Table 2.
Efficacy outcomes for first-line use of IO vs chemotherapy for the treatment of upper GI carcinomas
| Settinga | No. of studiesb | Study | Investigational agent | OS |
PFS |
Response rate |
|||
|---|---|---|---|---|---|---|---|---|---|
| HR (95% CI) | P c | HR (95% CI) | P c | OR (95% CI)d | P c | ||||
| IO vs Chemo | |||||||||
| PD-L1 CPS > 1% | 1 | KEYNOTE-062 | Pembrolizumab | 0.91(0.74 to 1.10)d | − | 1.66 (1.37 to 2.01) | − | 0.29 (0.19 to 0.45) | <.001 |
| PD-L1 CPS > 5% or 10% | 1 | KEYNOTE-062 | Pembrolizumab | 0.69(0.49 to 0.97) | − | 1.10 (0.79 to 1.51) | − | 0.55 (0.29 to 1.04) | .065 |
| IO+Chemo vs Chemo | |||||||||
| All studies | 6 | KEYNOTE-062 | Pembrolizumab | 0.85 (0.70 to 1.03) | − | 0.84 (0.70 to 1.02) | − | 1.60 (1.12 to 2.28) | − |
| KEYNOTE-590 | Pembrolizumab | 0.73 (0.62 to 0.86) | − | 0.65 (0.55 to 0.76) | − | 1.98 (1.47 to 2.68) | − | ||
| ATTRACTION-4 | Nivolumab | 0.90 (0.75 to 1.07) | − | 0.68 (0.54 to 0.86) | − | 1.48 (1.10 to 1.98) | − | ||
| CheckMate-649 | Nivolumab | 0.80 (0.68 to 0.94)e | − | 0.77 (0.68 to 0.87) | − | 1.84 (1.38 to 2.45) | − | ||
| CheckMate-648 | Nivolumab | 0.74 (0.59 to 0.94) | − | 0.81 (0.67 to 0.99) | − | — | − | ||
| ESCORT-1st | Camrelizumab | 0.70 (0.56 to 0.87) | − | 0.56 (0.46 to 0.68) | − | — | − | ||
| Pooled evidence | Any agent | 0.79 (0.74 to 0.85) | <.001 | 0.72 (0.64 to 0.81) | <.001 | 1.72 (1.48 to 2.00)(4) | <.001 | ||
| PD-L1 CPS > 1% | 4 | Pooled evidence | Any agent | 0.54 (0.37 to 0.78) | <.001 | 0.69 (0.58 to 0.83) | <.001 | 1.60 (1.12 to 2.28)(1) | .009 |
| PD-L1 CPS > 5 or 10% | 4 | Pooled evidence | Any agent | 0.69 (0.61 to 0.77) | <.001 | 0.62 (0.55 to 0.69) | <.001 | 1.83 (1.42 to 2.37)(2) | <.001 |
| Any PD-L1 status | 5 | Pooled evidence | Any agent | 0.78 (0.73 to 0.84) | <.001 | 0.69 (0.55 to 0.78) | <.001 | 1.70 (1.38 to 2.10) (2) | <.001 |
| Squamous histology | 3 | Pooled evidence | Any agent | 0.72 (0.64 to 0.81) | <.001 | 0.69 (0.65 to 0.79) | <.001 | NA | − |
| Adenocarcinoma histology | 4 | Pooled evidence | Any agent | 0.83 (0.76 to 0.90) | <.001 | 0.76 (0.70 to 0.83) | <.001 | NA | − |
| MSI-high population (PD-L1–positive population) | |||||||||
| AnyIO vs Chemo | 2 | Pooled evidence | Any agent | 0.33 (0.19 to 0.57) | <.001 | (PD-L1 CPS > 1% in KEYNOTE-062 and PD-L1 CPS > 5% in CheckMate-649) | |||
| IO+Chemo vs Chemo | 2 | Pooled evidence | Any agent | 0.35 (0.18 to 0.59) | .002 | − | NA | NA | − |
| IO vs Chemo | 1 | KEYNOTE-062 | Pembrolizumab | 0.29 (0.11 to 0.81) | <.001 | − | NA | NA | − |
Detailed description of the studies included in the pooled analysis is included in the Supplementary Methods (available online). Chemo = chemotherapy; CPS = combined positive score; IO = immunotherapy; MSI = MicroSatellite Instability; OS = overall survival; PD-L1 = programmed cell death ligand-1; PFS = progression free survival; SCC = squamous cell carcinoma.
The number of studies with relative data.
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.
Number of studies analyzed in the pooled analysis for response rate.
Confidence interval for the hazard ratio was 99.2% in the original manuscript and was recalculated as the 95% confidence interval for the present comparisons.