3559 - Thymic Composition on CT Associates with Peripheral Immune Phenotype in Lung Cancer

Purpose/Objective(s): Immune function influences NSCLC outcomes, yet scalable noninvasive immune assessment tools are limited. Building on CT-based thymic analysis literature, we developed percent thymic tissue (pTT), a CT-based measure of residual adult thymus. While pTT correlates with NSCLC outcomes and toxicity, the underlying mechanisms are unclear. We investigated whether pTT correlates with peripheral immune markers in NSCLC, hypothesizing that high pTT correlates with increased CD8? T cells and decreased exhaustion markers.

Materials/Methods: We included adults with stage II-III NSCLC planned for definitive radiotherapy with chemotherapy and/or immunotherapy at a single institution who had pre-treatment immune profiling. pTT was estimated via thymic bed segmentation and Gaussian mixture modeling. Patients were categorized as having preserved thymic tissue (=1% pTT, “functional” thymus) versus near-complete fatty replacement (<1% pTT, “involuted” thymus). T-cell and cytokine levels were compared using linear mixed-effects models (LMM) adjusted for age and sex. A random forest (RF) classifier with 5-fold cross-validation assessed discriminative performance of broad T-cell subsets and LMM-identified markers. Change-in-estimate sensitivity analysis evaluated potential confounding by tumor genomics.

Results: Pre-treatment blood samples from 89 patients were analyzed (median age 74, 44.9% female). 61.8% (n=55) had functional thymus. TIM3, an exhaustion marker, was higher in patients with involuted thymus. %TIM3? of CD4?Ki67? cells (ß = 0.35, 95% CI 0.06–0.65, p = 0.020), %TIM3? of CD4?PD1? cells (ß = 0.43, 95% CI 0.05–0.81, p = 0.029), and %TIM3? of CD4? T cells (ß = 0.36, 95% CI 0.003–0.72, p = 0.051) were higher in involuted patients. Involuted patients trended toward higher %CD3 and %CD4? cells, with lower %CD8? cells; however, these differences were not statistically significant. The RF model achieved moderate discriminative performance for identifying involution (AUC 0.744), driven by %CD4?, %CD3?, %TIM3? of CD4?, and %CD8?. Associations remained robust when controlling for tumor mutational burden, homologous recombination repair deficiency, and mutations in TP53 and RBM10. Controlling for KEAP1 strengthened signals for TIM3 subsets (e.g., %TIM3? of CD4?PD1? ß increased from 0.72 to 0.90, p=0.012). Higher pTT also correlated with increased IL-8 (p = 0.019) and IL-13 (p = 0.032).

Conclusion: Radiographic thymic involution is associated with a CD4?-skewed, exhausted T-cell compartment, independent of tumor genomics. NSCLC patients with involuted pTT exhibited increased expression of TIM3, an immune checkpoint protein, within CD4? cells, suggesting that NSCLC patients lacking functional thymus harbor a more exhausted, less cytotoxic immune profile. Higher pTT also correlated with increased IL-8 and IL-13. These findings suggest pTT captures distinct tumor-host immune interactions and warrants further investigation as a noninvasive marker of immune status.