Lymph Node Sampling Patterns and Completeness of Staging During Systematic Mediastinal Lymph Node Staging in Patients...

Introduction: Systematic endoscopic mediastinal lymph node (LN) staging in early stage non-small-cell lung cancer (NSCLC) is widely recommended for optimal pre-operative staging due to imperfect accuracy of PET/CT. The SEISMIC study utilized synoptic reporting of findings during systematic endoscopic staging to inform radiotherapy planning in patients with locally advanced NSCLC and established a significant rate of discordance between PET-identified extent of disease and observed findings following EBUS. International guidelines recommend systematic LN staging in patients with clinical nodal stage (cN) 1-3 disease as a quality indicator, though performance metrics for quality assessment of systematic endoscopic mediastinal staging in NSCLC have not been described. Methods: A post hoc analysis of 155 participants with locally advanced NSCLC recruited to the SEISMIC study who underwent systematic endoscopic mediastinal staging, with sampling of all LN ≥ 6 mm, prior to planning of radical or high-dose palliative radiation, was completed. Synoptic procedural reports and pre-procedure PET imaging were examined to determine the number of LN sampled, as well as the number of radiologically normal LN (<10 mm diameter, no PET-avidity) sampled, and to establish the completeness of mediastinal staging. Results: Sampling was performed from a median two LN per patient. Sampling from at least one radiologically normal LN was performed in 92/155 (59%) participants, with ≥2 radiologically normal LN sampled in 47 participants (30%). PET-occult LN disease was identified in 18 of 92 (20%) of patients who underwent sampling of radiologically normal LN. The number of LN sampled in individual patients was largely influenced by the number of observed LN ≥ 6 mm in the short-axis dimension. Among these 76 participants staged cN2a by PET, EBUS identified PET-occult metastases in seven (9%), with four (5%) upstaged to cN2b and three (4%) to cN3. In only three patients (2% of cohort) were LN > 6 mm at higher station than the highest LN identified by PET not sampled endoscopically, confirming a high rate of complete mediastinal assessment. Conclusions: Quality of systematic staging in the SEISMIC study was high. Synoptic reporting supports high quality systematic endoscopic mediastinal LN staging in NSCLC and enables performance monitoring of systematic EBUS staging. Novel quality metrics are proposed for quality assessment, and consistent use of these may encourage more extensive LN sampling.