Diagnostic Performance of Duplex Ultrasonography in Post-Radiation Common Carotid Artery Stenosis: A Cross-Sectional...

Whether established diagnostic criteria of carotid duplex ultrasonography (CDU) for atherosclerotic carotid artery stenosis are applicable to radiation vasculopathy in head and neck cancer (HNC) patients remains uncertain. This study aimed to validate the diagnostic accuracy of CDU for screening radiation-induced common carotid artery (CCA) stenosis.

We enrolled 134 HNC patients with 268 carotid arteries, all of whom underwent CDU and computed tomographic angiography. Diameter stenosis and hemodynamic parameters were assessed on CDU, with ≥50% stenosis on computed tomographic angiography serving as the reference standard. Diagnostic performance was evaluated using sensitivity, specificity, PPV, NPV, and receiver operating characteristic (ROC) analysis to identify optimal cutoff values.

Using a B-mode cutoff of 50% to detect ≥50% CCA stenosis, the sensitivity, specificity, PPV, and NPV were 85.2, 88.1, 45.1, and 98.1%, respectively. The optimal B-mode threshold of 53%, which yielded a sensitivity of 75.9%, specificity of 91.4%, PPV of 51.2%, and NPV of 96.9%, with an area under the ROC curve (AUROC) of 0.894. In comparison, the optimal PSV threshold of 203 cm/s showed lower sensitivity (51.7%) but higher specificity (99.2%), with a PPV of 88.2%, NPV of 94.6%, and an AUROC of 0.698. Combining B-mode and PSV criteria resulted in an AUROC of 0.891, offering minimal incremental diagnostic benefit compared with B-mode assessment alone.

B-mode CDU may be a feasible screening tool for radiation-induced CCA stenosis. A B-mode threshold of 53% achieved diagnostic performance comparable to combined B-mode and velocity criteria for detecting ≥50% stenosis.