Is central lung tumour location really predictive for occult mediastinal nodal disease in (suspected) non-small-cell lung cancer staged cN0 on 18F-fluorodeoxyglucose positron emission tomography–computed tomography?

Abstract

Current guidelines recommend preoperative invasive mediastinal staging in centrally located tumours with negative mediastinum on positron emission tomography-computed tomography, based on a 20-30% prevalence of occult mediastinal disease (pN2-3). However, a uniform definition of central tumour location is lacking. Our objective was to determine the best definition in predicting occult pN2-3. A single-institution database was queried for patients with (suspected) non-small-cell lung cancer staged cN0 after positron emission tomography-computed tomography and referred to invasive staging and/or primary surgery. We evaluated 5 definitions: inner 1/3, inner 2/3, contact with bronchovascular structures, ≤2 cm from bronchus or endobronchial visualization. Between 2005 and 2015, 813 patients were eligible (cT1: 42%, cT2: 28%, cT3: 17% and cT4: 11%). Invasive mediastinal staging and resection were performed in 30% and 97% of patients, respectively. Any nodal upstaging (pN+) was found in 21% of patients, of whom pN2-3 was found in 8%. Central tumour location demonstrated 4 times higher odds for any pN+ [for inner 1/3 vs outer 2/3, odds ratio 3.90 (95% confidence interval 2.24-6.77), P < 0.001], whereas no significantly different odds was observed for pN2-3. The discriminative ability for pN+ was not significantly different between the several definitions. The prevalence of occult pN2-3 was only 8% when modern fusion positron emission tomography-computed tomography imaging pointed at clinical N0 non-small-cell lung cancer. None of the 5 verified definitions of centrality was predictive for occult pN2-3. However, each definition of centrality was related to any pN+ at a prevalence of 21%, without significant differences in discriminative ability between definitions. These data question whether indication for preoperative invasive mediastinal staging should be based on centrality alone.