More efficient, smaller multicancer screening trials

Area of Science:

• Biostatistics
• Clinical Trial Design
• Cancer Screening

Background:

• The NHS-Galleri Trial demonstrated the feasibility of a screening design where only intervention arm samples are tested.
• This study evaluates analysis methods for trials where control arm samples might be tested retrospectively.

Purpose of the Study:

• To assess the efficiency of different analysis methods for blood-based cancer screening trials.
• To compare traditional, intended-effect, and targeted analysis methods.

Main Methods:

• Compared traditional (all events), intended-effect (positive tests), and targeted (test-positive events) analyses.
• Utilized approximate statistical methods and scenario analysis for comparison.

Main Results:

• Intended-effect and targeted analyses require smaller sample sizes for cancer-specific mortality evaluation than traditional methods.
• Intended-effect analysis requires a smaller sample size than targeted analysis, especially when cancer death risk in positive cases is high.

Conclusions:

• Intended-effect or targeted analysis can substantially reduce sample size needs for blood-based cancer screening trials.
• Targeted analysis minimizes retrospective testing and associated notification issues.
• Trialists should balance sample size reduction against retrospective testing costs when selecting an analysis method.