First-line (1L) nivolumab (NIVO) plus chemotherapy (chemo) vs chemo in patients (pts) with advanced gastric cancer, gastroesophageal junction cancer, and esophageal adenocarcinoma (GC/GEJC/EAC): CheckMate 649 Chinese subgroup analysis with 3-year follow-up.

318 Background: NIVO + chemo demonstrated clinically meaningful improvement in overall survival (OS) and an acceptable safety profile vs chemo in previously untreated Chinese pts from CheckMate 649, consistent with the overall study population with advanced GC/GEJC/EAC. 1L NIVO + chemo is approved for advanced GC/GEJC/EAC in multiple countries, including China. We report 4-yr results of NIVO + chemo vs chemo in Chinese pts from CheckMate 649. Methods: Adults with previously untreated, unresectable advanced or metastatic, non-HER2+ GC/GEJC/EAC were enrolled regardless of programmed death ligand 1 (PD-L1) expression. Randomized pts received NIVO (360 mg Q3W or 240 mg Q2W) + chemo (XELOX Q3W or FOLFOX Q2W), NIVO + ipilimumab, or chemo. Dual primary endpoints for NIVO + chemo vs chemo were OS and progression-free survival (PFS) by blinded independent central review (BICR) in pts with PD-L1 combined positive score (CPS) ≥ 5. Results: 208 Chinese pts were randomized to NIVO + chemo or chemo. At 49-month (mo) minimum follow-up, NIVO + chemo continued to demonstrate OS and PFS benefit vs chemo in pts with PD-L1 CPS ≥ 5 and in all randomized pts (Table). The 4-yr OS rate was 25% with NIVO + chemo vs 11% with chemo in pts with PD-L1 CPS ≥ 5 and 21% vs 9% in all randomized pts. Objective response rate (ORR; 95% CI) per BICR in pts with PD-L1 CPS ≥ 5 who had measurable lesions at baseline was 68% (56–79) with NIVO + chemo vs 48% (36–60) with chemo; corresponding ORR in all randomized pts was 66% (55–76) and 45% (35–56). Responses were more durable with NIVO + chemo vs chemo both in pts with PD-L1 CPS ≥ 5 (median duration of response [mDOR; 95% CI] 12.5 mo [7.2–23.4] vs 6.9 mo [3.9–8.5]) and in all randomized pts (mDOR [95% CI] 12.5 mo [7.2–17.7] vs 5.6 mo [4.4–8.3]). No new safety signals were identified (Table). Conclusions: After 4 yrs of follow-up, NIVO + chemo continued to demonstrate clinically meaningful survival benefit and more durable responses vs chemo in Chinese pts, with an acceptable safety profile. These results are consistent with previous reports and with the overall study population with advanced GC/GEJC/EAC and further support NIVO + chemo as a standard 1L treatment option for Chinese pts. Clinical trial information: NCT02872116 . [Table: see text]

392 Background: NIVO + chemo demonstrated clinically meaningful survival benefit and an acceptable safety profile vs chemo in previously untreated Chinese pts with advanced GC/GEJC/EAC from CheckMate 649, consistent with the overall study population. 1L NIVO + chemo is currently approved for pts with advanced non-HER2+ GC/GEJC/EAC in China and other countries. We report 5-y results of NIVO + chemo vs chemo in Chinese pts from CheckMate 649. Methods: Adults with previously untreated, unresectable advanced or metastatic, non-HER2+ GC/GEJC/EAC were enrolled regardless of programmed death ligand 1 (PD-L1) expression. Randomized pts received NIVO + chemo, NIVO + ipilimumab, or chemo. Dual primary endpoints for NIVO + chemo vs chemo were overall survival (OS) and progression-free survival (PFS) by blinded independent central review (BICR) in pts with PD-L1 combined positive score (CPS) ≥ 5. Results: 208 Chinese pts were randomized to NIVO + chemo or chemo. At 61-month (mo) minimum follow-up, NIVO + chemo continued to demonstrate OS and PFS benefit vs chemo in pts with PD-L1 CPS ≥ 5 and all randomized pts (Table). 5-y OS rate was 24% with NIVO + chemo vs 8% with chemo in pts with PD-L1 CPS ≥ 5 and 20% vs 7% in all randomized pts. Objective response rate (ORR) was higher and responses were more durable with NIVO + chemo vs chemo in pts with PD-L1 CPS ≥ 5 and all randomized pts (Table). No new safety signals were identified with longer follow-up. Conclusions: NIVO + chemo continued to demonstrate clinically meaningful long-term survival benefit, more durable responses, and acceptable safety vs chemo in Chinese pts after 5 y of follow-up, consistent with earlier reports and with the overall study population of pts with advanced non-HER2+ GC/GEJC/EAC. These results further support NIVO + chemo as a standard 1L treatment option for Chinese pts. Clinical trial information: NCT02872116 . Efficacy PD-L1 CPS ≥ 5 All randomized NIVO + chemo(n = 75) Chemo (n = 81) NIVO + chemo (n = 99) Chemo (n = 109) mOS (95% CI), mo 15.5(11.9–21.1) 9.6(8.0–12.1) 14.3(11.5–16.5) 10.3(8.1–12.1) HR (95% CI) 0.57 (0.40–0.82) 0.63 (0.46–0.85) mPFS a (95% CI), mo 8.5(6.0–14.0) 4.3(4.1–6.5) 8.3(6.2–12.4) 5.6(4.2–6.8) HR (95% CI) 0.51 (0.34–0.76) 0.57 (0.41–0.80) ORR a,b (95% CI), % 68 (56–79) 48 (36–60) 66 (55–76) 45 (35–56) mDOR a,c (95% CI), mo 12.5 (7.2–23.4) 6.9 (3.9–8.5) 12.5 (7.2–17.7) 5.6 (4.4–8.3) a Per BICR. b In pts with measurable target lesions at baseline. c In responders. DOR, duration of response; m, median.

Background: Overall survival (OS) for advanced or metastatic HER2-negative GC/GEJC with standard 1L chemo remains poor (median OS < 1 year). CheckMate 649 is the largest randomized, global phase 3 study of 1L programmed death (PD)-1 inhibitor-based therapies in GC/GEJC/EAC (NCT02872116). Results from this study demonstrated superior OS, along with progression-free survival (PFS) benefit and an acceptable safety profile with 1L NIVO + chemo vs chemo alone (Moehler et al. Ann Oncol 2020). We present the pre-planned subgroup analysis from CheckMate 649 in Chinese pts.Methods: Adults with previously untreated, unresectable advanced or metastatic GC/GEJC/EAC were enrolled regardless of PD ligand 1 (PD-L1) expression. Pts with known HER2-positive status were excluded. Pts were randomized to receive NIVO (360 mg Q3W or 240 mg Q2W) + chemo (XELOX Q3W or FOLFOX Q2W), NIVO + ipilimumab, or chemo. Dual primary endpoints for NIVO + chemo vs chemo were OS and PFS by blinded central review in pts with PD-L1 combined positive score (CPS) ≥ 5.Results: 208 Chinese pts were concurrently randomized to NIVO + chemo or chemo, including 156 pts (75%) with PD-L1 CPS ≥ 5. Median age was 60 years; 88% had GC; 12% had GEJC; no pts had EAC. At minimum follow-up of 12 months (mo), NIVO + chemo demonstrated clinically meaningful improvement in OS and PFS vs chemo in pts with PD-L1 CPS ≥ 5 (OS, HR 0.54 [95% CI 0.36-0.79]; PFS, HR 0.52 [0.34-0.77]; Table), consistent with the overall study population. OS benefit was also observed in pts with PD-L1 CPS ≥ 1 and the all-randomized population (Table). No new safety signals were identified. Conclusions: NIVO + chemo demonstrated a clinically meaningful improvement in OS and PFS and an acceptable safety profile vs chemo alone in previously untreated Chinese pts, consistent with the overall study population with advanced GC/GEJC/EAC from CheckMate 649. TableEfficacyNIVO + chemoChemoPD-L1 CPS ≥ 5N = 75N = 81Median OS (95% CI), mo15.5 (11.9-25.5)9.6 (8.0-12.1)HR (95% CI)0.54 (0.36-0.79)Median PFS (95% CI), mo8.5 (5.9-12.4)4.3 (4.1-6.5)HR (95% CI)0.52 (0.34-0.77)ORR (95% CI), %63 (51-74)43 (32-55)PD-L1 CPS ≥ 1N = 89N = 94Median OS (95% CI), mo14.3 (11.5-17.5)9.9 (8.1-12.1)HR (95% CI)0.62 (0.43-0.87)All randomizedN = 99N = 109Median OS (95% CI), mo14.3 (11.5-17.5)10.3 (8.1-12.1)HR (95% CI)0.61 (0.44-0.85)Safety: Treatment-related adverse events, n (%)All treatedN = 99N = 106Any grade98 (99)100 (94)Grade 3-464 (65)53 (50)Leading to discontinuation50 (51)27 (25)Deaths1 (1)1 (<1) Citation Format: Lin Shen, Yuxian Bai, Xiaoyan Lin, Wei Li, Jufeng Wang, Xiaochun Zhang, Hongming Pan, Chunmei Bai, Li Bai, Ying Cheng, Jingdong Zhang, Haijun Zhong, Yi Ba, Wenwei Hu, Ruihua Xu, Weijian Guo, Shukui Qin, Nong Yang, Jianwei Lu, Kohei Shitara, Ming Lei, Mingshun Li, Nicole Bao, Tian Chen, Tianshu Liu. First-Line (1L) nivolumab (NIVO) plus chemotherapy (chemo) versus chemo in patients (pts) with advanced gastric cancer/gastroesophageal junction cancer/esophageal adenocarcinoma (GC/GEJC/EAC): CheckMate 649 Chinese subgroup analysis [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr CT184.

398 Background: At 4-y follow-up, 1L NIVO + chemo continued to demonstrate clinically meaningful overall survival (OS) and progression-free survival (PFS) benefit vs chemo with acceptable safety in patients (pts) with advanced non-HER2+ GC/GEJC/EAC from CheckMate 649. We report efficacy and safety results of NIVO + chemo vs chemo at 5-y follow-up. Methods: Adults with previously untreated, unresectable, advanced or metastatic, non-HER2+ GC/GEJC/EAC were enrolled, regardless of programmed death ligand 1 (PD-L1) expression. Pts were randomized to NIVO (360 mg Q3W or 240 mg Q2W) + chemo (XELOX Q3W or FOLFOX Q2W), NIVO + ipilimumab, or chemo. Primary endpoints for NIVO + chemo vs chemo were OS and PFS by blinded independent central review (BICR) in pts with PD-L1 combined positive score (CPS) ≥ 5. Results: Pts were randomized to NIVO + chemo (n = 789) or chemo (n = 792). NIVO + chemo continued to show OS and PFS benefit vs chemo in pts with PD-L1 CPS ≥ 5, pts with PD-L1 CPS ≥ 1, and all randomized pts at 60-month (mo) minimum follow-up (Table). OS rates at 60-mo were higher with NIVO + chemo vs chemo in pts with PD-L1 CPS ≥ 5, pts with PD-L1 CPS ≥ 1, and all randomized pts (Table), and OS benefit with NIVO + chemo continued to be observed in most prespecified subgroups. Objective response rates (ORRs) were higher and responses were more durable with NIVO + chemo vs chemo in pts with PD-L1 CPS ≥ 5, pts with PD-L1 CPS ≥ 1, and all randomized pts (Table). No new safety signals were identified. Conclusions: These results represent the first report of 5-y follow-up for anti–PD-1 + chemo combination therapy in GC/GEJC/EAC to our knowledge. NIVO + chemo continued to provide sustained long-term survival vs chemo with an acceptable safety profile after 5 y of follow-up. These data continue to support the use of NIVO + chemo as a standard 1L treatment for advanced GC/GEJC/EAC. Clinical trial information: NCT02872116 . Efficacy PD-L1 CPS ≥ 5 PD-L1 CPS ≥ 1 All randomized NIVO + chemo(n = 473) Chemo (n = 482) NIVO + chemo (n = 641) Chemo (n = 656) NIVO + chemo (n = 789) Chemo (n = 792) mOS (95% CI), mo 14.4 (13.1–16.2) 11.1 (10.1–12.1) 13.8 (12.4–14.8) 11.4 (10.7–12.3) 13.7 (12.4–14.5) 11.6 (10.9–12.5) HR (95% CI) 0.71 (0.61–0.81) 0.76 (0.67–0.85) 0.79 (0.71–0.88) 60-mo OS rate (95% CI), % 16 (12–19) 6 (4–9) 13 (11–16) 5 (4–7) 12 (10–14) 6 (4–8) mPFS a (95% CI), mo 8.3 (7.0–9.4) 6.1 (5.6–6.9) 7.5 (7.0–8.5) 6.9 (6.2–7.1) 7.8 (7.1–8.6) 6.9 (6.7–7.2) HR (95% CI) 0.71 (0.61–0.82) 0.77 (0.68–0.87) 0.79 (0.71–0.89) ORR a,b (95% CI), % 60 (55–65) 45 (40–50) 60 (55–64) 46 (42–51) 58 (54–62) 46 (42–50) mDOR a,c (95% CI), mo 9.6 (8.3–12.4) 7.0 (5.7–8.0) 8.6 (7.9–10.5) 6.9 (5.8–7.6) 8.5 (7.7–9.9) 6.9 (5.9–7.6) a Per BICR. b In pts with measurable target lesions at baseline. c In all measurable responders. DOR, duration of response; m, median.

291 Background: NIVO + chemo demonstrated superior overall survival (OS) and clinically meaningful progression-free survival (PFS) benefit vs chemo and an acceptable safety profile in previously untreated patients (pts) with advanced GC/GEJC/EAC, leading to approvals in multiple countries including the US. NIVO + chemo continued to demonstrate clinically meaningful improvement in efficacy after 2 years of follow-up. We present efficacy and safety analyses from NIVO + chemo vs chemo from the 3-year follow-up of CheckMate 649. Methods: Adults with previously untreated, unresectable advanced, or metastatic GC/GEJC/EAC were enrolled, regardless of programmed death ligand 1 (PD-L1) expression. Pts with known HER2-positive status were excluded. Randomized pts received NIVO (360 mg Q3W or 240 mg Q2W) + chemo (XELOX Q3W or FOLFOX Q2W), NIVO + ipilimumab, or chemo. Dual primary endpoints for NIVO + chemo vs chemo were OS and PFS by blinded independent central review (BICR) in pts with PD-L1 combined positive score (CPS) ≥ 5. Results: 581 pts were concurrently randomized to NIVO + chemo or chemo. With 36-month (mo) minimum follow-up, NIVO + chemo continued to demonstrate OS and PFS benefit vs chemo in pts with PD-L1 CPS ≥ 5 and all randomized pts. The objective response rate (ORR) per BICR in pts with PD-L1 CPS ≥ 5 who had measurable lesions at baseline was 60% (95% CI 55–65) with NIVO + chemo vs 45% (95% CI 40–50) with chemo; in all randomized pts, ORR per BICR was 58% (95% CI 54–62) with NIVO + chemo vs 46% (95% CI 42–50) with chemo. Responses were more durable with NIVO + chemo vs chemo in pts with PD-L1 CPS ≥ 5 (median [m] duration of response [mDOR] 9.6 mo [95% CI 8.2–12.4] vs 7.0 mo [95% CI 5.6–7.9], respectively) and in all randomized pts (mDOR 8.5 mo [95% CI 7.7–9.9] vs 6.9 mo [95% CI 5.8–7.2], respectively). OS benefit with NIVO + chemo was observed across most prespecified subgroups. No new safety signals were identified. A summary of treatment-related adverse events (TRAEs) is shown here. Conclusions: After 3 years of follow-up, NIVO + chemo continued to demonstrate clinically meaningful long-term survival benefit with an acceptable safety profile, further supporting its use as a standard 1L treatment in previously untreated pts with advanced GC/GEJC/EAC. Clinical trial information: NCT02872116 . [Table: see text]

306 Background: NIVO + chemo demonstrated superior overall survival (OS) and clinically meaningful progression-free survival (PFS) benefit vs chemo and an acceptable safety profile in previously untreated, advanced non-HER2+ GC/GEJC/EAC, leading to approvals in many countries. NIVO + chemo continued to demonstrate clinically meaningful improvement in efficacy at 2 and 3-yr follow-ups. We present 4-yr follow-up results for NIVO + chemo vs chemo from CheckMate 649. Methods: Adults with previously untreated, unresectable advanced, or metastatic GC/GEJC/EAC were enrolled, regardless of programmed death ligand 1 (PD-L1) expression. HER2+ patients (pts) were excluded. Randomized pts received NIVO (360 mg Q3W or 240 mg Q2W) + chemo (XELOX Q3W or FOLFOX Q2W), NIVO + ipilimumab, or chemo. Dual primary endpoints for NIVO + chemo vs chemo were OS and PFS by blinded independent central review (BICR) in pts with PD-L1 combined positive score (CPS) ≥ 5. Results: 1581 pts were randomized to NIVO + chemo or chemo. At the 48-month (mo) minimum follow-up, NIVO + chemo continued to demonstrate OS and PFS benefit vs chemo in pts with PD-L1 CPS ≥ 5 and all randomized pts (Table) OS benefit with NIVO + chemo was observed in most prespecified subgroups. Objective response rates (ORR) were higher and responses were more durable with NIVO + chemo vs chemo in pts with PD-L1 CPS ≥ 5 and all randomized pts (Table). In the exploratory analysis of OS by response at the 18-week landmark timepoint, there were numerically more pts achieving response with NIVO + chemo vs chemo, and median OS (95% CI) with NIVO + chemo was numerically longer in responders vs non-responders both in pts with PD-L1 CPS ≥ 5 (20.5 [17.5–25.0] vs 14.0 [11.6–15.7]) and all randomized pts (19.4 [17.5–21.7] vs 13.1 [11.6–14.4]). No new safety signals were identified, consistent with the 3-yr follow-up. Conclusions: NIVO + chemo is the first PD-1 inhibitor/chemo combination to demonstrate long-term efficacy and acceptable safety after 4 yrs of follow up in previously untreated advanced GC/GEJC/EAC. These results are consistent with earlier follow-ups, further supporting NIVO + chemo as a standard 1L treatment in these pts. Clinical trial information: NCT02872116 .[Table: see text]

4040 Background: NIVO + chemo demonstrated superior overall survival (OS) and clinically meaningful progression-free survival (PFS) benefit vs chemo and an acceptable safety profile in previously untreated, advanced non-HER2+ GC/GEJC/EAC, leading to approvals in many countries. NIVO + chemo continued to demonstrate clinically meaningful improvement in efficacy at 2 and 3-yr follow-ups. We present 4-yr follow-up results for NIVO + chemo vs chemo from CheckMate 649. Methods: Adults with previously untreated, unresectable advanced, or metastatic GC/GEJC/EAC were enrolled, regardless of programmed death ligand 1 (PD-L1) expression. HER2+ patients (pts) were excluded. Randomized pts received NIVO (360 mg Q3W or 240 mg Q2W) + chemo (XELOX Q3W or FOLFOX Q2W), NIVO + ipilimumab, or chemo. Dual primary endpoints for NIVO + chemo vs chemo were OS and PFS by blinded independent central review (BICR) in pts with PD-L1 combined positive score (CPS) ≥ 5. Results: 1581 pts were randomized to NIVO + chemo or chemo. At the 48-month (mo) minimum follow-up, NIVO + chemo continued to demonstrate OS and PFS benefit vs chemo in pts with PD-L1 CPS ≥ 5 and all randomized pts (Table). OS benefit with NIVO + chemo was observed in most prespecified subgroups. Objective response rates (ORR) were higher and responses were more durable with NIVO + chemo vs chemo in pts with PD-L1 CPS ≥ 5 and all randomized pts (Table). In the exploratory analysis of OS by response at the 18-week landmark timepoint, there were numerically more pts achieving response with NIVO + chemo vs chemo, and median OS (95% CI) with NIVO + chemo was numerically longer in responders vs non-responders both in pts with PD-L1 CPS ≥ 5 (20.5 [17.5–25.0] vs 14.0 [11.6–15.7]) and all randomized pts (19.4 [17.5–21.7] vs 13.1 [11.6–14.4]). No new safety signals were identified, consistent with the 3-yr follow-up. Conclusions: NIVO + chemo is the first PD-1 inhibitor/chemo combination to demonstrate long-term efficacy and acceptable safety after 4 yrs of follow up in previously untreated advanced GC/GEJC/EAC. These results are consistent with earlier follow-ups, further supporting NIVO + chemo as a standard 1L treatment in these pts. Clinical trial information: NCT02872116 . [Table: see text]

4025 Background: NIVO + chemo demonstrated superior overall survival (OS) and clinically meaningful progression-free survival (PFS) benefit vs chemo and acceptable safety in previously untreated patients (pts) with advanced GC/GEJC/EAC, leading to approvals in multiple countries including the US. NIVO + chemo continued to demonstrate clinically meaningful improvement in efficacy after 2 years of follow-up. We present efficacy and safety analyses from NIVO + chemo vs chemo from the 3-year follow-up of CheckMate 649. Methods: Adults with previously untreated, unresectable advanced, or metastatic GC/GEJC/EAC were enrolled, regardless of programmed death ligand 1 (PD-L1) expression, excluding pts with known HER2-positive status. Pts were randomized to NIVO (360 mg Q3W or 240 mg Q2W) + chemo (XELOX Q3W or FOLFOX Q2W), NIVO + ipilimumab, or chemo. Primary endpoints for NIVO + chemo vs chemo were OS and PFS by blinded independent central review (BICR) in pts with PD-L1 combined positive score (CPS) ≥ 5. Results: 1581 pts were concurrently randomized to NIVO + chemo or chemo. With 36-month (mo) minimum follow-up, NIVO + chemo continued to demonstrate OS and PFS benefit vs chemo in pts with PD-L1 CPS ≥ 5 and all randomized pts. The objective response rate (ORR) per BICR in pts with PD-L1 CPS ≥ 5 who had measurable lesions at baseline was 60% (95% CI 55–65) with NIVO + chemo vs 45% (95% CI 40–50) with chemo; in all randomized pts, ORR per BICR was 58% (95% CI 54–62) with NIVO + chemo vs 46% (95% CI 42–50) with chemo. Responses were more durable with NIVO + chemo vs chemo in pts with PD-L1 CPS ≥ 5 (median [m] duration of response [mDOR] 9.6 mo [95% CI 8.2–12.4] vs 7.0 mo [95% CI 5.6–7.9], respectively) and in all randomized pts (mDOR 8.5 mo [95% CI 7.7–9.9] vs 6.9 mo [95% CI 5.8–7.2], respectively). OS benefit with NIVO + chemo was observed across most prespecified subgroups. No new safety signals were identified. A summary of treatment-related adverse events (TRAEs) is shown in the Table. Additional analyses will be presented. Conclusions: After 3 years of follow-up, NIVO + chemo continued to demonstrate clinically meaningful long-term survival benefit with acceptable safety, further supporting its use as a standard 1L treatment in previously untreated pts with advanced GC/GEJC/EAC. Clinical trial information: NCT02872116 . [Table: see text]

How low can you go? PD-L1 expression as a biomarker in trials of cancer immunotherapy

LBA4001 Background: NIVO demonstrated superior overall survival (OS) vs chemo in previously treated patients (pts) with ESCC (ATTRACTION-3). We report OS and progression-free survival (PFS) from CheckMate 648, the first global phase III study to evaluate both an immuno-oncology (I-O)/chemo combination and an I-O/I-O combination in advanced ESCC. Methods: Adults with previously untreated, unresectable advanced, recurrent or metastatic ESCC were enrolled regardless of tumor cell PD-L1 expression. Pts were randomized to NIVO (240 mg Q2W) + chemo (fluorouracil + cisplatin Q4W), NIVO (3 mg/kg Q2W) + IPI (1 mg/kg Q6W), or chemo alone. Primary endpoints for both comparisons were OS and PFS per blinded independent central review (BICR) in pts with tumor cell PD-L1 ≥ 1%. Hierarchically tested secondary endpoints included OS and PFS in all randomized pts. Results: 970 pts were randomized to NIVO + chemo, NIVO + IPI, and chemo arms (49% with tumor cell PD-L1 ≥ 1%). With 13 months (mo) minimum follow-up, NIVO + chemo and NIVO + IPI led to statistically significant improvement in OS vs chemo in pts with tumor cell PD-L1 ≥ 1% and all randomized pts (Table). Statistically significant PFS benefit was also observed for NIVO + chemo vs chemo (HR 0.65 [98.5% CI 0.46–0.92]; P = 0.0023) in pts with tumor cell PD-L1 ≥ 1%. PFS in NIVO + IPI vs chemo in pts with tumor cell PD-L1 ≥ 1% did not meet the prespecified boundary for significance. The objective response rate (per BICR) was 53% (NIVO + chemo), 35% (NIVO + IPI), and 20% (chemo) in pts with tumor cell PD-L1 ≥ 1% and in all randomized pts was 47%, 28%, and 27%, respectively; longer median (95% CI) duration of response was observed vs chemo for pts with tumor cell PD-L1 ≥ 1%: 8.4 (6.9–12.4), 11.8 (7.1–27.4), and 5.7 (4.4–8.7) mo and for all randomized pts: 8.2 (6.9–9.7), 11.1 (8.3–14.0), and 7.1 (5.7–8.2) mo, respectively. No new safety signals were identified (Table). Conclusions: NIVO plus chemo and NIVO plus IPI both demonstrated superior OS vs chemo, along with durable objective responses and acceptable safety, in pts with advanced ESCC, and each represents a potential new 1L treatment option. Clinical trial information: NCT03143153. [Table: see text]

290 Background: NIVO + chemo and NIVO + IPI demonstrated superior overall survival (OS) vs chemo in CheckMate 648 (NCT03143153), leading to approvals in the US, EU, Japan, and other countries. We report longer follow-up results. Methods: Adults with previously untreated, unresectable advanced, recurrent, or metastatic ESCC were randomized to NIVO (240 mg Q2W) + chemo (fluorouracil + cisplatin Q4W), NIVO (3 mg/kg Q2W) + IPI (1 mg/kg Q6W) or chemo. Primary endpoints were OS and progression-free survival (PFS) per blinded independent central review (BICR). Hierarchical testing was done first in patients (pts) with tumor cell programmed death ligand 1 (PD-L1) ≥ 1%, then in all randomized pts. Results: 970 pts were randomized to NIVO + chemo, NIVO + IPI, or chemo. With 29-mo minimum follow-up, NIVO + chemo and NIVO + IPI continued to show improvement in OS vs chemo, including higher 24-mo OS rates, in pts with tumor cell PD-L1 ≥ 1% and all randomized pts. Responses were more durable and a larger proportion of responders had a duration of response (DOR) ≥ 24 mo with NIVO + chemo and NIVO + IPI vs chemo in pts with tumor cell PD-L1 ≥ 1% (22%, 36%, 13%, respectively) and all randomized pts (21%, 29%, 13%). Additional efficacy data by PD-L1 status will be presented. Any-grade treatment-related adverse events (TRAEs) occurred in 96% (grade 3/4, 49%) of pts with NIVO + chemo, 80% (33%) with NIVO + IPI, and 90% (36%) with chemo. Any-grade TRAEs leading to discontinuation occurred in 35% of pts with NIVO + chemo, 19% with NIVO + IPI, and 21% with chemo. Treatment-related deaths occurred in 2% of pts in each arm. Conclusions: NIVO + chemo and NIVO + IPI continued to demonstrate clinically meaningful survival benefit vs chemo, durable objective responses, and acceptable safety profiles with longer follow-up. This further supports each regimen as a new 1L treatment option for advanced ESCC. Clinical trial information: NCT03143153 . [Table: see text]

TPS468 Background: An unmet medical need remains in patients (pts) with HER2+ gastric or GEJ cancer. HER2 is a validated target in up to 20% of pts with gastric or GEJ cancer. Results of the KEYNOTE-811 trial demonstrated that pembro + trastuzumab and chemo improved progression-free survival (PFS) and overall survival (OS) versus placebo + trastuzumab and chemo for first-line treatment of pts with HER2+ gastric or GEJ cancer with a programmed death ligand 1 (PD-L1) combined positive score (CPS) ≥1 (Janjigian Y et al. N Engl J Med. 2024;391:1360). In the DESTINY-Gastric03 trial, first-line combinations involving T-DXd, a HER2-directed antibody-drug conjugate, and fluoropyrimidine (5-fluorouracil [5-FU] or capecitabine [CAPE]) ± pembro showed acceptable safety and encouraging efficacy in pts with HER2+ gastric or GEJ cancer irrespective of tumor PD-L1 CPS (Janjigian Y et al. Ann Oncol . 2024;35:S878). Building on this evidence, the phase 3 DESTINY-Gastric05 trial aims to evaluate a potentially improved platinum-free treatment approach for all pts with HER2+ gastric or GEJ cancer. Methods: DESTINY-Gastric05 (NCT06731478) is an open-label, randomized, multicenter, phase 3 global trial designed to evaluate the efficacy and safety of T-DXd in combination with 5-FU (or CAPE) + pembro versus standard-of-care chemo with trastuzumab + pembro as first-line treatment in pts with unresectable, locally advanced or metastatic centrally confirmed HER2+ (immunohistochemistry [IHC] 3+ or IHC 2+/in situ hybridization+) gastric or GEJ cancer with PD-L1 CPS ≥1 (main cohort). Pts must have ≥1 measurable lesion by RECIST v1.1, a left ventricular ejection fraction ≥50%, and an Eastern Cooperative Oncology Group performance status of 0 or 1. In the main cohort, approximately 576 pts will be randomly assigned in a 1:1 ratio to receive: T-DXd 5.4 mg/kg + 5-FU or CAPE + pembro (arm M1); or trastuzumab + platinum-based chemo (either cisplatin + 5-FU or oxaliplatin + CAPE) + pembro (arm M2). The primary efficacy endpoint is PFS based on blinded independent central review (BICR), and the key secondary endpoint is OS. Other secondary endpoints include overall response rate, duration of response, and time to response per RECIST v1.1 assessed by BICR and investigator. Safety and tolerability will also be assessed. A separate exploratory cohort (approximately 150 pts) will evaluate the efficacy and safety of T-DXd in combination with 5-FU or CAPE versus trastuzumab plus standard-of-care chemo in pts with PD-L1 CPS <1. Recruitment started on February 27, 2025; as of August 6, 2025, 17 of a planned 726 pts had been enrolled. Clinical trial information: NCT06731478 .

4035 Background: NIVO + chemo and NIVO + IPI demonstrated significant overall survival (OS) benefit vs chemo in previously untreated patients (pts) with advanced ESCC in the phase 3 CheckMate 648 study. We report expanded results from the primary analysis with 13-month (mo) minimum follow-up. Methods: Pts with previously untreated, unresectable advanced, recurrent, or metastatic ESCC were randomized to NIVO (240 mg Q2W) + chemo (fluorouracil + cisplatin Q4W), NIVO (3 mg/kg Q2W) + IPI (1 mg/kg Q6W), or chemo. Primary endpoints were OS and progression-free survival (PFS) per blinded independent central review (BICR) in pts with tumor cell programmed death ligand 1 (PD-L1) ≥ 1%. Secondary endpoints planned for hierarchical testing included OS, PFS, and objective response rate (ORR) per BICR in all randomized pts. Duration of response (DOR) per BICR and PFS2 per investigator (time from randomization to progression after subsequent systemic therapy, initiation of second subsequent systemic therapy, or death) were exploratory endpoints. Results: Among all pts randomized to NIVO + chemo (n = 321), NIVO + IPI (n = 325), or chemo (n = 324), PFS2 favored NIVO + chemo (HR 0.64, 95% CI 0.54–0.77) and NIVO + IPI (HR 0.74, 95% CI 0.62–0.88) vs chemo. ORR (95% CI) was 47% (42–53), 28% (23–33), and 27% (22–32), respectively. More responders with NIVO + chemo or NIVO + IPI vs chemo had prolonged DOR (≥12 mo; 39%, 48%, and 23%, respectively). Efficacy data by tumor cell PD-L1 and PD-L1 combined positive score will be presented. Grade 3/4 treatment-related adverse events with potential immunologic etiology (select TRAEs) occurred in ≤ 6% of pts with NIVO + chemo and NIVO + IPI, and non-endocrine select TRAEs resolved in 57%–95% of pts across organ categories (Table). Conclusions: NIVO + chemo and NIVO + IPI demonstrated favorable PFS2 and a higher proportion of pts with prolonged DOR vs chemo, as well as acceptable safety profiles. These results provide further support for each regimen as a new potential 1L standard of care in advanced ESCC. Clinical trial information: NCT03143153. [Table: see text]

491 Background: CheckMate 274 demonstrated a significant improvement in disease-free survival (DFS) with nivolumab (NIVO) versus placebo (PBO) both in the intent-to-treat population (hazard ratio [HR], 0.70; 98.22% confidence interval [CI], 0.55–0.90; P < 0.001) and in patients (pts) with tumor programmed death ligand 1 (PD-L1) expression ≥ 1% assessed by the tumor proportion score (TPS) (HR, 0.55; 98.72% CI, 0.35–0.85; P < 0.001). An exploratory subgroup analysis showed a trend toward a DFS benefit with NIVO in pts with TPS < 1% (0.82; 95% CI, 0.63–1.06). To further characterize the relationship between PD-L1 expression and NIVO efficacy, we report an analysis of DFS based on PD-L1 expression in both tumor and immune cells using the combined positive score (CPS). Methods: CheckMate 274 is a phase 3, randomized, double-blind, multicenter trial of NIVO versus PBO in pts with high-risk muscle-invasive urothelial carcinoma after radical surgery. Pts were randomized 1:1 to NIVO 240 mg or PBO every 2 weeks intravenously for 1 year of adjuvant treatment. The primary endpoints of the study are DFS in the intent-to-treat population and in pts with TPS ≥ 1%. The Dako PD-L1 IHC 28-8 pharmDx assay was used to evaluate TPS. CPS was determined retrospectively from previously stained immunohistochemistry slides using the CPS algorithm. CPS was calculated as the number of both PD-L1 positive tumor and immune cells divided by the number of viable tumor cells in the evaluable tumor area, multiplied by 100; TPS was similarly calculated with the number of PD-L1 positive tumor cells as the numerator. This analysis only included pts with both quantifiable CPS and TPS. Results: Of the 629 pts with quantifiable TPS and CPS, 249 (40%) had TPS ≥ 1% (NIVO, n = 124; PBO, n = 125), 380 (60%) had TPS < 1% (NIVO, n = 191; PBO, n = 189), 557 (89%) had CPS ≥ 1 (NIVO, n = 281; PBO, n = 276), and 72 (11%) had CPS < 1 (NIVO, n = 34; PBO, n = 38). Within TPS < 1% pts, 81% (n = 309) had CPS ≥ 1. The number of pts and the DFS outcomes in pts with TPS ≥ 1% and CPS ≥ 1 are shown in the Table. In pts with TPS < 1% who also had CPS ≥ 1, median DFS (95% CI) was 19.2 (15.6–33.4) months with NIVO versus 10.1 (8.2–19.4) months with PBO. The HR for NIVO versus PBO in these pts was 0.73 (95% CI, 0.54–0.99). Conclusions: This exploratory analysis of PD-L1 expression by CPS showed a higher proportion of pts with CPS ≥ 1 than TPS ≥ 1%, and that most pts with TPS < 1% had CPS ≥ 1. In the CPS ≥ 1 subgroup, median DFS with NIVO was more than double that with placebo. These results support the conclusion that pts with TPS < 1% also benefit from adjuvant NIVO. Clinical trial information: NCT02632409. [Table: see text]

Background: Immune checkpoint inhibitors (ICIs) have become the standard of care in a range of tumors. However, a large proportion of patients treated with ICIs have treatment-refractory/resistant tumors requiring innovative approaches. Lymphocyte-activation gene 3 (LAG-3) is an immune checkpoint receptor that inhibits effector T-cell function. Preclinical data suggest that simultaneous activation of the LAG-3 and programmed death-1 pathways results in greater T-cell exhaustion than either pathway alone, leading to impaired antitumor immune response and increased tumor growth. Therefore, LAG-3 may be a promising new target for cancer immunotherapy. The objective of this study was to assess the distribution of LAG-3 expression on lymphocytes in melanoma, gastric cancer (GC), and gastroesophageal junction cancer (GEJC) samples using an immunohistochemistry (IHC) assay. Methods: Formalin-fixed, paraffin-embedded samples from resected melanoma, GC, or GEJC ≤ 5 years old were procured for analysis. The LAG-3 IHC assay was developed and performed by LabCorp (Research Triangle Park, NC) using the anti-LAG-3 17B4 mouse antibody clone on a Leica Bond III platform. LAG-3 expression was assessed as the percentage of positive immune cells (ICs) within the invasive margin, tumor, and stroma. Programmed death ligand 1 (PD-L1) expression on tumor cells (TCs) and ICs was evaluated on the same samples using the Dako PD-L1 IHC 28-8 pharmDx assay to determine the percentage of PD-L1-expressing TCs and the combined positive score (CPS). Results: A total of 102 melanoma samples, 256 GC samples, and 84 GEJC samples were included in this study. LAG-3 expression was detected in the majority of samples across all tumor types, with higher expression in GEJC than in melanoma or GC. LAG-3 prevalence was similar across subgroups based on disease stage (melanoma, GC) and grade (GC). LAG-3 prevalence was also consistent across subgroups based on age, sex, and ethnicity (all tumor types), as well as alcohol consumption and smoking history (melanoma and GC). LAG-3 expression was weakly correlated with PD-L1 expression on TCs (Spearman's ρ of 0.31 for GC, 0.41 for GEJC, and 0.46 for melanoma) and moderately correlated with PD-L1 CPS (Spearman's ρ of 0.63 for GC and 0.73 for GEJC). Similar correlations were observed after binning PD-L1 TC score and CPS. Conclusions: LAG-3-expressing lymphocytes were identified in all tumor types tested. LAG-3 expression did not differ with age, sex, ethnicity, alcohol consumption, and smoking history, and appeared to be associated, but not strongly correlated, with PD-L1 expression in all tumor types. PD-L1 and LAG-3 expression in procured samples may not be reflective of expression observed in clinical trials, and further studies are needed to characterize LAG-3 expression in a clinical setting. Citation Format: Lloye M. Dillon, John Wojcik, Keyur Desai, Ming Lei, Lori Johnson, Bryan McCune, Krystal Johnson, Jeffrey Shuster, Steven M. Anderson, Bin Li. Distribution and prevalence of LAG-3 expression in samples of melanoma and gastric/gastroesophageal junction cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 1625.