E2413. Fully Automatic Volume Measurement of the Adrenal Gland on CT Using Deep Learning to Classify Adrenal Hyperplasia
Authors
Taek Min Kim;
Seoul National University Hospital
Seung Jae Choi;
Seoul National University Hospital
Ji Yeon Ko;
Seoul National University Hospital
Jeong Yeon Cho;
Seoul National University Hospital
Young-Gon Kim;
Seoul National University Hospital
Sang Youn Kim;
Seoul National University Hospital
Objective:
Adrenal hyperplasia is typically diagnosed by radiologists based on adrenal thickening, but no clear quantitative standard has been established. Therefore, efforts have been made to use adrenal volume to characterize various pathological conditions that can cause adrenal hyperplasia. However, manual segmentation of the adrenal gland in three dimensions is time-consuming and not feasible in clinical practice. In this study, an automatic segmentation algorithm for the adrenal gland was developed and validated on abdominal CT scans. Furthermore, we tried to diagnose adrenal hyperplasia using the predicted adrenal volume and patient characteristics (sex, height, and weight).
Materials and Methods:
This retrospective study evaluated automated adrenal segmentation in 308 abdominal CT scans from 48 patients with adrenal hyperplasia and 260 patients with normal glands from 2010 - 2021 (mean age, 42 years; 156 women). The dataset was split into training, validation, and test sets at a ratio of 6:2:2. Contrast-enhanced CT images and manually drawn adrenal gland masks were used to develop a U-Net-based segmentation model. Predicted adrenal volumes were obtained by five-fold splitting of the dataset without overlapping the test set. Adrenal volumes and anthropometric parameters (height, weight, and sex) were utilized to develop an algorithm to classify adrenal hyperplasia, using multilayer perceptron, support vector classification, a random forest classifier, and a decision tree classifier. To measure the performance of the developed model, the dice coefficient and intraclass correlation coefficient (ICC) were used for segmentation, and area under the receiver-operating characteristic curve (AUC), accuracy, sensitivity, and specificity were used for classification.
Results:
The model for segmenting adrenal glands achieved a dice coefficient of 0.7009 for 308 cases and an ICC of 0.91 (95% CI, 0.90 - 0.93) for adrenal volume. The normal group showed significantly higher ICCs for bilateral adrenal volume than the adrenal hyperplasia group (ICC = 0.91 and 0.83, respectively. p = 0.01). The models for classifying hyperplasia had the following results: AUC, 0.98 – 0.99; accuracy, 0.948 – 0.961; sensitivity, 0.750 – 0.813; and specificity, 0.973 – 1.000.
Conclusion:
A deep learning segmentation method can accurately segment the adrenal gland on CT scans in a heterogeneous dataset including normal adrenal glands and adrenal hyperplasia. Furthermore, the machine learning algorithm to classify adrenal hyperplasia using adrenal volume and anthropometric parameters (height, weight, and sex) showed good performance. The proposed segmentation algorithm may help clinicians identifying possible cases of adrenal hyperplasia.
