suggests genetic factors can be attributed to 82% of the observed difference in moderate to severe BPD, but large genome-wide studies have yielded conflicting results. Comparative studies are required to determine which biomarker or imaging mode may most accurately diagnose early BPD development. Models which include the most predictive factors should be evaluated going forward.

Bridging the Gap: Exploring Bronchopulmonary Dysplasia through the Lens of Biomedical Informatics

Bronchopulmonary dysplasia (BPD), a chronic lung disease predominantly affecting premature infants, poses substantial clinical challenges. This review delves into the promise of biomedical informatics (BMI) in reshaping BPD research and care. We commence by highlighting the escalating prevalence and healthcare impact of BPD, emphasizing the necessity for innovative strategies to comprehend its intricate nature. To this end, we introduce BMI as a potent toolset adept at managing and analyzing extensive, diverse biomedical data. The challenges intrinsic to BPD research are addressed, underscoring the inadequacies of conventional approaches and the compelling need for data-driven solutions. We subsequently explore how BMI can revolutionize BPD research, encompassing genomics and personalized medicine to reveal potential biomarkers and individualized treatment strategies. Predictive analytics emerges as a pivotal facet of BMI, enabling early diagnosis and risk assessment for timely interventions. Moreover, we examine how mobile health technologies facilitate real-time monitoring and enhance patient engagement, ultimately refining BPD management. Ethical and legal considerations surrounding BMI implementation in BPD research are discussed, accentuating issues of privacy, data security, and informed consent. In summation, this review highlights BMI's transformative potential in advancing BPD research, addressing challenges, and opening avenues for personalized medicine and predictive analytics.

Transforming neonatal care with artificial intelligence: challenges, ethical consideration, and opportunities

Artificial intelligence (AI) offers tremendous potential to transform neonatology through improved diagnostics, personalized treatments, and earlier prevention of complications. However, there are many challenges to address before AI is ready for clinical practice. This review defines key AI concepts and discusses ethical considerations and implicit biases associated with AI. Next we will review literature examples of AI already being explored in neonatology research and we will suggest future potentials for AI work. Examples discussed in this article include predicting outcomes such as sepsis, optimizing oxygen therapy, and image analysis to detect brain injury and retinopathy of prematurity. Realizing AI's potential necessitates collaboration between diverse stakeholders across the entire process of incorporating AI tools in the NICU to address testability, usability, bias, and transparency. With multi-center and multi-disciplinary collaboration, AI holds tremendous potential to transform the future of neonatology.

Machine learning predictive models for grading bronchopulmonary dysplasia: umbilical cord blood IL-6 as a biomarker

Objectives

This study aimed to analyze the predictive value of umbilical cord blood Interleukin-6 (UCB IL-6) for the severity-graded BPD and to establish machine learning (ML) predictive models in a Chinese population based on the 2019 NRN evidence-based guidelines.

Methods

In this retrospective analysis, we included infants born with gestational age <32 weeks, who underwent UCB IL-6 testing within 24 h of admission to our NICU between 2020 and 2022. We collected their medical information encompassing the maternal, perinatal, and early neonatal phases. Furthermore, we classified the grade of BPD according to the 2019 NRN evidence-based guidelines. The correlation between UCB IL-6 and the grades of BPD was analyzed. Univariate analysis and ordinal logistic regression were employed to identify risk factors, followed by the development of ML predictive models based on XGBoost, CatBoost, LightGBM, and Random Forest. The AUROC was used to evaluate the diagnostic value of each model. Besides, we generated feature importance distribution plots based on SHAP values to emphasize the significance of UCB IL-6 in the models.

Results

The study ultimately enrolled 414 preterm infants, with No BPD group (n = 309), Grade 1 BPD group (n = 73), and Grade 2-3 BPD group (n = 32). The levels of UCB IL-6 increased with the grades of BPD. UCB IL-6 demonstrated clinical significance in predicting various grades of BPD, particularly in distinguishing Grade 2-3 BPD patients, with an AUROC of 0.815 (95% CI: 0.753-0.877). All four ML models, XGBoost, CatBoost, LightGBM, and Random Forest, exhibited Micro-average AUROC values of 0.841, 0.870, 0.851, and 0.878, respectively. Notably, UCB IL-6 consistently appeared as the most prominent feature across the feature importance distribution plots in all four models.

Conclusion

UCB IL-6 significantly contributes to predicting severity-graded BPD, especially in grade 2-3 BPD. Through the development of four ML predictive models, we highlighted UCB IL-6's importance.

Comprehensive assessment of the genetic characteristics of small for gestational age newborns in NICU: from diagnosis of genetic disorders to prediction of prognosis

BACKGROUND

In China, ~1,072,100 small for gestational age (SGA) births occur annually. These SGA newborns are a high-risk population of developmental delay. Our study aimed to evaluate the genetic profile of SGA newborns in the newborn intensive care unit (NICU) and establish a prognosis prediction model by combining clinical and genetic factors.

METHODS

A cohort of 723 SGA and 1317 appropriate for gestational age (AGA) newborns were recruited between June 2018 and June 2020. Clinical exome sequencing was performed for each newborn. The gene-based rare-variant collapsing analyses and the gene burden test were applied to identify the risk genes for SGA and SGA with poor prognosis. The Gradient Boosting Machine framework was used to generate two models to predict the prognosis of SGA. The performance of two models were validated with an independent cohort of 115 SGA newborns without genetic diagnosis from July 2020 to April 2022. All newborns in this study were recruited through the China Neonatal Genomes Project (CNGP) and were hospitalized in NICU, Children's Hospital of Fudan University, Shanghai, China.

RESULTS

Among the 723 SGA newborns, 88(12.2%) received genetic diagnosis, including 42(47.7%) with monogenic diseases and 46(52.3%) with chromosomal abnormalities. SGA with genetic diagnosis showed higher rates in severe SGA(54.5% vs. 41.9%, P=0.0025) than SGA without genetic diagnosis. SGA with chromosomal abnormalities showed higher incidences of physical and neurodevelopmental delay compared to those with monogenic diseases (45.7% vs. 19.0%, P=0.012). We filtered out 3 genes (ITGB4, TXNRD2, RRM2B) as potential causative genes for SGA and 1 gene (ADIPOQ) as potential causative gene for SGA with poor prognosis. The model integrating clinical and genetic factors demonstrated a higher area under the receiver operating characteristic curve (AUC) over the model based solely on clinical factors in both the SGA-model generation dataset (AUC=0.9[95% confidence interval 0.84-0.96] vs. AUC=0.74 [0.64-0.84]; P=0.00196) and the independent SGA-validation dataset (AUC=0.76 [0.6-0.93] vs. AUC=0.53[0.29-0.76]; P=0.0117).

CONCLUSION

SGA newborns in NICU presented with roughly equal proportions of monogenic and chromosomal abnormalities. Chromosomal disorders were associated with poorer prognosis. The rare-variant collapsing analyses studies have the ability to identify potential causative factors associated with growth and development. The SGA prognosis prediction model integrating genetic and clinical factors outperformed that relying solely on clinical factors. The application of genetic sequencing in hospitalized SGA newborns may improve early genetic diagnosis and prognosis prediction.

Risk stratification of hemodynamically significant patent ductus arteriosus by clinical and genetic factors

BACKGROUND

Hemodynamically significant patent ductus arteriosus (hsPDA) is associated with increased comorbidities in neonates. Early evaluation of hsPDA risk is critical to implement individualized intervention. The aim of the study was to provide a powerful reference for the early identification of high-risk hsPDA population and early treatment decisions.

METHODS

We enrolled infants who were diagnosed with PDA and performed exome sequencing. The collapsing analyses were used to find the risk gene set (RGS) of hsPDA for model construction. The credibility of RGS was proven by RNA sequencing. Multivariate logistic regression was performed to establish models combining clinical and genetic features. The models were evaluated by area under the receiver operating curve (AUC) and decision curve analysis (DCA).

RESULTS

In this retrospective cohort study of 2199 PDA patients, 549 (25.0%) infants were diagnosed with hsPDA. The model [all clinical characteristics selected by least absolute shrinkage and selection operator regression (all CCs)] based on six clinical variables was acquired within three days of life, including gestational age (GA), respiratory distress syndrome (RDS), the lowest platelet count, invasive mechanical ventilation, and positive inotropic and vasoactive drugs. It has an AUC of 0.790 [95% confidence interval (CI) = 0.749-0.832], while the simplified model (basic clinical characteristic model) including GA and RDS has an AUC of 0.753 (95% CI = 0.706-0.799). There was a certain consistency between RGS and differentially expressed genes of the ductus arteriosus in mice. The AUC of the models was improved by RGS, and the improvement was significant (all CCs vs. all CCs + RGS: 0.790 vs. 0.817, P < 0.001). DCA demonstrated that all models were clinically useful.

CONCLUSIONS

Models based on clinical factors were developed to accurately stratify the risk of hsPDA in the first three days of life. Genetic features might further improve the model performance. Video Abstract (MP4 86834 kb).

The past, current, and future of neonatal intensive care units with artificial intelligence: a systematic review

Machine learning and deep learning are two subsets of artificial intelligence that involve teaching computers to learn and make decisions from any sort of data. Most recent developments in artificial intelligence are coming from deep learning, which has proven revolutionary in almost all fields, from computer vision to health sciences. The effects of deep learning in medicine have changed the conventional ways of clinical application significantly. Although some sub-fields of medicine, such as pediatrics, have been relatively slow in receiving the critical benefits of deep learning, related research in pediatrics has started to accumulate to a significant level, too. Hence, in this paper, we review recently developed machine learning and deep learning-based solutions for neonatology applications. We systematically evaluate the roles of both classical machine learning and deep learning in neonatology applications, define the methodologies, including algorithmic developments, and describe the remaining challenges in the assessment of neonatal diseases by using PRISMA 2020 guidelines. To date, the primary areas of focus in neonatology regarding AI applications have included survival analysis, neuroimaging, analysis of vital parameters and biosignals, and retinopathy of prematurity diagnosis. We have categorically summarized 106 research articles from 1996 to 2022 and discussed their pros and cons, respectively. In this systematic review, we aimed to further enhance the comprehensiveness of the study. We also discuss possible directions for new AI models and the future of neonatology with the rising power of AI, suggesting roadmaps for the integration of AI into neonatal intensive care units.

Genetics of bronchopulmonary dysplasia: An update

Bronchopulmonary dysplasia (BPD) is a multi-factorial disease that results from multiple clinical factors, including lung immaturity, mechanical ventilation, oxidative stress, pulmonary congestion due to increasing cardiac blood shunting, nutritional and immunological factors. Twin studies have indicated that susceptibility to BPD can be strongly inherited in some settings. Studies have reported associations between common genetic variants and BPD in preterm infants. Recent genomic studies have highlighted a potential role for molecular pathways involved in inflammation and lung development in affected infants. Rare mutations in genes encoding the lipid transporter ATP-binding cassette, sub-family A, member 3 (ABCA3 gene) which is involved in surfactant synthesis in alveolar type II cells, as well as surfactant protein B (SFTPB) and C (SFTPC) can also result in severe form of neonatal-onset interstitial lung diseases and may also potentially affect the course of BPD. This chapter summarizes the current state of knowledge on the genetics of BPD.

Neonatal sequential organ failure assessment score within 72 h after delivery reliably predicts bronchopulmonary dysplasia in very preterm infants

Background

The neonatal sequential organ failure assessment (nSOFA) score is an operational definition of organ dysfunction employed to predict sepsis-associated mortality. However, the relationship between the nSOFA score and bronchopulmonary dysplasia (BPD) has not been investigated clearly. This study evaluates whether the nSOFA score within 72 h after delivery could be used to predict the occurrence of BPD in very preterm infants.

Methods

In this retrospective, single-center cohort study, preterm infants born between 2019 and 2021 were investigated, the nSOFA score was calculated from medical records after admission to the neonatal intensive care unit (NICU) within 72 h after delivery, and the peak value was used for calculation. A logistic regression model was used to evaluate the relationship between the nSOFA score and BPD. Propensity score matching and subgroup analysis were performed to verify the reliability of the results.

Results

Of 238 infants meeting the inclusion criteria, 93 infants (39.1%) were diagnosed with BPD. The receiver operating characteristic curve of the nSOFA score in predicting BPD was 0.790 [95% confidence interval (CI): 0.731-0.849]. The logistic regression model showed that an increment of one in the nSOFA score was related to a 2.09-fold increase in the odds of BPD (95% CI: 1.57-2.76) and 6.36-fold increase when the nSOFA score was higher than 1.5 (95% CI: 2.73-14.79).

Conclusions

The nSOFA score within 72 h after delivery is independently related to BPD and can be used to identify high-risk infants and implement early interventions.

Prediction Models for Bronchopulmonary Dysplasia in Preterm Infants: A Systematic Review and Meta-Analysis

OBJECTIVE

To review systematically and assess the accuracy of prediction models for bronchopulmonary dysplasia (BPD) at 36 weeks of postmenstrual age.

STUDY DESIGN

Searches were conducted in MEDLINE and EMBASE. Studies published between 1990 and 2022 were included if they developed or validated a prediction model for BPD or the combined outcome death/BPD at 36 weeks in the first 14 days of life in infants born preterm. Data were extracted independently by 2 authors following the Critical Appraisal and Data Extraction for Systematic Reviews of Prediction Modelling Studies (ie, CHARMS) and PRISMA guidelines. Risk of bias was assessed using the Prediction model Risk Of Bias ASsessment Tool (ie, PROBAST).

RESULTS

Sixty-five studies were reviewed, including 158 development and 108 externally validated models. Median c-statistic of 0.84 (range 0.43-1.00) was reported at model development, and 0.77 (range 0.41-0.97) at external validation. All models were rated at high risk of bias, due to limitations in the analysis part. Meta-analysis of the validated models revealed increased c-statistics after the first week of life for both the BPD and death/BPD outcome.

CONCLUSIONS

Although BPD prediction models perform satisfactorily, they were all at high risk of bias. Methodologic improvement and complete reporting are needed before they can be considered for use in clinical practice. Future research should aim to validate and update existing models.

High-Resolution and Multidimensional Phenotypes Can Complement Genomics Data to Diagnose Diseases in the Neonatal Population

Advances in genomic medicine have greatly improved our understanding of human diseases. However, phenome is not well understood. High-resolution and multidimensional phenotypes have shed light on the mechanisms underlying neonatal diseases in greater details and have the potential to optimize clinical strategies. In this review, we first highlight the value of analyzing traditional phenotypes using a data science approach in the neonatal population. We then discuss recent research on high-resolution, multidimensional, and structured phenotypes in neonatal critical diseases. Finally, we briefly introduce current technologies available for the analysis of multidimensional data and the value that can be provided by integrating these data into clinical practice. In summary, a time series of multidimensional phenome can improve our understanding of disease mechanisms and diagnostic decision-making, stratify patients, and provide clinicians with optimized strategies for therapeutic intervention; however, the available technologies for collecting multidimensional data and the best platform for connecting multiple modalities should be considered.

Applications of Artificial Intelligence in Neonatology

The development of artificial intelligence methods has impacted therapeutics, personalized diagnostics, drug discovery, and medical imaging. Although, in many situations, AI clinical decision-support tools may seem superior to rule-based tools, their use may result in additional challenges. Examples include the paucity of large datasets and the presence of unbalanced data (i.e., due to the low occurrence of adverse outcomes), as often seen in neonatal medicine. The most recent and impactful applications of AI in neonatal medicine are discussed in this review, highlighting future research directions relating to the neonatal population. Current AI applications tested in neonatology include tools for vital signs monitoring, disease prediction (respiratory distress syndrome, bronchopulmonary dysplasia, apnea of prematurity) and risk stratification (retinopathy of prematurity, intestinal perforation, jaundice), neurological diagnostic and prognostic support (electroencephalograms, sleep stage classification, neuroimaging), and novel image recognition technologies, which are particularly useful for prompt recognition of infections. To have these kinds of tools helping neonatologists in daily clinical practice could be something extremely revolutionary in the next future. On the other hand, it is important to recognize the limitations of AI to ensure the proper use of this technology.

Identification of genetic susceptibility in preterm newborns with bronchopulmonary dysplasia by whole-exome sequencing: BIVM gene may play a role

Bronchopulmonary dysplasia (BPD) is a common chronic respiratory disease in preterm infants caused by multifactorial etiology. Genetic factors are involved in the occurrence of BPD, but studies have found that candidate genes have poor reproducibility and are influenced by ethnic heterogeneity; therefore, more exploration is still needed. We performed whole-exon sequencing in 34 preterm infants with BPD and 32 non-BPD control neonates. The data were analyzed and interpreted by Fisher difference comparison, PLINK and eQTL association analysis, KEGG and GO enrichment analysis, STRING tool, Cytoscape software, ProtParam tool, HOPE online software, and GEOR2 analysis on NCBI GEO dataset. BPD has a highly heterogeneity in different populations, and we found 35 genes overlapped with previous whole-exon sequencing studies, such as APOB gene. Arterial and epithelial cell development and energy metabolism pathways affect BPD. In this study, 24 key genes were identified, and BIVM rs3825519 mutation leads to prolonged assisted ventilation in patients with BPD. A novel DDAH1 mutation site (NM_012137: exon1: c.89 T > G: p.L30R) was found in 9 BPD patients.

CONCLUSION

BIVM gene rs3825519 mutation may play a role in the pathogenesis of BPD by affecting cilia movement, and the DDAH1 and APOB genes mutations may have a pathogenic role in BPD.

WHAT IS KNOWN

• Genetic factors are involved in the occurrence of bronchopulmonary dysplasia. • The candidate genes have poor reproducibility and are influenced by ethnic heterogeneity, therefore, more exploration is still needed.

WHAT IS NEW

• We identified the role of susceptible SNPs in BPD in Shenzhen, China, and identified 24 key genes that influence the pathogenesis of BPD, and also found 35 genes overlapped with previous whole exon sequencing studies, such as APOB gene. • We found that BIVM and DDAH1 genes may play a pathogenic role in the pathogenesis of BPD.

Artificial intelligence in bronchopulmonary dysplasia- current research and unexplored frontiers

Provide an overview of bronchopulmonary dysplasia, its definitions, and their shortcomings. Explore the areas where machine learning may be used to further our understanding of bronchopulmonary dysplasia.

Precision medicine via the integration of phenotype-genotype information in neonatal genome project

The explosion of next-generation sequencing (NGS) has enabled the widespread use of genomic data in precision medicine. Currently, several neonatal genome projects have emerged to explore the advantages of NGS to diagnose or screen for rare genetic disorders. These projects have made remarkable achievements, but still the genome data could be further explored with the assistance of phenotype collection. In contrast, longitudinal birth cohorts are great examples to record and apply phenotypic information in clinical studies starting at the neonatal period, especially the trajectory analyses for health development or disease progression. It is obvious that efficient integration of genotype and phenotype benefits not only the clinical management of rare genetic disorders but also the risk assessment of complex diseases. Here, we first summarize the recent neonatal genome projects as well as some longitudinal birth cohorts. Then, we propose two simplified strategies by integrating genotypic and phenotypic information in precision medicine based on current studies. Finally, research collaborations, sociological issues, and future perspectives are discussed. How to maximize neonatal genomic information to benefit the pediatric population remains an area in need of more research and effort.

Application and potential of artificial intelligence in neonatal medicine

Neonatal care is becoming increasingly complex with large amounts of rich, routinely recorded physiological, diagnostic and outcome data. Artificial intelligence (AI) has the potential to harness this vast quantity and range of information and become a powerful tool to support clinical decision making, personalised care, precise prognostics, and enhance patient safety. Current AI approaches in neonatal medicine include tools for disease prediction and risk stratification, neurological diagnostic support and novel image recognition technologies. Key to the integration of AI in neonatal medicine is the understanding of its limitations and a standardised critical appraisal of AI tools. Barriers and challenges to this include the quality of datasets used, performance assessment, and appropriate external validation and clinical impact studies. Improving digital literacy amongst healthcare professionals and cross-disciplinary collaborations are needed to harness the full potential of AI to help take the next significant steps in improving neonatal outcomes for high-risk infants.

Recent research on the mechanism of mesenchymal stem cells in the treatment of bronchopulmonary dysplasia

Bronchopulmonary dysplasia (BPD) is a chronic lung disease due to impaired pulmonary development and is one of the main causes of respiratory failure in preterm infants. Preterm infants with BPD have significantly higher complication and mortality rates than those without BPD. At present, comprehensive management is the main intervention method for BPD, including reasonable respiratory and circulatory support, appropriate enteral nutrition and parenteral nutrition, application of caffeine/glucocorticoids/surfactants, and out-of-hospital management after discharge. The continuous advances in stem cell medicine in recent years provide new ideas for the treatment of BPD. Various pre-clinical trials have confirmed that stem cell therapy can effectively prevent lung injury and promote lung growth and damage repair. This article performs a comprehensive analysis of the mechanism of mesenchymal stem cells in the treatment of BPD, so as to provide a basis for clinical applications.

Urinary biomarkers for the early prediction of bronchopulmonary dysplasia in preterm infants: A pilot study

BACKGROUND

This study investigated whether 8-hydroxy-2'-deoxyguanosine (8-OHdG) and N-terminal pro-brain natriuretic peptide (NT-proBNP) concentrations in the urine could predict bronchopulmonary dysplasia (BPD) in preterm infants.

METHODS

This prospective cohort study enrolled 165 preterm infants, of whom 70 developed BPD. We measured urinary 8-OHdG and NT-proBNP concentrations from day of life (DOL) 7 to 28. Then, we evaluated the prediction efficiency by receiver operating characteristic curves and assessed correlations between the two biomarkers. Finally, we identified the predictive risk factors for BPD by multivariable logistic regression.

RESULTS

8-OHdG and NT-proBNP levels were significantly higher from DOL 7 to 28 in the BPD group than in the control group (P < 0.05). Additionally, the 8-OHdG level was positively correlated with the NT-proBNP level (r: 0.655-0.789, P < 0.001), and the 8-OHdG and NT-proBNP levels were positively correlated with mechanical ventilation duration and oxygen exposure time (r: 0.175-0.505, P < 0.05) from DOL 7 to 28. Furthermore, the 8-OHdG (DOL 14-28) and NT-proBNP (DOL 7-28) levels were significantly associated with BPD development (P < 0.05).

CONCLUSION

The urine 8-OHdG concentrations from DOL 14 to 28 and NT-proBNP concentrations from DOL 7 to 28 may be practical non-invasive predictors of BPD development in preterm infants.