A recently published online editorial in the Archives of Pathology & Laboratory Medicine highlights a hot topic in cancer care: To what lengths should we go to ensure consistency in immunohistochemistry (IHC) testing across laboratories? In practice, clinical IHC laboratories provide consistent reports on samples that are negative or strongly positive. However, IHC results differ among laboratories for tumor samples with biomarker expression levels between these two extremes.
The College of American Pathologists (CAP) HER2 proficiency testing data from 2022 (HER2A & B 2022) and 2023 (HER2A 2023) illustrate the point. Out of 60 samples from three IHC proficiency testing surveys, there was >95 percent interlaboratory agreement for the HER2 0/1 (negative) and HER2 3+ (positive) samples. However, not a single HER2 2+ sample achieved a minimum consensus for the set of samples used for scoring (>80 percent agreement). The literature is replete with many similar published studies.
The recent editorial’s authors, Barbarajean Magnani, MD, PhD, and Clive Taylor, MA, MD, DPhil, FRCP (Ir), FRCPath, examine the root cause by comparing and contrasting IHC to immunoassay testing (e.g., on serum or plasma). According to these experts, immunoassay testing generally achieves >99 percent analytic accuracy. IHC, in comparison, is in the 70–90 percent range. Why the difference?
Immunohistochemistry vs immunoassays
Clinical immunoassays are aligned globally through the use of calibrators and international reference standards. Immunoassay validation includes an assessment of linear range, limit of detection, and/or limit of quantification, as measures of analytical sensitivity. Laboratories verify day-to-day immunoassay consistency with Levey–Jennings charting and Westgard rules, a form of statistical process control. None of these quality assurance tools and methods exist in IHC. Magnani and Taylor explain that the current low levels of quality assurance aligns IHC with histologic stains rather than immunoassays.
Another difference between IHC and immunoassay is that the former is read by a pathologist while the latter is measured by an instrument. The readout on tissue samples in IHC involves microscopic evaluation of staining in the context of cells and tissues. The readout for immunoassays is from a homogeneous liquid in a cuvette.
Of course, accurate IHC testing depends on both the technical performance of the assay and the readout. In their editorial, Magnani and Taylor focus on the former as an important contributor to inter-laboratory discrepancies. Some of the best data on technical component variability are reported by major European proficiency testing agencies, such as NordiQC and UKNEQAS ICC & ISH, in their survey reports. They evaluate the IHC technical component via central review of participating lab-stained survey slides. Those slides, bearing serial sections of the exact same samples, often show striking variability in staining from one lab to another.
IHC as an immunoassay
IHC is at a fork in the road as a discipline of laboratory testing: Continue on the present path with 10–30 percent error rates or do something different?
Magnani and Taylor propose that the path forward is to start treating IHC as an immunoassay rather than a histologic stain. The same immunoassay quality assurance requirements, including calibration and statistical process control, should apply to IHC.
Their proposal would be a significant change for the industry, requiring discussion and phase-in over time. It would also require new quality assurance reagents and image analysis tools. Overcoming these hurdles is both possible and essential to ensuring that every cancer patient receives the best that modern medicine has to offer.