A prospective study comparing rapid assessment of smears and ThinPrep for endoscopic ultrasound-guided fine-needle aspirates.

ABSTRACT

BACKGROUND AND STUDY AIMS: ThinPrep is often used for endoscopic ultrasound fine-needle aspiration (EUS-FNA) samples but the sensitivity of this method is unknown. The objective of the study was to compare sensitivity and accuracy of ThinPrep versus the smear method in pancreas and lymph node samples obtained by EUS-FNA. PATIENTS AND METHODS: Patients with suspected malignancy in the pancreas or lymph node underwent EUS-FNA. On-site rapid assessment of all aspirates using the smear method was performed. After rapid assessment, three additional passes from each site were submitted into ThinPrep liquid medium. Cytopathologists interpreting the smear method and ThinPrep slides were blinded to each other. The gold standard was final cytology or pathology results. RESULTS: A total of 130 patients (36 % women, mean age 63 years) underwent EUS-FNA of 139 sites (50 pancreas, 89 lymph node). Malignancy was confirmed in 47 pancreas samples (94 %) and 48 lymph node samples (54 %). Mean +/- SD number of passes made for the smear method was 2.6 +/- 1.3. For pancreatic cancer, the sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV), and accuracy of the ThinPrep versus the smear method were 62 % versus 98 %, 100 % versus 100 %, 100 % versus 100 %, 14 % versus 75 %, and 64 % versus 98 %, respectively. For lymph nodes the values were 67 % versus 92 %, 100 % versus 98 %, 100 % versus 98 %, 72 % versus 72 %, and 82 % versus 94 %, respectively. CONCLUSIONS: The smear method is more sensitive and accurate than ThinPrep in detecting malignancy from EUS-FNA samples of the pancreas and lymph nodes. Smear method with on-site rapid assessment should be favored over ThinPrep in suspected malignancy.