Conventional ultrasound combined with contrast-enhanced ultrasound predicts lateral lymph node metastasis in papillary thyroid carcinoma.

Abstract

Background: This study aimed to investigate the correlation between conventional ultrasound and contrast-enhanced ultrasound (CEUS) features and lateral lymph node metastasis (LLNM) in papillary thyroid carcinoma (PTC), establish a predictive model, and provide imaging evidence for clinical diagnosis and treatment.

Methods: This study selected 428 patients with postoperative pathologically confirmed PTC, who had undergone cervical lymph node dissection, from September 2020 to August 2021 at the First Affiliated Hospital of Soochow University. According to the postoperative pathological results, the patients were divided into those with LLNM (n = 94) and those without LLNM (n = 334). The clinical characteristics and conventional ultrasound and CEUS characteristics of the two groups were retrospectively analyzed, and the differences between them were compared. Independent risk factors related to LLNM were screened, a prediction model was constructed, and its prediction efficiency and clinical practicality were evaluated.

Results: The independent risk factors for LLNM were nodules located in the upper thyroid (odds ratio [OR] = 2.640, 95% confidence interval [CI]: 1.488-4.682), maximum tumor diameter≥1.0 cm (OR = 2.027, 95% CI: 1.146-3.586), microcalcification (OR = 2.176, 95% CI: 1.153-4.106), central lymph node metastasis (OR = 3.091, 95% CI: 1.721-5.549), enhanced late hyperenhancement (OR = 2.440, 95% CI: 1.081-5.508), and membrane continuity interruption in early enhancement (OR = 3.988, 95% CI: 2.315-6.871) (P < 0.05 for all). The sensitivity and specificity of the combined index in predicting LLNM in PTC patients were 72.34% and 78.74%, respectively (best cut-off value: 0.511); the area under the curve (AUC) was 0.818 (95% CI: 0.778-0.853). Moreover, the AUC of the combined index in predicting LLNM in PTC patients was greater than that of conventional ultrasound alone. The calibration curve of the nomogram constructed based on the aforementioned six independent risk factors showed that the model could fit the actual probability of LLNM well with high calibration. Decision curve analysis revealed that the model has good clinical applicability.

Conclusions: The nomogram model constructed by conventional ultrasound combined with CEUS can effectively predict lateral cervical lymph node metastasis, providing an intuitive guide tool diagnosis and treatment.