88P Uncertain resection for localized cN0M0 non-small cell lung cancer: The crucial prognosis of suboptimal lymph node assessment

Abstract

Surgery remains the gold standard in the treatment of early stages of lung cancer.To achieve longer overall survival, complete resection of the tumor and careful lymphadenectomy are essential. The concept of uncertain resection R(un) was defined in 2005 by IASLC to highlight inadequate and limited resections and focused on intraoperative lymph node assessment. Although complete lymphadenectomy is recommended even in cN0M0 patients, the effect of R(un) in these patients has never been studied. We include cN0M0 patients who underwent an anatomical lung resection for NSCLC from 2008 to 2018 at Toulouse University Hospital. We retrospectively reclassified tumours according to IASLC R-status based on pathologist reports. The primary outcome was overall patient survival according to the R classification system. Secondary outcomes were to determine whether uncertain resection was an independent factor in mortality, to highlight heterogenity of R(un) to perform subcategorization predicting survival. Of the 1108 cN0M0 patients, we identified 732 (66.1%) R0 patients, 291 (26.2%) R(un) patients, and 85 (7.7%) R1 patients. There was a significant difference in survival according to R status (p<0.0001). The impact of complete resection was confirmed in the adjusted models, with an adjusted hazard ratio of 1.26 (95% CI: 1.01-1.59) between R0 and R(un) patients and 2.41 (95% CI: 1.76-3.30) between R0 and R1 patients. We found substantial heterogeneity among R(un) patients with significantly different survival according to the R(un) definition criteria. Therefore, we defined 3 new subgroups of R(un) to discriminate between populations of homogeneous survival. R(un)c included all patients with lymph node involvement of the highest mediastinal node. R(un)b included patients with less than 3 mediastinal lymph node sites examined or no station 7 nodes examined. All other R(un) patients who did not meet the above criteria were classified as R(un)a. Our study demonstrated the validity of the IASLC R classification in cN0M0 patients. By demonstrating the heterogeneity of R(un)patients, we proposed the creation of subcategories to highlight the critical problem of suboptimal lymph node assessment.