Multiparametric Magnetic Resonance Imaging for Assessing Thymic Epithelial Tumors: Correlation With Pathological Subtypes and Clinical Stages

Background: World Health Organization classification and Masaoka-Koga stage are widely used for thymic epithelial tumors (TETs). Reduced field-of-view (rFOV) diffusion-weighed imaging (DWI) proved to improve the image quality. Dynamic contrast-enhanced (DCE) MRI was commonly used in evaluating tumors. Purpose: To investigate the value of multiparametric MRI in evaluating TETs. Study Type: Retrospective. Subjects: Eighty-seven participants including 38 low risk (52.08 +/- 14.19 years), 30 high risk (52.40 +/- 11.35 years), and 19 thymic carcinoma patients (59.76 +/- 10.78 years). Field strength/Sequence: A 3 T, turbo spin echo imaging, echo planar imaging, volumetric interpolated breath-hold examination with radial acquisition trajectory. Assessment: DCE-MRI and apparent diffusion coefficient (ADC) variables were compared. Diagnostic performances of single significant factor and combined model were compared. Statistical Tests: Parameters were compared using one-way ANOVA or independent-samples t test. Logistic regression was employed to investigate the combined model. Receiver operating curves (ROC) and DeLong's test were used to compare the diagnostic efficiency. Results: ADC, K-trans, and k(ep), values were significantly different among low-risk, high-risk and carcinoma group (ADC, 1.279 +/- 0.345 x 10(-3) mm(2)/sec, 0.978 +/- 0.260 x 10(-3) mm(2)/sec, 0.661 +/- 0.134 x 10(-3) mm(2)/sec; K-trans 0.167 +/- 0.071 min(-1), 0.254 +/- 0.136 min(-1), 0.393 +/- 0.110 min(-1); k(ep), 0.345 +/- 0.113 min(-1), 0.560 +/- 0.269 min(-1), 0.872 +/- 0.149 min(-1)). They were significantly different for early stage and advanced stage (ADC, 1.270 +/- 0.356 x 10(-3) mm(2)/sec vs. 0.845 +/- 0.251 x 10(-3) mm(2)/sec; K-trans 0.179 +/- 0.092 min(-1) vs. 0.304 +/- 0.142 min(-1); k(ep) 0.370 +/- 0.181 min(-1) vs. 0.674 +/- 0.362 min(-1)). The combination of them had highest diagnostic efficiency for WHO classification (AUC, 0.925; sensitivity, 83.7%; specificity, 89.5%), clinical stage (AUC, 0.879; sensitivity, 80.9%; specificity, 82.5%). Data Conclusion: Multiparametric MRI model may be useful for discriminating WHO classification and clinical stage of TETs.