Hounsfield unit thresholds in differentiating adenoma from nonadenoma adrenal masses: a retrospective adrenalectomy-based study
Abstract
Background/aim: Adrenal incidentalomas are increasingly detected during imaging studies performed for various indications, and their frequency in endocrinology practice has risen in recent years. While adenomas account for approximately 80% of adrenal incidentalomas, the remaining 20% include lesions such as primary malignancies, metastases, and pheochromocytomas that should not be overlooked. Given the risks associated with adrenal biopsy, accurate diagnosis with noninvasive methods is essential. In this retrospective study, we evaluated the Hounsfield unit (HU) values and lesion sizes measured on noncontrast computed tomography (CT) to identify a more accurate threshold than the widely used 10 HU. Materials and methods: In this retrospective study, patients who underwent unilateral adrenalectomy between 2013 and 2022 were evaluated. Noncontrast CT images were assessed for lesion size and attenuation, and statistical analyses included tests for nonnormally distributed data, logistic regression, and ROC analysis to determine the diagnostic performance of mass size and density in differentiating adenoma from nonadenoma lesions. Results: A total of 106 patients who underwent adrenalectomy for various indications were included, of whom 55 had adenomas and 51 had nonadenoma masses. The mean size of the adenoma versus nonadenoma lesions was 37.2 ± 15 mm and 55.8 ± 30 mm, respectively (p < 0.05), while mean HU values were 11.4 ± 17 and 35.9 ± 14, respectively (p < 0.05). The ROC analysis showed that HU values exhibited strong discriminative ability in distinguishing adenoma from nonadenoma masses (AUC = 0.869). According to Youden’s index, 20 HU was identified as the optimal threshold value. The sensitivity/specificity values for 10 HU, 15 HU, and 20 HU were 98%/56%, 98%/61%, and 92%/82%, respectively. Conclusion: Within this selected adrenalectomy cohort, adopting a 20 HU threshold instead of the conventional 10 HU value significantly improved specificity in differentiating adenomas from nonadenomas. However, broader prospective validation in unselected incidentaloma populations is warranted before recommending a definitive change in routine clinical guidelines.