Prediagnostic evaluation of multicancer detection tests: design and analysis considerations.

Area of Science:

• Biostatistics
• Oncology
• Molecular Diagnostics

Background:

• Multicancer detection tests aim to identify preclinical cancers via blood-based molecular signals.
• Prediagnostic performance studies are crucial for evaluating these tests using stored specimens.

Purpose of the Study:

• To evaluate the effectiveness of the sample size trade-off method for prediagnostic studies of multicancer detection tests.
• To demonstrate sample size reduction and improved positive predictive value estimation.

Main Methods:

• The study compares the sample size trade-off method with standard calculations for prediagnostic performance studies.
• Key metrics include cancer-specific true-positive and false-positive rates, and positive predictive value.
• The sample size trade-off method prioritizes a zero-false-positive rate over precise true-positive rate targeting.

Main Results:

• The sample size trade-off method substantially reduces required sample sizes.
• For ovarian cancer (rarest case), it yields a sample size of 163,000 versus 720,000 with standard methods.
• This method increases the lower bound of the positive predictive value.

Conclusions:

• The sample size trade-off method offers a more efficient approach for prediagnostic evaluation of multicancer detection tests.
• Recommendations are provided for specimen repository planning and test evaluation.
• This method enhances the reliability of predictive performance metrics.