Histogram analysis in the differentiation of adrenal adenomas from pheochromocytomas: the value of a single measurement.
Number:27ESURABS0055
Type:Scientific Electronic Presentation
Authors: Ana Paula Teixeira, William Haddad Jr, Luan Barreto, Andre Secaf, Livia Mermejo, Fabiano Lucchesi, Carlos Augusto Molina, Jorge Elias Jr, Valdair Muglia
Keywords:Adrenal incidentalomas, pheochromocytomas, computed tomography
SECTIONS
Objectives

Objective

We aimed to assess the diagnostic accuracy of the histogram analysis (HA) using a single measurement from unenhanced computed tomography (CT) images for the differentiation between adrenal adenomas from pheochromocytomas (PCCs).

Flow chart of the study
Material and methods

Study population

This study was conducted in two distinct centers after approval by both institutional committees on human research with a waiver for written informed consent due to its retrospective nature. This work used the Radiologic Information System and Pathological Electronic data. Two radiologists (not involved with imaging analysis) searched for “pheochromocytomas” in both datasets and retrospectively identified the cases of all patients with such diagnosis proved by biopsy and/or surgery and histopathological confirmation from January 2009 to July 2019. Based on the date of PCCs, we retrieved up to two adenomas for each PCC within two weeks apart to prevent any chronologic bias starting from the closest date. This interval was set to ensure that patients were examined with the same contemporary CT.

Adenomas were confirmed if there was a...

...read more

Figure 2. 53-year-old female patient. A. A post-contrast, venous phase, CT axial image demonstrating a 2.8-cm homogenous right adrenal lesion. B. An unenhanced CT axial image at the same level with an ROI drawn at the center of the lesion showed a mean attenuation of 24.7 HU that could be a LPA (but no criteria at this point). However, at histogram analysis and calcP10 there was more than 10% of negative voxels suggesting the diagnosis of a LPA. Patient is asymptomatic, and the lesion has been stable for 62 months.
Figure 3. A 40-year-old male patient with symptoms and laboratory findings of pheochromocytoma. A. A post-contrast, arterial phase, CT image showing a hyper-vascular 1.5-cm lesion on left adrenal gland (arrow) with a central cystic area. B. An unenhanced CT axial image at the same level with an ROI drawn to avoid the cystic area. The mean attenuation is 12.5 HU. The histogram analysis and calc P10 showed more than 10% negative voxels. Aside from the clinical and laboratorial setting, this could be assumed to be an adenoma, based on HA and calc P10 criteria, were it not for the heterogeneity of the lesion.
Results

The demographic data are shown in Table 1. The mean age of patients with PCCs was significantly younger than those with adenoma (p=0.001). There was no significant difference for gender and laterality of lesions between the two groups.  Reader 1 measurements indicated that 25.0% (13/52) of adenomas were lipid-poor (mean attenuation > 10 HU); this proportion was 30.7% for reader 2 (16/52).

When assessing the imaging parameters on CT images (Table 2), the PCCs were significantly larger than adenomas. The largest axis was twice as large as those from adenoma (2.53 vs 5.32 cm for reader 1; 2.42 vs 5.03 cm for reader 2; p<0.0001 for both). Only three adenomas were larger than 4.0 cm (5.8%) whereas 12 PCCs measured less than 4.0 (41.4%). The mean attenuation value on unenhanced images was also significantly different between adenomas and PCCs: 4.84 vs 36.58 HU for reader...

...read more

Table 1. Demographic and clinical data
Table 2. Imaging parameters. Comparison between Adenomas and Pheochromocytomas group
Conclusions

Our study indicated that histogram analysis either using voxel counting or the 10th percentile formula may help differentiate LPA adenomas from PCCs using nonenhanced images without penalizing specificity. This criterion, if further confirmed by other studies, can be a powerful adjuvant technique when assessing adrenal incidentalomas by CT images.

 

We acknowledge some limitations. First, it is a retrospective study with an increased risk of biases especially selection bias, as cited for PCCs. We aimed to minimize the risk by using adequate inclusion and exclusion criteria. Second, there were a limited number of cases, which reflects the relatively low incidence of PCCs. Although adrenal adenomas are common lesions, we included a small portion of them in a proportion of 2:1 to not induce significant discrepancy between the two groups, which could be up to 50...

...read more