Skip to content
DraconiticResearch
Search Sources API
197,768 studies31 sourcesUpdated 21 min ago
PreprintCC BYmedRxiv

Systematic review finds keratoconus AI unproven in fellow eyes with normal topography and tomography

  • H. Jiang1
  • F. Gao
  • Y. Jie
  • L. Yang
  • Y. Li
  • Y. Jiang1
  1. 1Department of Ophthalmology, Peking Union Medical College Hospital, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China
Read full textPDFDOICite

Abstract

Published artificial intelligence for keratoconus is unproven in fellow eyes that topography and tomography call normal, the eyes where surgical decisions are made, although it reports near-perfect accuracy for early disease. In a registered systematic review of 573 studies (PROSPERO CRD420261441197), none of 2,035 claim-by-stratum units from 212 adjudicated studies established discrimination in these eyes. Most units (90.0%) answered a different question, never constructed the target population or certified its normality; external validation did not test label generation, and five systematic reviews did not assess it. In a public cohort, a frozen score yielded an area under the receiver operating characteristic curve of 0.996 against a same-eye index label and 0.549 against a contralateral proxy; its inputs reconstructed the label-generating index (r = 0.991), so accuracy there measured agreement with the index. Anchor-gap evaluation measures accuracy in these eyes against the contralateral clinical diagnosis.

Related

  1. Clinical outcomes of scleral lenses in keratoconus: a systematic review and Meta-analysis
    Mojgan Pakbin, Maryam Dashti, Ali Reza Akbarzadeh Baghban +4
    International Journal of OphthalmologyPeer-reviewedCC BY-NC-NDFull textClinical trials & evidence synthesisPMC
  2. Ocular surface symptoms and tear-film parameters in keratoconus: a systematic review and meta-analysis
    Yang Cao, Yue Yang, Wei Deng +1
    Frontiers in medicinePeer-reviewedCC BYFull textEpidemiology & public healthEurope PMCEurope PMCPMC
  3. Repeatability and Reliability of Corneal Tomography in Keratoconus Patients with Corneal Allogenic Intrastromal Ring Segments.
    Ugur Tunc, Dilan Colak, Burcu Yakut +4
    Beyoglu eye journalPeer-reviewedCC BY-NCFull textMedical imaging & radiologyEurope PMC
  4. Systematic review of spectral computed tomography quantitative parameters for predicting EGFR mutation status in non-small cell lung cancer
    Quantitative imaging in medicine and surgeryPeer-reviewedCC BY-NC-NDFull textMedical imaging & radiologyEurope PMC
  5. Optical Coherence Tomography Findings Following Intravascular Lithotripsy in Calcified Coronary Artery Disease: A Systematic Review and Meta-Analysis
    Giacomo Maria Cioffi, Stéphane Cook, Mario Togni +10
    CJC openPeer-reviewedCC BY-NC-NDFull textCardiovascular medicineEurope PMC

Cite

BibTeXRIS
BibTeX
@misc{jiang2026systematic,
  title = {{Systematic review finds keratoconus AI unproven in fellow eyes with normal topography and tomography}},
  author = {H. Jiang and F. Gao and Y. Jie and L. Yang and Y. Li and Y. Jiang},
  howpublished = {medRxiv},
  year = {2026},
  month = sep,
  doi = {10.64898/2026.09.29.26364068},
  url = {https://www.medrxiv.org/content/10.64898/2026.09.29.26364068v1},
  note = {CC BY},
}
RIS
TY  - UNPB
TI  - Systematic review finds keratoconus AI unproven in fellow eyes with normal topography and tomography
AU  - H. Jiang
AU  - F. Gao
AU  - Y. Jie
AU  - L. Yang
AU  - Y. Li
AU  - Y. Jiang
PY  - 2026
DA  - 2026/09/29
PB  - medRxiv
DO  - 10.64898/2026.09.29.26364068
UR  - https://www.medrxiv.org/content/10.64898/2026.09.29.26364068v1
AB  - Published artificial intelligence for keratoconus is unproven in fellow eyes that topography and tomography call normal, the eyes where surgical decisions are made, although it reports near-perfect accuracy for early disease. In a registered systematic review of 573 studies (PROSPERO CRD420261441197), none of 2,035 claim-by-stratum units from 212 adjudicated studies established discrimination in these eyes. Most units (90.0%) answered a different question, never constructed the target population or certified its normality; external validation did not test label generation, and five systematic reviews did not assess it. In a public cohort, a frozen score yielded an area under the receiver operating characteristic curve of 0.996 against a same-eye index label and 0.549 against a contralateral proxy; its inputs reconstructed the label-generating index (r = 0.991), so accuracy there measured agreement with the index. Anchor-gap evaluation measures accuracy in these eyes against the contralateral clinical diagnosis.
ER  -