Article Text
Abstract
Introduction Our translational study aims to identify prognostic and actionable biomarkers in early-staged cervical cancer from Senticol III cohort.
Methods Tumoral slides of the first 150 randomized patients were analyzed based on SEDLIS criteria, TIL’s infiltration, PD-L1 expression and HPV status. Sequencing of tumor DNA was carried out using an in-house targeted next-generation sequencing panel of 571 genes. Relapse-free survival (RFS) was defined as the time between randomization and relapse of any type. We performed univariate and multivariate analysis to compare distributions of categorical variables.
Results The sequencing quality control criteria were met for 135 samples. PIK3CA was the most commonly altered gene (17%). Loss-of-function variants in KMT2-family genes were found in 16% of samples, and in 6% for ARID1A. There were 6 relapses, of which 4 were considered as low-risk based on SEDLIS criteria. KMT2C (p=0,018) and ARID1A (p < 0,001) pathogen variants were significantly associated with RFS. On multivariate analysis, alteration of ARID1A was independently associated with relapse. We found 32% of genomic alterations actionable for a target therapy.
Conclusion/Implications Genes involved in chromatin remodeling (ARID1A and KMT2C) are prognostic biomarkers of clinical interest in ESCC. On multivariate analysis, only alteration of ARID1A remains an independent biomarker predictive of relapse. Actionable molecular alterations were identified in 32% of the patients, highlighting the relevance of personalized therapeutic approaches including epidrugs in ESCC.