Table 2

Diagnostic Accuracy for the Detection of Significant In-Stent Restenosis After Reconstruction Using Filtered Back Projection and Iterative Reconstruction Compared With Invasive Catheterization

AccuracyFiltered Back ProjectionIterative Reconstruction
Patient-based analysis
 Sensitivity97 (32/33) [84–100]100 (33/33) [89–100]
 Specificity53 (9/17) [28–77]65 (11/17) [28–86]
 PPV80 (32/40) [64–91]85 (33/39) [70–94]
 NPV90 (9/10) [56–100]100 (11/11) [72–100]
Vessel-based analysis
 Sensitivity89 (43/48) [78–97]96 (46/48) [86–100]
 Specificity79 (120/152) [72–85]84 (129/152) [78–90]
 PPV57 (42/74) [45–69]66 (47/71) [54–78]
 NPV96 (121/126) [91–99]98 (127/129) [95–100]
Stent-based analysis
 Sensitivity85 (12/14) [57–98]100 (14/14) [77–100]
 Specificity69 (51/73) [58–80]75 (55/73) [64–85]
 PPV32 (11/34) [20–54]44 (14/32) [26–62]
 NPV96 (51/53) [87–100]100 (55/55) [94–100]
 Unassessable stents26 (30)22
 Mean diameter of assessable stents, mm3.2 ± 0.33.2 ± 0.3
 Mean diameter of unassessable stents, mm2.7 ± 0.22.6 ± 0.2
 Unassessable bare-metal stents8/37 (22)8/37 (22)
 Unassessable drug-eluting stents18/50 (36)14/50 (28)

Values are % (n/N) [95% CI], n (%), or mean ± 1 SD. Filtered back projection versus iterative reconstruction: no significant differences according to chi-square test.

NPV = negative predictive value; PPV = positive predictive value.