Oncology Research for Medical Students

Oncology research for beginners typically uses registry or record data, reviews and case-based work to study diagnosis, treatment patterns, outcomes and patient experience. Because cancer outcomes involve survival and time-to-event methods, plan statistical support early.

Question types

• Epidemiology: incidence, stage at diagnosis, trends (for example from cancer registries).
• Treatment patterns and outcomes: use of therapies, survival, toxicity.
• Screening and early detection: uptake and outcomes.
• Supportive care and survivorship: quality of life, symptoms, access.
• Evidence synthesis: reviews of therapies, biomarkers or care models.
• Case reports of unusual responses or adverse events.

Key concepts

• Endpoints: overall survival, progression-free survival, response rate and patient-reported outcomes each answer different questions; define them as the source does.
• Time-to-event analysis: Kaplan–Meier curves, hazard ratios (see Hazard Ratio Explained).
• Confounding and selection: treatment choice depends on stage, fitness and preference.
• Data sources: registries such as SEER (US) or national equivalents; institutional databases with approval.

Illustrative example

Invented for teaching. "Among adults diagnosed with stage II colon cancer between 2015 and 2020 in a registry, is receipt of adjuvant chemotherapy associated with 5-year survival, adjusted for age and comorbidity?" Design: retrospective cohort with a statistician. Limitations: unmeasured confounding, registry variable definitions.

Cautions

Observational comparisons of treatments are prone to bias; use cautious language and consider the hierarchy of evidence. Cancer data are sensitive, so follow strict privacy rules.

Using real-world data carefully

Registry data are valuable because they cover many patients, but they may lack details such as performance status, comorbidity or reasons for treatment choice. State which variables are available, which are missing and how this limits interpretation. Avoid ranking therapies on observational survival alone, and say clearly when a comparison is descriptive. Ask a statistician about methods for time-to-event data before analysis.

Common pitfalls

• Comparing survival without addressing lead-time or immortal time bias.
• Using surrogate endpoints as proof of benefit.

Key takeaways

• Define endpoints, plan time-to-event analysis and expect confounding.
• Use registries and approvals appropriately.

This page is general educational information about research methods. It is not medical, legal or regulatory advice; follow your institution's ethics board and your target journal's instructions.