Table 4.

The performance of nomograms for the prediction of ≥ 3 metastatic lymph nodes

Cohort	Threshold	Sensitivity (%)	Specificity (%)	Accuracy (%)	PPV (%)	NPV (%)	AUC
Development cohort (n = 179)	30%	42.9 (9/21) [21.8, 66.0]	96.8 (153/158) [92.8, 99.0]	90.5 (162/179) [85.2, 94.4]	64.3 (9/14) [35.1, 87.2]	92.7 (153/165) [87.6, 96.2]	0.89 [0.83, 0.93]
	40%	33.3 (7/21) [14.6, 57.0]	98.1 (155/158) [94.6, 99.6]	90.5 (162/179) [85.2, 94.4]	70.0 (7/10) [34.8, 93.3]	91.7 (155/169) [86.5, 95.4]
	50%	23.8 (5/21) [8.2, 47.2]	98.7 (156/158) [95.5, 99.9]	89.9 (161/179) [84.6, 93.4]	71.3 (5/7) [29.0, 96.3]	90.7 (156/172) [85.3, 94.6]
Internal validation cohort (n = 90)	40%	40.0 (4/10) [12.2, 73.8]	97.5 (78/80) [91.3, 99.7]	91.11 (82/90) [83.23, 96.1]	66.7 (4/6) [22.3, 95.7]	92.9 (78/84) [85.1, 97.3]	0.91 [0.83, 0.96]
External validation cohort (n = 197)	40%	30.0 (6/20) [11.9, 54.3]	98.3 (174/177) [95.1, 99.7]	91.4 (180/197) [86.5, 94.9]	66.7 (6/9) [29.9, 92.5]	92.6 (174/188) [87.8, 95.9 ]	0.87 [0.78, 0.96]

Data in parentheses are numbers of patients used to calculate percentages. Data in brackets are 95% confidence intervals. Other data are reported as percentages

PPV positive predictive value, NPV negative predictive value, AUC area under the receiver operating characteristic curve