Radiologic characterization of adrenal masses: The role of computed tomography - Derived attenuation values

Fiemu E. Nwariaku, Julie Champine, Lawrence T. Kim, Shelby Burkey, Grant O'Keefe, William H. Snyder

Research output: Contribution to journalArticlepeer-review

21 Scopus citations


Background. Recent studies suggest that low computed tomography (CT) attenuation values can be used to differentiate benign adrenal adenomas from non-adenomas. We examined the utility of non-enhanced CT attenuation values of ≤ 10 Hounsfield units (HU) in differentiating adrenal adenomas from non-adenomas. Methods. During a 25-month period, patients who underwent adrenalectomy for adrenal masses visualized on abdominal CT were evaluated. Histopathologic examination of the excised adrenal mass was the gold standard for diagnosis. Patients who have undergone contrast-enhanced CT and those with bilateral adrenal masses were excluded. Statistical analysis was performed with the Student t test. Results. Fourteen patients met study criteria. Seven adenomas and 7 non-adenomas were identified. Mean CT attenuation values were 29 ± 18 HU. There were no differences in age or tumor weight and size between groups. Adrenal adenomas had significantly lower attenuation values compared with non-adenomas, 17 ± 14 HU versus 43 ± 11 HU, (P <.05). The sensitivity-to-specificity ratio for diagnosing an adrenal adenoma was 43%:100%, whereas the positive and negative predictive values were 100% and 64%, respectively. All adrenal masses with attenuation values of ≤ 10 HU were adenomas. Conclusions. The attenuation value of ≤ 10 HU on non-enhanced CT correctly identifies adrenal adenomas and is highly predictive of the diagnosis of adenoma.

Original languageEnglish (US)
Pages (from-to)1068-1071
Number of pages4
Issue number6
StatePublished - 2001

ASJC Scopus subject areas

  • Surgery


Dive into the research topics of 'Radiologic characterization of adrenal masses: The role of computed tomography - Derived attenuation values'. Together they form a unique fingerprint.

Cite this