Predicting high proliferative Ki-67 index in patients with adrenocortical carcinoma based on texture analysis of contrast-enhanced computed tomography images: a cross-sectional study

BACKGROUND: Adrenocortical carcinoma is characterized by a high risk of aggressive disease progression and limited effectiveness of available treatment. Therefore, early diagnosis and assessment of its potential development are crucial. Although computed tomography is highly accurate in detecting the structural characteristics of tumors, its ability to predict adrenocortical carcinoma progression is unclear. AIM: This study aimed to evaluate the accuracy of using texture analysis of contrast-enhanced computed tomography images in predicting a high index of proliferative activity (Ki-67) in patients with adrenocortical carcinoma. METHODS: The study examined four-phase contrast-enhanced computed tomography images of patients with histologically verified adrenocortical carcinoma (retrospective part) and reevaluated these images (prospective part). Computed tomography image analysis included labeling a tumor lesion, assessing and post-processing texture features, decreasing dimensions, and performing a cluster analysis to evaluate the discriminatory ability of computed tomography texture parameters. The predictive accuracy of texture analysis of computed tomography images was assessed based on its ability to identify increased proliferative activity (KI-67 10%). The index was derived from the results of immunohistopathological testing of adrenal tissue samples obtained during surgery. RESULTS: Texture analysis of contrast-enhanced computed tomography images of 24 patients with histologically diagnosed adrenocortical carcinoma was performed: Ki-67 ≤ 10% in 9 patients and Ki-67 10% in 15 patients. The analysis revealed a statistically significant association (p = 0.015, Fisher’s exact test) between two fuzzy clusters based on texture features and the Ki-67 classification of patients with adrenocortical carcinoma. These results indicate the ability to predict high proliferative activity with a 0.05 level of significance. CONCLUSION: Texture analysis of contrast-enhanced computed tomography images of patients with adrenocortical carcinoma enables noninvasive assessment of tumor progression risk following surgical removal.