Method Validation Edition

Detection capability software for method validation LoB, LoD and LoQ estimation — parametric and non-parametric approaches, probit regression and precision profile variance function methods. Covers the EP17-A2 study designs and the older approaches laboratories still use.

Download 15-day trial Buy from US$ 475

Every feature from all five editions for 15 days.
Method Validation edition from US$ 475 a year · 30-day money-back guarantee.

Microsoft Excel with the Analyse-it tab selected, showing the Detection Capability section of the estradiol report from CLSI EP17-A2 Appendix A: the Detection Decision table with the critical value (LoB) from the blank material, the Detection Limit table with the LoD from the pooled SD of the low-level material, the frequency density histogram of the blank and low-level distributions with the LoB and LoD lines drawn on it, and the task pane open. Handwritten notes: Runs inside Excel: every analysis is on the Analyse-it tab; Limit of blank and limit of detection, each from its own material; Blank and non-blank distributions with LoB and LoD marked; LoB and LoD methods, alpha and beta; The report is an ordinary Excel worksheet: share it, archive it, open it on any PC with Excel.
clsi logo CLSI
EP17-A2

Know what your assay can and cannot detect

We use Analyse-it frequently for our verification and pre-verification work, in accordance with CLSI guidelines for in-vitro diagnostics. It’s saved time and effort compared to the hodge-podge of applications we used before, JMP, SAS, etc...
Brian Noland, Ph.D.
Principal Scientist, Product Development
Biosite / Inverness Medical Innovations

Excel’s STDEV and PERCENTILE functions give the SD or a quantile of a set of blank results and stop. No limit of detection from the low-level samples, no variance function fitted to a precision profile, no probit regression. No plot of the blank and low-level distributions against the limits. Low-concentration results drive clinical decisions — ruling out disease, screening, monitoring drug levels. A measurement procedure that cannot reliably separate a true signal from blank noise reports false positives or misses genuine low-level results. Set the limits too high and genuine low results go unreported; set them too low and confidence in the assay is lost.

Where does blank noise end? What is the lowest concentration the procedure reliably detects? Where is the measurement uncertainty small enough for the result to be clinically actionable? EP17-A2 defines the framework: LoB, LoD and LoQ. Analyse-it covers the full EP17-A2 workflow, with more than one estimation approach for the LoB and LoD. Precision profiles from EP05-A3 studies feed directly into the variance function methods — no re-entering of data, no separate tools for LoB, LoD and LoQ.

Parametric and non-parametric limit of blank from blank material

The limit of blank is the noise floor: the highest result expected from a sample that contains no analyte. Estimate it parametrically from the SD of the blank measurements, or non-parametrically from their quantile. The blank and low-level replicates of an EP17-A2 study go straight into the analysis. The same design fitted to a second reagent lot gives that lot its own limits.

  • Parametric (SD of blank material)
  • Non-parametric (quantile of blank material)
  • LoB set to a known value
Microsoft Excel showing the Detection Capability section of the estradiol report for the second reagent lot from CLSI EP17-A2 Appendix A: the Detection Decision and Detection Limit tables, the frequency density histogram with the LoB and LoD lines, and the task pane open on the LoB estimation options. Handwritten notes: LoB from the blanks, LoD from the non-blank material, each with its table; Histogram of each material with LoB and LoD marked; LoB estimator: normal quantile or non-parametric.
Estradiol, second reagent lot: the LoB and LoD tables and the frequency density histogram, with the LoB method chosen in the task pane.

Limit of detection from the pooled SD of low-level samples

The limit of detection is the lowest concentration reliably distinguished from blank. Following EP17-A2, the LoD comes from the LoB and the pooled SD of the non-blank, low-level materials. The frequency density histogram overlays the blank and low-level distributions with the LoB and LoD lines. The separation between blank noise and true signal is then visible, not only tabulated.

  • Pooled SD of non-blank materials
  • Frequency density histogram with LoB and LoD

LoB and LoD from the precision profile variance function

When an EP05-A3 precision study already covers the low end of the measuring range, the detection limits come from it and no separate blank study is needed. One of seven variance functions — constant, mixed constant/proportional, the 3-parameter alternative power model and others — is fitted to the precision profile. The LoB and LoD are read from the fitted function. The same precision data then supports both the precision claim and the detection capability claim.

  • LoB from precision profile variance function
  • LoD from precision profile variance function
Microsoft Excel showing the New Marker report from CLSI EP17-A2 Appendix B: the 3-parameter alternative power variance function fitted to the precision profile of six pools, its parameter estimates, and the task pane open on Detection Capability with the LoD taken from the precision profile function. Handwritten notes: Variance function fitted to the precision profile: SD against concentration; 3-parameter alternative power model: beta0, beta1 and J; LoD: the SD from the precision profile function.
New Marker, EP17-A2 Appendix B: the 3-parameter alternative power variance function fitted to six pools; the LoD takes its SD from the fitted function.

Probit regression: LoD at a 95% probability of detection

For molecular and immunoassay methods where each result is detected or not detected, detection is probabilistic. Probit regression fits the detection rate against concentration across a dilution series. The LoD is the concentration at which the probability of detection reaches the required threshold, such as 95%. Fit each reagent lot separately, as EP17-A2 Appendix C does.

  • Probit regression
  • Probit regression curve
  • Multiple inverse predictions with confidence intervals new in v5.65
Microsoft Excel showing the probit regression report for reagent lot 1 from CLSI EP17-A2 Appendix C: the plot of detection rate against concentration on a log scale, the Fit Probit table with the parameter estimates and their 95% confidence intervals, the Predict for given Probability table with the concentration at a 95% probability of detection, and the task pane open. Handwritten notes: Probit fit: probability of detection against concentration; LoD: the concentration at a 95% probability of detection; Detection probability, set here.
Probit regression for reagent lot 1: detection rate against concentration, the parameter estimates and the concentration giving a 95% probability of detection.

Limit of quantitation from the precision profile variance function

The limit of quantitation is where a result becomes a number worth reporting, not only a detection. The LoQ is the concentration at which imprecision drops below the threshold for clinically reliable quantitative results, such as a CV of 10% or 20%. The estimate comes directly from the precision profile variance function. The analysis takes the EP05-A3 precision data directly.

  • LoQ from precision profile variance function
Microsoft Excel showing the Troponin I precision profile report from CLSI EP17-A2 Appendix D: the plot of CV against concentration with the fitted constant variance function, its parameter estimate, the range of the fit and the Inverse Prediction table giving the concentration at a 10% CV, with the task pane open. Handwritten notes: Precision profile: CV against concentration, with the fitted variance function; LoQ: the concentration at the chosen CV, by inverse prediction; Predict U given CV: the CV, typed in.
Troponin I: the precision profile with its fitted variance function, and the LoQ by inverse prediction at a 10% CV.

Example analyses

See detection capability results in detail — LoB, LoD, probit regression and LoQ — using CLSI example datasets you can download and follow along with.

EP17 A2 Example 1 2 pages EP17-A2 — Appendix A
Estradiol detection capability, two reagents. 5 blank and 5 low-level samples per reagent, 60 results each. Within-run and total precision for every sample, the limit of blank from the blank distribution and the limit of detection from the pooled SD of the low-level samples. A bi-histogram sets the blank against the low-level material.
EP17 A2 Example 2 4 pages EP17-A2 — Appendix B
New Marker, limit of detection from a precision profile. Six pools per reagent across the measuring interval. The profile is fitted with a three-parameter alternative variance function for one reagent and a mixed constant and proportional function for the other. The limit of detection is then read from the profile against a known limit of blank.
EP17 A2 Example 3 2 pages EP17-A2 — Appendix D
Troponin I, limit of quantitation. Nine pools per reagent, 720 results each. Precision profile fitted with a constant variance function, then inverse prediction of the concentration at which the CV meets the goal — the limit of quantitation.
EP17 A2 Example 4 3 pages EP17-A2 — Appendix C
Detection capability by probit analysis. A dilution series of detected and not-detected results, fitted separately for three reagent lots. Probit regression against log concentration with parameter estimates and Wald CIs, a likelihood ratio test for the model and the concentration detected with 95% probability.

Part of measurement system analysis

Detection capability is one part of measurement system analysis, alongside precision (EP05-A3), linearity (EP06-Ed2), bias/trueness verification (EP15-A3) and preliminary evaluation (EP10-A3-AMD). Precision profiles from EP05-A3 feed directly into detection capability estimation.

Related guides in the Learn section: LoB, LoD and LoQ explained, the LoQ from a precision profile and probit analysis for the LoD. A further guide covers performance studies for an IVD 510(k).

Software you can trust

Validated calculations you can defend at inspection Every calculation is performed by Analyse-it — no Excel formulas, no third-party functions. Results are validated against CLSI reference datasets, published datasets and thousands of internal test cases before every release. How Analyse-it is developed and validated →
Data stays in your facility Analyse-it runs entirely within Microsoft Excel on your PC. No cloud processing, no data transmission. Pre-submission data, data derived from patients and in-process results stay within your facility under your own data governance controls.
Standard Excel workbooks anyone can open Every analysis is an ordinary .xlsx workbook. Share with colleagues, submit to regulatory affairs, archive for audit, open on any PC with Excel. No proprietary format, no licence required to view results. Colleagues and auditors see exactly what you see.
Results that cannot be accidentally broken Analysis output contains computed values, not formulas. Nothing to accidentally overwrite, no cell references to break, no formula errors to introduce. The results you reported are exactly what you will find when you reopen the workbook months or years later for an audit.

Free trial and pricing

Try it on your own data first. The 15-day trial is every feature from all five editions — install it and start straight away.

Download 15-day trial See pricing or get a quote →

Method Validation edition: US$ 475 per year or US$ 1155 for a perpetual licence. Every purchase carries a 30-day money-back guarantee. Need a quote for purchasing? Add the licence to the cart and save it as a PDF quote.