Fig 3.

Study estimates of sensitivity and false positive rate (1−specificity) in receiver operating characteristic space by index test (artificial intelligence) and comparator (radiologist) for eight included studies. Comparators are defined as consensus of two readers and arbitration (radiologist consensus), or single reader decision/average of multiple readers (radiologist single/average). iVertical dashed lines represent specificity for screening programmes for Denmark (2% false positive rate),⁶¹ UK (3% false positive rate),^{62 63} and US (11% false positive rate).⁶⁴ Retrospective test accuracy studies: Salim et al,³⁵ Schaffter et al,³⁶ and McKinney et al.²⁹ Enriched test set multiple reader multiple case laboratory studies: Pacilè et al,³⁰ Watanabe et al,³⁷ Rodriguez-Ruiz et al³³ (Rodriguez-Ruiz 2019a in figure), Lotter 2021,²⁸ and Rodriguez-Ruiz et al³² (Rodriguez-Ruiz 2019b in figure)