EUVIMEDOdkrijte EUVIMED
Systematische ÜbersichtMetaanalyseOpen access

Artificial Intelligence Models for Pediatric Lung Sound Analysis: Systematic Review and Meta-Analysis

Modelle der künstlichen Intelligenz für die Analyse kindlicher Lungengeräusche: systematische Übersicht und Metaanalyse

Ji Soo Park iD, Sa-Yoon Park iD, Jae Won Moon iD, Kwangsoo Kim iD, Dong In Suh iD

Journal of Medical Internet Research · 2025-04-18 · 27 · e66491

DOI: 10.2196/66491 ↗

Auf einen Blick

Die Metaanalyse von 41 Studien zeigt hohe Genauigkeit von ML-Modellen bei der Analyse kindlicher Lungengeräusche. Für Wheezing und abnorme Lungengeräusche lagen Sensitivität und Spezifität hoch.

Fragestellung

Wie gut leisten ML-Modelle in der Analyse kindlicher Lungengeräusche?

Methode

Systematische Übersichtsarbeit und Metaanalyse; Suche in Medline via PubMed, Embase, Web of Science, OVID und IEEE Xplore bis 16. Dezember 2024; Risiko-of-Bias-Bewertung mit modifiziertem QUADAS-2.

Ergebnisse

41 Studien. Wheeze-Erkennung: gepoolte Sensitivität 0,902 (95 %-KI 0,726–0,970), Spezifität 0,955 (95 %-KI 0,762–0,993). Abnorme Lungengeräusche: Sensitivität 0,907 (95 %-KI 0,816–0,956), Spezifität 0,877 (95 %-KI 0,813–0,921).

Einschränkungen

Hohe Heterogenität bei Datensatzgröße, Annotation und Bewertung; meist kleine, monozentrische Datensätze; begrenzte externe Validierung.

Schlussfolgerung der Autor:innen

ML-Modelle zeigen hohe Genauigkeit in der Analyse kindlicher Lungengeräusche, sind aber durch heterogene Datensätze, fehlende Standards und begrenzte externe Validierung eingeschränkt.

Kernaussagen

  • 41 Studien wurden eingeschlossen.
  • Wheeze-Erkennung: Sensitivität 0,902, Spezifität 0,955.
  • Abnorme Lungengeräusche: Sensitivität 0,907, Spezifität 0,877.

Deutsche Zusammenfassung: KI-gestützt aus dem Original-Abstract erstellt; alle Zahlen wurden automatisch mit dem Original abgeglichen. Maßgeblich ist der Original-Abstract.

At a glance

Systematische Übersicht in Journal of Medical Internet Research, published online on 2025-04-18. Written by Ji Soo Park and 4 co-authors. Institution: Department of Pediatrics, Seoul National University College of Medicine, Seoul, Republic of Korea. Major MEDLINE topics: Respiratory Sounds, Artificial Intelligence, Auscultation. Cited 21 times so far (Crossref); 72 references. The full text is freely available in PubMed Central.

Auf einen BlickPovzetekSubjectsBibliografski podatkiCiteRelated studies

Abstract (original language)

Background

Pediatric respiratory diseases, including asthma and pneumonia, are major causes of morbidity and mortality in children. Auscultation of lung sounds is a key diagnostic tool but is prone to subjective variability. The integration of artificial intelligence (AI) and machine learning (ML) with electronic stethoscopes offers a promising approach for automated and objective lung sound.

Objective

This systematic review and meta-analysis assess the performance of ML models in pediatric lung sound analysis. The study evaluates the methodologies, model performance, and database characteristics while identifying limitations and future directions for clinical implementation.

Methods

A systematic search was conducted in Medline via PubMed, Embase, Web of Science, OVID, and IEEE Xplore for studies published between January 1, 1990, and December 16, 2024. Inclusion criteria are as follows: studies developing ML models for pediatric lung sound classification with a defined database, physician-labeled reference standard, and reported performance metrics. Exclusion criteria are as follows: studies focusing on adults, cardiac auscultation, validation of existing models, or lacking performance metrics. Risk of bias was assessed using a modified Quality Assessment of Diagnostic Accuracy Studies (version 2) framework. Data were extracted on study design, dataset, ML methods, feature extraction, and classification tasks. Bivariate meta-analysis was performed for binary classification tasks, including wheezing and abnormal lung sound detection.

Results

A total of 41 studies met the inclusion criteria. The most common classification task was binary detection of abnormal lung sounds, particularly wheezing. Pooled sensitivity and specificity for wheeze detection were 0.902 (95% CI 0.726-0.970) and 0.955 (95% CI 0.762-0.993), respectively. For abnormal lung sound detection, pooled sensitivity was 0.907 (95% CI 0.816-0.956) and specificity 0.877 (95% CI 0.813-0.921). The most frequently used feature extraction methods were Mel-spectrogram, Mel-frequency cepstral coefficients, and short-time Fourier transform. Convolutional neural networks were the predominant ML model, often combined with recurrent neural networks or residual network architectures. However, high heterogeneity in dataset size, annotation methods, and evaluation criteria were observed. Most studies relied on small, single-center datasets, limiting generalizability.

Conclusions

ML models show high accuracy in pediatric lung sound analysis, but face limitations due to dataset heterogeneity, lack of standard guidelines, and limited external validation. Future research should focus on standardized protocols and the development of large-scale, multicenter datasets to improve model robustness and clinical implementation.

Subjects

MeSH terms (MEDLINE)

Respiratory SoundsArtificial IntelligenceAuscultationHumansChildMachine LearningStethoscopes

Author keywords

abnormal lung sound detectionartificial intelligenceasthmaauscultationchildrendiagnosticlung sound analysismachine learningmel-spectrogrammorbiditymortalitypediatricpneumoniarespiratory disease classificationrespiratory pathology

Authors and institutions

  • Department of Pediatrics, Seoul National University College of Medicine, Seoul, Republic of Korea.
  • The Institute of Convergence Medicine with Innovative Technology, Seoul National University Hospital, Seoul, Republic of Korea.
  • Department of Physiology, College of Korean Medicine, Wonkwang University, Iksan, Republic of Korea.
  • Department of Medicine, Seoul National University College of Medicine, Seoul, Republic of Korea.

ORCID profiles: Ji Soo Park, Sa-Yoon Park, Jae Won Moon, Kwangsoo Kim, Dong In Suh

Cite

APA

Park, J. S., Park, S. Y., Moon, J. W., Kim, K., & Suh, D. I. (2025). Artificial Intelligence Models for Pediatric Lung Sound Analysis: Systematic Review and Meta-Analysis. Journal of Medical Internet Research, 27, e66491. https://doi.org/10.2196/66491

Vancouver

Park JS, Park SY, Moon JW, Kim K, Suh DI. Artificial Intelligence Models for Pediatric Lung Sound Analysis: Systematic Review and Meta-Analysis. Journal of Medical Internet Research. 2025;27:e66491. doi:10.2196/66491

BibTeX
@article{Park2025,
  author = {Park, Ji Soo and Park, Sa-Yoon and Moon, Jae Won and Kim, Kwangsoo and Suh, Dong In},
  title = {Artificial Intelligence Models for Pediatric Lung Sound Analysis: Systematic Review and Meta-Analysis},
  journal = {Journal of Medical Internet Research},
  year = {2025},
  volume = {27},
  pages = {e66491},
  doi = {10.2196/66491}
}
Start for free →