Malignancy risk associated with the EBUS-FNA diagnostic categories nondiagnostic, benign, atypical, suspicious for malignancy, and malignant for mediastinal lymph node aspirate specimens.

INTRODUCTION

Endobronchial ultrasonography-guided transbronchial fine-needle aspiration (EBUS-TBFNA) is used for preoperative staging of primary lung carcinomas. Published sensitivity and specificity are up to 86% and 100%, respectively. Diagnostic categories used by many cytopathologists are nondiagnostic, benign, atypical, suspicious, and malignant. Little information exists about the risk of malignancy associated with each of these categories.

MATERIALS AND METHODS

Records of the Department of Pathology at the University of Utah were searched for all EBUS-TBFNAs of mediastinal and pulmonary hilar lymph nodes. Only cases with surgical follow-up were included in this study. For each diagnostic category (nondiagnostic, benign, atypical, suspicious, and malignant), the percentage of cases proven to be malignant was calculated following correlation of cytologic and surgical diagnoses. Positive and negative predictive values were calculated. For calculation of accuracy statistics, atypical cases were considered benign and suspicious cases were classified as malignant.

RESULTS

For this study, 136 EBUS-TBFNAs of lymph nodes were obtained with adequate surgical follow-up. Risk of malignancy for nondiagnostic specimens was 42%, benign specimens 32%, atypical specimens 40%, suspicious specimens 83%, and malignant specimens 84%. Positive predictive value was 84%, and negative predictive value was 68%.

CONCLUSIONS

The categories stratified malignancy risk ranging from a low of 32% for benign to 84% for malignant. The categories suspicious and malignant had similar malignancy risks. Atypical aspirates had a higher malignancy risk than benign aspirates did. Nondiagnostic aspirates had a malignancy risk similar to that of atypical aspirates. This scoring system may aid in treatment planning and patient counselling.