Utility of EBUS-TBNA in PET-positive mediastinal lymph nodes in subjects with extra-thoracic malignancy

Patients with primary extra-thoracic malignancy (ETM) often have hyper-metabolic mediastinal lymph nodes (HM-MLN) in the PET-scan done for initial staging or post treatment follow-up. There is scant data on the etiology of HM-MLN in such patients, which can also be due to non-malignant causes. We used endobronchial ultrasound (EBUS) guided sampling to determine the etiology of HM-MLN in patients with ETM and study the relationship between PET-SUV values and a diagnosis of malignancy in this population.

65 consecutive patients, from March 2013 to March 2017 with either known ETM for primary staging or post-treatment follow-up, with PET CT showing HM-MLN (SUV > 2.5) were included in the study.

65 patients with ETM had EBUS-TBNA for HM-MLN. 20/65 (30.7%) were malignant, 45/65 (69.23%) were benign MLN. In patients with benign etiology of HM-MLN, 6/45 (13.3%) had necrotising granulomatous, 24/45 (53.3%) had non- necrotising granulomatous MLN and 15/45 (33.3%) had reactive MLN. We found discordance (i.e. primary ETM responded to treatment and a new HM-MLN was detected) in 21/65 (32.3%) patients with PET-CT done for initial ETM staging, and 44/65 (67.7%) with a post-treatment PET-CT. showed. Correlating SUV with diagnoses, the SUV values in EBUS-proven malignant MLN were 8.9 ± 4.1, while they were 10.2 ± 5.57 in benign MLN. There was no statistically significant difference between the SUV of benign and malignant MLNs.

This study shows a significant incidence of EBUS-TBNA proven benign diagnoses 45/65 (69.2%) in 'SUV-deemed-malignant MLN' and a poor relationship between high SUV and malignant MLN, in patients with known ETM. The ETM related HM-MLN have a significant chance of being benign, and a tissue diagnosis is imperative as it impacts on the treatment plan and prognosis.