How It Works
Study Design
The LoQ evaluation uses low-level samples with known assigned concentrations spanning the low-end region of the measuring interval. CLSI EP17-A2 specifies the following minimum design:
Limit of QuantitationVerification / EstablishmentCLSI EP17-A2≥ 2 reagent lots
| Reagent lots | Minimum 2 (LoQ calculated separately per lot; maximum reported) |
| Low-level samples | Minimum 4 unique samples with known assigned concentrations |
| Concentration levels | 9 levels spanning the low-end region recommended |
| Days | Minimum 3 test days |
| Runs per day | Multiple runs per day to assess within-laboratory precision |
| Replicates per sample | Minimum 3 per reagent lot / instrument / day combination |
| Total replicates per lot | Minimum 36 (across all low-level samples, runs, and days) |
Assigned values must be known (from reference materials or prior assignment) — they are required for fitting the power model. Concentration levels should span from near the expected LoQ down to as low as practically measurable, covering the full low-end region of interest.
Why 9 concentration levels? The power model regression requires enough points across the low-end concentration range to fit a reliable curve. A wider spread of concentration levels — particularly including levels both below and above the expected LoQ — gives the model sufficient data to accurately locate the CV%-threshold intersection. Studies with only a few levels, or levels clustered near a single concentration, produce poorly constrained fits and unreliable LoQ estimates.