Modelling RT-qPCR cycle-threshold using digital PCR data for implementing SARS-CoV-2 viral load studies

Area of Science:

• Virology
• Molecular Diagnostics
• Epidemiology

Background:

• Accurate quantification of SARS-CoV-2 viral load is crucial for observational studies and clinical management.
• Digital PCR offers advanced features for precise viral load determination.
• Existing methods like RT-qPCR provide cycle threshold (Ct) values that require conversion for standardized interpretation.

Purpose of the Study:

• To leverage digital PCR for reliable SARS-CoV-2 viral load quantification (copies/μL) in observational studies.
• To develop and validate a model for converting RT-qPCR Ct values to viral load.
• To utilize quantified viral load for retrospective analysis and development of case-fatality prediction tools.

Main Methods:

• A cohort of 51 COVID-19 positive samples was analyzed using both RT-qPCR and digital PCR.
• A linear regression model was trained to convert Ct values to viral load (copies/μL).
• The model was applied to 6208 diagnostic samples, and a logistic regression model was built to predict case-fatality using viral load, age, and gender.

Main Results:

• The Ct value to viral load conversion model demonstrated high accuracy.
• Retrospective analysis revealed low viral loads during Italian inter-epidemic periods.
• A predictive model for case-fatality showed high specificity (99.0%) and variable sensitivity (21.7% to 75%) depending on the threshold, with comparable performance using categorized viral load or raw Ct values.

Conclusions:

• Digital PCR-based modeling of Ct values facilitates SARS-CoV-2 research and the development of predictive tools.
• These tools can identify high-risk patients early, regardless of the specific RT-qPCR platform used.
• Achieving adequate sensitivity for clinical decision-making is feasible with acceptable specificity, adaptable to public health needs and available resources.