Reconciling Survival Differences in Lobar and Sublobar Resection for Non-Small Cell Lung Cancer

Randomized trials demonstrate noninferiority of sublobar resection in well-selected patients with non-small cell lung cancer, while retrospective cohort studies suggest worse survival. To investigate this discrepancy, we analyzed conditional death rate for both approaches, hypothesizing that while overall death rate for sublobar resection would be higher, the conditional death rate would be stable regardless of resection extent.

The National Cancer Database was queried for adult patients undergoing lobar versus sublobar resection for pathologic stage I and II disease from 2006 to 2017. Data were stratified by stage. Covariates included demographic and clinical variables. Conditional death rate (probability of death given survival at start of year) was plotted and analyzed for divergence.

A total of 139,965 patients were identified with 110,478 (78.9%) stage I patients and 29,487 (21.1%) stage II patients. Sublobar resection was associated with decreased 10-y overall survival (stage I 52% versus 60%; P < 0.001; stage II 28% versus 37%; P < 0.001). For stage I disease the conditional death rate for all approaches (wedge, segmentectomy, and lobectomy) remained proportional and positive throughout the study time-period (0.0002, 0.0004, and 0.0008, respectively). In contrast, for stage II disease, there was an accelerated drop for sublobar resection (wedge and segmentectomy) with a convergence with lobectomy at ∼84 mo (-0.0108, -0.0098, and -0.0074, respectively).

In stage I disease there continues to be decreased survival for patients that undergo sublobar resection in retrospective cancer registries; however, the conditional death rate differs minimally based on extent of resection. This suggests a similar pattern of disease-related death post resection.