Table 4.
Efficacy outcomes for the use of IO in second or more line of treatment for upper GI carcinomas compared with standard chemotherapy
| Setting | No. of studies | Study | Investigational agent | Overall survival |
PFS |
Response rate |
|||
|---|---|---|---|---|---|---|---|---|---|
| HR (95% CI) | P a | HR (95% CI) | P a | OR (95% CI) |
P
a
|
||||
| All studies | 6 | ATTRACTION-3 | Nivolumab | 0.77 (0.62 to 0.96) | − | 1.08 (0.87 to 1.34) | − | 0.87 (0.51 to 1.49) | − |
| ESCORT | Camrelizumab | 0.71 (0.57 to 0.87) | − | 0.69 (0.56 to 0.86) | − | 3.72 (1.98 to 6.99) | − | ||
| KEYNOTE-061 | Pembrolizumab | 0.94 (0.79 to 1.12) | − | 1.49 (1.25 to 1.57) | − | 0.77 (0.48 to 1.24) | − | ||
| KEYNOTE-063 | Pembrolizumab | NA | − | NA | − | 0.62 (0.20 to 1.90) | − | ||
| KEYNOTE-181 | Pembrolizumab | 0.89 (0.75 to 1.05) | − | 1.11 (0.94 to 1.31) | − | 2.09 (1.21 to 3.64) | − | ||
| RATIONALE-302 | Tislelizumab | 0.70 (0.57 to 0.85) | − | NA | − | NA | − | ||
| Any PD-L1 status | 5 | Pooled evidence | Any agent | 0.81 (0.74 to 0.88) | < .001 | 1.06 (0.79 to 1.42) | .70 | 1.31 (0.69 to 2.49) | .41 |
| All SCC | 4 | Pooled evidence | Any agent | 0.74 (0.68 to 0.82) | < .001 | 0.88 (0.69 to 1.14) | .34 | 1.98 (0.81 to 4.84) | .13 |
| All AdenoCa | 2 | Pooled evidence | Pembrolizumab | 0.99 (0.85 to 1.15) | .89 | 1.49 (1.25 to 1.77) | − | 0.75 (0.48 to 1.15) | .19 |
| PD-L1 CPS > 1% | 3 | Pooled evidence | Any agent | 0.73 (0.63 to 0.84) | < .001 | 0.88 (0.43 to 1.79) | .72 | 1.08 (0.66 to 1.77) | .76 |
| SCC | 2 | Pooled evidence | Cam/Nivo | 0.64 (0.51 to 079) | < .001 | 0.60 (0.43 to 0.84)b | NA | NA | − |
| Adeno | 1 | KEYNOTE-061 | Pembrolizumab | 0.81 (0.66 to 1.00) | 1.25 (1.02 to 1.54) | − | 1.08 (0.66 to 1.77) | .76 | |
| PD-L1 CPS > 10% | 5 | Pooled evidence | Any agent | 0.65 (0.55 to 0.78) | < .001 | 0.71 (0.56 to 0.89) | .003 | 3.82 (1.91 to 7.66) | <.001 |
| SCC | 4 | Pooled evidence | Any agent | 0.62 (0.54 to 0.81) | < .001 | 0.47 (0.24 to 0.88) | − | 3.74 (1.50 to 9.33) | .004 |
| AdenoCa | 2 | Pooled evidence | Pembrolizumab | 0.76 (0.54 to 1.07) | .11 | 0.79 (0.51 to 1.21) | − | 3.83 (1.42 to 10.32) | .008 |
| PD-L1 CPS < 1% | 3 | Pooled evidence | Any agent | 0.93 (0.79 to 1.10) | .41 | 1.27 (0.50 to 3.25) | .62 | 0.15 (0.03 to 0.72) | .02c |
| PD-L1 CPS < 1% SCC | 2 | Pooled evidence | Cam/Nivo | 0.83 (0.68 to 1.01) | .07 | 0.79 (0.59 to 1.05)b | NA | − | − |
| PD-L1 CPS < 1% AdenoCa | 1 | KEYNOTE-061 | Pembrolizumab | 1.20 (0.89 to 1.63) | − | 2.05 (1.50 to 2.79) | − | − | − |
| PD-L1 CPS < 5% SCC | 2 | Pooled evidence | Cam/Nivo | 0.76 (0.64 to 0.90) | .002 | 0.78 (0.61 to 0.99)b | NA | − | − |
| PD-L1 CPS < 10% SCC | 3 | Pooled evidence | Any agent | 0.78 (0.67 to 0.90) | .006 | 0.74 (0.59 to 0.94)b | NA | − | − |
| PD-L1 CPS < 10% all | 3 | Pooled evidence | Any agent | 0.83 (0.68 to 1.02) | .08 | NA | − | − | − |
| MSI-high in PD-L1 positive | 1 | KEYNOTE-061 | Pembrolizumab | 0.42 (0.13 to 1.31) | − | NA | − | 4.30 (0.70 to 27.16) | − |
Inverse-Variance and the Mantel-Haenszel statistical methods were applied for calculation of pooled hazard ratios and odds ratios, respectively. A 2-sided P value less than .05 was considered statistically significant. AdenoCa = adenocarcinoma; Cam = camrelizumab; Chemo = chemotherapy; CPS = combined positive score; GI = gastrointestinal; IO = immunotherapy; MSI = MicroSatellite Instability; Nivo = nivolumab; OS = overall survival; OR = odds ratio; PD-L1 = programmed cell death ligand-1; PFS = progression-free survival; SCC = squamous cell carcinoma.
These results come only from ESCORT trial.
Result is based on KEYNOTE-061 trial.