Metastases or benign adrenal lesions in patients with histopathological verification of lung cancer: Can CT texture analysis distinguish?. Issue 138 (May 2021)

• Record Type: Journal Article
• Authors: Andersen, Michael Brun; Bodtger, Uffe; Andersen, Iben Rahbek; Thorup, Kennet Sønderstgaard; Ganeshan, Balaji; Rasmussen, Finn
• Is Part Of: European journal of radiology. Issue 138(2021)
• Publication Date: 2021-05
• Subjects: NSCLC non-small cell lung cancer -- SCLC small cell lung cancer -- 18-FDG 18-FlouroDexoyGlucose -- EUS endoscopic ultrasound -- EUS-B endobronchial ultrasound -- CTTA computed tomography texture analysis -- ML machine learning -- AI artificial intelligence -- ROI region of interest -- ICC intraclass correlation; Computed tomography -- Texture analysis -- Lung cancer -- Adrenal gland; Medical radiology -- Periodicals; Radiology -- Periodicals; Radiologie médicale -- Périodiques; Medical radiology; Periodicals; 616.075705
• DOI: 10.1016/j.ejrad.2021.109664
• Languages: English
• ISSNs: 0720-048X

Abstract

Highlights: In a lung cancer population, Adrenal CTTA cannot discriminate benign from malignant lesion, despite several CTTA parameters differed between metastatic and benign lesions. The optimal model based on multivariate logistic regression showed AUC of 73 %. The use of adrenal CTTA cannot be recommended in clinical decision making. Abstract: Introduction: Distant metastases are found in the many of patients with lung cancer at time of diagnosis. Several diagnostic tools are available to distinguish between metastatic spread and benign lesions in the adrenal gland. However, all require additional diagnostic steps after the initial CT. The purpose of this study was to evaluate if texture analysis of CT-abnormal adrenal glands on the initial CT correctly differentiates between malignant and benign lesions in patients with confirmed lung cancer. Materials and methods: In this retrospective study 160 patients with endoscopic ultrasound-guided biopsy from the left adrenal gland and a contrast-enhanced CT in portal venous phase were assessed with texture analysis. A region of interest encircling the entire adrenal gland was used and from this dataset the slice with the largest cross section of the lesion was analyzed individually. Results: Several texture parameters showed statistically significantly difference between metastatic and benign lesions but with considerable between-groups overlaps in confidence intervals. Sensitivity and specificity were assessed using ROC-curves, and in univariate binary logistic regression the area under the curve ranged from 36 % (Kurtosis 0.5) to 69 % (Entropy 2.5) compared to 73 % in the best fitting model using multivariate binary logistic regression. Conclusion: In lung cancer patients with abnormal adrenal gland at imaging, adrenal gland texture analyses appear not to have any role in discriminating benign from malignant lesions.