Relationship between PD-L1 expression and the number of biopsy specimens in advanced gastric cancer.

481 Background: Programmed cell death ligand 1 (PD-L1) expression shows spatial heterogeneity in gastric cancer. Compared with single biopsies, multiple biopsies may provide reliable PD-L1 expression results. However, the optimal number of biopsy specimens remains unclear. This study aimed to assess the relationship between PD-L1 expression and the number of biopsy specimens in advanced gastric cancer. Methods: We retrospectively analyzed patients with advanced gastric cancer who received first-line chemotherapy at a single institution between December 2021 and June 2024. Those in whom PD-L1 expression was evaluated using tumor-containing biopsy specimens were included. PD-L1 expression was measured using the Dako PD-L1 IHC 28-8 pharmDx assay. PD-L1 positivity was defined as a combined positive score of ≥5. The association between PD-L1 positivity and the number of biopsy specimens was tested using the chi-square or Fisher’s exact test. Results: Of the 183 patients screened, 110 were included. The patient characteristics were as follows: median age, 71 (range: 34–87) years; sex (male/female), 75/35; primary site (gastric/gastroesophageal junction), 102/8; and histological type (diffuse/intestinal/other), 59/34/17. The PD-L1 positivity prevalence was 71.8%. In 97 patients, human epidermal growth factor receptor 2 (HER2) expression was detected in the same specimens as PD-L1, with a HER2-positive rate of 14.4%. The mean numbers of biopsy and tumor-containing biopsy specimens were 5.14 (range: 2–10) and 4.25 (range: 1–10), respectively. The PD-L1 positivity prevalence was significantly higher when the number of biopsy specimens was ≥5 compared with ≤4 (77.5% vs 56.7%, P =0.03; Table). The same trend was observed according to the histological type (diffuse type, 75.0% vs. 46.7%, P =0.04; intestinal type, 79.2% vs. 60.0%, P =0.39). In HER2-negative cases, PD-L1 positivity was significantly more prevalent when the number of biopsy specimens was ≥5 compared with ≤4 (83.6% vs 54.5%, P =0.006; Table). However, there was no such difference in HER2-positive cases (54.4% vs 66.7%, P =1). Conclusions: The data suggest that collecting at least five biopsy specimens increases the likelihood of identifying PD-L1 positivity in advanced gastric cancer, particularly in HER2-negative cases. This approach may improve PD-L1 evaluation accuracy and lead to better treatment decisions. Relationship between the number of biopsy specimens and PD-L1 positivity. Number of biopsy specimens PD-L1 positivity (%) PD-L1 negativity (%) P value Overall 0.03 ≥5 77.5 22.5 ≤4 56.7 43.3 HER2-negative 0.006 ≥5 83.6 16.4 ≤4 54.5 45.5