<Patients>

Study patients (87 oncology patients) 

• M : F = 59 (67.8%) : 28 (32.2%)
• Age: 64.7 ± 11.6 years (range, 30-88 years)
• Primary site of malignancies: lung cancer (n = 51), esophageal cancer (n = 9), breast cancer (n = 8), pancreas cancer (n = 7), gastric cancer (n = 6), prostate cancer (n = 5), hypopharyngeal cancer (n = 1)

<Clinical and Radiological factors>

Clinical factors

• Sex, age, lung disease (i.e., anthracofibrosis, pneumonia, interstitial lung disease), malignancy type (lung cancer vs. non-lung cancer)

US evaluation by two neuroradiologists using EPIQ 7 unit (Philips Medical System, Bothell, WA)

• Criteria for metastastic lymph node during US examination
• Size: short-axis diameter > 0.8 cm
• Shape: round, short/long axis ratio > 0.5
• Echogenicity: hypoechogenicity or heterogeneous internal echogenicity
• Loss of fatty hilum
• Presence of intranodal necrosis and calcification

US-FNA or US-CNB were performed in all included LNs.

<Statistical analysis>

Determining the diagnostic indices of subsequent neck US for LNs detected on chest CT

• Receiver operatic characteristic (ROC) analysis: sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), diagnostic accuracy

Comparing the differences between the false-positive and true-positive groups

• The Wilcoxon rank-sum test for continuous variables
• The chi-square or Fisher exact test for categorical variables

ROC curve analysis for obtaining the optimal cut-off value of age

• age cut-off: 65 years
• AUC (area under the ROC curve), 0.722 -> the largest AUC
• 95% CI [0.607-0.819], P-value < 0.001

Univariate and multivariable logistic regressions analyses

• odds ratios (ORs) to identify independent predictors for metastatic LNs

Statistical analyses by using SPSS and Medcalc softwares