Ying Chai ^* Jie Min ^* Fei Dong Yongyuan Chen ^* Wenshan Li ^* Yimin Wu ^* Yanbin Tan ^* Fan Yang ^* Pin Wu ^*

• Zhejiang University, Hangzhou, China

Identifying patients with Non-Small Cell Lung Cancer (NSCLC) who are optimal candidates for immunotherapy is a cornerstone in clinical decision-making. The Tumor Immune Microenvironment (TIME) is intricately linked with both the prognosis of the malignancy and the efficacy of immunotherapeutic interventions. CD8+ T cells, and more specifically, tissue-resident memory CD8+ T cells (CD8+ TRM cells), are postulated to be pivotal in orchestrating the immune system's assault on tumor cells. Nevertheless, the accurate quantification of immune cell infiltration-and by extension, the prediction of immunotherapeutic efficacy-remains a significant scientific frontier. In this study, we introduce a cutting-edge noninvasive radiomic model, grounded on TIME markers (CD3+ T, CD8+ T, and CD8+ TRM cells), to infer the levels of immune cell infiltration in NSCLC patients receiving immune checkpoint inhibitors (ICIs), and ultimately, to predict their response to immunotherapy. Our empirical evidence reveals that patients with substantial CD8+ T cell infiltration experience markedly improved Progress Free Survival (PFS) compared to those with minimal infiltration, asserting the status of the CD8+ T cell as an independent prognosticator of PFS in the context of immunotherapy. Although CD8+ TRM cells demonstrated the greatest predictive accuracy for immunotherapy response, their predictive strength for PFS was marginally surpassed by that of CD8+ T cells. These insights advocate for the application of the proposed noninvasive radiomic model, which utilizes TIME analysis, as a reliable predictor for immunotherapy outcomes and PFS in NSCLC patients.