Validation of a risk stratification model for patients with isolated mediastinal/hilar lymphadenopathy (IMHL) and a negative EBUS

Background: IMHL may be pathological (malignancy, TB, sarcoid) or 'reactive'. EBUS transbronchial needle aspiration (EBUS-TBNA) is the first line investigation but has a reported (Navani, N. et al. Am J Respir Crit Care Med 2012) low negative predictive value (NPV). We devised a risk stratification model to classify patients as high or low risk of false negative sampling after negative EBUS-TBNA (Evison, M et al. BMJ Open Respir Res 2014). The aim of this study was to validate our model in a 2nd cohort. Methods: A retrospective review of prospectively collected data on consecutive patients undergoing EBUS-TBNA for IMHL (Sept 2013-July 2014). Patients with negative EBUS-TBNA were classified high or low risk (low risk defined: max lymphnode diameter <20mm and at least 1 co-morbidity associated with IMHL). Reactive lymphadenopathy was diagnosed if all pathological/microbiological sampling and ≥6 months follow-up failed to yield an alternative diagnosis. Results: 61 patients with IMHL underwent EBUS-TBNA. Pathological diagnosis achieved in 27 (44%) (7 carcinoma, 14 sarcoidosis, 6 TB). Patients with negative EBUS-TBNA (n=34) were classified as low (n=19) or high (n=15) risk. In the low risk cohort 17/19 (89.5%) were diagnosed with reactive lymphadenopathy and 2/19 with sarcoidosis. In the high risk cohort 11/15 (73.3%) had pathological cause of IMHL (8 sarcoidosis, 2 TB, 1 MAI) and 4/15 reactive. NPV of EBUS alone was 61.8%, increasing to 89.5% with the risk stratification model (p=0.055). Conclusion: The risk stratification model improved the NPV of EBUS. This requires prospective validation to determine its potential clinical role in managing patients with IMHL.