More lymph nodes examined at segmentectomy tracked with longer survival in early-stage NSCLC

Five-year survival from 52.5% to 83.9% across four yield groups

The cohort draws on SEER data from 2000 to 2022 and was restricted to anatomical segmentectomy, so wedge resections and lobectomies were outside the analysis. Over a median follow-up of 58 months, 783 of the 2,217 patients died. Kaplan-Meier curves for the four yield groups separated with a log-rank P below 0.001.

Each step up in yield carried a higher five-year survival estimate. The gap between patients with no nodes examined and those with more than 10 was 31.4 percentage points. The largest single step, of 13.4 points, lay between no nodes and 1 to 5 nodes.

A hazard ratio of 0.40 with more than ten nodes, after adjustment

In multivariable Cox regression, against patients with no nodes examined, the hazard ratio for all-cause death was 0.79 (95% CI 0.67 to 0.94) with 1 to 5 nodes, 0.55 (95% CI 0.44 to 0.70) with 6 to 10 and 0.40 (95% CI 0.30 to 0.52) with more than 10. Each comparison reached P≤0.006, and the test for trend gave P<0.001.

The authors also modelled lung cancer-specific death with Fine-Gray competing-risk regression, which treats deaths from other causes as a competing event. Their conclusion states that higher yield was associated with lower lung cancer-specific mortality as well as with longer overall survival.

No clear plateau in the spline, and the same pattern by age and T stage

Grouped categories can suggest a cut-off where none exists, so the investigators also treated node count as a continuous variable using restricted cubic splines. The overall association was significant (P<0.001), and the test for non-linearity gave P=0.27, which the authors read as no strong evidence that the relationship bends or levels off. The spline did not show a clear plateau, although a non-significant test cannot rule one out.

The association held in patients younger than 70 and in those aged 70 or older, and across T1a, T1b and T1c tumours, with no significant interaction in either comparison. Age and tumour size can influence both the extent of a resection and survival, and the association persisted within each of these strata.

“These findings support lymph node yield as a pragmatic marker of nodal evaluation during segmentectomy.”Tian et al., Translational Lung Cancer Research (2026)

A registry association that cannot separate staging from surgical quality

Segmentectomy has been used more widely for small peripheral tumours since the JCOG0802/WJOG4607L trial reported longer overall survival with segmentectomy than with lobectomy in peripheral tumours of 2 cm or less, and CALGB 140503 found sublobar resection non-inferior to lobectomy for disease-free survival. A smaller resection leaves less hilar and segmental tissue in the specimen, so the adequacy of nodal assessment during these operations is an open practical question.

The paper describes a dose-response association and proposes node yield as a marker of nodal evaluation; it does not claim that removing more nodes extends life. A retrospective registry cannot make that distinction. More nodes examined can identify occult nodal disease, moving patients with positive nodes out of the node-negative group and improving the apparent survival of those who remain. A higher count may also reflect the hospital, the surgeon, the pathology work-up or a patient fit enough for a fuller dissection, and adjustment for recorded variables does not remove differences the registry does not capture.

The group with no nodes examined deserves particular caution. Its five-year survival of 52.5% is low for clinically node-negative tumours of 30 mm or less, and a segmentectomy recorded with no nodal assessment may mark patients or procedures that differed from the rest of the cohort in ways the model could not see. The hazard ratios for every other group are expressed against this reference, so its composition affects how large each effect looks.

Node yield proposed as a pragmatic marker of nodal evaluation

The authors frame their result as support for lymph node yield as a pragmatic marker of nodal evaluation during segmentectomy. A count of examined nodes is already recorded in routine pathology reports, so it could be collected without additional work. Whether a specific minimum count should be set, and whether raising yield in practice changes outcomes, would need prospective data or analyses that capture the reasons behind low yields.