Background and objectiveSignal changes in MRI for Kienböck’s disease have only been qualitatively assessed so far. In light of this, we proposed a new grading system for quantitative analysis with an ordinal scale.MethodsThe study included 31 patients (17 men, 14 women) with Kienböck's disease. By referring to Nakamura’s MRI grading system, we devised a grading system with five grades (Grades 1-5) using proton density-weighted (PDW) coronal images with respect to the signal intensity of the lunate. All cases were examined by using the MRI grading system by three hand surgeons, both preoperatively and postoperatively. We evaluated the inter-rater reliability of our grading system by using the interclass correlation coefficient. After surgery, we implemented annual MRI evaluation for as long as possible and quantitatively assessed changes in MRI grades. We also investigated the correlation between postoperative MRI grades, Mayo Wrist Scores (MWS), and age at the surgery by using Pearson’s coefficient.ResultsThe MRI evaluation was performed 2-15 years after surgery. The reliability of our grading system was high; inter-rater interclass correlation coefficients were 0.783 (examiners 1-2), 0.780 (examiners 1-3), and 0.825 (examiners 2-3), representing a substantial agreement. The correlation coefficient between the MRI grade and MWS was -0.31, suggesting a mild negative correlation; postoperative MRI grade also correlated with age at surgery (Pearson’s coefficient: 0.447).ConclusionsOur proposed MRI grading system has high reliability and could be used to assess the regeneration of a necrotic lunate for quantitative analysis on an ordinal scale. Improvements were observed one to four years postoperatively, demonstrating a mild correlation with the clinical results.