Impact of Using Median vs. Mean in Calculating ERBB2 FISH Results in Breast Cancer

Erb-b2 receptor tyrosine kinase 2 (ERBB2) testing is used to measure the status of ERBB2 gene expression and DNA amplification. Test results have been reported with a discrepancy as high as 20%. The purpose of this study was to improve ERBB2 fluorescence in situ hybridization (FISH) sensitivity by evaluating results generated by median as well as mean calculations.

We retrospectively identified breast cancer cases at our institution in which ERBB2 FISH testing was performed in-house from June 2018 to May 2020. FISH results were classified using the 2018 American Society of Clinical Oncology/College of American Pathologists guidelines: groups 1 and 5 are FISH positive and negative, respectively, and groups 2-4 are equivocal requiring additional work-up. FISH counting sheets were collected and regrouped by median ERBB2 copy number counts and median ERBB2/CEP17 ratio and compared with the mean ERBB2 and mean ERBB2/CEP17 ratio. Intra-tumor genetic heterogeneity and CEP17 copy number gain (CEP17 ≥3) were assessed to see if they affect the discrepancy between median and mean groups.

Seventy-two breast cancer cases were collected and evaluated. Eleven cases (11 of 72 [15%]) had discrepant grouping by mean and median calculations. A significant number of discrepancies were found for CEP17 copy number gain (p = 0.027) but not for ERBB2 (p = 0.411), the ERBB2/CEP17 ratio (p = 0.445), FISH results (p = 0.194), or genetic heterogeneity (p = 0.465). Among the four cases regrouped to median group 1, 2 were from mean group 3 and underwent anti-ERBB2 targeted therapy and 2 were from mean groups 4 and 5 may have benefitted from targeted therapy with more than 30% amplified cells. The median may be better to reflect the monosomy subclone within tumor tissues for the case 387 moved from mean group 5 to median group 2. The 6 cases moved from mean group 5 to median group 4 with CEP17 copy number gain may have had a poor prognosis based on other study result.

Including the median calculation may increase ERBB2 sensitivity and identification of CEP17 copy number gain. Further clinical studies are necessary to examine the outcome of including median in calculating ERBB2/CEP17 values.