摘要
目的探讨多期增强CT扫描的动脉期强化幅度、强化类型预测胰腺神经内分泌肿瘤（PNETs）病理分级的临床应用价值。方法 29个病理证实的PNETs纳入回顾性研究。根据WHO 2010标准PNETs病理分级为低级别（G1、G2）、高级别（G3）。通过Spearman秩相关分析动脉期强化幅度与病理分级的相关性,通过受试者工作特征（ROC）曲线,确定动脉期强化幅度诊断G3级病变的截断值并分析其诊断效能。强化类型分为流出型（Ⅰ型）、平台型（Ⅱ型）、渐进型（Ⅲ型）,通过Spearman秩相关分析强化类型与病理级别的相关性,分析各强化类型预测病理分级的效能。结果 G1、G2、G3级PNETs分别为18个、5个、6个。动脉期强化幅度与病理分级之间存在一定程度的负相关（r=-0.528,P=0.003）;动脉期强化幅度的截断值取21 HU（以〈21 HU为低幅强化）,判定G3级PNETs可获得极高的诊断效能（特异度、灵敏度均达100%）。流出型15个、平台型4个、渐进型10个,强化类型与病理分级之间存在一定程度的正相关（r=0.521,P=0.004）;流出型强化判定G1/G2级病灶的灵敏度65.2%、特异度100%、准确度72.4%、阳性预测值100%、阴性预测值42.9%,流出型强化病灶中未出现G3级病灶。结论不同病理分级的PNETs多期增强CT扫描的动脉期强化幅度、强化类型有一定的特征。以21 HU为截断界值的动脉期低幅强化判定恶性的G3级PNETs有极高的诊断效能,流出型强化是排除G3级PNETs较可靠的指标。
Objective To investigate the value of contrast enhancement magnitude and pattern on multidetector CT in predicting grade of pancreatic endocrine tumors（ p NETs）. Methods 29 pathologically confirmed lesions,with preoperative contrast CT examinations,were included. According to 2010 revised version of the World Health Organization（ WHO）classification,p NETs were divided into G1,G2,G3. The correlation between contrast enhancement magnitude and pathological grades was analysed by Spearman rank correlation. Through receiver operating characteristic（ ROC） curve,contrast enhancement magnitude in arterial phase for diagnosing G3-p NETs was investigated. Contrast enhancement patterns were classified as wash-out（ Ⅰ） pattern,even（ Ⅱ） pattern,late enhancement（ Ⅲ） pattern,and the correlation with pathologic grades was analysed,and the efficacy of diagnosis was evaluated. Results The counts of G1,G2,G3-PNETs were 18,5,6. Contrast enhancement magnitude was negatively correlated with pathological grades（ r =-0. 528,P = 0. 003）. When cut-off value was 21 HU（ 〈 21 HU was defined as low-magnitude enhancement）,the sensitivity and specificity for diagnosing G3-p NETs both achieved 100%. There were 15 wash-out pattern lesions,4 even pattern lesions and 10 late enhancement pattern lesions. Contrast enhancement patterns were positively correlated with pathological grades（ r = 0. 521,P =0. 004）. The sensitivity,specificity,accuracy,positive predictive value,negative predictive value for diagnosing low grade lesions（ G1/G2） through wash-out pattern were 65. 2%,100%,72. 4%,100%,42. 9%,while none of wash-out pattern lesions were G3 p NETs. Conclusion Contrast enhancement magnitude and pattern of p NETs with different pathological grades were characteristic. 21 HU was the optimal cut-off value for diagnosis of G3 p NETs with an extremely high efficacy,while wash-out pattern suggested excluding G3 PNETs.
作者
张兰
韩萍
李欣
孙文刚
ZHANG Lan;HAN Ping;LI Xin(Department of Radiology,Union Hospital,Tongji Medical College,Huazhong University of Science and Technology,430022,P.R.Chin)
出处
《临床放射学杂志》
CSCD
北大核心
2018年第6期957-962,共6页
Journal of Clinical Radiology
基金
2016年湖北省分子影像重点实验室基金项目(编号:02.03.2017-200)
武汉协和医院2016年药剂护专项基金项目(编号:02.03.2017-117)
作者简介
通讯作者:李欣
