Cytology of endobronchial ultrasound-guided transbronchial needle aspiration: a retrospective study with histology correlation.

BACKGROUND

Endobronchial ultrasound (EBUS) is a relatively new modality that can be used to guide transbronchial needle aspiration (TBNA) of mediastinal and hilar lymph nodes and peripheral lung lesions. Few studies have investigated the cytological profile of EBUS-TBNA specimens. In this study, we have reviewed the cytological profile of 135 consecutive cases, including 71 lymph node cases, 4 lung cases, and 60 cases of both lymph node and lung sampling. Our study contains the largest number of cases in the evaluation of cytomorphology.

METHODS

The cytological specimens were collected using an ultrasound bronchofibervideoscope with a 22-gauge needle and core biopsies were obtained with a 19-gauge needle. An experienced cytotechnologist performed an immediate on-site evaluation of adequacy. An immediate assessment was given to the clinician after each pass. In many patients, multiple sites were sampled. The average slides of each case were 9.9 (median of 12), with a range from 2 to 24.

RESULTS

Of 131 cases of lymph node sampling, 45 cases (34.6%) were diagnosed as malignant, 73 cases (55.7%) as benign process, 5 cases (3.8%) as suspicious for malignancy, and 1 case (0.8%) as atypical cells. Of the 64 cases of lung lesion sampling, 21 cases (32.8%) were diagnosed as malignant, 35 cases (54.7%) as benign process, 1 case (1.5%) as suspicious for malignancy, and 4 cases (6.3%) as atypical cells. The lymph node nondiagnostic rate was 5.3%, whereas the nondiagnostic rate for lung lesions was 4.7%. Eighty-eight cases (65.2%, 88/135) had corresponding core biopsies (with a 19-gauge needle) or follow-up surgery. When histology was taken as the gold standard, the sensitivity, specificity, and positive and negative predictive values for EBUS-TBNA were 85.0%, 100%, and 100% and 89.7%, respectively. However, when both histology and clinical follow-up were considered together, the overall sensitivity and negative predictive values were increased to 94.7% (P < .05) and 96.6% (P < .05), respectively.

CONCLUSIONS

This study shows that EBUS-TBNA is an accurate and sensitive method for diagnosing and staging lung cancer. The constant challenge that we as cytopathologists are now facing is how to improve our diagnostic ability and accuracy for lung cancer. We believe that this optimal goal can be achieved with the effective use of EBUS-TBNA sampling and collaboration with our clinical colleagues.