KEY POINTS
- This retrospective study analyzed two independent cohorts of locally advanced NSCLC patients treated with definitive concurrent chemoradiotherapy at Copenhagen University Hospital and Princess Margaret Cancer Centre. The final analytical cohorts contained 165 patients (cohort A) and 159 patients (cohort B), all receiving 60–66 Gy with concurrent platinum-based chemotherapy. Patients required baseline and serial absolute lymphocyte counts (ALC), completed radiotherapy without interruption, and had available treatment plans. The study examined whether radiation dose or underlying tumor burden better explained lymphocyte depletion and its association with overall survival.
- The investigators compared radiation exposure to uninvolved mediastinal tissue with established immune-dose metrics. An automatically segmented mediastinal structure, excluding gross tumor and major organs, served as a surrogate for irradiation of uninvolved mediastinal lymphatic tissue. Mean mediastinum dose (MMD) was compared with mean lung dose, mean heart dose, mean body dose, and estimated dose to immune cells (EDIC). Multivariable models incorporated baseline lymphocyte count and clinical factors, while bootstrap resampling assessed predictor stability.
- Mediastinal dose predicted end-of-treatment lymphocyte counts as well as EDIC. In cohort A, models incorporating MMD and EDIC achieved adjusted R² values of 0.332 and 0.337, respectively; corresponding values in cohort B were 0.253 and 0.252. The similar explanatory performance suggests that irradiation of lymphatic tissue may contribute to lymphocyte depletion alongside irradiation of circulating blood. However, no model explained more than approximately one-third of lymphocyte-count variation, and dose to multiple thoracic structures was strongly correlated.
- Severe lymphopenia was associated with substantially worse survival before adjustment. Grade ≥3 lymphopenia, defined as end-of-RT ALC <0.5 × 10⁹/L, occurred in 52/165 patients (31.5%) in cohort A and 103/159 (64.8%) in cohort B. Unadjusted mortality hazard ratios for patients with severe lymphopenia were 2.29 (95% CI 1.39–3.77; p<0.001) and 1.68 (95% CI 1.04–2.71; p=0.03), respectively. The large difference in lymphopenia prevalence also highlights the heterogeneity between treatment populations.
- The survival association weakened after matching for disease-related characteristics. In cohort A, propensity score matching based on tumor volume, nodal stage, and baseline ALC produced 39 matched patients per group; the mortality HR decreased to 1.42 (95% CI 0.72–2.80). In cohort B, matching on nodal stage, baseline ALC, and sex produced 42 patients per group, with HR 1.41 (95% CI 0.73–2.71). Survival differences were no longer statistically significant in either cohort (p=0.30), suggesting that tumor burden and lymphatic spread may account for part of the initially observed association.
- A similar confounding pattern emerged for estimated immune-cell dose. In cohort A, EDIC above the median of 6.1 Gy was initially associated with worse overall survival (p=0.007). After matching for gross tumor volume and nodal stage, the association was no longer significant (p=0.31). Cohort B showed no significant survival difference when stratified by its median EDIC of 9.7 Gy (p=0.14). These results indicate that larger radiation exposure to immune-related structures may partly reflect more extensive disease rather than an independently modifiable cause of inferior survival.
- The findings do not exclude a causal effect of radiation-induced lymphopenia. Propensity matching substantially reduced the analytical populations, leaving wide confidence intervals compatible with potentially meaningful residual risk. Treatment eras, chemotherapy timing, staging systems, and baseline disease characteristics differed between cohorts, and most patients did not receive post-CRT immunotherapy. The investigators also could not isolate irradiation of lymph nodes from adjacent vessels and circulating blood, or identify a validated structure whose sparing improves clinical outcomes.
CLINICAL TAKEAWAY
Severe lymphopenia after concurrent chemoradiotherapy was associated with inferior survival in locally advanced NSCLC, but this relationship weakened substantially after accounting for tumor burden and nodal disease. The findings challenge causal interpretations of existing lymphocyte-sparing dose models and suggest that mediastinal lymphatic tissue deserves further investigation. Prospective studies are required before prioritizing additional immune-related dose constraints for the purpose of improving survival.