Histogram analysis in the differentiation between adrenal adenomas and pheochromocytomas: the value of a single...

To assess the diagnostic accuracy of histogram analysis on unenhanced computed tomography (CT) for differentiating between adrenal adenomas and pheochromocytomas (PCCs).

We retrospectively identified patients with proven PCCs who had undergone CT examinations between January 2009 and July 2019 at one of two institutions. For each PCC, we selected one or two adenomas diagnosed within two weeks of the date of diagnosis of the PCC. For each lesion, two readers scored the size, determined the mean attenuation, and generated a voxel histogram. The 10th percentile (P10) was obtained from the conventional histogram analysis, as well as being calculated with the following formula: P10 = mean attenuation - (1.282 × standard deviation). The mean attenuation threshold, histogram analysis (observed) P10, and calculated P10 (calcP10) were compared in terms of their diagnostic accuracy.

We included 52 adenomas and 29 PCCs. The sensitivity, specificity, and accuracy of the mean attenuation threshold were 75.0%, 100.0%, and 82.5%, respectively, for reader 1, whereas they were 71.5%, 100.0%, and 81.5%, respectively, for reader 2. The sensitivity, specificity, and accuracy of the observed P10 and calcP10 were equal for both readers: 90.4%, 96.5%, and 92.6%, respectively, for reader 1; and 92.3%, 93.1%, and 92.6%, respectively, for reader 2. The increase in sensitivity was significant for both readers (p = 0.009 and p = 0.005, respectively).

For differentiating between adenomas and PCCs, the histogram analysis (observed P10 and calcP10) appears to outperform the mean attenuation threshold as a diagnostic criterion.