Endoscopic ultrasound-guided fine-needle aspiration in 179 cases: the M. D. Anderson Cancer Center experience

Abstract

Background: Recently, endoscopic ultrasound-guided fine-needle aspiration (EUS-FNA) has emerged as a diagnostic adjunct for small pancreatic lesions and abdominal and mediastinal lymph node diseases.

Design: During a 21-month period, we performed 179 EUS-FNAs in 166 consecutive patients; these data are the subject of this study. An average of 2.6 needle passes were obtained and aspiration was performed most commonly in the pancreas (162 cases, 91%). The FNA smears were reviewed using six diagnostic categories (negative for malignancy/nondiagnostic [NND], atypia, suspicious for malignancy, benign tumor/cyst, neuroendocrine neoplasm [NEN], and carcinoma). The review diagnosis was correlated with the histologic diagnosis made on resection or surgical biopsy specimens in 70 cases. Up to 17 months of clinical follow-up were sought for the cases with a negative or inconclusive FNA diagnosis and no diagnostic tissue confirmation (81 cases).

Results: The review FNA diagnoses were as follows: NND (49 cases), atypia (17 cases), suspicious for malignancy (12 cases), benign tumor/cyst excluding NEN (10 cases), NEN (6 cases), carcinoma (85 cases). Follow-up methods included resection (49 cases), surgical biopsy (21 cases), repeat FNA or brushing cytology (28 cases), and clinical follow-up only (81 cases). Of the 49 NND cases, 23 (47%) had positive follow-up results (i.e., false-negative diagnosis) that were confirmed by tissue diagnosis (resection/surgical biopsy in 11 cases [48%] and repeat FNA/brushing in 12 cases [52%]). These included pancreatic/ampullary adenocarcinoma in 20 cases, esophageal squamous carcinoma in 1 case, and NEN in 2 cases. Follow-up also revealed carcinoma in all 12 suspicious cases and 13 pancreatic adenocarcinomas and 1 microcystic adenoma in 14 of the 17 atypical cases. Overall, repeat computed tomography (CT)-guided FNA samples yielded a definite diagnosis in four atypical and seven NND cases, whereas EUS-FNA results provided a definite diagnosis in three cases in which CT-guided FNA failed and in two cases in which ampullary biopsy failed. No false-positive cases were identified. The false-negative rate due to inadequate sampling was 13.2%. Sensitivity (including cases with inadequate cellularity and nondiagnostic aspirates) was 81.7% and specificity was 100%. None of the factors evaluated (lesion characteristics, aspiration site, and tumor type) significantly influenced diagnostic results.

Conclusion: EUS-FNA is a valuable diagnostic and staging tool with high specificity and sensitivity. Negative or nondiagnostic cases on EUS-FNA require further diagnostic work for a definitive diagnosis when clinical or radiographic findings do not correlate with the FNA results.