The germinal center-tertiary lymphoid structure (GC-TLS) after neoadjuvant chemo-immunotherapy for locally advanced lung squamous cell carcinoma can predict the disease progression

Lili Jiang^1*Li Shuang¹ping zhou¹Yan Huang²Min Chen¹Chan Yang¹

• ¹Department of Pathology, West China Hospital, Sichuan University, Chengdu, China
• ²Department of Medical Oncology, Cancer Center, West China Hospital, Sichuan University, Chengdu, Sichuan Province, China

Background: The tertiary lymphoid structures (TLSs) are the anti-tumor immune hubs in the tumor microenvironment. The germinal center (GC) (a marker of maturation) and spatial distribution of TLS may determine the responsiveness of immunotherapy. However, the regulatory mechanism of neoadjuvant chemotherapy (NACT) and combined immunotherapy (NACT-IO) on the dynamic remodeling of TLS has not been elucidated. Methods: The NACT-IO group (72 patients), NACT group (50 patients), UT group (50 patients, un-neoadjuvant therapy) were included. Multiple immunofluorescence (mIF) was used to analyze the difference of microenvironment in paired samples (the same case) pre and post neoadjuvant therapy. To further analyze the effect of treatment on the maturity and spatial distribution pattern of TLS (within/outside tumor bed) in postoperative samples, and to establish a quantitative method of TLS based on hot spot area to evaluate its prognostic value. Results: Spatial heterogeneity analysis that the density of total TLSs (t-TLSs) and GC-positive TLSs (GC-TLSs) in the tumor bed of NACT (p<0.01, p<0.01) group and NACT-IO (p<0.001, p<0.001) group were significantly higher than that outside the tumor bed. Compared with the UT group, NACT and NACT-IO significantly increased the density of t-TLSs (p<0.01, p<0.001) and GC-TLSs (p<0.01, p<0.01) in the tumor bed. In addition, there was an inverted U-shaped correlation between GC-TLS and treatment cycle: the density of GC-TLSs reaches the peak value after receiving two or less (≤ 2) cycles of NACT and NACT-IO, and decreased significantly after receiving more than two (> 2) cycles of NACT and NACT-IO (p<0.05). Multivariate Cox regression model confirmed that low GC-TLS burden (≤2/20×HPF) within tumor bed hotspots (HR=3.99, 95%CI=1.10-14.5, p=0.036) was superior to the traditional prognostic factor of pathological remission in ≤ 2-cycles of NACT-IO subgroup (HR=3.44, 95%CI=1.03-11.47, p=0.044), and became the strongest independent factor for predicting disease free survival (DFS). Conclusions: This study reveals for the first time that NACT and NACT-IO enhance anti-tumor efficacy through multidimensional (abundance, spatial distribution and maturity) dynamic remodeling of TLS, and proposes the short course of ≤ 2 cycles of NACT-IO can maximize the prognostic value of GC-TLS, providing key evidence for optimizing the treatment ' time window '.