CoChat
Sign Up
Cited by 18Retinal Imaging and AnalysisRetinal Diseases and TreatmentsArtificial Intelligence in Healthcare

Performance and limitation of machine learning algorithms for diabetic retinopathy screening and its application in health management: a meta-analysis

Open access
Moannaei, Mehrsa, Jadidian, Faezeh, Doustmohammadi, Tahereh, Kiapasha, Amir Mohammad, Bayani, Romina, Rahmani, Mohammadreza, Jahanbazy, Mohammad Reza, Sohrabivafa, Fereshteh +4 more·2025·BioMedical Engineering OnLine
DOI 10.1186/s12938-025-01336-1PMC PMC11909973
Open this paper in CoChat — free account

Abstract

BACKGROUND: In recent years, artificial intelligence and machine learning algorithms have been used more extensively to diagnose diabetic retinopathy and other diseases. Still, the effectiveness of these methods has not been thoroughly investigated. This study aimed to evaluate the performance and limitations of machine learning and deep learning algorithms in detecting diabetic retinopathy. METHODS: This study was conducted based on the PRISMA checklist. We searched online databases, including PubMed, Scopus, and Google Scholar, for relevant articles up to September 30, 2023. After the title, abstract, and full-text screening, data extraction and quality assessment were done for the included studies. Finally, a meta-analysis was performed. RESULTS: We included 76 studies with a total of 1,371,517 retinal images, of which 51 were used for meta-analysis. Our meta-analysis showed a significant sensitivity and specificity with a percentage of 90.54 (95%CI [90.42, 90.66], P < 0.001) and 78.33% (95%CI [78.21, 78.45], P < 0.001). However, the AUC (area under curvature) did not statistically differ across studies, but had a significant figure of 0.94 (95% CI [- 46.71, 48.60], P = 1). CONCLUSIONS: Although machine learning and deep learning algorithms can properly diagnose diabetic retinopathy, their discriminating capacity is limited. However, they could simplify the diagnosing process. Further studies are required to improve algorithms.

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1186/s12938-025-01336-1

Cites

98 of 98 references verified0 citation errors · 0 not checkable

Statistics

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Model generalizability across diverse populations and variability in imaging quality due to differences in medical equipment and patient demographics are notable challenges for ML-based PAS diagnosis.

    “Despite these promising results, notable challenges included model generalizability across diverse populations and variability in imaging quality due to differences in medical equipment and patient demographics.”

    From Machine Learning in Maternity: Early Diagnosis of Placenta Accreta Spectrum (2026) · cites this paper as [14]

    The Discussion says that image quality and training-dataset diversity can affect ML performance in diabetic retinopathy screening, and that diverse, high-quality datasets are important for generalizability across populations. It does not address PAS diagnosis or say that differences in medical equipment and patient demographics cause variability in imaging quality. No page number is available in our copy. (LLM, conf=0.87)

    Experimental — accuracy not yet measured

  • The efficacy of ML algorithms for PAS diagnosis can vary across different populations, as disparities in patient demographics and health backgrounds may affect performance and generalizability.

    “The efficacy of ML algorithms can vary across different populations, as disparities in patient demographics and health backgrounds may affect the program's performance and generalizability [14].”

    From Machine Learning in Maternity: Early Diagnosis of Placenta Accreta Spectrum (2026) · cites this paper as [14]

    The Discussion says that ML performance in diabetic retinopathy screening can depend on training-dataset diversity and disease prevalence, and that diverse datasets are important for generalizability across populations. It also notes possible bias from poorly described patient characteristics. The paper does not study PAS diagnosis or specifically establish that differences in patient demographics and health backgrounds affect PAS-model performance. (LLM, conf=0.87)

    Experimental — accuracy not yet measured

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Cites

2 retracted

Flagged first, then by citations

Research works printed but not linked: 1

  1. [7]Shukla UV, Tripathy K. Diabetic retinopathy. Treasure Island: StatPearls Publishing; 2024. 32809640

Cited by

Sorted by citations

Related

Authors
Moannaei, M., Jadidian, F., Doustmohammadi, T., Kiapasha, A. M., Bayani, R., Rahmani, M., Jahanbazy, M. R., Sohrabivafa, F., Asadi Anar, M., Magsudy, A., Sadat Rafiei, S. K., & khakpour, Y.
Affiliations
  • Hormozgan University of Medical Sciences (IR)
  • Shahid Beheshti University (IR)
  • Zanjan University of Medical Sciences (IR)
  • Isfahan University of Medical Sciences (IR)
  • Islamic Azad University of Tabriz (IR)
  • Guilan University of Medical Sciences (IR)
ORCID
Venue
BioMedical Engineering OnLine
Publisher
Springer Science and Business Media LLC
Published
2025-03-14
Volume
24
Issue
1
Pages
34-34
ISSN
1475-925X (ISSN-L)
DOI
10.1186/s12938-025-01336-1
PMC
PMC11909973
OpenAlex
W4408438178
Cited by
18
References
98
Citations by year
2026: 10 · 2025: 8
FWCI
2.82
Type
journal article
Language
en
Open access
Licence
Locations
  • BioMedical Engineering OnLine (publishedVersion, cc-by-nc-nd)
  • Europe PMC (PubMed Central) (publishedVersion, cc-by-nc-nd)
  • PubMed Central (publishedVersion, cc-by-nc-nd)
  • PubMed (publishedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Topics
  • Health Sciences › Medicine › Radiology, Nuclear Medicine and Imaging › Retinal Imaging and Analysis
  • Health Sciences › Medicine › Ophthalmology › Retinal Diseases and Treatments
  • Health Sciences › Health Professions › Health Information Management › Artificial Intelligence in Healthcare
Keywords
diabetic retinopathy, automated screening, machine learning, deep learning, retinal image analysis, diagnostic accuracy, sensitivity and specificity, artificial intelligence in healthcare, health management
MeSH terms
Machine Learning, Algorithms, Diabetic Retinopathy, Humans

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review