Machine learning models for identifying preterm infants at risk of cerebral hemorrhage

Predicting severe intraventricular hemorrhage or early death using machine learning algorithms in VLBWI of the Korean Neonatal Network Database

Severe intraventricular hemorrhage (IVH) in premature infants can lead to serious neurological complications. This retrospective cohort study used the Korean Neonatal Network (KNN) dataset to develop prediction models for severe IVH or early death in very-low-birth-weight infants (VLBWIs) using machine-learning algorithms. The study included VLBWIs registered in the KNN database. The outcome was the diagnosis of IVH Grades 3-4 or death within one week of birth. Predictors were categorized into three groups based on their observed stage during the perinatal period. The dataset was divided into derivation and validation sets at an 8:2 ratio. Models were built using Logistic Regression with Ridge Regulation (LR), Random Forest, and eXtreme Gradient Boosting (XGB). Stage 1 models, based on predictors observed before birth, exhibited similar performance. Stage 2 models, based on predictors observed up to one hour after birth, showed improved performance in all models compared to Stage 1 models. Stage 3 models, based on predictors observed up to one week after birth, showed the best performance, particularly in the XGB model. Its integration into treatment and management protocols can potentially reduce the incidence of permanent brain injury caused by IVH during the early stages of birth.

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.

Clinical Variables Associated With Grade III and IV Intraventricular Hemorrhage (IVH) in Preterm Infants Weighing Less Than 750 Grams

BACKGROUND

Despite innovative advances in neonatal medicine, intraventricular hemorrhage (IVH) continues to be a significant complication in neonatal intensive care units globally.

OBJECTIVE

The study aimed to discern the variables heightening the risk of severe IVH (Grade III and IV) in extremely premature infants weighing less than 750 grams. We postulated that a descending hematocrit (Hct) trend during the first week of life could serve as a predictive marker for the development of severe IVH in this vulnerable population.

METHODS

This retrospective case-control study encompassed infants weighing less than 750 grams at birth, diagnosed with Grade III and/or IV IVH, and born in a tertiary center from 2009 to 2014. A group of 17 infants with severe IVH was compared with 14 gestational age-matched controls. Acid-base status, glucose, fluid goal, urine output, and nutrient (caloric and protein) intake during the first four days of life were meticulously evaluated. Statistically significant variables from baseline data were further analyzed via univariable and multivariable logistic regression analyses, ensuring control for potential confounding variables.

RESULTS

The univariate logistic regression model delineated odds ratios (ORs) of 0.842 for day 2 average Hct (confidence interval [CI], 0.718-0.987) and 0.16 for urine output on day 3 (CI, 0.024-1.056), with the remaining six variables demonstrating no significant association. In the post-multivariable regression analysis, day 2 Hct was the only significant variable (OR, 0.731; 95% CI, 0.537-0.995; P=0.04). The receiver operating characteristic (ROC) curve analysis portrayed an area under the curve of 71% for the day 2 Hct variable.

CONCLUSION

The study revealed that a dip in Hct on day 2 of life augments the likelihood of Grade III and IV IVH among extremely premature infants with a birth weight of less than 750 grams. This insight amplifies our understanding of risk factors associated with severe IVH development in extremely preterm infants, potentially aiding in refining preventive strategies and optimizing clinical management and treatment of these affected infants.

Machine learning to predict major bleeding during anticoagulation for venous thromboembolism: possibilities and limitations

Predictive tools for major bleeding (MB) using machine learning (ML) might be advantageous over traditional methods. We used data from the Registro Informatizado de Enfermedad TromboEmbólica (RIETE) to develop ML algorithms to identify patients with venous thromboembolism (VTE) at increased risk of MB during the first 3 months of anticoagulation. A total of 55 baseline variables were used as predictors. New data prospectively collected from the RIETE were used for further validation. The RIETE and VTE-BLEED scores were used for comparisons. External validation was performed with the COMMAND-VTE database. Learning was carried out with data from 49 587 patients, of whom 873 (1.8%) had MB. The best performing ML method was XGBoost. In the prospective validation cohort the sensitivity, specificity, positive predictive value and F1 score were: 33.2%, 93%, 10%, and 15.4% respectively. F1 value for the RIETE and VTE-BLEED scores were 8.6% and 6.4% respectively. In the external validation cohort the metrics were 10.3%, 87.6%, 3.5% and 5.2% respectively. In that cohort, the F1 value for the RIETE score was 17.3% and for the VTE-BLEED score 9.75%. The performance of the XGBoost algorithm was better than that from the RIETE and VTE-BLEED scores only in the prospective validation cohort, but not in the external validation cohort.

Prediction of extubation failure among low birthweight neonates using machine learning

OBJECTIVE

To develop machine learning models predicting extubation failure in low birthweight neonates using large amounts of clinical data.

STUDY DESIGN

Retrospective cohort study using MIMIC-III, a large single-center, open-source clinical dataset. Logistic regression and boosted-tree (XGBoost) models using demographics, medications, and vital sign and ventilatory data were developed to predict extubation failure, defined as reintubation within 7 days.

RESULTS

1348 low birthweight (≤2500 g) neonates who received mechanical ventilation within the first 7 days were included, of which 350 (26%) failed a trial of extubation. The best-performing model was a boosted-tree model incorporating demographics, vital signs, ventilator parameters, and medications (AUROC 0.82). The most important features were birthweight, last FiO2, average mean airway pressure, caffeine use, and gestational age.

CONCLUSIONS

Machine learning models identified low birthweight ventilated neonates at risk for extubation failure. These models will need to be validated across multiple centers to determine generalizability of this tool.

The use of machine learning and artificial intelligence within pediatric critical care

The field of pediatric critical care has been hampered in the era of precision medicine by our inability to accurately define and subclassify disease phenotypes. This has been caused by heterogeneity across age groups that further challenges the ability to perform randomized controlled trials in pediatrics. One approach to overcome these inherent challenges include the use of machine learning algorithms that can assist in generating more meaningful interpretations from clinical data. This review summarizes machine learning and artificial intelligence techniques that are currently in use for clinical data modeling with relevance to pediatric critical care. Focus has been placed on the differences between techniques and the role of each in the clinical arena. The various forms of clinical decision support that utilize machine learning are also described. We review the applications and limitations of machine learning techniques to empower clinicians to make informed decisions at the bedside. IMPACT: Critical care units generate large amounts of under-utilized data that can be processed through artificial intelligence. This review summarizes the machine learning and artificial intelligence techniques currently being used to process clinical data. The review highlights the applications and limitations of these techniques within a clinical context to aid providers in making more informed decisions at the bedside.

Glycocalyx Sensing with a Mathematical Model of Acoustic Shear Wave Biosensor

The article deals with an idea of exploiting an acoustic shear wave biosensor for investigating the glycocalyx, a polysaccharide polymer molecule layer on the endothelium of blood vessels that, according to recent studies, plays an important role in protecting against diseases. To test this idea, a mathematical model of an acoustic shear wave sensor and corresponding software developed earlier for proteomic applications are used. In this case, the glycocalyx is treated as a layer homogenized over the thin polymer "villi". Its material characteristics depend on the density, thickness, and length of the villi and on the viscous properties of the surrounding liquid (blood plasma). It is proved that the model used has a good sensitivity to the above parameters of the villi and blood plasma. Numerical experiments performed using real data collected retrospectively from premature infants show that the use of acoustic shear wave sensors may be a promising approach to investigate properties of glycocalyx-like structures and their role in prematurity.

Biomarker und Neuromonitoring zur Entwicklungsprognose nach perinataler Hirnschädigung

Das sich entwickelnde Gehirn ist in der Perinatalperiode besonders empfindlich für eine Vielzahl von Insulten, wie z. B. Extremfrühgeburtlichkeit und perinatale Asphyxie. Ihre Komplikationen können zu lebenslangen neurokognitiven, sensorischen und psychosozialen Einschränkungen führen; deren Vorhersage bleibt eine Herausforderung. Eine Schlüsselfunktion kommt der möglichst exakten Identifikation von Hirnläsionen und funktionellen Störungen zu. Die Prädiktion stützt sich auf frühe diagnostische Verfahren und die klinische Erfassung der Meilensteine der Entwicklung. Zur klinischen Diagnostik und zum Neuromonitoring in der Neonatal- und frühen Säuglingsperiode stehen bildgebende Verfahren zur Verfügung. Hierzu zählen zerebrale Sonographie, MRT am errechneten Termin, amplitudenintegriertes (a)EEG und/oder klassisches EEG, Nah-Infrarot-Spektroskopie, General Movements Assessment und die frühe klinische Nachuntersuchung z. B. mithilfe der Hammersmith Neonatal/Infant Neurological Examination. Innovative Biomarker und -muster (Omics) sowie (epi)genetische Prädispositionen sind Gegenstand wissenschaftlicher Untersuchungen. Neben der Erfassung klinischer Risiken kommt psychosozialen Faktoren im Umfeld des Kindes eine entscheidende Rolle zu. Eine möglichst akkurate Prognose ist mit hohem Aufwand verbunden, jedoch zur gezielten Beratung der Familien und der Einleitung von frühen Interventionen, insbesondere vor dem Hintergrund der hohen Plastizität des sich entwickelnden Gehirns, von großer Bedeutung. Diese Übersichtsarbeit fokussiert die Charakterisierung der oben genannten Verfahren und ihrer Kombinationsmöglichkeiten. Zudem wird ein Ausblick gegeben, wie innovative Techniken in Zukunft die Prädiktion der Entwicklung und Nachsorge dieser Kinder vereinfachen können.

Application of Machine Learning Approaches to Predict Postnatal Growth Failure in Very Low Birth Weight Infants

PURPOSE

The aims of the study were to develop and evaluate a machine learning model with which to predict postnatal growth failure (PGF) among very low birth weight (VLBW) infants.

MATERIALS AND METHODS

Of 10425 VLBW infants registered in the Korean Neonatal Network between 2013 and 2017, 7954 infants were included. PGF was defined as a decrease in Z score >1.28 at discharge, compared to that at birth. Six metrics [area under the receiver operating characteristic curve (AUROC), accuracy, precision, sensitivity, specificity, and F1 score] were obtained at five time points (at birth, 7 days, 14 days, 28 days after birth, and at discharge). Machine learning models were built using four different techniques [extreme gradient boosting (XGB), random forest, support vector machine, and convolutional neural network] to compare against the conventional multiple logistic regression (MLR) model.

RESULTS

The XGB algorithm showed the best performance with all six metrics across the board. When compared with MLR, XGB showed a significantly higher AUROC (p=0.03) for Day 7, which was the primary performance metric. Using optimal cut-off points, for Day 7, XGB still showed better performances in terms of AUROC (0.74), accuracy (0.68), and F1 score (0.67). AUROC values seemed to increase slightly from birth to 7 days after birth with significance, almost reaching a plateau after 7 days after birth.

CONCLUSION

We have shown the possibility of predicting PGF through machine learning algorithms, especially XGB. Such models may help neonatologists in the early diagnosis of high-risk infants for PGF for early intervention.

Machine Learning Models for Predicting Mortality in 7472 Very Low Birth Weight Infants Using Data from a Nationwide Neonatal Network

Statistical and analytical methods using artificial intelligence approaches such as machine learning (ML) are increasingly being applied to the field of pediatrics, particularly to neonatology. This study compared the representative ML analysis and the logistic regression (LR), which is a traditional statistical analysis method, using them to predict mortality of very low birth weight infants (VLBWI). We included 7472 VLBWI data from a nationwide Korean neonatal network. Eleven predictor variables (neonatal factors: male sex, gestational age, 5 min Apgar scores, body temperature, and resuscitation at birth; maternal factors: diabetes mellitus, hypertension, chorioamnionitis, premature rupture of membranes, antenatal steroid, and cesarean delivery) were selected based on clinical impact and statistical analysis. We compared the predicted mortality between ML methods—such as artificial neural network (ANN), random forest (RF), and support vector machine (SVM)—and LR with a randomly selected training set (80%) and a test set (20%). The model performances of area under the receiver operating curve (95% confidence interval) equaled LR 0.841 (0.811−0.872), ANN 0.845 (0.815−0.875), and RF 0.826 (0.795−0.858). The exception was SVM 0.631 (0.578−0.683). No statistically significant differences were observed between the performance of LR, ANN, and RF (i.e., p > 0.05). However, the SVM model was lower (p < 0.01). We suggest that VLBWI mortality prediction using ML methods would yield the same prediction rate as the traditional statistical LR method and may be suitable for predicting mortality. However, low prediction rates are observed in certain ML methods; hence, further research is needed on these limitations and selecting an appropriate method.

A Setup for Camera-Based Detection of Simulated Pathological States Using a Neonatal Phantom

Premature infants are among the most vulnerable patients in a hospital. Due to numerous complications associated with immaturity, a continuous monitoring of vital signs with a high sensitivity and accuracy is required. Today, wired sensors are attached to the patient's skin. However, adhesive electrodes can be potentially harmful as they can damage the very thin immature skin. Although unobtrusive monitoring systems using cameras show the potential to replace cable-based techniques, advanced image processing algorithms are data-driven and, therefore, need much data to be trained. Due to the low availability of public neonatal image data, a patient phantom could help to implement algorithms for the robust extraction of vital signs from video recordings. In this work, a camera-based system is presented and validated using a neonatal phantom, which enabled a simulation of common neonatal pathologies such as hypo-/hyperthermia and brady-/tachycardia. The implemented algorithm was able to continuously measure and analyze the heart rate via photoplethysmography imaging with a mean absolute error of 0.91 bpm, as well as the distribution of a neonate's skin temperature with a mean absolute error of less than 0.55 °C. For accurate measurements, a temperature gain offset correction on the registered image from two infrared thermography cameras was performed. A deep learning-based keypoint detector was applied for temperature mapping and guidance for the feature extraction. The presented setup successfully detected several levels of hypo- and hyperthermia, an increased central-peripheral temperature difference, tachycardia and bradycardia.

Effect of Minimally Invasive Puncture Drainage and Conservative Treatment on Prognosis of Patients with Cerebral Hemorrhage

The objective of this study was to explore the effect of minimally invasive puncture drainage under unsupervised learning algorithm and conservative treatment on the prognosis of patients with cerebral hemorrhage. Fifty patients with cerebral hemorrhage were selected as the research objects. The CT images of patients were segmented by unsupervised learning algorithm, and the application value of unsupervised learning algorithm on CT images of patients with cerebral hemorrhage was evaluated. According to the treatment wishes of the patients themselves and their authorizers, they were divided into 30 patients with cerebral hemorrhage in the minimally invasive group and 20 patients with cerebral hemorrhage in the conservative group. The incidence rate of complications of cerebral hemorrhage, the length of hospitalization of the two groups, hematoma volume at admission, 3 days and 7 days after operation, and the hematoma dissipation rate on the 3rd and 7th day after operation were used as the evaluation index of therapeutic effect. MRS and ADL scores were used as prognostic indicators. The results show that K-means clustering algorithm has high quality and short time for CT image segmentation. The overall incidence rate of complications in minimally invasive group was 10%, lower than that in conservative group (25%) (P < 0.05), and the length of hospitalization in minimally invasive group was longer than that in conservative group (P < 0.05). The hematoma volume of minimally invasive group was 16.5 ± 2.4 mL on the 3rd day after operation, and that of conservative group was 27.4 ± 1.8 mL. There was significant difference between the two groups (P < 0.05). In addition, CT showed that the hematoma reduction degree of minimally invasive group was higher than that of conservative group, and the hematoma dissipation rate was higher than that of conservative group on the 3rd and 7th day (P < 0.05). The good MRS score in minimally invasive group was 3.15 times that in conservative group, and the good ADL score was 1.6 times that in conservative group, and there was significant difference in the total score between the two groups (P < 0.05). Minimally invasive puncture drainage is better than conservative treatment in the clearance of hematoma, which is conducive to the recovery of neurological function and daily life of patients with cerebral hemorrhage and is of great help to the prognosis of patients.

Machine Learning Approaches to Predict In-Hospital Mortality among Neonates with Clinically Suspected Sepsis in the Neonatal Intensive Care Unit

Background: preterm and critically ill neonates often experience clinically suspected sepsis during their prolonged hospitalization in the neonatal intensive care unit (NICU), which can be the initial sign of final adverse outcomes. Therefore, we aimed to utilize machine learning approaches to predict neonatal in-hospital mortality through data-driven learning. Methods: a total of 1095 neonates who experienced clinically suspected sepsis in a tertiary-level NICU in Taiwan between August 2017 and July 2020 were enrolled. Clinically suspected sepsis was defined based on clinical features and laboratory criteria and the administration of empiric antibiotics by clinicians. The variables used for analysis included patient demographics, clinical features, laboratory data, and medications. The machine learning methods used included deep neural network (DNN), k-nearest neighbors, support vector machine, random forest, and extreme gradient boost. The performance of these models was evaluated using the area under the receiver operating characteristic curve (AUC). Results: the final in-hospital mortality of this cohort was 8.2% (90 neonates died). A total of 765 (69.8%) and 330 (30.2%) patients were randomly assigned to the training and test sets, respectively. Regarding the efficacy of the single model that most accurately predicted the outcome, DNN exhibited the greatest AUC (0.923, 95% confidence interval [CI] 0.953–0.893) and the best accuracy (95.64%, 95% CI 96.76–94.52%), Cohen’s kappa coefficient value (0.74, 95% CI 0.79–0.69) and Matthews correlation coefficient value (0.75, 95% CI 0.80–0.70). The top three most influential variables in the DNN importance matrix plot were the requirement of ventilator support at the onset of suspected sepsis, the feeding conditions, and intravascular volume expansion. The model performance was indistinguishable between the training and test sets. Conclusions: the DNN model was successfully established to predict in-hospital mortality in neonates with clinically suspected sepsis, and the machine learning algorithm is applicable for clinicians to gain insights and have better communication with families in advance.

Precision Medicine in Neonates: A Tailored Approach to Neonatal Brain Injury

Despite advances in neonatal care to prevent neonatal brain injury and neurodevelopmental impairment, predicting long-term outcome in neonates at risk for brain injury remains difficult. Early prognosis is currently based on cranial ultrasound (CUS), MRI, EEG, NIRS, and/or general movements assessed at specific ages, and predicting outcome in an individual (precision medicine) is not yet possible. New algorithms based on large databases and machine learning applied to clinical, neuromonitoring, and neuroimaging data and genetic analysis and assays measuring multiple biomarkers (omics) can fulfill the needs of modern neonatology. A synergy of all these techniques and the use of automatic quantitative analysis might give clinicians the possibility to provide patient-targeted decision-making for individualized diagnosis, therapy, and outcome prediction. This review will first focus on common neonatal neurological diseases, associated risk factors, and most common treatments. After that, we will discuss how precision medicine and machine learning (ML) approaches could change the future of prediction and prognosis in this field.

Heart Rate Variability in the Perinatal Period: A Critical and Conceptual Review

Neonatal intensive care units (NICUs) greatly expand the use of technology. There is a need to accurately diagnose discomfort, pain, and complications, such as sepsis, mainly before they occur. While specific treatments are possible, they are often time-consuming, invasive, or painful, with detrimental effects for the development of the infant. In the last 40 years, heart rate variability (HRV) has emerged as a non-invasive measurement to monitor newborns and infants, but it still is underused. Hence, the present paper aims to review the utility of HRV in neonatology and the instruments available to assess it, showing how HRV could be an innovative tool in the years to come. When continuously monitored, HRV could help assess the baby’s overall wellbeing and neurological development to detect stress-/pain-related behaviors or pathological conditions, such as respiratory distress syndrome and hyperbilirubinemia, to address when to perform procedures to reduce the baby’s stress/pain and interventions, such as therapeutic hypothermia, and to avoid severe complications, such as sepsis and necrotizing enterocolitis, thus reducing mortality. Based on literature and previous experiences, the first step to efficiently introduce HRV in the NICUs could consist in a monitoring system that uses photoplethysmography, which is low-cost and non-invasive, and displays one or a few metrics with good clinical utility. However, to fully harness HRV clinical potential and to greatly improve neonatal care, the monitoring systems will have to rely on modern bioinformatics (machine learning and artificial intelligence algorithms), which could easily integrate infant’s HRV metrics, vital signs, and especially past history, thus elaborating models capable to efficiently monitor and predict the infant’s clinical conditions. For this reason, hospitals and institutions will have to establish tight collaborations between the obstetric, neonatal, and pediatric departments: this way, healthcare would truly improve in every stage of the perinatal period (from conception to the first years of life), since information about patients’ health would flow freely among different professionals, and high-quality research could be performed integrating the data recorded in those departments.

Advances in cerebral palsy biomarkers

Cerebral palsy (CP), defined as a group of nonprogressive disorders of movement and posture, is the most common cause of severe neurodisability in children. The prevalence of CP is the same across the globe, affecting approximately 17 million people worldwide. Cerebral Palsy is an umbrella term used to describe the disease due to its inherent heterogeneity. For instance, CP has multiple (1) causes; (2) clinical types; (3) patterns of neuropathology on brain imaging and (4) it's associated with several developmental pathologies such as intellectual disability, autism, epilepsy, and visual impairment. Understanding its physiopathology is crucial to developing protective strategies. Despite its importance, there is still insufficient progress in the areas of CP prediction, early diagnosis, treatment, and prevention. Herein we describe the current risk factors and biomarkers used for the diagnosis and prediction of CP. With the advancement in biomarker discovery, we predict that our understanding of the etiopathophysiology of CP will also increase, lending to more opportunities for developing novel treatments and prognosis.