Machine learning applications on neonatal sepsis treatment: a scoping review

Management of neonates at risk of early onset sepsis: a probability-based approach and recent literature appraisal : Update of the Swiss national guideline of the Swiss Society of Neonatology and the Pediatric Infectious Disease Group Switzerland

In Switzerland and other high-income countries, one out of 3000 to 5000 term and late preterm neonates develops early onset sepsis (EOS) associated with a mortality of around 3%, while incidence and mortality of EOS in very preterm infants are substantially higher. Exposure to antibiotics for suspected EOS is disproportionally high compared to the incidence of EOS with consequences for future health and antimicrobial resistance (AMR). A safe reduction of unnecessary antibiotic treatment has to be a major goal of new management strategies and guidelines. Antibiotics should be administered immediately in situations with clinical signs of septic shock. Group B streptococcus (GBS) and Escherichia coli (E. coli) are the leading pathogens of EOS. Amoxicillin combined with an aminoglycoside remains the first choice for empirical treatment. Serial physical examinations are recommended for all neonates with risk factors for EOS. Neonates without any clinical signs suggestive of EOS should not be treated with antibiotics. In Switzerland, we do not recommend the use of the EOS calculator, a risk stratification tool, due to its unclear impact in a population with an observed antibiotic exposure below 3%. Not all neonates with respiratory distress should be empirically treated with antibiotics. Isolated tachypnea or respiratory distress starting immediately after delivery by elective caesarean section or a clearly assessed alternative explanation than EOS for clinical signs may point towards a low probability of sepsis. On the other hand, unexplained prematurity with risk factors has an inherent higher risk of EOS. Before the start of antibiotic therapy, blood cultures should be drawn with a minimum volume of 1 ml in a single aerobic blood culture bottle. This standard procedure allows antibiotics to be stopped after 24 to 36 h if no pathogen is detected in blood cultures. Current data do not support the use of PCR-based pathogen detection in blood as a standard method. Lumbar puncture is recommended in blood culture-proven EOS, critical illness, or in the presence of neurological symptoms such as seizures or altered consciousness. The accuracy of a single biomarker measurement to distinguish inflammation from infection is low in neonates. Therefore, biomarker guidance is not a standard part of decision-making regarding the start or stop of antibiotic therapy but may be used as part of an algorithm and after appropriate education of health care teams. Every newborn started on antibiotics should be assessed for organ dysfunction with prompt initiation of respiratory and hemodynamic support if needed. An elevated lactate may be a sign of poor perfusion and requires a comprehensive assessment of the clinical condition. Interventions to restore perfusion include fluid boli with crystalloids and catecholamines. Neonates in critical condition should be cared for in a specialized unit. In situations with a low probability of EOS, antibiotics should be stopped as early as possible within the first 24 h after the start of therapy. In cases with microbiologically proven EOS, reassessment and streamlining of antibiotic therapy in neonates is an important step to minimize AMR.

CONCLUSION

 This guideline, developed through a critical review of the literature, facilitates a probability-based approach to the management of neonates at risk of early onset sepsis.

WHAT IS KNOWN

• Neonatal exposure to antibiotics is disproportionally high compared with the incidence of early onset sepsis with implications for future health and antimicrobial resistance.

WHAT IS NEW

• A probability-based approach may facilitate a more balanced management of neonatal sepsis and antibiotic stewardship.

Early detection of neonatal sepsis and reduction of overall antibiotic exposure: Towards precision medicine

Infections claim the lives of over half a million newborns annually and expose survivors to the risk of lifelong disability. The challenge to clinicians is to identify newborns with invasive infections rapidly, promptly initiate antimicrobial treatment, and take measures to prevent and treat organ dysfunction. Moreover, excessive antibiotic use is a global public health problem. Despite considerable research on clinical and laboratory markers of neonatal sepsis, the effective translation into clinical practice remains limited. There is no single clinical or laboratory marker, nor any combination of markers that definitively confirms or rules out neonatal sepsis. The interpretation of these markers should take into account their diagnostic value for a given patient, along with their added value to the clinical decision-making process. The digitalization of health care systems, combined with increased computational power and advances in machine learning, offers the possibility of developing accurate predictive algorithms for early detection of neonatal sepsis.

AI Algorithms for Modeling the Risk, Progression, and Treatment of Sepsis, Including Early-Onset Sepsis-A Systematic Review

Sepsis remains a significant contributor to neonatal mortality worldwide. However, the nonspecific nature of sepsis symptoms in neonates often leads to the necessity of empirical treatment, placing a burden of ineffective treatment on patients. Furthermore, the global challenge of antimicrobial resistance is exacerbating the situation. Artificial intelligence (AI) is transforming medical practice and in hospital settings. AI shows great potential for assessing sepsis risk and devising optimal treatment strategies. Background/Objectives: This review aims to investigate the application of AI in the detection and management of neonatal sepsis. Methods: A systematic literature review (SLR) evaluating AI methods in modeling and classifying sepsis between 1 January 2014, and 1 January 2024, was conducted. PubMed, Scopus, Cochrane, and Web of Science were systematically searched for English-language studies focusing on neonatal sepsis. Results: The analyzed studies predominantly utilized retrospective electronic medical record (EMR) data to develop, validate, and test AI models to predict sepsis occurrence and relevant parameters. Key predictors included low gestational age, low birth weight, high results of C-reactive protein and white blood cell counts, and tachycardia and respiratory failure. Machine learning models such as logistic regression, random forest, K-nearest neighbor (KNN), support vector machine (SVM), and XGBoost demonstrated effectiveness in this context. Conclusions: The summarized results of this review highlight the great promise of AI as a clinical decision support system for diagnostics, risk assessment, and personalized therapy selection in managing neonatal sepsis.

Struggling to Understand the NEC Spectrum-Could the Integration of Metabolomics, Clinical-Laboratory Data, and Other Emerging Technologies Help Diagnosis?

Necrotizing enterocolitis (NEC) is the most prevalent and potentially fatal intestinal injury mainly affecting premature infants, with significant long-term consequences for those who survive. This review explores the scale of the problem, highlighting advancements in epidemiology, the understanding of pathophysiology, and improvements in the prediction and diagnosis of this complex, multifactorial, and multifaced disease. Additionally, we focus on the potential role of metabolomics in distinguishing NEC from other conditions, which could allow for an earlier and more accurate classification of intestinal injuries in infants. By integrating metabolomic data with other diagnostic approaches, it is hoped to enhance our ability to predict outcomes and tailor treatments, ultimately improving care for affected infants.

Neonatal Infectious Disease: A Major Contributor to Infant Mortality Requiring Advances in Point-of-Care Diagnosis

Neonatal infectious disease continues to result in high rates of infant morbidity and mortality. Early- and late-onset disease represent difficult to detect and difficult to treat illnesses, particularly when antimicrobial resistant pathogens are present. Newborns are immunodeficient and are at increased risk of vertical and horizontal infection, with preterm infants increasingly susceptible. Additional risk factors associated with infection include prolonged use of a central catheter and/or ventilation, congenital abnormalities, admittance to intensive care units, and the use of broad-spectrum antibiotics. There is increasing recognition of the importance of the host microbiome and dysbiosis on neonatal infectious disease, including necrotising enterocolitis and sepsis in patients. Current diagnostic methods rely on blood culture, which is unreliable, time consuming, and can result in false negatives. There is a lack of accurate and reliable diagnostic tools available for the early detection of infectious disease in infants; therefore, efficient triage and treatment remains challenging. The application of biomarkers, machine learning, artificial intelligence, biosensors, and microfluidics technology, may offer improved diagnostic methodologies. Point-of-care devices, such diagnostic methodologies, may provide fast, reliable, and accurate diagnostic aids for neonatal patients. This review will discuss neonatal infectious disease as impacted by antimicrobial resistance and will highlight novel point-of-care diagnostic options.

Machine Learning: A Potential Therapeutic Tool to Facilitate Neonatal Therapeutic Decision Making

Bacterial infection is one of the major causes of neonatal morbidity and mortality worldwide. Finding rapid and reliable methods for early recognition and diagnosis of bacterial infections and early individualization of antibacterial drug administration are essential to eradicate these infections and prevent serious complications. However, this is often difficult to perform due to non-specific clinical presentations, low accuracy of current diagnostic methods, and limited knowledge of neonatal pharmacokinetics. Although neonatal medicine has been relatively late to embrace the benefits of machine learning (ML), there have been some initial applications of ML for the early prediction of neonatal sepsis and individualization of antibiotics. This article provides a brief introduction to ML and discusses the current state of the art in diagnosing and treating neonatal bacterial infections, gaps, potential uses of ML, and future directions to address the limitations of current studies. Neonatal bacterial infections involve a combination of physiologic development, disease expression, and treatment response outcomes. To address this complex relationship, future models could consider appropriate ML algorithms to capture time series features while integrating influences from the host, microbes, and drugs to optimize antimicrobial drug use in neonates. All models require prospective clinical trials to validate their clinical utility before clinical use.

Identification of key regulatory genes in the pathogenesis of COVID-19 and sepsis: An observational study

Patients with severe COVID-19 and those with sepsis have similar clinical manifestations. We used bioinformatics methods to identify the common hub genes in these 2 diseases. Two RNA-seq datasets from the Gene Expression Omnibus were used to identify common differentially expressed genes (DEGs) in COVID-19 and sepsis. These common genes were used for analysis of functional enrichment; pathway analysis; identification of associated transcription factors, metabolites, and miRNAs; and mapping of protein-protein interaction networks. The major hub genes of COVID-19 and sepsis were identified, and validation datasets were used to assess the value of these hub genes using receiver operating characteristic (ROC) curves. Analysis of the 800 common DEGs for COVID-19 and sepsis, as well as common transcription factors, miRNAs, and metabolites, demonstrated that the immune response had a key role in both diseases. DLGAP5, BUB1, CDK1, CCNB1, and BUB1B were the most important common hub genes. Analysis of a validation cohort indicated these 5 genes had significantly higher expression in COVID-19 patients and sepsis patients than in corresponding controls, and the area under the ROC curves ranged from 0.832 to 0.981 for COVID-19 and 0.840 to 0.930 for sepsis. We used bioinformatics tools to identify common DEGs, miRNAs, and transcription factors for COVID-19 and sepsis. The 5 identified hub genes had higher expression in validation cohorts of COVID-19 and sepsis. These genes had good or excellent diagnostic performance based on ROC analysis, and therefore have potential use as novel markers or therapeutic targets.

Tackling the Antimicrobial Resistance "Pandemic" with Machine Learning Tools: A Summary of Available Evidence

Antimicrobial resistance is recognised as one of the top threats healthcare is bound to face in the future. There have been various attempts to preserve the efficacy of existing antimicrobials, develop new and efficient antimicrobials, manage infections with multi-drug resistant strains, and improve patient outcomes, resulting in a growing mass of routinely available data, including electronic health records and microbiological information that can be employed to develop individualised antimicrobial stewardship. Machine learning methods have been developed to predict antimicrobial resistance from whole-genome sequencing data, forecast medication susceptibility, recognise epidemic patterns for surveillance purposes, or propose new antibacterial treatments and accelerate scientific discovery. Unfortunately, there is an evident gap between the number of machine learning applications in science and the effective implementation of these systems. This narrative review highlights some of the outstanding opportunities that machine learning offers when applied in research related to antimicrobial resistance. In the future, machine learning tools may prove to be superbugs' kryptonite. This review aims to provide an overview of available publications to aid researchers that are looking to expand their work with new approaches and to acquaint them with the current application of machine learning techniques in this field.

Early detection of late-onset neonatal sepsis from noninvasive biosignals using deep learning: A multicenter prospective development and validation study

BACKGROUND

Neonatal sepsis is responsible for significant morbidity and mortality worldwide. Its accurate and timely diagnosis is hindered by vague symptoms and the urgent necessity for early antibiotic intervention. The gold standard for diagnosing the condition is the identification of a pathogenic organism from normally sterile sites via laboratory testing. However, this method is resource-intensive and cannot be conducted continuously.

OBJECTIVE

This study aimed to predict the onset of late-onset sepsis (LOS) with good diagnostic value as early as possible using non-invasive biosignal measurements from neonatal intensive care unit (NICU) monitors.

METHODS

In this prospective multicenter study, we developed a multimodal machine learning algorithm based on a convolutional neural network (CNN) structure that uses the power spectral density (PSD) of recorded biosignals to predict the onset of LOS. This approach aimed to discern LOS-related pathogenic spectral signatures without labor-intensive manual artifact removal.

RESULTS

The model achieved an area under the receiver operating characteristic score of 0.810 (95 % CI 0.698-0.922) on the validation dataset. With an optimal operating point, LOS detection had 83 % sensitivity and 73 % specificity. The median early detection was 44 h before clinical suspicion. The results highlighted the additive importance of electrocardiogram and respiratory impedance (RESP) signals in improving predictive accuracy. According to a more detailed analysis, the predictive power arose from the morphology of the electrocardiogram's R-wave and sudden changes in the RESP signal.

CONCLUSION

Raw biosignals from NICU monitors, in conjunction with PSD transformation, as input to the CNN, can provide state-of-the-art prediction performance for LOS without the need for artifact removal. To the knowledge of the authors, this is the first study to highlight the independent and additive predictive potential of electrocardiogram R-wave morphology and concurrent, sudden changes in the RESP waveform in predicting the onset of LOS using non-invasive biosignals.