Novel visualized quantitative epigenetic imprinted gene biomarkers diagnose the malignancy of ten cancer types.

BACKGROUND: Epigenetic alterations are involved in most cancers, but its application in cancer diagnosis is still limited. More practical and intuitive methods to detect the aberrant expressions from clinical samples using highly sensitive biomarkers are needed. In this study, we developed a novel approach in identifying, visualizing, and quantifying the biallelic and multiallelic expressions of an imprinted gene panel associated with cancer status. We evaluated the normal and aberrant expressions measured using the imprinted gene panel to formulate diagnostic models which could accurately distinguish the imprinting differences of normal and benign cases from cancerous tissues for each of the ten cancer types.

RESULTS: The Quantitative Chromogenic Imprinted Gene In Situ Hybridization (QCIGISH) method developed from a 1013-case study which provides a visual and quantitative analysis of non-coding RNA allelic expressions identified the guanine nucleotide-binding protein, alpha-stimulating complex locus (GNAS), growth factor receptor-bound protein (GRB10), and small nuclear ribonucleoprotein polypeptide N (SNRPN) out of five tested imprinted genes as efficient epigenetic biomarkers for the early-stage detection of ten cancer types. A binary algorithm developed for cancer diagnosis showed that elevated biallelic expression (BAE), multiallelic expression (MAE), and total expression (TE) measurements for the imprinted gene panel were associated with cell carcinogenesis, with the formulated diagnostic models achieving consistently high sensitivities (91-98%) and specificities (86-98%) across the different cancer types.

CONCLUSIONS: The QCIGISH method provides an innovative way to visually assess and quantitatively analyze individual cells for cancer potential extending from hyperplasia and dysplasia until carcinoma in situ and invasion, which effectively supplements standard clinical cytologic and histopathologic diagnosis for early cancer detection. In addition, the diagnostic models developed from the BAE, MAE, and TE measurements of the imprinted gene panel GNAS, GRB10, and SNRPN could provide important predictive information which are useful in early-stage cancer detection and personalized cancer management.