EUVIMEDEUVIMED entdecken
ÜbersichtsarbeitSystematische ÜbersichtMetaanalyseOpen Access · CC BY 4.0

The Efficacy of Electronic Health Record-Based Artificial Intelligence Models for Early Detection of Pancreatic Cancer: A Systematic Review and Meta-Analysis

Die Wirksamkeit von auf elektronischen Gesundheitsakten basierenden KI-Modellen zur Früherkennung von Pankreaskarzinom: systematische Übersichtsarbeit und Metaanalyse

George G Makiev iD, Igor V Samoylenko iD, Valeria V Nazarova iD, Zahra R Magomedova iD, Alexey A Tryakin iD, Tigran G Gevorkyan iD

Cancers · 20.01.2026 · 18 (2) · 315

DOI: 10.3390/cancers18020315 ↗

Auf einen Blick

19 Studien zu KI-Modellen auf Basis strukturierter EHR-Daten zeigten eine gute Gesamttrennschärfe für die Vorhersage von Pankreaskarzinom. Neuronale Netze hatten die höchste gepoolte AUC.

Fragestellung

Wie gut sagen KI-Modelle auf Basis strukturierter EHR-Daten ein Pankreaskarzinom frühzeitig vorher?

Methode

Systematische Suche in PubMed, MedRxiv, BioRxiv und Google Scholar von 2010–2025; Metaanalyse mit Random-effects-Modell für AUC, Sensitivität und Spezifität.

Ergebnisse

19 Studien. Gepoolte AUC über alle Modelle 0,785 (95 %-KI 0,759–0,810). NN: AUC 0,826; LogReg 0,799; RF 0,762; XGB 0,779 (alle p < 0,001).

Einschränkungen

Extrem hohe Heterogenität bei Sensitivität und Spezifität (I2 ≥ 99,9 %); positive prädiktive Werte meist < 1 %; wenige Daten für einige Modelle.

Schlussfolgerung der Autor:innen

KI-Modelle auf Basis von EHR-Daten sind für die frühe Erkennung von Pankreaskarzinom vielversprechend. Wegen hoher Heterogenität und meist niedrigem PPV wird ein gezieltes Risikostratifikationskonzept statt allgemeinem Screening betont.

Kernaussagen

  • Gepoolte AUC über alle Modelle: 0,785 (95 %-KI 0,759–0,810)
  • Neuronale Netze: höchste AUC mit 0,826
  • PPV in den Studien meist < 1 %

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

Kurz gefasst

Übersichtsarbeit in der Zeitschrift Cancers, online erschienen am 20.01.2026. Verfasst von George G Makiev und 5 weiteren Autor:innen. Beteiligte Einrichtung: Predictive Modeling Department, Research Center for Artificial Intelligence in Healthcare, N.N. Blokhin National Medical Research Center of Oncology,…. Die Arbeit wurde bisher 2-mal zitiert (Crossref) und nennt 33 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: "Development of predictive models based on artificial intelligence for early detection of oncological diseases based on multimodal medical and socio-demographi…, The Ministry of Economic Development of the Russian Federation.

Auf einen BlickAbstractSchlagwörterBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Background

The persistently low 5-year survival rate for pancreatic cancer (PC) underscores the critical need for early detection. However, population-wide screening remains impractical. Artificial Intelligence (AI) models using electronic health record (EHR) data offer a promising avenue for pre-symptomatic risk stratification.

Objective

To systematically review and meta-analyze the performance of AI models for PC prediction based exclusively on structured EHR data.

Methods

We systematically searched PubMed, MedRxiv, BioRxiv, and Google Scholar (2010–2025). Inclusion criteria encompassed studies using EHR-derived data (excluding imaging/genomics), applying AI for PC prediction, reporting AUC, and including a non-cancer cohort. Two reviewers independently extracted data. Random-effects meta-analysis was performed for AUC, sensitivity (Se), and specificity (Sp) using R software version 4.5.1. Heterogeneity was assessed using I2 statistics and publication bias was evaluated.

Results

Of 946 screened records, 19 studies met the inclusion criteria. The pooled AUC across all models was 0.785 (95% CI: 0.759–0.810), indicating good overall discriminatory ability. Neural Network (NN) models demonstrated a statistically significantly higher pooled AUC (0.826) compared to Logistic Regression (LogReg, 0.799), Random Forests (RF, 0.762), and XGBoost (XGB, 0.779) (all p < 0.001). In analyses with sufficient data, models like Light Gradient Boosting (LGB) showed superior Se and Sp (99% and 98.7%, respectively) compared to NNs and LogReg, though based on limited studies. Meta-analysis of Se and Sp revealed extreme heterogeneity (I2 ≥ 99.9%), and the positive predictive values (PPVs) reported across studies were consistently low (often < 1%), reflecting the challenge of screening a low-prevalence disease.

Conclusions

AI models using EHR data show significant promise for early PC detection, with NNs achieving the highest pooled AUC. However, high heterogeneity and typically low PPV highlight the need for standardized methodologies and a targeted risk-stratification approach rather than general population screening. Future prospective validation and integration into clinical decision-support systems are essential.

Schlagwörter

Stichwörter der Autor:innen

artificial intelligenceearly detectionelectronic health recordsmachine learningpancreatic cancer

Autor:innen und Institutionen

  • Predictive Modeling Department, Research Center for Artificial Intelligence in Healthcare, N.N. Blokhin National Medical Research Center of Oncology, Ministry of Health of Russia, 24 Kashirskoe Shosse, Moscow 115522, Ru…
  • Department of Medical Oncology for Gastrointestinal Tumors, N.N. Blokhin National Medical Research Center of Oncology, Ministry of Health of Russia, 24 Kashirskoe Shosse, Moscow 115522, Russia.

ORCID-Profile: George G Makiev, Igor V Samoylenko, Valeria V Nazarova, Zahra R Magomedova, Alexey A Tryakin, Tigran G Gevorkyan

Förderung

  • "Development of predictive models based on artificial intelligence for early detection of oncological diseases based on multimodal medical and socio-demographi…
  • The Ministry of Economic Development of the Russian Federation

Zitieren

APA

Makiev, G. G., Samoylenko, I. V., Nazarova, V. V., Magomedova, Z. R., Tryakin, A. A., & Gevorkyan, T. G. (2026). The Efficacy of Electronic Health Record-Based Artificial Intelligence Models for Early Detection of Pancreatic Cancer: A Systematic Review and Meta-Analysis. Cancers, 18(2), 315. https://doi.org/10.3390/cancers18020315

Vancouver

Makiev GG, Samoylenko IV, Nazarova VV, Magomedova ZR, Tryakin AA, Gevorkyan TG. The Efficacy of Electronic Health Record-Based Artificial Intelligence Models for Early Detection of Pancreatic Cancer: A Systematic Review and Meta-Analysis. Cancers. 2026;18(2):315. doi:10.3390/cancers18020315

BibTeX
@article{Makiev2026,
  author = {Makiev, George G and Samoylenko, Igor V and Nazarova, Valeria V and Magomedova, Zahra R and Tryakin, Alexey A and Gevorkyan, Tigran G},
  title = {The Efficacy of Electronic Health Record-Based Artificial Intelligence Models for Early Detection of Pancreatic Cancer: A Systematic Review and Meta-Analysis},
  journal = {Cancers},
  year = {2026},
  volume = {18},
  number = {2},
  pages = {315},
  doi = {10.3390/cancers18020315}
}
Kostenlos starten →