To construct a prediction model of central lymph node metastasis in thyroid cancer by combining elastography parameters and ultrasound image features / 中华内分泌外科杂志

ABSTRACT

Objective:To establish a nomogram model based on elastic imaging parameters and ultrasound image features, and evaluate its predictive value in central lymph node metastasis (CLNM) in papillary thyroid carcinoma (PTC) .Methods:The clinical data of 168 patients (the research group) with papillary thyroid carcinoma who underwent thyroid surgery in our hospital from Jan. 2019 to Dec. 2021 were retrospectively collected, including gender, age, ultrasound elastography parameters (elasticity ratio, blue area ratio), and ultrasound examination indicators (nodule diameter, nodule number, internal echo, border, edge, aspect ratio, microcalcification, capsule invasion). Another 150 patients who underwent thyroid surgery in our hospital during the same period were selected as the validation group.According to the results of postoperative pathological examination, the the research group were divided into two groups: 64 cases (38.10%) of CLNM and 104 cases (61.90%) of non-CLNM. Binary logistic regression analysis was used to explore the influencing factors of CLNM in PTC patients, and a nomogram model based on elastic imaging parameters and ultrasound image features was established. The nomogram model was drawn to predict the receiver operating characteristic (ROC) curve of CLNM in PTC patients.Results:There were statistically significant differences in nodule diameter, edge, microcalcification, capsule invasion, blue area ratio, and elasticity ratio ( P<0.05). Most of the nodules in the CLNM group were ≥10 mm in diameter, with uneven margins, an aspect ratio of <1, microcalcifications and capsular invasion. Logistic regression analysis showed that nodule diameter, capsule invasion, blue area ratio and elastic ratio were risk factors for CLNM ( P<0.05). The AUC of the combined detection was 0.857 (0.777-0.937), and the sensitivity and specificity were 78.1% and 86.5%, respectively, and the AUC and sensitivity were significantly higher than the individual detection of each index ( P<0.05). In the research group, the sensitivity and specificity of the ultrasound parameter prediction model in predicting CLNM were 81.25% (52/64) and 84.62% (88/104), respectively. In the validation group, the sensitivity and specificity of the ultrasound parameter prediction model in predicting CLNM were 79.17% (38/48) and 85.29% (87/102), respectively. Conclusion:Elastography parameters (blue area ratio, elasticity ratio) and ultrasound image features (nodule diameter, capsular invasion) are the influencing factors of CLNM in PTC patients, and the combined prediction based on the above four indicators has good application value.