Usefulness of Texture Analysis Using MRI in Thymic Epithelial Tumors

Area of Science:

• Radiology and Medical Imaging
• Oncology
• Computational Pathology

Background:

• Thymic epithelial tumors (TETs) present a spectrum from low-risk thymoma to thymic carcinoma, posing diagnostic challenges.
• Differentiating TET subtypes non-invasively is crucial for accurate prognosis and treatment.
• Current imaging modalities struggle to reliably distinguish between these subtypes.

Purpose of the Study:

• To evaluate the effectiveness of MRI texture analysis in differentiating TET subtypes based on the 2021 WHO classification.
• To assess the diagnostic performance of various MRI sequences and texture features.

Main Methods:

• Forty-seven patients with histologically confirmed TETs underwent preoperative MRI (T1WI, T2WI, DWI, ADC).
• Texture analysis was performed using LIFEx software, extracting 41 features from 3D tumor volumes.
• Statistical analysis included Kruskal-Wallis, Mann-Whitney U tests, and ROC analysis for diagnostic performance.

Main Results:

• Significant differences in texture parameters were observed between low-risk thymoma group (LRG), high-risk thymoma group (HRG), and thymic carcinoma (TC).
• Highest AUCs were 0.785 (DWI) for LRG vs. HRG, 0.933 (T1WI) for LRG vs. TC, and 0.867 (ADC) for HRG vs. TC.
• Sequence-specific texture features demonstrated moderate to high diagnostic performance.

Conclusions:

• MRI-based texture analysis is a promising noninvasive tool for differentiating TET subtypes.
• Texture parameters show potential for preoperative risk stratification and guiding treatment decisions in TET patients.
• Further validation may enhance the clinical utility of MRI radiomics in thymic malignancies.