Surgical or medical strategy for locally-advanced, stage IIIA/B-N2 non-small cell lung cancer: Reproducibility of decision-making at a multidisciplinary tumor board

International audienceBackground: Stage IIIA/B-N2 is a very heterogeneous group of patients and accounts for one third of NSCLC at diagnosis. The best treatment strategy is established at a Multidisciplinary Tumor Board (MTB): surgical resection with neoadjuvant or adjuvant therapy versus definitive chemoradiation with immune checkpoint inhibitors consolidation. Despite the crucial role of MTBs in this complex setting, limited data is available regarding its performances and the reproducibility of the decision-making. Methods: Using a large cohort of IIIA/B-N2 NSCLC patients, we described patient's characteristics and treatment strategies established at the initial MTB: with a “surgical strategy” group, for potentially resectable disease, and a “medical strategy” group for non-resectable patients. A third group consisted of patients who were not eligible for surgery after neoadjuvant treatment and switched from the surgical to the medical strategy. We randomly selected 30 cases (10 in each of the 3 groups) for a blinded re-discussion at a fictive MTB and analyzed the reproducibility and factors associated with treatment decision. Results: Ninety-seven IIIA/B-N2 NSCLC patients were enrolled between June 2017 and December 2019. The initial MTB opted for a medical or a surgical strategy in 44% and 56% of patients respectively. We identified histology, tumor size and localization, extent of lymph node involvement and the presence of bulky mediastinal nodes as key decision-making factors. Thirteen patients were not eligible for surgical resection after neoadjuvant therapy and switched for a medical strategy. Overall concordance between the initial decision and the re-discussion was 70%. The kappa correlation coefficient was 0.43. Concordance was higher for patients with limited mediastinal node invasion. Survival did not appear to be impacted by conflicting decisions. Conclusions: Reproducibility of treatment decision-making for stage IIIA/B-N2 NSCLC patients at a MTB is moderate but does not impact survival