Clinical significance of a 10-mm cutoff size for adrenal lesions: a retrospective study with 547 non-oncologic patients undergoing adrenal computed tomography.

ABSTRACT

PURPOSE: To investigate the proportion of clinically significant adrenal lesions in patients with a subcentimeter adrenal lesion, and the sensitivity of a cutoff size of 10 mm on computed tomography (CT). METHODS: This retrospective study included consecutive 547 non-oncologic patients who underwent adrenal CT. Clinically significant adrenal lesions were defined as those that were biochemically abnormal (n = 99) or surgically resected according to the clinician's decision (n = 23). Long-axis diameters (LDs) and short-axis diameters (SDs) of the lesions were measured on CT by two independent readers. Likelihood of the focal lesion was analyzed using a five-point scale (1 = very low; 5 = very high). 66 Sensitivities for clinically significant lesions were analyzed according to cutoff size. Proportions of the clinically significant lesions for subcentimeter lesions were analyzed according to the visual score. RESULTS: Sensitivities for clinically significant lesions for cutoffs of 10, 15, and 20 mm were 93%, 79%, and 63% for LD and 85%, 61%, and 49% for SD for Reader 1 and 89%, 78%, and 65% for LD and 80%, 65%, and 48% for SD for Reader 2, respectively (p < 0.001 for 10 mm versus the other cutoffs). In subcentimeter lesions with visual scores of 1-3, the proportions of clinically significant lesions were 5.4% for LD or SD for Reader 1 and 6.6% for LD and 7.7% for SD for Reader 2, respectively. CONCLUSION: A lesion LD of ≥ 10 mm was a reasonable cutoff for determining adrenal abnormality. Subcentimeter lesions without visually high suspicion had a low risk of clinical significant lesions in our study cohort. Higher cutoffs significantly decreased sensitivity.