ECE2014 Oral Communications Adrenal clinical (5 abstracts)
1Service dendocrinologie, Assistance Publique Hôpitaux de Paris, Hôpital Cochin, Paris, France; 2Institut Cochin, Inserm U1016, CNRS UMR8104, Université Paris Descartes, Paris, France; 3Departments of Pathology, Erasmus MC University Medical Center, Rotterdam, The Netherlands; 4Service doncogénétique, Assistance Publique Hôpitaux de Paris, Hôpital Cochin, Paris, France; 5Endocrinology and Diabetes Unit, University Hospital of Würzburg, Würzburg, Germany; 6Department of Surgery, Philipps University Marburg, Baldingerstrasse, Marburg, Germany; 7Charite University, Dept. of Medicine, Berlin, Germany; 8Medizinische Klinik und Poliklinik IV, Ludwig-Maximilians-Universität München, München, Germany; 9Endocrinology Unit, Department of Clinical Pathophysiology, University of Florence and Istituto Toscano Tumori, Florence, Italy; 10Department of Medicine, Endocrinology Unit, University of Padova, Padova, Italy; 11Department of Nuclear Medicine and Endocrine Oncology, Institut Gustave Roussy, Université Paris-Sud, Villejuif, France; 12Centre de référence des maladies rares de la surrénale, Service dEndocrinologieAssistance Publique Hôpitaux de Paris, Hôpital Cochin, Paris, France.
Background: The prognosis of adrenocortical carcinomas (ACCs) is heterogeneous. Genome-wide methylation analysis of tumor DNA identified a sub-group of ACCs from the French COMETE network with CpG island hypermethylation evoking a CpG island methylator phenotype (CIMP). These ACCs are associated with a poorer prognosis (Barreau, JCEM 2013).
The aim was to validate the prognostic value of CIMP in a large multicentric independent cohort from ENSAT (European Network for the Study of Adrenal Tumors).
Experimental design: The CpG island methylation was measured for 27 genes by methylation-specific multiplex-ligation-dependent probe amplification (MS-MLPA) using the ME002-B1 kit (MRC-Holland, Amsterdam). The best discrimination rule and thresholds were determined on the set of 50 ACCs previously studied by pangenomic methylation. The methylation was then measured on a validation cohort of 149 ACCs from 21 centers, using the same MS-MLPA kit. Survival was analyzed using Cox models.
Results: The best classifier for the CIMP was the mean methylation of 4 MS-MLPA probes (PAX5, GSTP1, PYCARD, PAX6), with a threshold of 12% of methylation.
In the validation cohort, the sex ratio (F/M) was 1.9. The median age was 49 years. Sixty-seven percent had hormonal hypersecretion. Fifty-eight percent were localized (ENSAT stages III) and 42% were locally advanced or metastatic (stages IIIIV). The median follow-up was 37 months.
The methylation level based on the 4-probes classifier was 21.6% (0 to 96%), comparable to the initial cohort (24.5%, P=0.51).
Hypermethylation was associated with decreased overall survival (HR=2.19 (1.283.76), P=0.0035) and event-free survival (HR=2.89 (1.824.57), P<105). The prognostic value of methylation on overall survival was independent of tumor extension (HR=2.02, P=0.011).
Conclusion: Tumor methylation measurement by MS-MLPA could provide a simple molecular tool for predicting the prognosis of ACCs.