Myocardial CT perfusion imaging for the detection of obstructive coronary artery disease: multisegment reconstruction does not improve diagnostic performance

European Radiology Experimental

Table 3 All patients: diagnostic performance of MSR and HSR of myocardial CTP

	All 93 patients
	Per-patient level			Per-territory level
Reconstruction/performance	HSR	MSR	p-value	HSR	MSR	p-value
Area under the curve^a	0.79 [0.69, 0.88]	0.65 [0.53, 0.78]	0.011	0.87 [0.83, 0.92]	0.71 [0.65, 0.78]	< 0.001
Sensitivity	88 (51/58) [77, 95]	67 (39/58) [54, 79]	0.001	79 (68/86) [69, 87]	50 (43/86) [39, 61]	< 0.001
Specificity	49 (17/35) [31, 66]	66 (23/35) [48, 81]	0.114	83 (171/206) [77, 88]	87 (179/206) [82, 91]	0.118
Positive predictive value	74 (51/69) [62, 84]	76 (39/51) [63, 87]	0.541	66 (68/103) [56, 75]	61 (43/70) [49, 73]	0.263
Negative predictive value	71 (17/24) [49, 87]	55 (23/42) [39, 70]	0.027	90 (171/189) [85, 94]	81 (179/222) [75, 86]	< 0.001

Reference: ≥ 50% diameter vessel stenosis detected in quantitative invasive coronary angiography. Data are the results of consensus reading of two readers. Unless otherwise stated, data are percentages, data in parentheses are raw data, and data in brackets are 95% confidence intervals. The 95% confidence intervals were estimated for unclustered data (see reference [31]). ^aData are the area under the curve and data in brackets are 95% confidence intervals. CTP Computed tomography perfusion, HSR Halfscan reconstruction, MSR Multisegment reconstruction