Quantitative IHC Assays May Bridge Gaps in Breast Cancer IHC Test Concordance

Improving the accuracy of FDA-approved immunohistochemistry (IHC) tests for breast cancer through the development of more quantitative assays is necessary given the divide between testing limitations and pathologists’ abilities to reproduce their results, and will likely require the incorporation of automated technology, according to David Rimm, MD, PhD.

“I think the solution is to move from reading to measuring,” Rimm said in an interview with OncLive® during the 2024 San Antonio Breast Cancer Symposium. “Reading means that pathologists are scoring [these assays] and measuring means that you’re using an analytic test.”

In the interview, Rimm discussed pathology limitations that have been exacerbated by the current lineup of breast cancer IHC tests, the growing importance of developing quantitative IHC assays, and ongoing research that may resolve some discrepancies in the concordance of results between current assays.

Rimm is the Anthony M. Brady Professor of Pathology and a professor of medicine (medical oncology) at the Yale School of Medicine, as well as director of the Yale Cancer Center Tissue Microarray Facility, director of Yale Pathology Tissue Services, and director of the Physician Scientist Training Program, Pathology Research, at the Yale Cancer Center in New Haven, Connecticut.

OncLive: What challenges have arisen for pathologists regarding the use of commercially available IHC tests?

Rimm: Recently, there have been several IHC tests approved that are not possible for the pathologist to perform. However, through various mechanisms, the companies that want to have as many drugs sold as possible have made tests that have been approved by the FDA. The pathologist is in the awkward position of having to perform a test that they may not be able to perform well, or maybe they can’t accurately reproduce, to get a drug prescribed.

[Pathologists] want to be team players. We don’t want to tell our oncologist colleagues that we can’t [perform these FDA-approved tests]. However, we know from the literature and from work in our lab [at Yale] and other labs around the world that it’s not humanly possible to reproducibly provide [some of the] results that are revealed by [some of the available] IHC tests.

How might pathologists work with medical oncologists to address issues of low IHC test concordance rates in practice?

Should we as pathologists play along, try to do the best we can, and use best practices? Best practices are what you use when you don’t have evidence. However, the evidence, [which can be found in several] different papers, indicates that pathologists’ [IHC tests are], at best, approximately 80% concordant. At worst, [they] are approximately 25% concordant. [Conversely], EGFR sequence mutations, a good biomarker, are 99% concordant. If I’m the patient, I want closer to 99% [concordance] than 80%, 60%, or 25% [concordance], so we try to use best practices. The pathologists who want to make [these tests] work because the FDA approved them, the drug companies [are promoting them], and our oncologist colleagues want [us to] provide accurate results. However, because they’re not so deep in the weeds on this, they might not know how inaccurate the tests are.

What should we do to solve that problem? One solution is to potentially try to move toward tests that are quantitative. Methods of quantitative assessment using automated technology are more highly concordant than pathologists’ readings.

What ongoing research may help improve IHC testing in the future?

Several organizations are trying to make a quantitative assay. One approach is to use the existing IHC assay with digitization and quantification of the optical density. Many groups are working on that.

Other groups are using artificial intelligence to try to figure out how much staining is in the assay. My group has used quantitative fluorescence, where we can now, instead of giving a [HER2] score of 0, 1, 2, or 3, give a concentration such as blood glucose, and return the data in attomoles per square mm. We’re still in the early stages of that. I’ve just started producing it and offering it in our CLIA lab. Ultimately, we hope that this will be licensed and be commonly available. [That way] we can bring pathology on slides to the level of pathology in the blood where it’s fully quantitative, we know the limits of detection and quantification, and it’s a true analytic test. That’s the solution to the nonconcordance of pathologists’ readings, and it’ll get pathologists out of the conundrum they’re in.

What is your main message for colleagues regarding the future of IHC testing in breast cancer?

I don’t think there’s any malice in this problem; it’s just a problem that has arisen. [However], the results of these assays not performing well and not being reproducible makes the whole [pathology] field look bad. I do not want to blame anybody, but [I want to] find a solution.