Machine Learning Approach to Predict Outcomes for the Optimization of Surgical Management of Severe Digital Trauma

• Applicant: Dr. Maximilian Mayrhofer-Schmid
• Subject Area: Orthopaedics, Traumatology, Reconstructive Surgery
• Term: since 2025
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 566657538

Severe finger injuries, especially complex traumas involving multiple structures or complete amputations, present a significant challenge in medical care. Patients often experience lasting functional deficits, chronic pain, and psychological impairments, all of which can greatly diminish their quality of life and ability to work. These impairments also can lead to increased costs for surgical treatment, hospital stays, and lost productivity. The initial treatment approach and the decision between different surgical methods for reconstruction or amputation strongly influence long-term outcomes. However, to date, there is no standardized tool or instrument that assists in clinical decision-making based on a long-term prognosis for postoperative function and quality of life. The recently published Mangled Digit Severity Score (MDSS) might be helpful for deciding between reconstruction and primary surgical amputation in severe digital injuries without traumatic amputation, but there are currently no larger prospective validation studies. Additionally, the potential development of painful neuromas following digital amputation highlights the need for optimized nerve management during stump formation. This research project aims to improve surgical decision-making in severe finger injuries using a machine learning (ML) approach. The data used to develop this model will partly come from a retrospective analysis designed to identify factors that affect the long-term course, such as hand function, pain, and quality of life, following reconstruction or amputation. In addition, two prospective studies will be conducted to integrate the generated data into the ML model: a validation study of the MDSS and a randomized controlled trial to investigate three different methods for nerve management in finger amputations. The findings from all three studies will be integrated continuously into a dynamic ML model that, in the future, is intended to guide and improve initial treatment decisions based on patient- and injury-specific characteristics, ultimately aiming to achieve optimized long-term outcomes in terms of hand function, pain, and quality of life. In the long run, the project aims to enhance initial surgical care, reduce the need for subsequent procedures, and mitigate pain, physical limitations, and psychological stress after severe finger injuries.

• DFG Programme: WBP Fellowship
• International Connection: USA
• Host: Dr. Kyle Eberlin