Examqo › Examqo library › Oncology › Staging, markers and research
Clinical trials and biostatistics in oncology
A phase 1 trial defines safety and dose, phase 2 tests activity, and phase 3 compares with standard treatment by randomisation; results are interpreted by the endpoint (overall survival is the gold standard), the design (superiority or non-inferiority) and measures such as the hazard ratio and predictive value.
Why it matters in the exam
Every exam has several trial questions: the goal of a phase 1 trial, what sets sample size in phase 3, when a non-inferiority trial is chosen, why stratify, what a basket trial is, which studies enter a meta-analysis, and what characterises progression-free survival. These are short questions based on precise definitions.
Many distractors confuse sensitivity and specificity, which are properties of the test, with positive predictive value, which depends on disease prevalence in the population. Another confusion is between a basket trial (one drug, one marker, many tumours) and an umbrella trial (one tumour type, many markers and drugs). Also remember the advantage of patient-reported outcomes.
Key points
- Phase 1 tests safety and dose, phase 2 tests activity, and phase 3 compares with standard treatment by randomisation.
- Sample size is set by α, power, expected effect size and event rate.
- Overall survival is the gold standard, and progression-free survival includes progression or death from any cause and depends on imaging frequency.
- A non-inferiority trial suits a new treatment with an advantage other than efficacy, and stratification balances prognostic factors.
- Positive predictive value depends on disease prevalence, and patient-reported outcomes better reflect symptoms.
Summary table
| Concept | Definition | Exam point |
|---|---|---|
| Phase 1 | Safety, maximum tolerated dose and the recommended phase 2 dose; dose-limiting toxicity | Usually a 3+3 dose-escalation design |
| Phase 2 and phase 3 | Phase 2: activity (response rate), sometimes single-arm; phase 3: randomised comparison with standard treatment | Sample size is set by α, power, expected effect size and event rate |
| Endpoints | Overall survival — the gold standard; progression-free survival — from randomisation to progression or death from any cause | PFS is not affected by later treatments, but depends on imaging frequency |
| Non-inferiority | Tests that the new treatment is not worse than standard by more than a pre-set margin | Chosen when the new treatment has another advantage: less toxicity, convenience, cost or shorter duration |
| Stratification and randomisation | Splitting into strata by prognostic factors before allocation | Its aim is balance between arms in known factors |
| Basket and umbrella trials | Basket: one drug for one marker across many tumour types; umbrella: one tumour type with many markers and drugs | A meta-analysis includes all relevant randomised trials, including unpublished ones |
| Test measures and patient-reported outcomes | Positive predictive value — the proportion with disease among positive results, depends on prevalence | Patient-reported outcomes capture symptoms and toxicity better than clinician assessment |
In the full version for subscribers
- A simple algorithm
- Memory aid
- Exam traps
- 5 review cards
- Enough for IMLE
The full version is open to subscribers of: Oncology, Public health.
Subscriber? Read the full topic Plans and prices 25 free questions
Sources
- Piantadosi, Clinical Trials: A Methodologic Perspective, 3rd ed. (2017) — Phases of development; Sample size; Non-inferiority designs
- US FDA Guidance: Clinical Trial Endpoints for the Approval of Cancer Drugs and Biologics, 2018 — Overall survival and tumour-based endpoints
- DeVita, Hellman and Rosenberg's Cancer: Principles and Practice of Oncology, 12th ed. (2023) — Design and analysis of oncology clinical trials
Sources are given for further reading; the text is written afresh and not copied from them.
Reviewed by the Examqo AI Professor · 2026-11-06. The text may contain errors; when in doubt, check the source.