Test Validation and Performance Assessment

Test validation and performance assessment is the body of work by which a molecular test is shown to measure what it claims, reproducibly, and to support the clinical use for which it is intended. It separates analytical validation (does the assay measure the analyte correctly?) from clinical validation (does the result reflect the clinical state?), and quantifies performance with metrics such as accuracy, precision, sensitivity, specificity, and limits of detection.

Definition

Test validation is the documented process of demonstrating, through systematic experiments, that a laboratory test performs accurately and reproducibly and is fit for its intended clinical purpose; performance assessment is the ongoing measurement and verification of those characteristics.

Scope

The entry covers the concepts and metrics used to establish and document molecular test performance before clinical use and to verify it afterward: analytical versus clinical validity, the standard performance characteristics, and the special considerations for high-complexity assays and their bioinformatics pipelines. It is a methodological topic and gives no instructions for diagnosing or treating patients.

Core questions

• What is the difference between analytical and clinical validity?
• Which performance characteristics must be established for a molecular test?
• How are limits of detection and quantitation determined?
• How are sequencing and the bioinformatics pipeline validated, not just the wet-bench assay?

Key concepts

• Analytical validity
• Clinical validity
• Accuracy and trueness
• Precision (repeatability and reproducibility)
• Analytical sensitivity and limit of detection
• Analytical specificity
• Reportable and measurable range
• Bioinformatics pipeline validation
• Fitness for intended use

Mechanisms

Validation proceeds by designed experiments using characterized samples. Accuracy is assessed by comparison to a reference method or known specimens; precision by repeated measurement under repeatability and reproducibility conditions; analytical sensitivity by determining the limit of detection; and analytical specificity by testing for interferences and cross-reactivity (Jennings et al., 2009). Quantitative molecular assays can be validated with the same clinical-chemistry framework used for other measurands, as shown for droplet digital PCR liquid-biopsy assays (Milosevic et al., 2018). For next-generation sequencing, the computational pipeline that calls and filters variants must be validated alongside the laboratory steps (Roy et al., 2018). Clinical validity, distinct from these, asks how well the result classifies the clinical condition.

Clinical relevance

Whether a molecular result can be trusted in patient care rests on the validation and performance work behind it; undocumented or weak performance characteristics undermine interpretation. This entry explains how laboratories establish that trust and is not a guide to ordering or acting on specific tests.

Evidence & guidelines

Validation practice is codified in professional guidelines, including the AMP/CAP principles for validating molecular pathology tests and sequencing pipelines (Jennings et al., 2009; Roy et al., 2018) and CLSI consensus documents (CLSI EP28). These provide the experimental designs and acceptance criteria that laboratories follow under accreditation.

History

As clinical molecular testing expanded, professional organizations adapted established clinical-chemistry and laboratory-medicine validation principles to molecular assays, then extended them to massively parallel sequencing and its bioinformatics, producing the layered validation standards used today (Jennings et al., 2009; Roy et al., 2018).

Debates

How should validation scale to high-complexity, multi-target assays?

Establishing performance for every possible variant detectable by broad sequencing panels is impractical, so guidelines rely on representative validation and pipeline validation; how much is sufficient remains an active discussion.

Seminal works

• jennings-2009
• roy-2018

Frequently asked questions

What is the difference between analytical validity and clinical validity?

Analytical validity is whether the assay measures the analyte accurately and reproducibly; clinical validity is whether the result actually corresponds to the clinical condition the test is meant to detect.

Why does next-generation sequencing need pipeline validation?

Because the variants reported depend on the software that aligns reads and calls variants, the bioinformatics pipeline can introduce errors and must be validated alongside the laboratory procedure.