Skip to main content

Molecular Biology

Assay Validation & Verification

In the regulatory framework of the clinical laboratory (CLIA ’88), the introduction of a new test is not as simple as purchasing a kit and running patient samples. The laboratory must generate empirical data proving the test performs accurately and reliably within that specific facility’s environment. The scope of this study depends entirely on the regulatory status of the method. The distinction between Verification and Validation is the single most important administrative concept in assay implementation

The Regulatory Distinction: Verification vs. Validation

The Center for Medicare and Medicaid Services (CMS) and accrediting bodies (CAP/The Joint Commission) classify methods into two categories, each requiring a different level of rigor to implement

  • Verification (Unmodified FDA-Cleared Tests)
    • Definition: The process of confirming that an FDA-approved In Vitro Diagnostic (IVD) kit performs according to the manufacturer’s claims when used in your laboratory
    • Scope: Since the manufacturer has already proven the test works (during FDA clearance), the lab only needs to demonstrate that their staff and environment don’t introduce error. This is a smaller, less expensive study
    • Required Elements: Accuracy, Precision, Reportable Range, and Reference Interval
  • Validation (Laboratory Developed Tests - LDTs)
    • Definition: The comprehensive process of establishing the performance specifications of a test from scratch. This applies to LDTs, “Home-Brew” assays, Research Use Only (RUO) reagents, or Modified IVDs (e.g., using a urine collection kit for a CSF sample)
    • Scope: The manufacturer has not guaranteed performance for this specific use. The laboratory takes full responsibility for proving the test works. This is a massive, costly study requiring many samples
    • Required Elements: All elements of Verification PLUS Analytical Sensitivity (LOD), Analytical Specificity (Interferences), and Clinical Utility

Required Performance Specifications

Whether performing a Verification or a Validation, specific performance characteristics must be statistically determined. These metrics form the “Validation Summary” signed by the Medical Director

  • Accuracy (Trueness)
    • Goal: To determine if the test gives the “correct” answer
    • Study Design: Compare the results of the new method against a “Reference Method” (Gold Standard) using at least 20–40 samples spanning the entire reportable range (Low, Medium, High)
    • Evaluation: For qualitative molecular tests, calculating Concordance (e.g., 98% agreement with the reference lab). For quantitative tests, performing Linear Regression to check slope and correlation
  • Precision (Reproducibility)
    • Goal: To determine if the test gives the same answer when repeated
    • Study Design: Running the same Positive and Negative samples multiple times
    • Intra-Run Precision (Repeatability): Running the sample 20 times in a single run. Checks for pipetting/instrument stability
    • Inter-Run Precision (Reproducibility): Running the same sample once a day for 20 days (or across different shifts/technologists). Checks for stability over time and operator variance
  • Reportable Range (Linearity)
    • Goal: To define the span of values for which the instrument can accurately report results without dilution
    • Study Design: Creating a “Linearity Panel” (usually 5 to 7 levels) by taking a high concentration sample and performing serial dilutions down to zero. The Observed values are plotted against the Expected values. The range where the line remains straight (\(R^2 > 0.95\)) is the Reportable Range
    • Hook Effect: In molecular biology, extremely high viral loads can inhibit PCR, causing a paradoxically low signal. The linearity study confirms the upper limit where this does not occur
  • Reference Interval (Normal Range)
    • Goal: To define what is “Normal” for the healthy population
    • Molecular Context: For infectious diseases (e.g., Influenza), the reference interval is typically “Not Detected.” For quantitative assays (e.g., CMV Viral Load), the lab must verify the “limit of quantification” below which a patient is considered negative or clinically insignificant

Additional Validation Requirements (LDTs Only)

If the test is an LDT or a modified IVD, the laboratory must prove the fundamental biochemistry works. These studies are not required for standard FDA-cleared Verification

  • Analytical Sensitivity (Limit of Detection - LOD)
    • Goal: To determine the absolute lowest concentration of DNA/RNA that can be detected 95% of the time
    • Study Design: The lab creates a dilution series of a known standard. Each dilution is run in 20 replicates. The concentration where 19 out of 20 (95%) are positive is established as the LOD
    • Importance: This is critical for diagnosing pathogens found in low numbers, such as viral meningitis in CSF
  • Analytical Specificity (Interference and Cross-Reactivity)
    • Goal: To ensure the primers do not bind to the wrong target
    • Interference Study: Spiking the sample with potential inhibitors (Hemoglobin, Bilirubin, Triglycerides, or Mucus) to ensure the extraction chemistry can handle “dirty” samples
    • Cross-Reactivity (Exclusivity): Testing the assay against genetically similar organisms. Example: A SARS-CoV-2 validation must test samples containing Coronavirus 229E, HKU1, and OC43 (common cold coronaviruses) to prove the primers do not cross-react and cause False Positives
  • Sample Stability
    • Goal: To determine how long a specimen is valid before processing
    • Study Design: A positive sample is split into aliquots. One is tested immediately (Time 0). Others are stored at Room Temp, 4°C, and -20°C, and tested at 24, 48, and 72 hours. This data dictates the rejection criteria for “aged” samples

Molecular-Specific Studies

Due to the nature of amplification technology, molecular operations require specific safety validations that are not necessary in chemistry or hematology

  • Carryover Studies (Contamination Check)
    • Goal: To assess the risk of the instrument or workflow contaminating a negative sample with DNA from a previous positive sample
    • Study Design: A “Checkerboard” pattern is run. High Positive samples are alternated with True Negative samples (High Pos / Neg / High Pos / Neg)
    • Pass Criteria: All Negative samples must remain Negative. If a Negative sample following a High Positive turns positive, the instrument has a carryover problem (e.g., inefficient tip washing or aerosol generation)
  • Matrix Validation
    • Goal: Proving the test works in the specific body fluid collected
    • Issue: FDA clearance usually specifies the matrix (e.g., “Nasopharyngeal Swab”). If the doctor wants to test Bronchoalveolar Lavage (BAL) fluid, the lab must perform a “Modified IVD” Validation to prove that the viscosity or inhibitors in the BAL fluid do not interfere with the assay

Documentation & Approval

The administrative conclusion of the process is the Validation Summary Report. This is a legal document required for accreditation

  • Medical Director Sign-Off: The Laboratory Director (MD or PhD) must review the statistical data and sign the summary before the test is used on a single patient. Testing patients prior to this date is a major regulatory violation
  • SOP Implementation: The validation data is used to write the Standard Operating Procedure (SOP). For example, the “Sample Stability” data is written directly into the “Specimen Rejection” section of the SOP
  • Proficiency Testing (PT): Once the test is live, the lab must enroll in an external PT program (e.g., CAP Surveys) to monitor ongoing accuracy. If no commercial PT exists for a rare LDT, the lab must set up an “Alternative Assessment” (e.g., swapping samples with another hospital) twice a year