A Web-Based Prediction Tool to Improve Identification of Patients With Undersized Hamstring Tendon Autograft in Anterior Cruciate Ligament Reconstruction

Abstract

Background:

An undersized hamstring tendon (HT) autograft is significantly associated with a higher graft failure rate in anterior cruciate ligament reconstruction (ACLR) surgery. The ability to accurately predict inadequate HT graft diameter is critical, as it could assist surgeons in making better graft choices and surgical plans.

Purpose:

To develop a web-based prediction tool to better assess the size of HT autograft and to help clinicians accurately identify patients with potentially undersized HT grafts in order to make appropriate clinical decisions.

Study Design:

Cross-sectional study; Level of evidence, 3.

Methods:

A total of 588 patients who received primary arthroscopic single-bundle ACLR surgery with gracilis tendon (GT) and semitendinosus tendon (ST) autograft were retrospectively reviewed. According to the size of 4-strand HT graft, patients were divided into diameter ≥8 mm and <8 mm groups. The least absolute shrinkage and selection operator method and logistic regression were used to identify the independent factors associated with HT graft diameter and establish the models. The prediction performance of the model was evaluated by concordance index and calibration combined with external validation. The diagnostic performance of the prediction model was assessed by sensitivity, specificity, predictive values, and likelihood ratios. Decision curve analysis was used to evaluate the clinical utility of the model.

Results:

Among the numerous indicators, sex, weight, height, thigh length, and ST-GT diameter (measured on plane 1 of a magnetic resonance imaging scan) were identified to be highly correlated predictors that could provide satisfactory prediction performance in determining the HT graft diameter. Based on these predictors, a prediction model named the HTD model was developed with satisfactory discrimination (concordance index, 0.932) and calibration (mean absolute error, 0.039). When the probability calculated by the HTD model was >65%, the sensitivity and specificity of predicting 4-strand HT graft diameter ≥8 mm were 86.7% and 90.2%, respectively.

Conclusion:

As a useful supplementary prediction tool, the HTD model could accurately predict the diameter of HT autograft during preoperative planning.