RESUMO
BACKGROUND: Ultrasound-guided transthoracic core needle biopsy (US-TCNB) is a promising method for establishing the correct diagnosis of mediastinal masses. However, the existing studies in this area are scant and with small samples. PURPOSE: To evaluate the diagnostic value and the complication rate of US-TCNB, particularly large bore cutting biopsy in patients with mediastinal lesions. MATERIAL AND METHODS: This retrospective study includes 566 patients with mediastinal lesions suspicious of malignancy evaluated between March 2004 and December 2018. Inclusion criteria: 1. Patients with mediastinal lesions detected on thoracic CT scan; 2. Lesions more than 15 mm; 3. Negative histological diagnosis after bronchoscopic biopsy; 4. Normal coagulation status; 5. Cooperative patient; 6. Written informed consent. US visualization of the mediastinal lesions was successful in 308 (54.4%). In all of them, US-TCNB was performed. All patients with mediastinal lesions unsuitable for US visualization were evaluated for a CT-guided transthoracic needle biopsy (CT-TTNB), which was done if the presence of a safe trajectory was available (n = 41, 7.2%). All patients inappropriate for image-guided TTNB were referred to primary surgical diagnostic procedures (n = 217, 38.3%). RESULTS: The US-TCNB is a highly effective (accuracy 96%, sensitivity 95%) and safe tool (2.6% complications) in the diagnosis of all subgroups mediastinal lesions. It is non-inferior to CT-TTNB (90%) and comes close to the effectiveness of surgical biopsy techniques (98.4%), but is less invasive and with a lower complication rate. CONCLUSION: US-TCNB of mediastinal lesions is highly effective and safe tool which is particularly helpful in critically ill patients.