Automated Pipeline for Shoulder Cartilage Segmentation and Extraction of T2 and Texture Features

Motivation: Automated evaluation of the quantitative MRI T2 maps in the shoulder joint cartilage can improve the efficiency, reproducibility and accuracy of the clinical studies. Goal(s): To design and validate automated shoulder cartilage pipeline, including fully automated cartilage segmentation as well as three T2 modalities, co-registered with the morphological 3D MRI. Approach: Automated shoulder cartilage segmentation provided sixteen cartilage regions, on which the three T2 modalities have been registered. Additionally, texture analysis has been performed on all three T2 maps. Results: While the absolute T2 values were statistically significantly different between three methods, the texture features were not. Impact: This automated pipeline for shoulder cartilage segmentation and extraction of T2 and texture features offers a reliable, efficient approach to assess cartilage health. The integration of T2 mapping techniques are paving the way for quantitative cartilage evaluation in musculoskeletal imaging.