Opening the book…
The limit of detection (LOD), also called the detection limit, is the lowest level that can be reliably detected under defined conditions, without necessarily being quantified accurately. A result below this capability should not be described as proof that the analyte is absent.
The limit of quantitation (LOQ) is the lowest level that can be quantified with suitable accuracy and precision. It is generally higher than the detection limit. Results between these levels may support detection without a reliable numerical concentration, according to the method’s reporting rules.
The term sensitivity can refer to different concepts. In analytical calibration, it often refers to the change in response per unit change in analyte concentration, represented by the slope for a linear model. Clinical diagnostic sensitivity is a different concept. Avoid treating sensitivity and LOD as exact synonyms.
LOD and LOQ can be evaluated using suitable approaches such as signal-to-noise, low-level performance studies, or the response standard deviation and calibration slope. One approach described in ICH Q2(R2) uses the following estimates:
Estimated LOD = 3.3 × σresponse / |m|
Estimated LOQ = 10 × σresponse / |m|
In these equations, σresponse is an appropriate standard deviation of the analytical response and m is the calibration slope. The choice and estimation of response variability must be justified. Dividing by the slope converts response units to concentration units. Simply multiplying a blank’s standard deviation by 3 or 10 does not generally yield a concentration limit.
Calculated estimates need appropriate experimental confirmation. These equations are not universal substitutes for the procedure specified for a particular assay or regulatory application.
This page titled 15.9: Detection and Quantitation Limits was last modified on Mon, 09 Nov 2026 00:18:42 GMT and is shared under a CC BY-SA 4.0 license and was authored, remixed, and/or curated by Emalee MacKenzie.
The book could not be loaded just now.