Sideroudi 2017.

Study characteristics
Patient Sampling	A cross‐sectional, observational study. Study participants were recruited from the cornea outpatient service in a consecutive if eligible basis; controls were refractive surgery candidates. 80 eyes formed the keratoconus group, 55 eyes formed the subclinical keratoconus group, and 50 normal eyes populated the control group.
Patient characteristics and setting	The keratoconus group included people diagnosed with stage I keratoconus according to the Amsler–Krumeich classification system. Inclusion criteria were as follows. Eccentric steepening Myopia or induced astigmatism < 5.00 D Mean central keratometry readings < 48.00 D The subclinical keratoconus group included people diagnosed with subclinical keratoconus. Inclusion criteria were as follows. Diagnosis of keratoconus in the fellow eye according to the Amsler‐Krumleich criteria KISA% index between 60% and 100% in the subclinical keratoconus eye Lack of any keratoconus‐related findings/signs in the slit‐lamp biomicroscopy Refractive surgery candidates with no evidence of ectasia populated the control group. Exclusion criteria: previous incisional eye surgery, corneal scars and opacities, history of herpetic keratitis, severe eye dryness, current corneal infection, glaucoma, suspicion of glaucoma, intraocular pressure‐lowering treatment, pregnancy or nursing, contact lens use, and underlying autoimmune disease.
Index tests	A self‐developed algorithm based on logistic regression in Visual Basic for Microsoft Excel performed a Fourier series harmonic analysis for the posterior corneal sagittal curvature data and evaluated the derived parameters in the diagnosis of Subclinical Keratoconus and Keratoconus.
Target condition and reference standard(s)	The diagnosis of keratoconus was based on the Amsler‐Krumeich classification and tomography images; however, it was unclear who made the diagnosis. The diagnosis was made before the analysis with the AI algorithm.
Flow and timing	It was unclear whether all included participants were diagnosed by the same person(s). All participants were included in the analyses.
Comparative	Different algorithms were developed; it was unclear whether the results were interpreted separately. All data were included in the different analyses.
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?	Yes
Did the study avoid inappropriate exclusions?	Yes
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?		Unclear 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?	Unclear
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?		Unclear 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?	Unclear
Were all patients included in the analysis?	Yes
Could the patient flow have introduced bias?		Unclear risk
DOMAIN 5: Comparative
Were different AI tests were developed and interpreted without knowledge of each other.	Unclear
Are the proportions and reasons for missing data similar for all index tests?	Yes
		Unclear risk