Xu 2017.

Study characteristics
Patient Sampling	Prospective, single‐centre, cross‐sectional study. People with keratoconus were enroled at the Affiliated Eye Hospital of Wenzhou Medical University in China. Complete ocular examinations were performed by 2 experienced doctors, including a review of medical and family history, corrected distance visual acuity, slit‐lamp biomicroscopy, fundus examination, and corneal topography using Medmont E300 (Medmont, Inc., Nunawading Melbourne, Australia).
Patient characteristics and setting	Participants were divided into the following 3 groups. Normal group (healthy eyes from normal subjects) were enroled from the hospital staff and university students if they met the following screening criteria. Central average keratometry < 45.0D Myopia < −6.00 D and astigmatism < −2.00 D No clinical signs or suggestive topographic patterns for suspicious subclinical keratoconus, keratoconus, or pellucid marginal degeneration No history of ocular surgery or trauma Stopped contact lens wear for ≥ 8 weeks for rigid gas‐permeable contact lenses and ≥ 4 week for soft contact lenses. Keratoconus group (mild or moderate keratoconus eyes) had to meet the following criteria. Central average keratometry > 47.0D ≥ 1 slit‐lamp sign (stromal thinning, Vogt's striae, or Fleischer's ring > 2‐mm arc) Asymmetric topographical features with I‐S ≥ 1.9 D of the vertical gradient power across the 6‐mm region No history of contact lens wear, ocular surgery, or extensive scarring Subclinical keratoconus group (fellow eye of unilateral keratoconus) had to meet the following criteria. Central average keratometry < 45.0D Diagnosis of keratoconus in the contralateral eye No clinical signs of keratoconus on slit‐lamp biomicroscopy, retinoscopy, and ophthalmoscopy Corneal topographical features with I‐S values < 1.4D of the vertical gradient power across the 6‐mm region No history of contact lens wear, ocular surgery, or trauma
Index tests	Participants were divided into a training set (normal, subclinical keratoconus group, and keratoconus group) used to build the discrimination function, and a validation set (normal and subclinical group) used to test the diagnostic power. The goal of the present study was to apply the Zernike fitting method to describe the 3D varying complexity of corneal shapes and the 3D distribution of corneal thickness, and to characterize the entire corneal topography and tomography data in subclinical eyes, keratoconus eyes, and normal eyes using Pentacam tomography. Furthermore, the metrics constructed from Zernike polynomials were compared to improve the diagnostic sensitivity and specificity for the detection of subclinical keratoconus corneas.
Target condition and reference standard(s)	2 experienced doctors performed complete ocular examinations, including a review of medical and family history, corrected distance visual acuity, slit‐lamp biomicroscopy, fundus examination, and corneal topography using Medmont E300 (Medmont, Inc., Nunawading Melbourne, Australia).
Flow and timing	The data from all participants were checked by 3 ophthalmologists. The diagnosis was made before the analysis with the AI algorithm.
Comparative	Not applicable
Notes	This study was supported by the National Natural Science Foundation of China (81400441 to Shen), the National Key Research and Development Program of China (2016YFC0102500 to Wang), and the Zhejiang provincial Natural Science Foundation of China (LQ17H120008 to Xu).
Methodological quality
Item	Authors' judgement	Risk of bias	Applicability concerns
DOMAIN 1: Patient selection
Was a consecutive or random sample of patients enrolled?	Unclear
Was a case‐control design avoided?	Unclear
Did the study avoid inappropriate exclusions?	Unclear
Could the selection of patients have introduced bias?		Unclear risk
Are there concerns that the included patients and setting do not match the review question?			Unclear
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