12.3.3. Method quantitation limit and accuracy profile

The assessment of a method’s quantitation limit is based on the construction and interpretation of an accuracy profile to estimate the precision and reliability of the values. Three independent trials must be performed on three independent 10-fold serial dilutions, including two replicate RNA extractions for each level of dilution. For each dilution series and each target amount, various parameters are determined from estimated target amounts, such as the inter-series variance and the repeatability variance, the sum of both giving the reliability variance. The standard deviation of the reliability (SDrl) is then obtained by the square root of the reliability variance. The mean bias is determined (difference between the theoretical value and the mean of the observed values). To construct the accuracy profile, the lower and upper tolerance interval limits of the quantitation method are determined using the following formula:

mean bias +/- 2 × SDrl

and compared to the acceptability limits defined by the laboratory, e.g.  +/- 0.5 log10 (Blanchard et al., 2012). The tolerance interval limits of the accuracy profile have to be within the acceptability limits, validating the method for the thus defined calibration range. The quantitation limit of the method is then determined by the first level load of the validated calibration range. An example of the confidence and acceptability limits of an RT-qPCR calibration curve is given Fig. 8, where the evaluated method is validated for a calibration range between 103 and 106 copies, with a quantitation limit of 103 copies.

Fig. 8. Example of the mean bias, confidence interval, acceptability limits and quantitation limit for a RT-qPCR calibration curve. After Blanchard et al., 2012.

12116PN revised Fig 8