CORR Insights®: Development of Machine Learning–based Algorithms to Predict the 2- and 5-year Risk of TKA After Tibial Plateau Fracture Treatment

Where Are We Now?

Posttraumatic arthritis is a concern for patients with periarticular fractures. Although the painstaking work of detail-oriented anatomic restoration is often pursued for these injuries, achieving perfect reductions isn’t always possible, depending on the complexity of the injury or the technical execution of the surgery. Furthermore, factors beyond the surgeon’s control, including patient characteristics and soft tissue considerations, can compromise the overall outcome. The critical challenges are how to prognosticate and how to mitigate the various surgical and nonsurgical risk factors that influence outcomes in injuries at risk for posttraumatic arthritis [3].

I commend Assink et al. [1] for their study in this month’s issue of Clinical Orthopaedics and Related Research®, in which they aim to address those challenges for patients with tibial plateau fractures. The authors’ prior study established a threshold of up to 4 mm of step or gap that is inconsequential to patient outcomes in nonoperatively treated plateau fractures [5]. The study of Assink and colleagues [1] extends on what was known before and seeks to predict conversion to TKA after fracture treatment using several factors, including articular incongruity and functional alignment. By leveraging machine learning, the authors provide a robust, personalized risk assessment tool for prognostication. The results are very promising; however, the question remains whether this tool can be implemented in a generalizable manner and integrated into routine clinical workflows across various settings, particularly in busy trauma centers.

Where Do We Need To Go?

We know that reduction quality matters for outcomes in periarticular injuries. This is evident in many anatomic locations, especially in the lower extremities. The tolerance for malreductions predisposing to arthritis in the tibial plateau has been debated. Various authors cite that a step or gap of 2 to 10 mm is a threshold that leads to an increased risk of posttraumatic arthritis [2]. The question then becomes how the severity of articular malreduction and alignment affects the patient’s prognosis for treatment failure in the form of posttraumatic arthritis that results in conversion to arthroplasty. Furthermore, what threshold is high enough to justify the rarely indicated acute TKA in the right patient demographic (such as older or less active patients)?

How Do We Get There?

Using AI-driven image analysis, we can process large data sets to establish more granular assessment of reduction quality, enabling more precise outcome prediction and surgical assessment. Additionally, AI could help weigh the clinical importance of malreductions based on their anatomic location—for example, differentiating between an incongruity in a low-weight–bearing area versus a critical load-bearing region on the tibial plateau. While we have yet to develop a robust algorithm that integrates these factors comprehensively, AI presents a promising avenue for refining our ability to assess and improve on our surgical outcomes. The article of Assink et al. [1] represents an important step forward, contributing to the ongoing effort to improve prognostication after periarticular injury.

Read This Next

• This is a comprehensive article highlighting the different considerations that contribute to posttraumatic arthritis from the basic science perspective [3].

• This clinical review provides considerations for preventing posttraumatic arthritis, with a particular emphasis on important surgical details [4].