Skip to main content

Table 4 Performance of multi-modality RStumor, multi-modality RS1.0 mm, clinical model and nomogram in the training, internal test and external test cohorts

From: Predicting lymph node metastasis in papillary thyroid carcinoma: radiomics using two types of ultrasound elastography

Cohort

Model

AUC (95% CI)

ACC (%)

SEN (%)

SPE (%)

PPV (%)

NPV (%)

Training cohort

Multimodality RStumor

0.678 (0.625, 0.730)

63.3 (247/390)

73.9 (153/207)

51.4 (94/183)

63.2 (153/242)

63.5 (94/148)

 

Multimodality RS1.0 mm

0.862 (0.825, 0.899)

80.0 (312/390)

81.2 (168/207)

78.7 (144/183)

81.2 (168/207)

78.7 (144/183)

 

Clinical model

0.750 (0.701, 0.798)

70.5 (275/390)

74.9 (155/207)

65.6 (120/183)

71.1 (155/218)

69.8 (120/172)

 

Nomogram

0.883 (0.849, 0.918)

80.8 (315/390)

81.2 (168/207)

80.3 (147/183)

82.4 (168/204)

79.0 (147/186)

Internal test cohort

Multimodality RStumor

0.682 (0.575, 0.789)

67.0 (65/97)

74.5 (38/51)

58.7 (27/46)

66.7 (38/57)

67.5 (27/40)

 

Multimodality RS1.0 mm

0.798 (0.708, 0.887)

76.3 (74/97)

68.6 (35/51)

84.8 (39/46)

83.3 (35/42)

70.9 (39/55)

 

Clinical model

0.777 (0.686, 0.869)

66.0 (64/97)

68.6 (35/51)

63.0 (29/46)

67.3 (35/52)

64.4 (29/45)

 

Nomogram

0.873 (0.799, 0.946)

83.5 (81/97)

86.3 (44/51)

80.4 (37/46)

83.0 (44/53)

84.1 (37/44)

External test cohort

Multimodality RStumor

0.665 (0.583, 0.747)

63.6 (110/173)

33.8 (24/71)

84.3 (86/102)

60.0 (24/40)

64.7 (86/133)

 

Multimodality RS1.0 mm

0.789 (0.720, 0.857)

72.8 (126/173)

54.9 (39/71)

85.3 (87/102)

72.2 (39/54)

73.1 (87/119)

 

Clinical model

0.781 (0.708, 0.854)

73.4 (127/173)

64.8 (46/71)

79.4 (81/102)

68.7 (46/67)

76.4 (81/106)

 

Nomogram

0.841 (0.782, 0.900)

77.5 (134/173)

78.9 (56/71)

76.5 (78/102)

70.0 (56/80)

83.9 (78/93)

  1. ACC: accuracy; AUC: area under the ROC curve; RS: radiomics signature; SEN: sensitive; SPE: specificity