A simple method based on qualitative MRI features for characterizing clear cell renal cell carcinoma in small renal masses: comparison with the clear cell likelihood score

Abstract

Purpose

To evaluate the efficacy of a simple method based on qualitative MRI features for characterizing clear cell renal cell carcinoma (ccRCC) in small renal masses (SRMs).

Materials and methods

This retrospective multicenter study included pathologically confirmed SRM patients who underwent multiparametric MRI between March 2017 and November 2023 at three institutions. Univariable logistic regression and Fleiss κ coefficient were employed to determine features with significant diagnostic value and high consistency for ccRCC. A simple method was developed based on the selected features using multivariable logistic regression. The performance of the method was compared with the clear cell likelihood score (ccLS) using DeLong test and McNemar test.

Results

A total of 200 SRMs from 194 patients (116 men; median age: 54 years) were included. Intense corticomedullary enhancement, microscopic fat, and pseudocapsule were selected to construct the simple method, which considered a mass to be ccRCC if any two of the aforementioned three signs were present. Compared with ccLS, our method demonstrated similar sensitivity (0.824 versus 0.725, P = 0.227) and specificity (0.840 versus 0.860, P > 0.999). The AUC for the simple method and ccLS was 0.832 (95% CI 0.744, 0.899) and 0.793 (95% CI 0.701, 0.867), respectively (P = 0.864). For ccRCC cases assigned a score of 1 to 3 by the ccLS, 57.1% (8/14) were diagnosed correctly by the simple method.

Conclusion

The simple method can accurately characterize ccRCC in SRM with comparable efficacy to ccLS. Atypical ccRCC scored 1 to 3 by ccLS may benefit from the method.