Enhanced HER2 status detection in breast and gastric cancers using surrogate DNA methylation markers

Abstract

There is a limited understanding of specific DNA methylation patterns associated with HER2 overexpression in breast and gastric cancers. Here we aim to solve the problem using inferred DNA methylation markers. DNA methylation data from The Cancer Genome Atlas (TCGA) were analyzed for breast and gastric cancers regarding HER2 status. We further applied a targeted bisulfite sequencing approach to elaborate the DNA methylation profile of the HER2 region, covering 7635 CpG sites. Based on these two sets of data, we selected specific DNA methylation markers inferring HER2 status for both breast and gastric cancers and validated their performance in assisting HER2-status determination on a retrospective cohort with 496 breast cancer and 372 gastric cancer. HER2-Meth could well distinguish HER2 IHC0/1+ from HER2 IHC3+ cases in both breast cancer (AUC = 0.983, n = 130) and gastric cancer (AUC = 0.974, n = 63), also could effectively discriminate HER2 IHC2+/FISH+ from HER2 IHC2+/FISH- cases in equivocal situations for both breast cancer (test set AUC = 0.879, n = 74; validation set AUC = 0.875, n = 75) and gastric cancer (test set AUC = 0.910, n = 70; validation set AUC = 0.941, n = 71), outperforming regular HER2 copy number test (An AUC of 0.793 for breast cancer and an AUC of 0.759 for gastric cancer) on HER2 IHC2+ cases. Furthermore, HER2-Meth demonstrated its potential for stratifying HER2-positive patients, enabling predictions regarding overall survivals, and the potential benefits of HER2-targeted therapies in breast cancer. The strong agreement observed between the methylation qPCR test and the results of IHC and FISH indicates significant potential for this approach as a complementary tool in guiding HER2-targeted therapies for patients with breast and gastric cancers.