AJSM Webinar Series: ACL Injury and PTOA: Bridging Biology, Imaging, and Machine Learning to Improve Outcomes-Lu Y et al. Establishing Clinically Distinct Patient Treatment Subgroups Following ACL Reconstruction: A Machine Learning Clustering Analysis. AJSM 2025. PMID: 40815848

Lu Y et al. Establishing Clinically Distinct Patient Treatment Subgroups Following ACL Reconstruction: A Machine Learning Clustering Analysis. AJSM 2025. PMID: 40815848

Back to course

Pdf Summary

This study used a large population registry and machine learning to identify clinically distinct subgroups of patients after anterior cruciate ligament (ACL) injury and to compare outcomes after ACL reconstruction (ACLR) versus nonoperative care.<br /><br />The researchers reviewed 923 patients with complete ACL tears and at least 7.5 years of follow-up. An unsupervised random forest clustering algorithm divided patients into two groups: an “optimal outcome” subgroup (653 patients) and a “suboptimal outcome” subgroup (270 patients). The suboptimal group was older, had higher body mass index, more smoking, more diabetes/inflammatory disease, lower sports participation, more laborer/sedentary occupations, and more concomitant medial meniscus injury. They also had worse clinical findings and higher rates of later arthritis and total knee arthroplasty (TKA).<br /><br />Using targeted maximum likelihood estimation, the authors found that ACLR had different benefits across subgroups. In the optimal subgroup, ACLR significantly reduced the risk of secondary meniscal injury, contralateral ACL injury, symptomatic posttraumatic osteoarthritis (PTOA), and progression to TKA. In the suboptimal subgroup, ACLR still lowered the risk of PTOA and TKA, but did not significantly reduce secondary meniscal injury or contralateral ACL injury.<br /><br />Regression analysis showed that younger age, lower BMI, student occupation, sports participation, and noncontact injury mechanism predicted membership in the optimal subgroup, while medial or bicompartmental meniscal injury predicted worse outcomes.<br /><br />Overall, the study suggests that ACL patients are not a uniform population. ACLR appears protective against long-term joint degeneration in both subgroups, but its ability to prevent other secondary injuries is reduced in older, heavier patients and those with meniscal damage.

Keywords

anterior cruciate ligament injury

ACL reconstruction

machine learning clustering

patient subgroups

population registry

posttraumatic osteoarthritis

total knee arthroplasty

meniscal injury

clinical outcomes

targeted maximum likelihood estimation