基于PET影像组学鉴别鼻咽癌及原发鼻咽淋巴瘤  

Differential Diagnosis of Nasopharyngeal Carcinoma and Primary Nasopharyngeal Lymphoma Based on PET Imaging

作　　者：林莉莉 李昂 季仲友[1] 蒋韩 LIN Li-li;LI Ang;JI Zhong-you;JIANG Han(Department of PET/CT,Fujian Medical University Union Hospital,Fujian Fuzhou 350001)

机构地区：[1]福建医科大学附属协和医院PET/CT室,福建福州350001

出　　处：《中国医疗器械信息》2022年第1期61-63,共3页China Medical Device Information

摘　　要：目的:探讨PET影像组学方法对鼻咽癌及原发鼻咽淋巴瘤鉴别诊断的意义。方法:回顾性分析89例鼻咽癌(NPC)原发鼻咽淋巴瘤(PNL)患者,提取PET影像组学特征,进行Logistic回归建模,得到模型参数,并建立模型方程。结果:NPC患者58例,PNL患者31例(NK/T细胞淋巴瘤10例,弥漫大B细胞淋巴瘤21例),采用低方差滤波法及LASSO法,将PET影像组学特征缩减到12个,共有4个特征(Small Dependence High GrayLevel Emphasis、Run Percentage、Short Run Low Gray Level Emphasis、Short Run Low Gray Level Emphasis)的系数具有统计学意义(P<0.05)。模型AUC为0.913(95%CI:0.857~0.969),灵敏度0.935,特异度0.724,准确度0.798。结论:PET影像组学方法可以作为鉴别NPC和PNL的可用方法。Objective:To explore the significance of PET radiomics in the differential diagnosis of nasopharyngeal carcinoma and primary nasopharyngeal lymphoma.Methods:A retrospective analysis of 89 patients with nasopharyngeal carcinoma(NPC)and primary nasopharyngeal lymphoma(PNL)was performed.The PET radiomics features were extracted,and Logistic regression modeling was performed to obtain model parameters and establish model equations.Results:58 patients with NPC and 31 patients with PNL(10 cases of NK/T-cell lymphoma and 21 cases of diffuse large B-cell lymphoma).Low variance filtering and LASSO were used to reduce PET imaging features to 12,4 features of which(Small Dependence High GrayLevel Emphasis,Run Percentage,Short Run Low Gray Level Emphasis,Short Run Low Gray Level Emphasis)was statistically significant(P<0.05).The AUC of the model was 0.913(95%CI:0.857~0.969),the sensitivity was 0.935,the specificity was 0.724,and the accuracy was 0.798.Conclusion:PET radiomics can be used to identify NPC and PNL.

关 键 词：PET 影像组学 鼻咽癌 原发鼻咽淋巴瘤 

分 类 号：R445.3[医药卫生—影像医学与核医学]