超声内镜诊断恶性胰腺囊性病变的效果评价

Efficacy of endoscopic ultrasound for the diagnosis of malignant pancreatic cystic lesions

摘要:

目的 分析恶性胰腺囊性病变（pancreatic cystic lesions，PCL）的超声内镜（endoscopic ultrasound，EUS）特征，评估EUS引导细针穿刺抽吸术（EUS⁃guided fine needle aspiration，EUS⁃FNA）诊断恶性PCL的效能。方法 回顾性分析2006年1月至2021年12月间在海军军医大学第一附属医院确诊且行EUS⁃FNA的130例PCL患者的临床特点及EUS特征，通过logistic回归分析明确恶性PCL的EUS高危特征。比较手术确诊病例中CT、MRI及EUS⁃FNA诊断恶性PCL的准确率。结果 130例确诊PCL病例中包括27例（20.77%）黏液性囊腺瘤（其中1例伴浸润癌）、19例（14.62%）导管内乳头状黏液瘤（其中8例伴浸润癌或高度异型增生）、15例（11.54%）浆液性囊腺瘤、10例（7.69%）实性假乳头状瘤、42例（32.31%）胰腺癌和17例（13.08%）非肿瘤性PCL。51例纳入恶性组，余79例为良性组。多因素logistic回归分析显示，EUS显示囊壁增厚（P=0.018，OR=3.603，95%CI：1.274~10.883）、实性成分（P=0.002，OR=33.041，95%CI：5.572~671.106）及胰周组织浸润（P=0.004，OR=12.810，95%CI：2.747~101.812）为预测恶性PCL的独立危险因素，其受试者工作特征曲线下面积为0.910。99例手术确诊病例中，CT、MRI及EUS⁃FNA诊断恶性PCL的准确率分别为39.8%（31/80）、49.3%（34/69）和56.6%（56/99）。结论 EUS有助于鉴别诊断恶性PCL，EUS⁃FNA诊断恶性PCL的准确率优于CT、MRI。

Abstract:

Objective To study the endoscopic ultrasound (EUS) characteristics of malignant pancreatic cystic lesions (PCL) and to evaluate the diagnostic efficacy of EUS-guided fine needle aspiration (EUS-FNA). Methods A retrospective analysis was performed on data of patients who underwent EUS-FNA and diagnosed as having PCL at the First Affiliated Hospital of Naval Medical University from January 2006 to December 2021. Clinical and EUS characteristics between benign and malignant PCL were compared. High-risk characteristics of EUS for malignant PCL were identified by logistic regression analysis. The diagnostic efficacy of EUS-FNA, computed tomography (CT), and magnetic resonance imaging (MRI) for PCL was compared in the surgically confirmed cases. Results A total of 130 PCL cases included 27 (20.77%) mucinous cystic neoplasms (in which 1 case had invasive carcinoma), 19 (14.62%) intraductal papillary mucinous neoplasm (in which 8 cases had invasive carcinoma or high dysplasia), 15 (11.54%) serous cystic neoplasms, 10 (7.69%) solid pseudopapillary neoplasms, 42 (32.31%) pancreatic cancer, and 17 (13.08%) non-neoplastic PCL. Fifty-one cases were included in the malignant group and 79 cases were in the benign group. In multivariate logistic analysis, EUS showing thickened cystic wall (P=0.018, OR=3.603, 95%CI: 1.274-10.883), solid component (P=0.002, OR=33.041, 95%CI: 5.572-671.106), and peripancreatic invasion (P=0.004, OR=12.810, 95%CI: 2.747-101.812) were independent risk factors for predicting malignant PCL, with an area under receiver operating characteristic curve of 0.910. Of the 99 surgically confirmed cases, the diagnostic accuracy of CT, MRI and EUS-FNA for malignant PCL was 39.8% (31/80), 49.3% (34/69) and 56.6% (56/99) respectively. Conclusion EUS characteristics may help to identify malignant PCL. The diagnostic accuracy of EUS-FNA for malignant PCL is superior to those of CT and MRI.