Prognostic performance of the N category in the 9th edition of lung cancer staging

Objectives To compare the prognostic performance of the N category of lung cancer in the 9th edition with previous editions (7th edition and 8th edition's proposal). Methods Patients who underwent lobectomy or pneumonectomy for lung cancer from January 2015 to December 2021 were retrospectively analyzed. Clinical and pathologic N categories were reclassified according to the 9th edition (N0, N1, N2a, and N2b), the 8th edition's proposal (N0, N1a, N1b, N2a1, N2a2, and N2b), and the 7th edition (N0, N1, and N2). Concordance index (C-index) and calibration were assessed for each edition. Results A total of 3864 patients were included (962 pN positive and 513 cN positive). The 9th edition demonstrated clear hazard stratification between neighboring pN categories after multivariable adjustment, whereas multiple overlaps were observed in the 8th edition's proposal. It had superior discrimination performance compared with the 7th edition in pathologic staging (all p < 0.05). Compared with the 8th edition's proposal, the 9th edition showed comparable performance in pN2 and overall patients (C-index, 0.560 vs 0.569 [p = 0.163]; 0.666 vs 0.668 [p = 0.396]), In clinical staging, there was no difference in discrimination across 7th to 9th editions (all p > 0.05). N1 dichotomization in the 8th edition's proposal showed discrimination ability (C-index, 0.539 [95% confidence interval: 0.502-0.576]) only in pathologic staging. The calibration was acceptable across the clinical 7th to 9th editions for 5-year survival. Conclusion The revision of the N category in the 9th edition appears reasonable, offering enhanced prognostic discrimination compared with the 7th edition and comparability to the 8th edition's proposal.