Lodde · European journal of cancer (Oxford, England : 1990) 2024 · prospective multicenter cohort study · n=522

Early versus late response to PD-1-based immunotherapy in metastatic melanoma.

Level 3 - non-randomized controlled study

Prospective multicenter observational cohort study

PubMed 39213786 · doi:10.1016/j.ejca.2024.114295 · record verified 2026-08-26

What was done

A prospective multicenter observational cohort study (DeCOG registry ADOREG-TRIM, NCT05750511) analyzed 522 patients receiving PD-1-based immunotherapy in any therapy line for non-resectable stage-IV melanoma. Patients were categorized by response kinetics into early responders (complete response [CR] or partial response [PR] within the first 3 months; EarlyR), late responders (CR/PR after 3 months; LateR), stable disease (SD), and non-responders (progressive disease; NonR). Multivariable log-binomial regression adjusted for age and sex evaluated predictors of early response versus non-response, and progression-free survival (PFS) and overall survival (OS) were determined.

What was found

Of 522 patients, 8.2% were EarlyR (n = 43), 19.0% were LateR (n = 99), 37.0% had SD (n = 193), and 35.8% were NonR (n = 187). Independent predictors of early response compared to NonR were positive tumor PD-L1 (RR = 1.99, 95% CI 1.14–3.46, p = 0.015) and normal serum CRP (RR = 1.59, 95% CI 0.93–2.70, p = 0.036). Median PFS and OS were: - EarlyR: PFS 46.0 months (95% CI 19.1–NR); OS 47.8 months (95% CI 36.9–NR) - LateR: PFS NR (95% CI NR–NR); OS NR (95% CI NR–NR) - SD: PFS 8.1 months (95% CI 7.0–10.4); OS 35.4 months (95% CI 29.2–NR) - NonR: PFS 2.0 months (95% CI 1.9–2.1); OS 6.1 months (95% CI 4.6–8.8)

Why it matters

Fewer than 10% of metastatic melanoma patients achieve an objective response within 3 months, but late responders demonstrate durable response and survival at least equivalent to early responders. This supports maintaining PD-1 therapy past 3 months in patients without disease progression.

Limits

Observational registry design vulnerable to selection bias and immortal-time/guarantee-time bias inherent in defining late responders. Patients across any treatment line were pooled without line-stratified data in the abstract, and the reported confidence interval for normal CRP spans 1.0 (0.93–2.70) despite a reported p-value of 0.036.