1082O KEYMAKER-U02 substudy 02C: Neoadjuvant pembrolizumab (pembro) and investigational agents followed by adjuvant pembro for stage IIIB-D melanoma
Background
Substudy 02C of the phase 1/2 KEYMAKER-U02 trial (NCT04303169) is evaluating neoadjuvant pembro with or without investigational agents followed by adjuvant pembro for stage IIIB-D melanoma. We present initial results from arm 4 (pembro + MK-4830 [anti-ILT4]) and arm 5 (favezelimab [anti-LAG-3] coformulated with pembro [fave/pembro]) and updated results from arm 1 (pembro + vibostolimab [vibo; anti-TIGIT]), arm 2 (pembro + gebasaxturev [geba; coxsackievirus A21]), and arm 3 (pembro alone).
Methods
Adults with resectable stage IIIB-D melanoma were randomly assigned to open arms. Patients (pts) received 2 doses of pembro 200 mg Q3W + vibo 200 mg Q3W in arm 1, 1 dose of pembro 400 mg + 5 doses of geba 3 × 108 TCID50 in arm 2, 1 dose of pembro 400 mg in arm 3, 2 doses of pembro 200 mg + MK-4830 800 mg Q3W in arm 4, and 2 doses of fave 800 mg/pembro 200 mg Q3W in arm 5. Resection occurred at week 6. At week 12, pts began adjuvant pembro 400 mg Q6W for ≤8 cycles. Primary end points were safety and pathologic complete response (pCR). Secondary end points included near pCR, partial pathologic response (pPR), and RFS. EFS was exploratory.
Results
117 pts were assigned to arm 1 (n = 26), arm 2 (n = 25), arm 3 (n = 15), arm 4 (n = 25), and arm 5 (n = 26). Median (range) follow-up was 29.9 (25.9-42.5) mo in arm 1, 29.7 (24.5-40.1) mo in arm 2, 34.3 (29.6-41.5) mo in arm 3, 15.2 (9.6-18.5) mo in arm 4, and 9.4 (4.1-12.6) mo in arm 5. Grade 3 or 4 treatment-related adverse events (TRAEs) occurred in 8% of pts in arm 1, 28% in arm 2, 7% in arm 3, 16% in arm 4, and 15% in arm 5. No grade 5 TRAEs occurred. Efficacy is reported in the table. Median RFS and EFS were not reached in any arm.
| Arm 1 pembro + vibo n = 26 | Arm 2 pembro + geba n = 25 | Arm 3 pembro alone n = 15 | Arm 4 pembro + MK-4830 n = 25 | Arm 5 fave/pembro n = 26 | |
|---|---|---|---|---|---|
| Major pathologic response (pCR + near pCR), % (95% CI) | 50 (30-70) | 40 (21-61) | 47 (21-73) | 32 (15-54) | 58 (37-77) |
| pCR, % (95% CI) | 38 (20-59) | 28 (12-49) | 40 (16-68) | 28 (12-49) | 38 (20-59) |
| Near pCR, % (95% CI) | 12 (2-30) | 12 (3-31) | 7 (<1-32) | 4 (<1-20) | 19 (7-39) |
| pPR, % (95% CI) | 31 (14-52) | 12 (3-31) | 27 (8-55) | 8 (1-26) | 19 (7-39) |
| Objective response rate per RECIST v1.1, % (95% CI) | 50 (30-71) | 32 (15-54) | 27 (8-55) | 44 (24-65) | 35 (17-56) |
| RFS rate, % | |||||
| 6 mo | 95 | 95 | 100 | 94 | 93 |
| 18 mo | 90 | 90 | 82 | – | – |
| EFS rate, % | |||||
| 6 mo | 86 | 76 | 87 | 88 | 92 |
| 18 mo | 81 | 72 | 80 | 78 | – |
Conclusions
All arms had manageable safety in stage IIIB-D melanoma. With the promising antitumor activity observed in this study, further investigation of pembro + vibo and fave/pembro in this setting is warranted. RFS by major pathologic response will be presented.