Endoscopic ultrasound-guided fine-needle aspiration biopsy of liver lesions: histological and cytological assessment.

BACKGROUND AND STUDY AIMS: EUS-guided fine-needle aspiration biopsy (EUS-FNA) is used increasingly for the diagnosis of mediastinal, biliopancreatic, and gastric tumors. However, little is known about EUS-FNA in hepatic lesions and the best method for tissue analysis. We assessed EUS-FNA combined with histological and cytological evaluation in selected patients. PATIENTS AND METHODS: 41 patients (66 +/- 7 years) were prospectively studied, 33 of whom had clinical findings suggestive of liver malignancies. Selection for EUS-FNA was based on an increased risk of bleeding from percutaneous biopsy (coagulopathy, cirrhosis, ascites, aspirin intake; n = 15), presence of small liver tumors < 2 cm (n = 12), or liver lesions found incidentally (n = 14). Transgastric EUS-FNA of lesions located in accessible liver segments was performed using the Hitachi FG-34UX longitudinal echo endoscope and a 22-G aspiration needle. Specimens were submitted separately for standard cytological and histological evaluation. In the case of malignancies, findings at surgery with histological examination, endoscopy, or computed tomography (CT)-guided biopsy of the primary cancer served as reference results (n = 33), while in benign disorders, a combination of imaging studies (Magnetic Resonance Tomography <MRT>, scintigraphy) and the clinical follow-up, as summarized in the physician's report, was used as reference.
RESULTS: EUS-FNA provided appropriate biopsy specimens in 40/41 patients. It was not possible to aspirate sufficient material in one patient. On average, 1.4 needle passes were necessary to obtain sufficient amounts of tissue. With regard to malignancy, the combination of histological and cytological examination had a sensitivity of 94%, specificity of 100%, negative predictive value (NPV) of 78%, and positive predictive value (PPV) of 100%. Tissue diagnoses were in agreement in 27/41 patients (65%). In the remaining patients, only the cytological examination identified six lesions correctly, while the histological assessment was correct in another seven patients. Malignant lesions were correctly identified by cytology in 24/33 (73%) patients, while histology alone was diagnostic for malignancy in 27/33 (82%) patients. When both modalities were combined, 31 out of 33-malignancies (94%) were correctly diagnosed. Minor complications occurred in two patients and consisted of self-limiting local bleeding.
CONCLUSIONS: EUS-FNA of liver tumors is a powerful, reliable, and safe procedure for the diagnosis of malignant liver lesions. Optimal diagnostic results are achieved by combining cytological with histological assessment. Hence, EUS-FNA is an alternative to percutaneous biopsy, particularly in patients at risk of bleeding or with small lesions of the liver.