Retrospective Study
Copyright ©The Author(s) 2024.
Artif Intell Med Imaging. Sep 28, 2024; 5(1): 93993
Published online Sep 28, 2024. doi: 10.35711/aimi.v5.i1.93993
Table 1 Statistical analysis results of clinical characteristics, n (%)
CharacteristicsTraining set (n = 242)
P valueTest set (n = 61)
P value
PNI (n = 148)
PNI+ (n = 94)
PNI (n = 40)
PNI+ (n = 21)
Age, (mean ± SD) (years)67.01 ± 10.5764.26 ± 12.06 0.06367.33 ± 8.1863.104 ± 8.66 0.065
Gender0.2480.86
Male85 (58.2)54 (39.1)20 (64.5) 11 (35.5)
Female63 (65.6)33 (34.4)20 (66.7) 10 (33.3)
Smoking0.8820.396
No101 (60.8)65 (39.2)29 (69.0)13 (31.0)
Yes47 (51.8)29 (38.2)11 (57.9)8 (42.1)
HGB (g/L)126.98 ± 20.67130.69 ± 20.000.071128.881 ± 14.16128.67 ± 22.280.965
RBC (1012/L)4.28 ± 0.624.40 ± 0.440.1264.40 ± 0.474.28 ± 0.480.339
WBC (109/L)6.55 ± 1.796.84 ± 2.190.2656.18 ± 1.686.54 ± 1.670.391
PLT (109/L)229.14 ± 77.31240.92 ± 76.730.248231.05 ± 70.18231.62 ± 50.560.974
Lymphocyte(109/L)1.61 ± 0.591.63 ± 0.680.8081.59 ± 0.681.71 ± 0.950.567
Monocyte(109/L)0.46 ± 0.230.47 ± 0170.6670.40 ± 0.140.70 ± 1.020.1
Neutrophil(109/L)4.27 ± 1.554.54 ± 1.940.2433.96 ± 1.514.47 ± 1.700.228
TG1.47 ± 1.111.31 ± 0.600.181.43 ± 0.721.59 ± 0.980.502
Cholesterol4.58 ± 0.884.72 ± 0.980.2324.82 ± 0.894.91 ± 1.080.707
HDL1.12 ± 0.281.19 ± 0.320.0881.42 ± 1.511.13 ± 0.240.392
LDL2.79 ± 0.692.93 ± 0.850.1642.96 ± 0.843.18 ± 1.090.392
AproA1.24 ± 0.191.26 ± 0.200.4061.29 ± 0.171.22 ± 0.150.126
AproB0.89 ± 0.190.90 ± 0.220.7950.94 ± 0.180.89 ± 0.190.279
CEA (≥ 5 ng/mL) 0.0160.173
No94 (67.6)45 (32.4)28 (71.8)11 (28.2)
Yes54 (52.4)49 (47.6)12 (54.5)10 (45.5)
CA19-9 (≥ 37 U/mL)0.0030.052
No136 (64.8)74 (35.2)37 (71.2)15 (28.8)
Yes12 (37.5)20 (62.5)3 (33.3)6 (66.7)
CT T stage0.0000.006
1/229 (85.3)5 (14.7)16 (88.6)2 (11.1)
373 (66.4)37 (33.6)17 (68.0)8 (32.0)
446 (46.9)52 (53.1)7 (38.9)11 (61.1)
Table 2 Performance of four machine learning classifiers (support vector machine, multi-layer perceptron, k-nearest neighbor and logistic regression)
ClassifiersTraining set
Test set
AUC
95%CI
Sensitivity
Specificity
AUC
95%CI
Sensitivity
Specificity
ASVM0.9040.865-0.9430.8400.8850.8900.794-0.9870.8570.850
AKNN0.7900.736-0.8440.6380.7910.7620.640-0.8840.6670.725
AMLP0.8210.769-0.8730.8400.6490.8140.691-0.9370.8100.750
ALR0.7880.731-0.8460.7230.7300.7500.607-0.8930.7140.750
VSVM0.8900.850-0.9300.9260.7030.8670.778-0.9560.8100.800
VKNN0.8340.783-0.8840.7020.8380.7900.676-0.9040.6670.800
VMLP0.8000.744-0.8560.7660.7300.7690.648-0.8900.5710.850
VLR0.7600.698-0.8220.6280.7770.7350.606-0.8630.9990.375
Table 3 Prediction performance of four models (arterial support vector machine, venous support vector machine, CT-Tstage and Nomogram)
Model
Dataset
AUC
95%CI
Sensitivity
Specificity
Recall
Accuracy
Precision
F1-score
ASVMTrain0.9040.865-0.9430.8400.8850.8400.8630.8800.860
Test0.8900.794-0.9870.8570.8500.8570.8540.8510.854
VSVMTrain0.8900.850-0.9300.9260.7030.9260.8150.7570.786
Test0.8670.778-0.9560.8100.8000.8100.8050.8020.806
CT-TstageTrain0.6470.583-0.7100.5530.6890.5530.6210.6400.593
Test0.7300.607-0.8540.5250.8250.5250.6750.7500.618
NomogramTrain0.9640.944-0.9830.8000.7890.8000.7950.7910.795
Test0.9550.900-0.9990.9520.9000.9520.9280.9050.919
Table 4 Delong-test results of four models (Arterial support vector machine, venous support vector machine, CT-T stage and Nomogram)
Model
Training set
Test set
Delong-test
ASVM
VSVM
CT-T stage
Nomogram
ASVM
VSVM
CT-T stage
Nomogram
ASVM-0.63041.934e-110.000271-0.69970.05270.05137
VSVM--1.737e-102.15e-05--0.06920.03611
CT-Tstage---2.2e-16---0.000305
Nomogram--------