Limits of blank, detection and quantitation
In one sentence
Limits of blank, detection and quantitation describe background, the ability to detect a signal and the ability to measure its amount reliably under specified conditions.
The intuition
In a quiet room, you may hear a faint voice without being able to judge its loudness accurately. First you need to know the room's background noise; then whether a voice is distinguishable; then how well its level can be measured.
The analogy has limits: an assay has specimen-specific background, calibration and decision rules. These limits are evaluated statistically rather than chosen by listening once.
How it works
Limit of blank (LoB) describes a high-end background result expected from samples without the target, at a stated probability. A blank can still produce measurement noise, contamination or nonspecific response. Its distribution helps distinguish background from signal; a single blank reading is not enough.
Limit of detection (LoD) describes the smallest amount or concentration detected with a stated probability under stated conditions, allowing a specified false-positive risk. LoD95 denotes a 95% detection probability in those conditions. It does not mean that every sample above the limit is positive, every sample below it is negative, or the patient's diagnosis is 95% certain.
Limit of quantitation (LoQ) describes the lowest amount reliably measured to specified accuracy requirements. Those requirements include acceptable bias and imprecision, or a defined total-error goal. A detected signal can be below the quantitative range. LoD and LoQ can coincide in some methods; they are different performance questions rather than a mandatory fixed spacing.
Thresholds for clinical decisions are another job. A clinically important concentration may differ from the LoD or LoQ. A positive qualitative assay also has its own classification rule; it need not report a quantitative value.
How it is measured
| Assay-card item | Validation question |
|---|---|
| Measures / how | Repeat blanks and low-level samples across specified conditions; evaluate detection and quantitative error |
| Input and cost | Representative clinical or suitably equivalent reference specimens; repeated tests consume material |
| Output / units | Background response, amount/concentration and detection probability; accuracy criteria for LoQ |
| Thresholds | Prespecified false-positive/false-negative and quantitative-error requirements |
| Failure modes | Wrong matrix, unrepresentative targets, contamination, calibration bias, lot/operator variation or inadequate input |
| Cannot tell | Clinical sensitivity, tumor location, prognosis or benefit from acting on the result |
| Validation | Exact analyte, specimen matrix, workflow, input range and assay version; independent verification where appropriate |
“Matrix” means the surrounding specimen material, such as plasma or tissue lysate. Spiked material can support testing, but similarity to real specimens must be justified. A claimed limit should identify whether it starts at purified material or covers collection through the final report. Compare conditions before comparing numbers: targeted protein quantitation and blood molecular detection have different inputs and error sources.
Why it matters in cancer
A lower nominal limit can be useful analytically without showing better recurrence detection or better care. Clinical sensitivity uses people with a defined condition as its denominator. It is not a unit conversion from laboratory LoD. A value below LoQ should not be used for an exact fold-change story without supporting quantitative performance.
Common confusions
- LoB concerns background response; it is not itself the lowest reliably detected target concentration.
- “Sensitivity” is ambiguous: ask whether the claim concerns a calibration slope, detection capability or clinical classification.
- Another assay's limit cannot be transferred across analytes, units, input or specimen preparation.
Try it
A fictional test has LoD95 of 2 units and LoQ of 8 units under its validated workflow. A sample produces a qualifying detected result near 4 units. Can its amount be reported as precisely established, then compared with 3 units as an exact 33% increase?
Answer: No. Detection can be supported while reliable quantitation is not. Preserve the qualification instead of interpreting the small numerical difference as an established biological change.
Explain it back
Which question does each of LoB, LoD and LoQ answer?
Takeaway
Keep the performance requirement, specimen and input beside every claimed limit.
Related concepts
Sources and scope
Source check: October 10, 2026. Fictional quantities; expert and learner review remain pending.
- CLSI harmonized terminology database — authoritative definitions, probability and accuracy requirements.
- CLSI EP17, second edition — evaluation and verification scope; public sample checked, full standard not accessed.
- FDA 2024 circulating tumor DNA (ctDNA) drug-development guidance — specimen equivalence, workflow and clinical-material confirmation for molecular assays.