Logistic 回归模型在超声评价甲状腺结节良恶性及淋巴转移的应用研究

Logistic regression mo delf or ultrason ic evaluation fo thyroid nodule benign an d malignancy and lym-pha tic metastasis

目的 应用Logistic回归模型 ,筛选出甲状腺恶性结节的超声影像学特征危险因素. 方法 240例甲状腺结节患者经超声检查,并手术治疗有明确病理诊断的患者纳入本研究,将术前结节的超声特征引入Logistic回归模型, 筛选甲状腺恶性结节的超声影像学特征危险因素. 结果 经过Logistic逐步回归模型分析,筛选出纵横比≥1、微钙化、低回声是甲状腺癌患病的影响因素. 本次研究同时发现微钙化可做为预测甲状腺癌淋巴转移的危险因素. 结论 超声医生可以在临床诊断中,参考本研究筛选出的危险因素作为诊断甲状腺恶性结节及转移的辅助指标. 在未来,前瞻性的更大规模的同类研究仍有必要.

Objective To explore the ultrasonic imaging features of malignant thyroid nodules risk fac-tors by logistic regression.Methods Two hundreds and forty patients with thyroid nodules by ultrasonic examina-tion and surgical treatment for patients with definited pathological diagnosis were included in this study.The pre-operative sonographic features of the nodules were put into the logistic regression model to explore the risk factors of ultrasonic imaging features of malignant thyroid nodules.Results By logistic regression model analysis,hypoe-cho,taller than wide≥1 and micro calcification were found to be significantly statistical variables for thyroid car-cinoma.Micro calcification could predict thyroid cancer for lymph node metastases.Conclusoin During clinical diagnosis,hypoecho,taller than wide more than 1 and micro calcification could be used as indicators by ultrasound doctors to judge malignant thyroid nodules and lymph node metastases.Prospective larger similar research is still very necessary to explore the indicator for thyroid carcinoma in the future.