Gene signatures and expression of miRNAs associated with efficacy of panitumumab in a head and neck cancer phase II trial

Publication

Gene signatures and expression of miRNAs associated with efficacy of panitumumab in a head and neck cancer phase II trial

Journal Paper/Review - May 28, 2018

Siano Marco, Espeli Vittoria, Mach Nicolas, Bossi Paolo, Licitra Lisa, Ghielmini Michele, Frattini Milo, Canevari Silvana, De Cecco Loris

Citation

Siano M, Espeli V, Mach N, Bossi P, Licitra L, Ghielmini M, Frattini M, Canevari S, De Cecco L. Gene signatures and expression of miRNAs associated with efficacy of panitumumab in a head and neck cancer phase II trial. Oral Oncol 2018; 82:144-151.

Type

Journal Paper/Review (English)

Journal

Oral Oncol 2018; 82

Publication Date

May 28, 2018

Issn Electronic

1879-0593

Pages

144-151

Brief description/objective

OBJECTIVE
Platinum-based chemotherapy plus the anti-EGFR monoclonal antibody (mAb) cetuximab is used to treat recurrent/metastatic (RM) head-neck squamous cell carcinoma (HNSCC). Recently, we defined Cluster3 gene-expression signature as a potential predictor of favorable progression-free survival (PFS) in cetuximab-treated RM-HNSCC patients and predictor of partial metabolic FDG-PET response in an afatinib window-of-opportunity trial. Another anti-EGFR-mAb (panitumumab) was used as the treatment agent in RM-HNSCC patients in the phase II PANI01trial. PANI01 tumor samples were analyzed using functional genomics to explore response predictors to anti-EGFR therapy.

MATERIALS AND METHODS
Whole-gene expression and real-time PCR analyses were applied to pre-treatment samples from 25 PANI01 patients. Three gene signatures (Cluster3 score, RAS onco-signature, microenvironment score) and seven selected miRNAs were separately analyzed for association with panitumumab efficacy.

RESULTS
Cluster3 expression levels had a profile with a significant bimodal separation of samples (P =  3.08 E-13). Higher RAS activation, microenvironment score, and miRNA expression were associated with low-Cluster3 patients. The same biomarkers were separately associated with PFS. Patients with high-Cluster3 had significantly longer PFS than patients with low-Cluster3 (median PFS: 174 versus 51 days; log-rank P = 0.0021). ROC analysis demonstrated accuracy in predicting PFS (AUC = 0.877).

CONCLUSIONS
Despite differences in clinical settings and anti-EGFR inhibitors used for treatment, response prediction by the Cluster3 signature and selected miRNAs was essentially the same. Translation into a useful clinical assay requires validation in a broader setting.