Skip to main content
. 2017 Jan 20;7:41216. doi: 10.1038/srep41216

Table 3. Diagnostic performances of B-mode US and B-mode US combined with SWE.

Parameter Threshold* AUC p value# Sensitivity Specificity p value* PPV NPV
B-mode US categorya >4a 0.9064 NA 67/73 (91.78) 110/137 (80.29) NA 67/94 (71.28) 110/116 (94.83)
+2D SWE
 ESDU >10.73 0.9571 0.0005 63/73 (86.30) 130/137 (94.89) 0.0002 63/70 (90.00) 130/140 (92.86)
 ESDE >13.4 0.9683 0.0003 66/73 (90.41) 132/137 (96.35) <0.0001 66/71 (92.96) 132/139 (94.96)
 Emax >80.8 0.9767 <0.0001 65/73 (89.04) 132/137 (96.35) <0.0001 65/70 (92.86) 132/140 (94.29)
+3D SWE
 ESDU >10.23 0.9454 0.0007 60/73 (82.19) 129/137 (94.16) 0.0006 60/68 (88.24) 129/142 (90.85)
 ESDE >13 0.9563 0.0007 62/73 (84.93) 131/137 (95.62) <0.0001 62/68 (91.18) 131/142 (92.25)
 Emax >131.7 0.9562 0.0016 66/73 (90.41) 122/137 (89.05) 0.0442 66/81 (81.48) 122/129 (94.57)

aB-mode US category was defined according to the BI-RADS classification for US. NA = not available. The + before 2D and 3D SWE indicates the effect of adding that SWE feature to the original BI-RADS category 4a and 4b masses. #p value of testing the null hypothesis that there is no change in AUC with the addition of the SWE feature (McNemar test). *p value of testing the null hypothesis that there is no change in specificity with the addition of the SWE feature (McNemar test). *Threshold as the optimal cut off, yielding the maximal sum of the sensitivity and specificity.