AI Summary & Key Takeaways
AI InsightHUmin, HUmean, HUmedian, percentile values, and variance can identify adrenal masses.
📌 3 Key Takeaways
Abstract
Aims: The rate of adrenal mass detection has increased due to the development of imaging modalities. It is vital to differentiate benign adrenal adenomas from other adrenal masses in order to establish whether an active management strategy is essential. Volumetric CT histogram analysis calculates the percentage of covered pixels in the negative attenuation region. The goal of this research was to evaluate the diagnostic utility of volume histogram analysis for adrenal tumors confirmed histopathologically as well as the ideal slice thickness for CT histogram analysis to differentiate between benign and malignant lesions with a density greater than 10 Hounsfield units (HU). Methods: The research analyzed the CT images of 127 individuals with 136 adrenal masses that were verified histopathologically after resection (57 lipid-poor adenomas, 21 pheochromocytomas, 47 metastases, and 11 adrenocortical carcinomas). For imaging, a 40-row MDCT device (Siemens Medical Solution, Erlanger, Germany) was utilized. 1 mm and 5 mm unenhanced CT images were obtained. Two separate radiologists manually assessed the Hounsfield units (HU) of the masses. The 5th to 95th percentiles of HU values, as well as the minimum, mean, and maximum values, skewness, kurtosis, and variance, were calculated. Interobserver agreement was determined by means of the interclass correlation coefficient (ICC). Results: The HU parameters for the malignant group were all higher than those of the benign group, and the difference in the 5 mm slice thickness was more significant than the 1 mm slice thickness. The difference between HUmin (P=0.007), HUmean and HUmedian (P <0.001), 5th to 50th (P <0.001), 75th (P=0.004), 90th (P=0.016), and 95th (P=0.049) percentiles was statistically significant. The malignant group had higher skewness and kurtosis than the benign group, while the benign group had higher variance. Statistically, the disparity between the variances was significant (P=0.046). The area under the curve (AUC) of the 25th percentile of the HU value was the highest (AUC=0.932; cut-off value=15; sensitivity=90.0%; specificity=85.7%). Conclusion: Noninvasive volumetric CT histogram analysis can detect malignant adrenal masses from benign tumors before an operation. Histogram analysis benefits from thicker slices. HUmin, HUmean, HUmedian, percentile values, and variance can identify adrenal masses.