Sensitivity, Specificity, and Quality Controls in Forensic Immunoassays

Sensitivity in an immunoassay context refers to analytical sensitivity: the smallest amount of antigen that generates a signal the test can reliably distinguish from background. This is expressed as the LOD or LOQ, both of which are experimentally determined from repeated measurements of blank samples and low-concentration standards. A common approach is the signal-to-noise method: the LOD is set at the blank mean plus three standard deviations of blank replicates, and the LOQ at ten standard deviations. Other approaches use a calibration curve and define the LOD as the concentration corresponding to 10% of maximum signal in a competitive assay, or 10% above background in a sandwich ELISA.

In forensic practice, sensitivity requirements depend on the expected specimen condition. A semen identification assay applied to fresh reference samples needs to detect prostate-specific antigen (PSA) or semenogelin at concentrations present in a diluted stain extract. The same assay applied to a three-month-old dried stain from a hot climate may be working at concentrations an order of magnitude lower because degradation has reduced the available antigen. Validation studies should include spiked-substrate samples, aged stains, and samples processed through the laboratory's standard extraction protocol, not only fresh neat biological material. The LOD reported from fresh samples alone overstates operational sensitivity.

The distinction between qualitative and quantitative sensitivity also matters in casework reporting. A lateral-flow strip gives a binary result (present or absent above a threshold), and its sensitivity is the lowest concentration that consistently produces a visible line. An ELISA plate gives a quantitative absorbance value, and its sensitivity can be expressed as a calibration curve with a LOD and LOQ. When reporting a positive result from a lateral-flow strip, the analyst can state that the target antigen was detected at or above the validated threshold, not that it was present at a specific concentration. When reporting from a quantitative ELISA, the concentration can be stated if the sample was in the linear range of the calibration curve.

Specificity is measured by challenging the assay with a panel of substances that might be present in a forensic sample alongside the target antigen. For a human blood identification assay, this panel would include blood from common non-human species (dog, cat, pig, cow, horse, rabbit, chicken, and at least one non-human primate), other human body fluids (saliva, semen, vaginal secretion, urine, sweat), plant material, and common household substances that might contaminate a sample. The assay is run on each substance individually, and cross-reactivity is reported as a percentage of the positive control signal or as a positive/negative classification at a given concentration.

Inhibition is a separate specificity concern from cross-reactivity. Some substances do not generate a false-positive signal but instead suppress the true-positive signal, producing a false negative. Bleach is the most common example in forensic casework: hypochlorite denatures proteins and destroys epitopes, so a blood deposit cleaned with bleach may produce a negative result even if haemoglobin is detectable by alternative chemical methods. Inhibition is identified during validation by spiking known concentrations of the target into extracts prepared from substrate treated with the potentially inhibitory substance.

Forensic immunoassays target protein antigens: haemoglobin and glycophorin A in blood identification, semenogelin and PSA in semen identification, amylase in saliva identification, and species-specific serum proteins in precipitin tests. Proteins are susceptible to degradation by several mechanisms, each of which can alter or destroy the epitope that the antibody recognises.

Hydrolysis breaks peptide bonds between amino acids, fragmenting the protein. If the epitope spans multiple residues, fragmentation destroys it. Oxidation, driven by UV light and atmospheric oxygen, modifies amino acid side chains, particularly tryptophan, methionine, and cysteine residues that often contribute to antibody-binding surfaces. Microbial proteolysis adds enzymatic cleavage on top of physical degradation: bacteria and fungi present on the substrate or deposited from the environment produce proteases that can efficiently digest protein antigens before the sample reaches the laboratory. Temperature amplifies all three mechanisms, which is why warm, humid climates accelerate degradation.

The practical consequence is that the effective LOD for a field-collected stain can be 10 to 100 times higher than the validated LOD measured on fresh samples. This means an assay that reliably detects 1 nanogram per millilitre in validation may fail to detect 50 nanograms per millilitre in a badly degraded extract. Laboratories working routinely with aged material often run a parallel protein quantitation step before the immunoassay to assess whether sufficient extractable protein is present. If total protein is below a threshold, a negative immunoassay result is reported with a caveat about insufficient recoverable material rather than as a straightforward negative.

Validation for degraded samples is not universal across all laboratory protocols. SWGMAT guidelines recommend that validation studies include substrate controls (blank substrate of the same material as the case sample), aged stain samples, and mixtures, but the specific aging conditions (temperature, humidity, UV exposure time) are left to the laboratory to define based on their casework population. Laboratories in hot climates should validate under conditions that reflect local environmental exposure, not only the temperate-climate conditions used in kit manufacturer testing.

Quality controls in forensic immunoassays serve two purposes simultaneously: they confirm that the assay is performing within its validated parameters on the day of testing, and they provide the documented evidence that the analyst and the court need to assess the reliability of the results. Controls are not optional or a formality. They are the mechanism by which an analyst can claim that a result is valid.

A minimum control set for forensic casework includes three types. The reagent blank (sometimes called the buffer blank) contains only the assay buffer and no biological material, confirming that the reagents themselves are not generating signal. The negative control contains a biological matrix that is known to be free of the target antigen, prepared in the same way as case samples, confirming that the extraction and assay procedure does not produce false positives. The positive control contains the target antigen at a concentration that should reliably produce a defined signal, typically at or just above the LOQ, confirming that the assay reagents are active and the procedure was performed correctly.

• Reagent blank: buffer only, no biological material. Tests for signal from assay reagents themselves. Expected result: no signal.
• Negative biological control: substrate or matrix known to lack the target. Tests for false positives from extraction or matrix effects. Expected result: no signal above background.
• Positive control: known concentration of target antigen, processed identically to case samples. Tests that reagents and procedure are working. Expected result: signal within the validated range.
• Calibration standards (quantitative assays): a series of known concentrations used to construct the calibration curve. Required for ELISA and RIA; not applicable to lateral-flow strips.

If any control falls outside its acceptance criteria, the entire assay run is invalid. Case sample results from that run cannot be reported. The run must be repeated after the source of failure is identified and corrected. This is not a cautious overreaction: it is the only defensible position. A court that asks why only some controls were failed and some results were retained anyway will expose a methodological flaw that can undermine all results from the laboratory.

Validation is the documented process of demonstrating that an immunoassay method performs as intended for its specific forensic application. Three bodies set the dominant frameworks for forensic immunoassay validation. SWGMAT in the United States published guidelines for the validation of forensic biology methods that are widely referenced even though SWGMAT was formally dissolved in 2014 and its functions absorbed into OSAC (Organization of Scientific Area Committees for Forensic Science) within NIST. ENFSI, coordinating forensic science institutes across Europe, has published best-practice manuals for body-fluid identification that incorporate validation requirements. ISO 17025:2017, the international standard for testing and calibration laboratory competence, sets the accreditation framework within which both SWGMAT and ENFSI requirements must be implemented.

National accreditation frameworks apply these requirements locally. In the United Kingdom, the Forensic Science Regulator's Codes of Practice and Conduct (legally enforceable since 2021 under the Forensic Science Regulator Act 2021) require that all forensic science activities in the criminal justice system meet ISO 17025. In India, forensic laboratories accredited by the National Accreditation Board for Testing and Calibration Laboratories (NABL) operate under ISO 17025 requirements; the Bharatiya Sakshya Adhiniyam 2023 (which replaced the Indian Evidence Act 1872) does not specify immunoassay validation standards but requires that expert opinion evidence be based on validated methods. In the United States, the Uniform Language for Testimony and Reports (ULTR) published by OSAC sets the terminology analysts must use when reporting immunoassay results in court. The common thread across all jurisdictions is that the analyst must be able to demonstrate, from documented validation data, that the method was fit for the specific purpose it was applied to.

The principles of sensitivity, specificity, and controls apply to all immunoassay formats, but their practical implementation differs by format. Understanding the format-specific considerations allows an analyst to identify where performance problems are most likely to arise.

ELISA, whether sandwich or competitive format, is the highest-sensitivity format in routine forensic use. A sandwich ELISA for PSA (semen identification) can achieve an LOD in the picogram-per-millilitre range on fresh samples. Controls on an ELISA plate include the blank wells, the calibration standard series, and at least one positive and one negative biological control per plate. Because ELISA is a plate-based format, inter-plate variability is a documented source of imprecision: validation must include between-plate reproducibility studies, and calibration standards must be run on every plate, not assumed to carry over from a previous run.

Lateral-flow immunochromatographic strips, widely used for scene-side body-fluid screening (including semenogelin strips and haemoglobin strips), are qualitative by design. Their LOD is fixed by the manufacturer at the concentration that reliably produces a visible test line, typically in the nanogram-per-millilitre range. The built-in control line on a lateral-flow strip confirms that the antibody conjugate has migrated correctly, but it does not confirm that the sample was processed correctly or that the target was present in the extract before it was applied to the strip. Laboratories using lateral-flow strips in casework must run a separate positive control extract alongside case samples to confirm that the extraction procedure and strip performance are intact.

Precipitin tests for species identification use an anti-species antibody to identify the source of a blood or tissue stain. The sensitivity of precipitin tests (Ouchterlony double diffusion, single radial immunodiffusion) is lower than ELISA, typically in the microgram-per-millilitre range, and their performance is especially sensitive to sample degradation because precipitation requires sufficient intact antigen to form a visible precipitate band. Cross-reactivity between closely related species is a documented limitation: anti-human precipitin sera may react with non-human primate samples, and this cross-reactivity must be characterised in validation. The precipitation reactions topic covers the mechanics of these formats in detail.