Statistical considerations for some issues in clinical bridging studies evaluating companion diagnostic devices

Area of Science:

• Biomarkers
• Clinical Trials
• In Vitro Diagnostics

Background:

• Companion diagnostic devices (CDx) are essential for the safe and effective use of associated therapeutics.
• Clinical trials ideally enroll patients using the final CDx, but this is often impractical due to unavailability.
• Clinical trial assays (CTA) are frequently used for enrollment, necessitating a bridge to the CDx.

Purpose of the Study:

• To review challenges in clinical bridging studies for companion diagnostic devices.
• To explore statistical methodologies for assessing CDx effectiveness, particularly with binary endpoints.
• To address issues such as missing data, local enrollment tests, and low positive rate biomarkers.

Main Methods:

• Literature review of clinical bridging studies.
• Analysis of challenges including missing data, local tests, and prescreening.
• Focus on statistical methodologies for binary endpoints and low-prevalence biomarkers.

Main Results:

• Clinical bridging studies face challenges like missing data and the use of non-marketable assays.
• Alternative statistical methods can assess CDx effectiveness, especially for low positive rate biomarkers.
• Bridging studies are vital for validating CDx when CTA is used for enrollment.

Conclusions:

• Addressing challenges in clinical bridging studies is critical for accurate CDx validation.
• Appropriate statistical methodologies can ensure the reliable assessment of therapeutic product efficacy.
• Successful bridging ensures the safety and effectiveness of both the therapeutic and companion diagnostic device.