CheckMate 648
In advanced oesophageal squamous-cell carcinoma, do nivolumab plus chemotherapy or nivolumab plus ipilimumab improve survival over chemotherapy?
One underlying study; aliases and follow-up reports consolidated. Primary discipline: Medical.
970 patients with previously untreated unresectable, advanced, recurrent or metastatic ESCC.
Advanced ESCC; tumour-cell PD-L1 ≥1% was a primary analysis population. Tumour-cell score and CPS are not interchangeable.
970; three arms randomised/evaluable as reported.
Nivolumab plus fluoropyrimidine/platinum chemotherapy, or nivolumab plus ipilimumab, per randomised arm.
Fluoropyrimidine/platinum chemotherapy alone.
Endpoints and results
Primary OS and BICR-assessed PFS in tumour-cell PD-L1 ≥1%; all-randomised OS/PFS key secondary endpoints.
- At ≥5 years, all-randomised OS HR 0.77 (95% CI 0.65–0.92) for each nivolumab regimen vs chemotherapy.
- 5-year OS all-randomised: nivolumab+chemo 13%, nivolumab+ipilimumab 16%, chemo 9%.
- In PD-L1 tumour-cell ≥1%, PFS benefit was seen with nivolumab+chemo (HR 0.67, 95% CI 0.51–0.88), not nivolumab+ipilimumab (HR 1.03, 95% CI 0.78–1.35).
- Grade 3–4 treatment-related AEs: 49%, 33%, 37%, respectively.
Absolute Effect and outcome visualisation
5-year OS, all randomised
% alive at 5 years · 5 years. Bars show reported fixed-time percentages.
Interpretation: Rounded OS differences are +4 and +7 percentage points vs chemotherapy; descriptive fixed-time estimates, not a direct comparison between experimental regimens.
Toxicity, discontinuation and quality of life
Trial-specific harms: Grade 3–4 treatment-related AEs 49% with nivolumab+chemo, 33% with nivolumab+ipilimumab and 37% with chemo.
Treatment burden/discontinuation: Regimen-specific discontinuation and exposure require full-text table extraction; immune toxicity differs from chemotherapy toxicity.
Quality of life: No numerical QoL effect asserted here without detailed PRO report.
Subgroups, maturity and limitations
Subgroups: Five-year OS benefit maintained in tumour-cell PD-L1 ≥1%; do not confuse tumour-cell scoring with CPS.
Maturity: Five-year follow-up published 2026.
Limitations: Three-arm trial with differing PFS patterns; nivolumab+ipilimumab did not improve PFS in the key PD-L1 population.
Both nivolumab regimens improved long-term OS vs chemotherapy, but PFS evidence and toxicity differ; treatment choice should not conflate the arms.
Current-practice context must be checked against regional guidelines, patient fitness, biomarkers, alternatives and regulatory status before clinical application. Clinical freshness review and browser-rendered QA remain pending.
Primary sources and registry
Registry: NCT03143153
The Absolute Effect · Updated 26 October 2026 · Never derive NNT from median survival.