Drug-diagnostics co-development in oncology

Area of Science:

• Genomic Medicine
• Clinical Trial Design
• Biostatistics

Background:

• Clinical observations of diverse disease progression and treatment responses are now supported by genomic data.
• Molecular characterization of diseases offers new therapeutic avenues but complicates clinical trial design and analysis.
• In oncology, broad patient treatment with targeted therapies is unsustainable if only a few benefit.

Purpose of the Study:

• To review prospective designs for developing novel therapeutics and predictive biomarkers.
• To address the need for new paradigms in randomized clinical trial (RCT) design and analysis for predictive medicine.
• To present a prediction-based analysis approach for RCTs that maintains statistical integrity.

Main Methods:

• Review of prospective clinical trial designs, ranging from single-biomarker development to genome-wide classifier discovery and validation.
• Outline of a prediction-based analysis framework for RCTs.
• Focus on designs that integrate biomarker development with therapeutic evaluation.

Main Results:

• The proposed prediction-based analysis preserves the type I error rate.
• The approach provides a reliable, internally validated method for identifying patient subgroups likely to benefit from new treatments.
• Designs accommodate various scenarios, from single biomarker-drug pairs to complex genome-wide predictive models.

Conclusions:

• Genomic heterogeneity demands innovative clinical trial designs and analysis methods.
• Predictive medicine requires robust statistical frameworks to ensure targeted therapies are effectively evaluated and applied.
• The outlined approach supports the development of personalized medicine by accurately predicting treatment response.