The Effectiveness of Computed Tomography Texture Analysis in Distinguishing Wilms Tumor from Neuroblastoma

Abstract

Objective: This study aimed to determine the effectiveness of computed tomography (CT) texture analysis parameters in distinguishing Wilms tumor (WT) from neuroblastoma (NB). Method: The research included forty-three patients (23 with WT and 20 with NB) with confirmed diagnoses and CT images. Texture analysis involved seven first-order and nine second-order parameters, with receiver operating characteristic (ROC) analysis performed to differentiate between WT and NB. Results: Results showed that the median tumor volume was significantly higher in the WT group. Significant differences were found in entropy and homogeneity among first-order parameters. ROC analysis revealed sensitivity and specificity values: 78% and 45% for uniformity, 56% and 85% for entropy and 47% and 75% for sum entropy. Conclusion: Findings suggest that texture analysis has potential in distinguishing WT from NB in pediatric CT scans, which may reduce the need for repeated scans, additional imaging and invasive procedures, thereby improving patient care and reducing healthcare costs

Keywords

• computed tomography
• texture analysis
• wilms tumor
• neuroblastoma