Performance of artificial intelligence on cervical vertebral maturation assessment: a systematic review and meta-analysis

Table 6 Results of grading of recommendations assessment, development and evaluation (GRADE)

Number of models/actual cervical stages	Study design	Factors that may decrease certainty of evidence						Certainty of Evidence
Number of models/actual cervical stages	Study design	Risk of bias	Indirectness	Inconsistency	Imprecision	Publication bias	Other consideration	Certainty of Evidence
28 models/ 10,865 stages	Case–control	Serious^a	Not serious	Very serious^b	Not serious	Not detected	Very strong association^c	⨁⨁⨁◯ Moderate

Explanations
^aEvidence was downgraded by one level. Seventeen studies had high or unclear risk of bias
^bEvidence was upgraded by 2 levels. Different imaging modalities, different artificial intelligence tasks, different data curation and different age and sex groups created inconsistency
^cEvidence was upgraded by 2 levels because of “very large” effect size (DOR > 5)

ISSN: 1472-6831