Biomarker-Stratified Design — SCE Medical Oncology MCQ

Instant feedback + full explanation. One question, done properly.

HardClinical Trials & StatisticsBiomarker-Stratified DesignSCE Medical Oncology

A researcher is designing a biomarker-stratified trial in NSCLC. Patients are randomised to treatment A or B, with stratification by PD-L1 status (≥50% vs <50%). The protocol specifies that if the overall population result is positive, the co-primary analysis will test PD-L1 ≥50% and PD-L1 <50% subgroups separately. What statistical design is this?

Educational content. Not a substitute for clinical judgement or local policy.

Reveal the answer and explanation

Correct answer: E — A prospective-retrospective biomarker-stratified design with hierarchical testing — the pre-specified testing order controls the family-wise error rate

Biomarker-stratified trial designs prospectively collect biomarker data and pre-specify subgroup analyses with formal statistical testing. Hierarchical testing (e.g. test overall population first → if positive, test PD-L1 ≥50% → if positive, test PD-L1 <50%) controls the family-wise error rate without requiring alpha-splitting. This design is used in many ICI trials (e.g. KEYNOTE-042 testing PD-L1 ≥50%, ≥20%, ≥1% hierarchically). It balances broad patient inclusion with biomarker-specific hypothesis testing.

Reference: JRCPTB Curriculum; FDA Enrichment Guidelines; Freidlin et al JCO 2010