How accurate is computed tomography in predicting the real size of adrenal tumors? A retrospective study

The ability to accurately assess tumor size is an important consideration during the preoperative evaluation of adrenal tumors, particularly solid nonfunctioning masses (incidentalomas or adrenalomas). Does the histological size of the adrenal tumor correspond to the preoperative computed tomography (CT)-estimated size?

To evaluate the accuracy of CT in predicting the real size of adrenal tumors.

Retrospective review of all clinical records of patients who underwent adrenalectomy from 1984 through 1995. The mean tumor size reported from CT examinations was compared with the corresponding size obtained from the pathology reports and the statistical difference was evaluated.

University and private hospitals in Athens, Greece.

Seventy-six patients who underwent adrenalectomy for various adrenal diseases and who met strict entry criteria.

For the entire population, the mean diameter of the tumors was estimated (CT reports) at 4.64 cm, but the real value (pathology reports) was 5.96 cm (P < .001). The underestimation held true for all of the studied subgroups that were defined by the different proposed cutoffs for malignancy. Three patients were incidentally found to have adrenal cancer, with the tumors measuring from 2.6 to 2.9 cm on CT. In addition, 4 pheochromocytomas were clinically and laboratory "silent" at the time of their discovery. The regression line (y = 0.85 + 1.09x) relating CT-estimated and histological tumor size was linear (r = 0.90, P < .001).

Computed tomography underestimates the real size of adrenal tumors. The CT-estimated value should be corrected accordingly to obtain the real size. The size of an adrenal tumor, even when corrected, cannot predict the tumor's clinical behavior in many cases. Surgeons should always cautiously interpret the proposed diagnostic cutoffs, especially when considering surgical or conservative management of small nonfunctioning adrenal tumors.