Web-based newborn screening system for metabolic diseases: machine learning versus clinicians

Improving newborn screening in India: Disease gaps and quality control

In India, newborn screening (NBS) is essential for detecting health problems in infants. Despite significant progress, significant gaps and challenges persist. India has made great strides in genomics dueto the existence of the National Institute of Biomedical Genomics in West Bengal. The work emphasizes the challenges NBS programs confront with technology, budgetary constraints, insufficient counseling, inequality in illness panels, and a lack of awareness. Advancements in technology, such as genetic testing and next-generation sequencing, are expected to significantly transform the process. The integration of analytical tools, artificial intelligence, and machine learning algorithms could improve the efficiency of newborn screening programs, offering a personalized healthcare approach. It is critical to address gaps in information, inequities in illness incidence, budgetary restrictions, and inadequate counseling. Strengthening national NBS programs requires increased public awareness and coordinated efforts between state and central agencies. Quality control procedures must be used at every level for implementation to be successful. Additional studies endeavor to enhance NBS in India through public education, illness screening expansion, enhanced quality control, government incentive implementation, partnership promotion, and expert training. Improved neonatal health outcomes and the viability of the program across the country will depend heavily on new technology and counseling techniques.

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.

Application of machine learning tools and integrated OMICS for screening and diagnosis of inborn errors of metabolism

INTRODUCTION

Tandem mass spectrometry (TMS) has emerged an important screening tool for various metabolic disorders in newborns. However, there is inherent risk of false positive outcomes. Objective To establish analyte-specific cutoffs in TMS by integrating metabolomics and genomics data to avoid false positivity and false negativity and improve its clinical utility.

METHODS

TMS was performed on 572 healthy and 3000 referred newborns. Urine organic acid analysis identified 23 types of inborn errors in 99 referred newborns. Whole exome sequencing was performed in 30 positive cases. The impact of physiological changes such as age, gender, and birthweight on various analytes was explored in healthy newborns. Machine learning tools were used to integrate demographic data with metabolomics and genomics data to establish disease-specific cut-offs; identify primary and secondary markers; build classification and regression trees (CART) for better differential diagnosis; for pathway modeling.

RESULTS

This integration helped in differentiating B12 deficiency from methylmalonic acidemia (MMA) and propionic acidemia (Phi coefficient=0.93); differentiating transient tyrosinemia from tyrosinemia type 1 (Phi coefficient=1.00); getting clues about the possible molecular defect in MMA to initiate appropriate intervention (Phi coefficient=1.00); to link pathogenicity scores with metabolomics profile in tyrosinemia (r2=0.92). CART model helped in establishing differential diagnosis of urea cycle disorders (Phi coefficient=1.00).

CONCLUSION

Calibrated cut-offs of different analytes in TMS and machine learning-based establishment of disease-specific thresholds of these markers through integrated OMICS have helped in improved differential diagnosis with significant reduction of the false positivity and false negativity rates.

Validation of a targeted metabolomics panel for improved second-tier newborn screening

Improved second-tier assays are needed to reduce the number of false positives in newborn screening (NBS) for inherited metabolic disorders including those on the Recommended Uniform Screening Panel (RUSP). We developed an expanded metabolite panel for second-tier testing of dried blood spot (DBS) samples from screen-positive cases reported by the California NBS program, consisting of true- and false-positives from four disorders: glutaric acidemia type I (GA1), methylmalonic acidemia (MMA), ornithine transcarbamylase deficiency (OTCD), and very long-chain acyl-CoA dehydrogenase deficiency (VLCADD). This panel was assembled from known disease markers and new features discovered by untargeted metabolomics and applied to second-tier analysis of single DBS punches using liquid chromatography-tandem mass spectrometry (LC-MS/MS) in a 3-min run. Additionally, we trained a Random Forest (RF) machine learning classifier to improve separation of true- and false positive cases. Targeted metabolomic analysis of 121 analytes from DBS extracts in combination with RF classification at a sensitivity of 100% reduced false positives for GA1 by 83%, MMA by 84%, OTCD by 100%, and VLCADD by 51%. This performance was driven by a combination of known disease markers (3-hydroxyglutaric acid, methylmalonic acid, citrulline, and C14:1), other amino acids and acylcarnitines, and novel metabolites identified to be isobaric to several long-chain acylcarnitine and hydroxy-acylcarnitine species. These findings establish the effectiveness of this second-tier test to improve screening for these four conditions and demonstrate the utility of supervised machine learning in reducing false-positives for conditions lacking clearly discriminating markers, with future studies aimed at optimizing and expanding the panel to additional disease targets.

Random forest classifier improving phenylketonuria screening performance in two Chinese populations

Phenylketonuria (PKU) is a genetic disorder with amino acid metabolic defect, which does great harms to the development of newborns and children. Early diagnosis and treatment can effectively prevent the disease progression. Here we developed a PKU screening model using random forest classifier (RFC) to improve PKU screening performance with excellent sensitivity, false positive rate (FPR) and positive predictive value (PPV) in all the validation dataset and two testing Chinese populations. RFC represented outstanding advantages comparing several different classification models based on machine learning and the traditional logistic regression model. RFC is promising to be applied to neonatal PKU screening.

Artificial Intelligence Applications in Health Care Practice: A Scoping Review (Preprint)

Background

Artificial intelligence (AI) is often heralded as a potential disruptor that will transform the practice of medicine. The amount of data collected and available in health care, coupled with advances in computational power, has contributed to advances in AI and an exponential growth of publications. However, the development of AI applications does not guarantee their adoption into routine practice. There is a risk that despite the resources invested, benefits for patients, staff, and society will not be realized if AI implementation is not better understood.

Objective

The aim of this study was to explore how the implementation of AI in health care practice has been described and researched in the literature by answering 3 questions: What are the characteristics of research on implementation of AI in practice? What types and applications of AI systems are described? What characteristics of the implementation process for AI systems are discernible?

Methods

A scoping review was conducted of MEDLINE (PubMed), Scopus, Web of Science, CINAHL, and PsycINFO databases to identify empirical studies of AI implementation in health care since 2011, in addition to snowball sampling of selected reference lists. Using Rayyan software, we screened titles and abstracts and selected full-text articles. Data from the included articles were charted and summarized.

Results

Of the 9218 records retrieved, 45 (0.49%) articles were included. The articles cover diverse clinical settings and disciplines; most (32/45, 71%) were published recently, were from high-income countries (33/45, 73%), and were intended for care providers (25/45, 56%). AI systems are predominantly intended for clinical care, particularly clinical care pertaining to patient-provider encounters. More than half (24/45, 53%) possess no action autonomy but rather support human decision-making. The focus of most research was on establishing the effectiveness of interventions (16/45, 35%) or related to technical and computational aspects of AI systems (11/45, 24%). Focus on the specifics of implementation processes does not yet seem to be a priority in research, and the use of frameworks to guide implementation is rare.

Conclusions

Our current empirical knowledge derives from implementations of AI systems with low action autonomy and approaches common to implementations of other types of information systems. To develop a specific and empirically based implementation framework, further research is needed on the more disruptive types of AI systems being implemented in routine care and on aspects unique to AI implementation in health care, such as building trust, addressing transparency issues, developing explainable and interpretable solutions, and addressing ethical concerns around privacy and data protection.

Machine and cognitive intelligence for human health: systematic review

Brain informatics is a novel interdisciplinary area that focuses on scientifically studying the mechanisms of human brain information processing by integrating experimental cognitive neuroscience with advanced Web intelligence-centered information technologies. Web intelligence, which aims to understand the computational, cognitive, physical, and social foundations of the future Web, has attracted increasing attention to facilitate the study of brain informatics to promote human health. A large number of articles created in the recent few years are proof of the investment in Web intelligence-assisted human health. This study systematically reviews academic studies regarding article trends, top journals, subjects, countries/regions, and institutions, study design, artificial intelligence technologies, clinical tasks, and performance evaluation. Results indicate that literature is especially welcomed in subjects such as medical informatics and health care sciences and service. There are several promising topics, for example, random forests, support vector machines, and conventional neural networks for disease detection and diagnosis, semantic Web, ontology mining, and topic modeling for clinical or biomedical text mining, artificial neural networks and logistic regression for prediction, and convolutional neural networks and support vector machines for monitoring and classification. Additionally, future research should focus on algorithm innovations, additional information use, functionality improvement, model and system generalization, scalability, evaluation, and automation, data acquirement and quality improvement, and allowing interaction. The findings of this study help better understand what and how Web intelligence can be applied to promote healthcare procedures and clinical outcomes. This provides important insights into the effective use of Web intelligence to support informatics-enabled brain studies.

Using random forest to detect multiple inherited metabolic diseases simultaneously based on GC-MS urinary metabolomics

Inborn errors of metabolism, also known as inherited metabolic diseases (IMDs), are related to genetic mutations and cause corresponding biochemical metabolic disorder of newborns and even sudden infant death. Timely detection and diagnosis of IMDs are of great significance for improving survival of newborns. Here we propose a strategy for simultaneously detecting six types of IMDs via combining GC-MS technique with the random forest algorithm (RF). Clinical urine samples from IMD and healthy patients are analyzed using GC-MS for acquiring metabolomics data. Then, the RF model is established as a multi-classification tool for the GC-MS data. Compared with the models built by artificial neural network and support vector machine, the results demonstrated the RF model has superior performance of high specificity, sensitivity, precision, accuracy, and matthews correlation coefficients on identifying all six types of IMDs and normal samples. The proposed strategy can afford a useful method for reliable and effective identification of multiple IMDs in clinical diagnosis.

Severity modeling of propionic acidemia using clinical and laboratory biomarkers

Purpose

To conduct a proof-of-principle study to identify subtypes of propionic acidemia (PA) and associated biomarkers.

Methods

Data from a clinically diverse PA patient population (https://clinicaltrials.gov/ct2/show/NCT02890342) were used to train and test machine learning models, identify PA-relevant biomarkers, and perform validation analysis using data from liver-transplanted participants. k-Means clustering was used to test for the existence of PA subtypes. Expert knowledge was used to define PA subtypes (mild and severe). Given expert classification, supervised machine learning (support vector machine with a polynomial kernel, svmPoly) performed dimensional reduction to define relevant features of each PA subtype.

Results

Forty participants enrolled in the study; five underwent liver transplant. Analysis with k-means clustering indicated that several PA subtypes may exist on the biochemical continuum. The conventional PA biomarkers, plasma total 2-methylctirate and propionylcarnitine, were not statistically significantly different between nontransplanted and transplanted participants motivating us to search for other biomarkers. Unbiased dimensional reduction using svmPoly revealed that plasma transthyretin, alanine:serine ratio, GDF15, FGF21, and in vivo 1-¹³C-propionate oxidation, play roles in defining PA subtypes.

Conclusion

Support vector machine prioritized biomarkers that helped classify propionic acidemia patients according to severity subtypes, with important ramifications for future clinical trials and management of PA.

Graphical Abstract

Improving the Diagnosis of Phenylketonuria by Using a Machine Learning-Based Screening Model of Neonatal MRM Data

Phenylketonuria (PKU) is a common genetic metabolic disorder that affects the infant's nerve development and manifests as abnormal behavior and developmental delay as the child grows. Currently, a triple–quadrupole mass spectrometer (TQ-MS) is a common high-accuracy clinical PKU screening method. However, there is high false-positive rate associated with this modality, and its reduction can provide a diagnostic and economic benefit to both pediatric patients and health providers. Machine learning methods have the advantage of utilizing high-dimensional and complex features, which can be obtained from the patient's metabolic patterns and interrogated for clinically relevant knowledge. In this study, using TQ-MS screening data of more than 600,000 patients collected at the Newborn Screening Center of Shanghai Children's Hospital, we derived a dataset containing 256 PKU-suspected cases. We then developed a machine learning logistic regression analysis model with the aim to minimize false-positive rates in the results of the initial PKU test. The model attained a 95–100% sensitivity, the specificity was improved 53.14%, and positive predictive value increased from 19.14 to 32.16%. Our study shows that machine learning models may be used as a pediatric diagnosis aid tool to reduce the number of suspected cases and to help eliminate patient recall. Our study can serve as a future reference for the selection and evaluation of computational screening methods.

Reducing False-Positive Results in Newborn Screening Using Machine Learning

Newborn screening (NBS) for inborn metabolic disorders is a highly successful public health program that by design is accompanied by false-positive results. Here we trained a Random Forest machine learning classifier on screening data to improve prediction of true and false positives. Data included 39 metabolic analytes detected by tandem mass spectrometry and clinical variables such as gestational age and birth weight. Analytical performance was evaluated for a cohort of 2777 screen positives reported by the California NBS program, which consisted of 235 confirmed cases and 2542 false positives for one of four disorders: glutaric acidemia type 1 (GA-1), methylmalonic acidemia (MMA), ornithine transcarbamylase deficiency (OTCD), and very long-chain acyl-CoA dehydrogenase deficiency (VLCADD). Without changing the sensitivity to detect these disorders in screening, Random Forest-based analysis of all metabolites reduced the number of false positives for GA-1 by 89%, for MMA by 45%, for OTCD by 98%, and for VLCADD by 2%. All primary disease markers and previously reported analytes such as methionine for MMA and OTCD were among the top-ranked analytes. Random Forest's ability to classify GA-1 false positives was found similar to results obtained using Clinical Laboratory Integrated Reports (CLIR). We developed an online Random Forest tool for interpretive analysis of increasingly complex data from newborn screening.

Applications of Machine Learning in miRNA Discovery and Target Prediction

MicroRNA (miRNA) is a small non-coding molecule that is involved in gene regulation and RNA silencing by complementary on their targets. Experimental methods for target prediction can be time-consuming and expensive. Thus, the application of the computational approach is implicated to enlighten these complications with experimental studies. However, there is still a need for an optimized approach in miRNA biology. Therefore, machine learning (ML) would initiate a new era of research in miRNA biology towards potential diseases biomarker. In this article, we described the application of ML approaches in miRNA discovery and target prediction with functions and future prospective. The implementation of a new era of computational methodologies in this direction would initiate further advanced levels of discoveries in miRNA.

Simultaneous detection of multiple inherited metabolic diseases using GC-MS urinary metabolomics by chemometrics multi-class classification strategies

Metabonomics has been widely used in disease diagnosis and clinically practical methods often require the detection of multi-class bio-samples. In this work, multi-class classification methods were investigated to simultaneously discriminate among 6 inherited metabolic diseases (IMDs) and the normal instances using gas chromatography-mass spectrometry (GC-MS) of urine samples. Two common multi-class classification strategies, one-against-all (OAA) and one-against-one (OAO) were compared and enhanced using a novel ensemble classification strategy (ECS), which developed a set of sequential sub-classifiers by fusion of OAA and OAO and made the final classification decisions using softmax function. GC-MS data of 240 instances of 6 IMDs and healthy controls were classified by different strategies based on orthogonal partial least squares discriminant analysis (OPLS-DA) and particle swarm optimization (PSO) algorithm was performed for feature selection. By OAA and OAO, the classification accuracies were 70.00% and 82.86%, respectively. Using the two methods based on ECS, the total classification accuracies were 0.9143 and 0.9429. The newly proposed ECS will provide a useful multi-class classification tool for simultaneous detection of clinically similar IMDs and promote practical and reliable diagnosis of IMDs using metabonomics data.

Screening newborns for metabolic disorders based on targeted metabolomics using tandem mass spectrometry

The main purpose of newborn screening is to diagnose genetic, metabolic, and other inherited disorders, at their earliest to start treatment before the clinical manifestations become evident. Understanding and tracing the biochemical data obtained from tandem mass spectrometry is vital for early diagnosis of metabolic diseases associated with such disorders. Accordingly, it is important to focus on the entire diagnostic process, including differential and confirmatory diagnostic options, and the major factors that influence the results of biochemical analysis. Compared to regular biochemical testing, this is a complex process carried out by a medical physician specialist. It is comprised of an integrated program requiring multidisciplinary approach such as, pediatric specialist, expert scientist, clinical laboratory technician, and nutritionist. Tandem mass spectrometry is a powerful tool to improve screening of newborns for diverse metabolic diseases. It is likely to be used to analyze other treatable disorders or significantly improve existing newborn tests to allow broad scale and precise testing. This new era of various screening programs, new treatments, and the availability of detection technology will prove to be beneficial for the future generations.

A telesurveillance system with automatic electrocardiogram interpretation based on support vector machine and rule-based processing

Background

Telehealth care is a global trend affecting clinical practice around the world. To mitigate the workload of health professionals and provide ubiquitous health care, a comprehensive surveillance system with value-added services based on information technologies must be established.

Objective

We conducted this study to describe our proposed telesurveillance system designed for monitoring and classifying electrocardiogram (ECG) signals and to evaluate the performance of ECG classification.

Methods

We established a telesurveillance system with an automatic ECG interpretation mechanism. The system included: (1) automatic ECG signal transmission via telecommunication, (2) ECG signal processing, including noise elimination, peak estimation, and feature extraction, (3) automatic ECG interpretation based on the support vector machine (SVM) classifier and rule-based processing, and (4) display of ECG signals and their analyzed results. We analyzed 213,420 ECG signals that were diagnosed by cardiologists as the gold standard to verify the classification performance.

Results

In the clinical ECG database from the Telehealth Center of the National Taiwan University Hospital (NTUH), the experimental results showed that the ECG classifier yielded a specificity value of 96.66% for normal rhythm detection, a sensitivity value of 98.50% for disease recognition, and an accuracy value of 81.17% for noise detection. For the detection performance of specific diseases, the recognition model mainly generated sensitivity values of 92.70% for atrial fibrillation, 89.10% for pacemaker rhythm, 88.60% for atrial premature contraction, 72.98% for T-wave inversion, 62.21% for atrial flutter, and 62.57% for first-degree atrioventricular block.

Conclusions

Through connected telehealth care devices, the telesurveillance system, and the automatic ECG interpretation system, this mechanism was intentionally designed for continuous decision-making support and is reliable enough to reduce the need for face-to-face diagnosis. With this value-added service, the system could widely assist physicians and other health professionals with decision making in clinical practice. The system will be very helpful for the patient who suffers from cardiac disease, but for whom it is inconvenient to go to the hospital very often.