Impact of HER2 assay sensitivity on the ability to detect tumors with low to ultra-low HER2 expression: A TMA study on more than 10,000 tumors from more than 100 tumor entities.

Abstract

3137 Background: Detection of low level HER2 expression (HER2 low) is a prerequisite for therapy with new HER2 inhibitors in advanced breast cancer. Currently, tumors with immunohistochemical 1+ expression, and possibly also those with lower expression, are considered HER2 low. Therefore, the question of optimal sensitivity of HER2 assays is of increasing importance for clinical HER2 testing. In addition, there is hope that tumor types other than those with typical 3+ positivity may also benefit from these therapies. The aim of this study was to systematically investigate the influence of HER2 immunohistochemistry assay parameters on the detection rate of tumors with HER2 low expression in different tumor types. Methods: A tissue microarray containing >10,000 tissue samples from more than 100 different tumor types and subtypes was analyzed for immunohistochemical expression of HER2 using different immunohistochemistry protocols and anti-HER2 antibodies, including the HercepTest and laboratory developed tests designed for high sensitivity. HER2 IHC evaluation included the recording of HER2 low (1+), ultralow (any staining, less than 1+), and 0 (completely negative). Results: Irrespective of the assay sensitivity all assays showed the expected/tolerable association with HER2 fluorescence in situ data in a cohort of 356 breast cancers. For the HercepTest, HER2 amplification was seen in 93% of 28 3+ cases, 55% of 11 2+ cases and in 1.5% of 274 0 and 1+ cases. For the most sensitive assay, amplification was seen in 90% of 30 3+ cases, 19% of 37 2+ cases and in 0.8% of 249 0 and 1+ cases. In non-breast cancers, the use of different assays resulted in highly variable results regarding the range of tumors with low or ultra-low HER2 expression in different tumor entities. For example, low (1+) and ultra-low (+) expression was observed in 7-24% (1+) and 11-28% (+) of serous ovarian cancers, 6.2-24% (1+) and 3.4-19% (+) of endometrioid endometrial cancers, 5.5-14% (1+) and 7.7-13% (+) of intestinal gastric cancers, 1.8-9.2% (1+) and 4.5-11% (+) of ductal adenocarcinomas of the pancreas, 19-33% (1+) and 5.5-11% (+) of muscle invasive urinary bladder cancers, 2.9-16% (1+) and 4.3-12% (+) of adenocarcinomas of the colon, 0-7.7% (1+) and 2.4-4.6% (+) of clear cell renal cell carcinomas, 4.8-6.5% (1+) and 6.0-8.6% (+) squamous cell carcinomas (SQCC) of the larynx, 6.3-7.1% (1+) and 2.4-6.3% (+) SQCC of the esophagus, 0.8-26% (1+) and 2.4-12% (+) SQCC of the cervix, as well as 1.5-5.1% (1+) and 1.5-2.2% (+) SQCC of the vulva. Conclusions: Low level HER2 expression occurs commonly in many cancer types. The rate of identifiable tumors with HER2 low or HER2 ultra-low expression greatly varies depending on assay parameters that can easily be modified. Clinical trials are needed to identify the HER2 expression thresholds to identify patients who may benefit from anti-HER2-low therapies.