Zeboulon 2020b.

Study characteristics
Patient Sampling	Retrospective, machine‐learning, experimental study. 22,066 Orbscan (Bausch&Lomb, USA) examinations were randomly extracted from the Orbscan database using the batch export functionality. The last examination of the first visit for each eye of each person was selected. This process reduced the number of examinations to 13,705. The cases were divided into the following groups: normal, keratoconus, history of myopic refractive surgery, Fuchs' corneal dystrophy, and other.
Patient characteristics and setting	Normal Keratoconus group: if the anterior curvature map showed 1 of the classic keratoconus patterns described by Rabinowitz et al. associated with corneal thinning Refractive surgery group: cases that underwent myopic laser surgery and had an oblate anterior surface (flat in its centre), a prolate posterior surface (steep in its centre), a central corneal thinning and lower central curvature values compared to the periphery. Fuchs' examinations had a central pachymetry > 600 microns with reduced peripheral thickness in the corneal periphery and an oblate posterior surface. Other: all examinations that could not be assigned to a specific class with a good level of confidence, including bad quality examinations
Index tests	The efficiency of unsupervised algorithms was tested to extract and sort usable examinations from a large unlabelled corneal topography database into different diagnostic clusters, with little human intervention, data cleaning or feature selection. Convolutional neural network (CNN) was used.
Target condition and reference standard(s)	All 13,705 examinations were manually labelled and checked by 2 corneal topography experts (with at least 5 years of practice in a corneal and refractive surgery department) in a random order. The cases were diagnosed before the convolutional neural network analysis.
Flow and timing	All participants were diagnosed by 2 cornea specialists. All cases were included in the analysis.
Comparative	Not applicable
Notes	No funding source mentioned.
Methodological quality
Item	Authors' judgement	Risk of bias	Applicability concerns
DOMAIN 1: Patient selection
Was a consecutive or random sample of patients enrolled?	Yes
Was a case‐control design avoided?	No
Did the study avoid inappropriate exclusions?	No
Could the selection of patients have introduced bias?		High risk
Are there concerns that the included patients and setting do not match the review question?			High
DOMAIN 2: Index test (All tests)
Were the index test results interpreted without knowledge of the results of the reference standard?	Yes
If a threshold was used, was it pre‐specified?	Unclear
Was the model designed in an appropriate manner?	Yes
Could the conduct or interpretation of the index test have introduced bias?		Low risk
Are there concerns that the index test, its conduct, or interpretation differ from the review question?			Low concern
DOMAIN 3: Reference standard
Is the reference standard likely to correctly classify the target condition?	Yes
Were the reference standard results interpreted without knowledge of the results of the index tests?	Yes
Could the reference standard, its conduct, or its interpretation have introduced bias?		Low risk
Are there concerns that the target condition as defined by the reference standard does not match the question?			Low concern
DOMAIN 4: Flow and timing
Did all patients receive the same reference standard?	Yes
Were all patients included in the analysis?	Yes
Could the patient flow have introduced bias?		Low risk
DOMAIN 5: Comparative

AI: artificial intelligence; AS‐OCT: anterior segment optical coherence tomography; AST: astigmatism index; BAD‐D: Belin‐Ambrósio Enhanced Ectasia Display total deviation; D: dioptre; I‐S: inferior‐superior; KISA% index: keratoconus percentage index, derived from central keratometry, the inferior‐superior value, the astigmatism index, and the SRAX index, an expression of irregular astigmatism occurring in keratoconus; LASIK: laser‐assisted in situ keratomileusis; OCT: optical coherence tomography; PPK: percent probability of keratoconus; SD: standard deviation; TMS: Topographic Modeling System.