【摘要】 目的探讨18F-氟代脱氧葡萄糖正电子成像术/计算机断层扫描(¹⁸F-FDG PET/CT)对甲状腺结节鉴别诊断的价值及使用标准化摄取值（SUVmax）评价甲状腺结节性质的可行性。方法回顾性分析101例术前行¹⁸F-FDG PET/CT显像检查,术后病理明确诊断的甲状腺结节患者,其中男性30例,女性71例;年龄17⁷8岁,平均（48.2±12.9）岁。计算PET/CT对甲状腺结节良恶性鉴别效能（敏感度、特异度、准确度、阳性预测值及阴性预测值）,并对病灶组织、健侧甲状腺组织或周围组织分别勾画出感兴趣区,测量SUVmax值进行相互比较。所获得患者的SUVmax值用均数±标准差的形式表述,均数差异使用t检验。结果 PET/CT对甲状腺结节良恶性鉴别的敏感度为89.1%,特异度为60.0%,准确度为62.3%,阴性预测值为35.2%,阳性预测值为90.1%;甲状腺恶性结节组SUVmax为:6.70±4.57（1.3²1.7）,高于良性结节组的3.87±2.41（1.3¹2）,差异有统计学意义（t=3.14,P<0.05）。结论 PET/CT对于甲状腺结节的良恶性鉴别具有一定的意义,且通过SUVmax值来判断良恶性肿瘤也具有可行性。更多还原

【Abstract】 Objective To explore the diagnosis of the¹⁸F-FDG PET / CT and evaluate the feasibility of SUVmax in the differential diagnosis of benign and malignant thyroid nodules. Methods The diagnostic efficacy of PET / CT in differentiating benign and malignant thyroid nodules was retrospectively analyzed by calculating the sensitivity,specificity,accuracy,positive predictive value and negative predictive value in one hundred and one patients with thyroid who had¹⁸F-FDG PET / CT before surgy,including thirty males and seventy-one females,age 17 78（ 48. 2 ± 12. 9）. The ROI of the contralateral nodules,the affected side of thyroid tissues or the surrounding tissues delineated,the SUVmax values were measured,SPSS software package was used to detect differences between the benign and the malignant lesions. Results The sensitivity of¹⁸F-FDG PET / CT examination for benign and malignant thyroid nodule was 89. 1%; the specificity was60. 0%; the accuracy was 62. 3%; negative predictive value was 35. 2% and positive predictive value was90. 1%. SUVmax value was statistically significant between the two groups of benign and malignant thyroid（ t = 3. 14,P< 0. 05）. The SUXmax in malignant group[6. 70 ± 4. 57（ 1. 3 21. 7） ]was higher than that in the benign group[3. 95 ± 2. 58（ 1. 3 12） ]. Conclusions It is meaningful for the¹⁸F-FDG PET / CT in identification of benign and malignant thyroid nodules. And it is feasible to distinguish between the benign and malignant nodules with reference to the SUVmax value.更多还原