Endocrine Abstracts

JOINT1284

Background: Ruling out malignancy in adrenal masses (AM) is a clinical challenge. We demonstrated that circulating cell-free DNA concentrations (ccfDNA-C) are higher in patients with adrenocortical carcinoma (ACC) compared to healthy subjects (HS). However, ccfDNA-C have not been compared among different types of AM.

Objectives: To assess the potential role of ccfDNA-C for AM differentiation.

Methods: We enrolled 97 adult patients (58 females) with a final diagnosis of adrenocortical adenoma (ACA, n=61), other benign AM (OB, n=7), ACC (n=17), pheochromocytoma (n=6) or adrenal metastases from other primary tumours (MET, n=6: 2 renal cell cancer, 1 melanoma, 1 unknown primary, 1 leiomyosarcoma, 1 papillary thyroid cancer). Blood samples, clinical, hormonal and radiological data were collected at first clinic review. AM with heterogeneous radiological appearance or plain Hounsfield Units >10 and not associated with overt adrenal hormone excess were labelled as “undefined AM” (n=35/92, 18 ACA, 4 ACC, 7 OB, 6 MET). ccfDNA was isolated with a commercial kit and ccfDNA-C were measured with fluorometer. Tumour size adjusted univariate analysis was conducted to assess ccfDNA-C distribution. We tested the diagnostic performance of our previously published HS-derived cut-off (>0.146 ng/μL) with logistic regression, positive (PPV) and negative predictive value (NPV) for ACC recognition.

Results: Unadjusted statistics showed higher ccfDNA-C in ACC than each group within the entire cohort (P=0.003 vs ACA, P=0.055 vs OB, P=0.001 vs MET, P=0.042 vs pheochromocytoma), but not within undefined AM (P=0.129 vs ACA, P=0.075 vs OB, P=0.004 vs MET). Tumour size-adjusted univariate analysis showed that ccfDNA-C were higher in ACC than each group in the entire cohort (P=0.039 vs ACA, P=0.005 vs OB, P=0.003 vs MET, P=0.021 vs pheochromocytoma) but also within undefined AM (P<0.001 vs ACA, P<0.001 vs OB, P<0.001 vs MET). In the entire cohort, the ccfDNA-C HS-derived cut-off was confirmed to predict ACC and showed Odds Ratio 14.982 (95% Confidence of Interval: 3.888–57.741), P<0.001. HS-derived cut-off predicted ACC with PPV=42.4% and NPV=95.3% in the entire cohort and with PPV=44.4% and NPV=100% in the undefined cohort. Approximately the same cut-off (≧0.148 ng/μL) was confirmed by Receiver-Operating Characteristic curve analysis showing sensitivity 82.4% and specificity 76.2% with same PPV and NPV as the HS-derived cut-off.

Conclusions: High ccfDNA-C in AM seem to be ACC-specific, and our ccfDNA-C HS-derived cut-off is useful for ACC discrimination. Further studies with larger cohorts of malignant adrenal lesions are needed to confirm our results.