A sequential classification rule based on multiple quantitative tests in the absence of a gold standard

Area of Science:

• Biostatistics
• Medical Diagnostics
• Machine Learning

Background:

• Combining multiple biomarkers can enhance diagnostic accuracy over single tests.
• A classification algorithm is needed to integrate marker data without a gold standard.

Purpose of the Study:

• To develop a method for constructing a composite test from multiple biomarkers.
• To derive an optimal classification rule for case/non-case status in the absence of a gold standard.

Main Methods:

• A sequential composite test approach was developed, treating tests as a sequence.
• The method utilizes a mixture of two multivariate normal latent models for case and non-case groups.
• An optimal classification rule was derived to maximize sensitivity at a given specificity.

Main Results:

• The proposed composite test demonstrated statistical properties and predictive accuracy in simulations.
• The method was successfully applied to a real-data example.
• The approach is also amenable to nonparametric implementation.

Conclusions:

• The developed sequential composite test offers an effective strategy for biomarker combination in diagnostic settings lacking a gold standard.
• This method provides a robust approach to derive optimal classification rules, enhancing diagnostic performance.