Quantitative computed tomography texture analysis: can it improve diagnostic accuracy to differentiate malignant lymph nodes?

Table 5 Receiver operating characteristic curve analysis of chest computed tomography (CT), textural features, 18-fluorodeoxyglucose positron emission tomography/CT (FDG PET/CT) and sonographic findings

Variable	AUC	SE (AUC)	95% CI	Se (%)	Sp (%)
Short-axis diameter, ≥ 10 mm	0.659	0.038	0.585–0.732	93	58.7
Compactness on CTTA	0.666	0.051	0.566–0.766	10.6	100
Normalized SD on CTTA^†	0.742	0.046	0.652–0.833	8.2	100
FDG PET/CT
SUV_max ≥ 2.5	0.639	0.042	0.556–0.721	94.4	55.8
SUV_max ≥ 5	0.739	0.042	0.657–0.821	74.7	94.3
SUV_Peak ≥ 4	0.739	0.041	0.658–0.820	73	95.6
SUV_max ≥ 2.5 and HU < 70	0.738	0.044	0.652–0.823	87.5	82.5
Size on EBUS, ≥10 mm	0.654	0.036	0.583–0.724	95.3	54.6
Round shape	0.598	0.043	0.513–0.682	71.4	72
Heterogeneous echogenicity	0.632	0.037	0.561–0.704	51	95
Coagulation necrosis sign	0.542	0.021	0.502–0.582	20.2	100

AUC, Area under the receiver operating characteristic curve; CI, Confidence interval; EBUS, Endobronchial ultrasound; SD, Standard deviation; Se, Sensitivity; Sp, Specificity; SUVmax, Standard maximal uptake value
^†Analyses was performed using 1 = benign and 0 = malignant

ISSN: 1470-7330