Skip to main content
. 2020 Sep 30;9(23):8809–8820. doi: 10.1002/cam4.3490

Table 3.

Predictors for metastatic LN status in the training set‐univariate analyses

Predictors LN‐ (n = 82) LN+ (n = 280) Statistics p
Mean age, y (SD) 53.06 (12.57) 59.41 (11.97) −4.179 a <0.001
Male, n (%) 56 (68.3) 170 (60.7) 1.553 b 0.213
Metastatic LNs by MRI, Median (Q) 0 (2) 2 (3) −8.703 c <0.001
Metastatic LNs by Faster R‐CNN, Median (Q) 0 (2) 2 (4) −8.967 c <0.001
Differentiation degree, n (%)
High 58 (70.7) 84 (30.0)
Moderate 22 (26.8) 162 (57.9) 44.844 b <0.001
Low 2 (2.4) 34 (12.1)
CEA Level, n (%)
Normal 66 (80.5) 185 (66.1) 6.200 b 0.013
Abnormal 16 (19.5) 95 (33.9)
CA199 Level, n (%)
Normal 76 (92.7) 248 (88.6) 1.141 b 0.285
Abnormal 6 (7.3) 32 (11.4)

N = 362.

Abbreviations: CA199, carbohydrate antigen 199; CEA, carcinoembryonic antigen; Faster R‐CNN, faster region‐based convolutional neural network; LN, lymph node; MRI, magnetic resonance imaging; SD, standard deviation.

a

Statistics for Student's t test.

b

Statistics for Pearson's Chi‐squared test.

c

Statistics for Mann‐Whitney U test.