This single‑center retrospective study quantified intratumoral CD8+ T cell subpopulations in primary non‑small cell lung cancer (NSCLC) using tissue microarrays from both the tumor center (TC) and invasive margin (IM). Multicolor immunofluorescence distinguished total CD8+ T cells (CK‑ CD8+), tissue‑resident memory T cells (TRM, CK‑ CD103+ CD8+), and bystander CD8+ T cells (Tbys, CK‑ CD103‑ CD8+).
Measured proportions in the TC region were: TRM/CD8+ 34.3%, Tbys/CD8+ 67.9%, and a TRM/Tbys ratio of 50.5%. In the IM region, TRM/CD8+ was 29.6%, Tbys/CD8+ 70.7%, and TRM/Tbys 41.9%.
The authors retrospectively collected 256 patients with stage IA–IIIB NSCLC who underwent radical surgical resection at Shandong Provincial Cancer Hospital between January 1, 2014 and December 31, 2018. Tissue microarrays were prepared to include both TC and IM areas for each tumor. Multicolor immunofluorescence staining quantified densities of CK‑ CD8+ cells, CK‑ CD103+ CD8+ (TRM), and CK‑ CD103‑ CD8+ (Tbys). Recurrence‑free survival (RFS) was assessed using Kaplan‑Meier curves and Cox proportional hazards models.
Median follow‑up was 37.9 months. During that interval, 87 of 256 patients (34.0%) developed recurrence. In analyses that considered the entire cohort without stratification, neither TRM/CD8+, Tbys/CD8+, nor the TRM/Tbys ratio measured in either the IM or TC regions were significantly associated with RFS (all P > 0.05).
When the full set of 256 patients was evaluated together, the densities and relative proportions of the measured CD8+ T cell subsets in the tumor center and invasive margin showed no statistically significant relationship with recurrence‑free survival. The authors therefore proceeded to stratify patients by lymph node involvement to search for subgroup‑specific associations.
Among the 172 patients staged N0 (no regional lymph node metastasis), several statistically significant associations between intratumoral CD8+ subpopulations and RFS were observed:
High Tbys/CD8+ in the IM region was associated with improved RFS compared with the low Tbys/CD8+ group (P = 0.050).
In the TC region, high TRM/CD8+ and high TRM/Tbys groups had lower RFS than corresponding low‑level groups (both P < 0.05).
Also in the TC region, the high Tbys/CD8+ group had better RFS than the low group (P = 0.005).
Multivariate Cox proportional hazards analysis restricted to N0 patients identified three independent predictors of postoperative recurrence: high TC TRM/CD8+ (hazard ratio [HR] = 2.772, 95% confidence interval [CI]: 1.260–6.098) and high TC TRM/Tbys (HR = 2.656, 95% CI: 1.205–5.855) were independent risk factors, while high TC Tbys/CD8+ was an independent protective factor (HR = 0.324, 95% CI: 0.147–0.711).
Among the 84 patients with N1–2 disease (regional lymph node metastasis present), neither TRM/CD8+, Tbys/CD8+, nor TRM/Tbys measured in the TC or IM regions were associated with RFS (all P > 0.05).
In adjusted models among N0 patients, the study reports:
These findings indicate that relative enrichment of TRM within the tumor center, and a higher TRM/Tbys balance, correlated with increased recurrence risk after curative resection in node‑negative patients, while enrichment of bystander CD8+ T cells in the tumor center correlated with lower recurrence risk.
The authors conclude that specific CD8+ T cell subsets localized to the tumor center of primary NSCLC are associated with postoperative recurrence only in N0 patients. Elevated TC TRM/CD8+ and TRM/Tbys ratios together with decreased TC Tbys/CD8+ ratios identify N0 patients at higher risk of recurrence after curative surgery. By contrast, these intratumoral T cell measures did not predict recurrence in node‑positive (N1–2) patients in this cohort.
If externally validated, these intratumoral immunophenotypes could support postoperative risk stratification and inform individualized surveillance or adjuvant therapy decisions for node‑negative NSCLC. The source reports no conflicts of interest declared by the authors.
Note: The article text available from the source supplied the cohort, methods, measured percentages, follow‑up duration, recurrence counts, P values, hazard ratios, and confidence intervals. No additional numerical or methodological details beyond those reported in the source were available in the provided abstract.