Evaluating the MRI-based clear cell likelihood score: is it clinically adequate for predicting clear cell carcinoma?

BackgroundAccurate differentiation of clear cell renal cell carcinoma (ccRCC), the most aggressive subtype of renal masses, is crucial for guiding management decisions. Magnetic resonance imaging (MRI)-based Clear Cell Likelihood Score (ccLS) has recently emerged as a useful tool in this regard.PurposeTo evaluate the diagnostic performance and inter-observer reliability of the MRI-based ccLS in distinguishing ccRCC from other renal tumors that could not be clearly classified with conventional imaging.Material and MethodsThis single-center, retrospective study included 176 patients with renal masses who underwent preoperative dynamic contrast-enhanced MRI. Two radiologists independently reviewed the images, applying the ccLS scoring system based on T2 signal intensity, corticomedullary phase enhancement, and other imaging features as previously described. The histopathological results were used as the reference standard. The diagnostic performance of ccLS, with varying thresholds, was assessed, and inter-observer agreement was evaluated.ResultsThe study found that the ccLS system demonstrated high sensitivity (93.3%) but low specificity (47.9%) at a threshold of ≥3 and balanced accuracy (sensitivity = 81%, specificity = 70.4%) at a threshold of ≥4. Larger tumors (≥4 cm) showed superior diagnostic performance. MRI features such as T2 hyperintensity and corticomedullary hypervascularity were significantly more frequent in ccRCC compared to non-ccRCC (P <0.001). The inter-observer agreement for ccLS and key MRI features including T2 hyperintensity and corticomedullary hypervascularity were substantial (weighted kappa = 0.71-0.74).ConclusionAlthough highly reproducible, the current ccLS algorithm, should be used cautiously in distinguishing ccRCC from other renal masses that cannot be easily classified with conventional imaging.