Diagnostic accuracy of image-guided percutaneous fine needle aspiration biopsy of the mediastinum.

Interpreting a fine needle aspiration biopsy (FNAB) from the mediastinum is challenging as this location may harbor many lesions, including primary and metastatic tumors. Image-guided transthoracic (percutaneous) FNAB is less invasive than mediastinoscopy or endoscopic-guided FNAB. The aim of this study was to determine the diagnostic accuracy of FNAB performed percutaneously for evaluating mediastinal lesions.A retrospective study of 157 consecutive CT-guided transthoracic FNAB of the mediastinum was performed (1988-2004). Direct smears (N = 145; average 13 slides/case), ThinPrep slides (N = 25), and adequate cell blocks (N = 131) were prepared from procured cytologic material. When needed, ancillary studies included immunocytochemistry (N = 53) and flow cytometry (N = 8). Subsequent histologic tissue diagnoses available for 68 cases were also reviewed. Patients were of average age 57 yr (range 1-88 yr), including 75 males and 82 females. A definitive diagnosis was rendered in 128 (82%) cases. Primary neoplasms (N = 38) included 24 lymphomas (6 Hodgkin and 18 non-Hodgkin), 7 thymomas, 1 thymic carcinoma, and 6 peripheral nerve sheath tumors. Metastases (N = 72) were mainly carcinomas (N = 71) and 1 melanoma. There were 4 non-neoplastic lesions (1 granulomatous process; 2 bronchogenic and 1 pericardial cyst), 1 case of undifferentiated malignant large cell neoplasm, 13 cases negative for malignancy, and 29 (18%) that were indeterminate, due largely to insufficient cellularity. Subsequent histologic diagnoses were concordant with FNAB diagnoses in 53/68 cases (78%). Nine FNAB were inadequate/nondiagnostic. There were 6 discordant cases, including 5 FNAB that were of adequate cellularity but interpreted as negative for malignant cells (on subsequent histology 2 turned out to be Hodgkin lymphoma, 2 carcinomas, and 1 diffuse large cell lymphoma), and 1 diagnosed as thymoma that on histologic evaluation was a thymic large cell lymphoma. Adequate diagnostic cytologic material was obtained by image-guided percutaneous FNAB of mediastinal lesions in 82% of our cases. Sufficient material was available to make cell blocks and perform ancillary studies when necessary. These data also show a high proportion of agreement (78%) between FNAB and subsequent histologic diagnoses for a wide variety of mediastinal lesions. The majority of discordant cases were primarily interpretive, with a final cytologic diagnosis negative for malignancy. Only one problematic case misdiagnosed on FNAB as thymoma was found on subsequent surgical excision to be a thymic large B cell lymphoma. Cases with nondefinitive FNAB diagnoses were largely due to sampling error and/or insufficient cellularity. Therefore, percutaneous FNAB of the mediastinum is a diagnostically helpful, minimally invasive procedure that can be performed in patients of all ages as part of the evaluation of a mediastinal mass lesion.