Antibody-drug conjugates in advanced small-cell lung cancer: a systematic review and single-arm meta-analysis of later-line monotherapy.

Objective: Antibody-drug conjugates (ADCs) are being investigated in advanced small-cell lung cancer (SCLC), but the evidence is dominated by early-phase studies conducted in non-exchangeable treatment settings. We evaluated the activity and toxicity of ADCs while separating later-line monotherapy from frontline combination therapy, later-line combination therapy, and post-induction maintenance. Methods: PubMed, Embase, the Cochrane Library, and Web of Science Core Collection were searched from inception and in 28 August 2026. ClinicalTrials.gov and the WHO International Clinical Trials Registry Platform (ICTRP) were searched on 28 August 2026. We constructed an arm-level dataset that recorded regimen, dose, clinical context, target, outcome definition, numerator, denominator, analysis set, and source. The primary quantitative synthesis was restricted to second-line or later ADC monotherapy; other clinical contexts were summarized descriptively. Single-arm proportions were synthesized using a binomial-normal random-intercept generalized linear mixed model on the logit scale. We reported 95% confidence intervals (CIs), between-study variance, heterogeneity, and 95% prediction intervals. An inverse-variance logit random-effects model with restricted maximum likelihood estimation and Hartung-Knapp inference was used as a sensitivity analysis. Fixed-time Kaplan-Meier survival estimates and inconsistently defined common adverse events were not pooled. Results: 16 trial families evaluating eight ADCs across five target classes met the eligibility criteria. In the primary later-line monotherapy analysis, the pooled ORR was 25.7% (95% CI, 16.4%-37.9%), and the pooled DCR was 72.7% (95% CI, 60.7%-82.2%). The pooled incidences of grade ≥3 TEAEs and grade ≥3 TRAEs were 62.8% (95% CI, 59.6%-65.9%) and 45.5% (95% CI, 35.5%-56.0%), respectively. The prediction intervals for ORR, DCR, and grade ≥3 TRAEs were wide. Proportions observed in a new study may therefore differ substantially from the pooled estimates. Frontline combinations, later-line combinations, and maintenance therapy yielded context-specific signals but were not compared with later-line monotherapy. Target-stratified analyses were exploratory; B7-H3 and DLL3 estimates were based on small, non-comparative groups, whereas Trop-2 and SEZ6 were not pooled because fewer than three studies were available. Conclusions: Later-line ADC monotherapy showed antitumor activity but frequent clinically important toxicity, with substantial between-study variation. The evidence does not establish superiority by treatment line, regimen, agent, or molecular target; single-arm response estimates also do not demonstrate a survival benefit. Randomized, context-specific trials using harmonized efficacy and safety definitions are needed.