Enhancing Clinical-Decision Making in Posttraumatic Rotational Limitations of the Forearm Through Deep-Learning Based Segmentation Techniques 

​

W.D.M. de Reus,  J.W. Colaris, J. Hirvasniemi, E.M. van Es 

​

Introduction:

Fractures in the forearm are common and frequently result in limitations of pronation/supination. Besides malunion as a possible causesoft tissue involvement may play a more significant role as well. More insight in both causes of impaired forearm rotation could help to treat patient in the least invasive way a possible, probably avoiding invasive corrective osteotomies in some patients. 

​

Objectives:

This study aimed to provide a robust framework for efficient identification of soft tissue structures potentially involved in posttraumatic rotational limitations of the forearm, enhancing clinical decision-making. To achieve this, nnU-Net was employed as a deep learning-based method for automated segmentation of anatomical structures involved in forearm rotation on magnetic resonance (MR) images. 

​

Methods and materials:

Manual ground truth annotations of six anatomical structures (radius, ulna, interosseous membrane, m. pronator quadratus, m. pronator teres, m. supinator) were performed on 24 axial in-phase sequence MR images of the forearm, equally distributed between left and right, and affected and unaffected forearms. Two nnU-Net configurations (2D and 3D) were trained on 20 forearms using 5-fold cross-validation, and an ensemble was created by combining predictions from both models. A test set of 4 forearms was used to evaluate segmentation performance using the Dice similarity coefficient (DSC) and the average symmetric surface distance (ASSD) metrics. Additionally, relative volume difference (Δrel) between ground truth and predicted segmentations were computed to assess under- or oversegmentation.  

​

Results:

The 3D model achieved the best segmentation performance, with a median DSC score of 0.894 (IQR=0.094) and a median ASSD of 0.324 (IQR=0.386) mm. It slightly undersegmented the anatomy, with a median relative volume difference of -2.7% (IQR=7.1%). Qualitative results revealed that the 3D model produced segmentation masks that contained fewer and less severe segmentation errors compared to the 2D model and ensemble. Visible segmentation errors of the 3D model occurred in the interosseous membrane, the proximal part of the m. pronator quadratus and the insertion of the m. pronator teres in some cases. 

​

Conclusion:

The 3D nnU-Net model has proven its suitability for clinical use, enabling fast, reproducible and precise segmentation of forearm structures involved in pronation/supination. This approach facilitates bilateral comparisons of soft tissue structures through visual assessment and quantitative analysis, supporting patient-specific and minimally invasive decision-making. 

​

​