Kojima 2021.

Study characteristics
Patient Sampling	Single centre, retrospective, case‐control study that included 179 eyes of 99 consecutive people with keratoconus and suspected keratoconus (68 males and 31 females, mean age 33.48 (SD 15.41) years), who visited the Nagoya Eye Clinic from January 2019 to December 2020 and were tested with an auto‐keratometer (ARK‐1s, Gamagori, Japan, NIDEK). During the same period, 468 eyes of 235 consecutive people (125 men and 110 women; mean age 37.55 (SD 22.70) years) examined for refractive correction were included as normal controls. The control group included people who had no abnormalities on slit‐lamp biomicroscopy examination and corneal topography.
Patient characteristics and setting	People already diagnosed with keratoconus or suspected keratoconus were included. The exact reason why healthy controls visited the centre is unclear.
Index tests	Regression algorithm, modified keratoconus keratometer index (KKI) using 3 variables: steep K‐value, flat K‐value, and astigmatism. logit = 1.284 × steep K (dioptre) − 0.618 × flat K (dioptre) − 3.163 × (0: non‐with‐the‐rule astigmatism; 1: with‐the‐rule astigmatism) − 28.662, KKI = exp(logit)/(1 + exp[logit]). The cut‐off value of 0.461 had been determined previously.
Target condition and reference standard(s)	2 cornea specialists diagnosed keratoconus through slit‐lamp microscopy and corneal topography. Keratoconus signs found in both slit‐lamp microscopy and corneal topography were classified as keratoconus, while signs found only in corneal topography were classified as suspected keratoconus. Forme fruste keratoconus was defined as an eye with normal corneal topography in the contralateral eye of the keratoconus. The severity of keratoconus was based on the Amsler–Krumeich classification.
Flow and timing	All cases were included in the reference standard and index test. All data were included in a 2 × 2 table.
Comparative	Not applicable
Notes	This research received no external funding.
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?	Yes
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
Were different AI tests were developed and interpreted without knowledge of each other.
Are the proportions and reasons for missing data similar for all index tests?