Abstract Background: Tumor immune microenvironment (TIME), including programmed death-1 ligand-1 (PD-L1) and CD8-positive tumor-infiltrating lymphocytes (TIL), has been suggested as a potential biomarker to predict the efficacy of anti-PD-1 antibody. Postoperative recurrence in patients with non-small cell lung cancer (NSCLC) is treated with therapeutic strategies similar to those for metastatic disease. However, in postoperative recurrence cases, PD-L1 evaluation is generally based on archival resection specimens, because there is little opportunity for a new biopsy of the recurrent site. Therefore, it remains unclear whether TIME of recurrent lesion is similar to TIME of primary lesion at surgery, and the efficacy of anti-PD-1 antibody for postoperative recurrent NSCLC patients based on TIME status at surgery is unknown. Materials and Methods: Cohort A included 59 patients with advanced and 25 patients with postoperative recurrence NSCLC who treated with first-line pembrolizumab monotherapy. PD-L1 expression and CD8-TIL were evaluated immunohistochemically in tumor specimens obtained at diagnosis in the advanced cases and at surgery in the postoperative recurrence cases. Tumor proportion score for PD-L1 and the extent of CD8-TIL were evaluated and classified as high or low. TIME was categorized into 4 subclasses, according to the respective PD-L1 and CD8-TIL, as follows: (class I) high/high; (class II) low/low; (class III) high/low; and (class IV) low/high. Progression-free survival (PFS) of pembrolizumab was compared based on TIME. Cohort B consisted of 50 NSCLC patients who underwent radical resection, and whose tumor specimens were obtained from recurrent site. As in cohort A, TIME was assessed for paired resection and recurrent tumor samples. Results: In cohort A, the median PFS of pembrolizumab for the advanced cases were 25.7 months in class I (n=25), 2.3 months in class II (n=8), 8.0 months in class III (n=20), and 6.5 months in class IV (n=6) (p<0.001), respectively; while those for the postoperative recurrence cases were 9.0 months (n=7), 4.3 months (n=7), 7.6 months (n=6), and 2.5 months (n=5) (p=0.55), respectively. In cohort B, TIME for resected primary lesion were observed in 9 with class I, 20 with class II, 3 with class III and 18 patients with class IV, respectively. The concordance of PD-L1 category (high vs low) between resected primary lesion and recurrent lesion was 88.0%, while that of TIME classification was 54.0%. Conclusion: TIME classification using tumor specimens just before administration of anti-PD-1 antibody was useful predictive biomarker. However, TIME alterations between resection and postoperative recurrence were observed in many cases, suggesting that TIME based on tumor specimens at surgery may not be useful for predicting the efficacy of anti-PD-1 antibody for postoperative recurrent NSCLC. Citation Format: Yoshiya Matsumoto, Tetsuya Watanabe, Megumi Mizutani, Akira Sugimoto, Hiroaki Nagamine, Yoko Tani, Kenji Sawa, Takuma Tsukioka, Hiroyasu Kaneda, Shigeki Mitsuoka, Noritoshi Nishiyama, Tomoya Kawaguchi. Alteration of tumor immune microenvironment after recurrence and its impact on the efficacy of anti-PD-1 antibody in resected non-small cell lung cancer patients [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2024; Part 1 (Regular Abstracts); 2024 Apr 5-10; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2024;84(6_Suppl):Abstract nr 2519.