Patient Selection for Genicular Artery Embolization

To educate the IR community on the value of patient specific factors, osteoarthritis (OA) phenotypes, and imaging and laboratory biomarkers in the treatment of knee OA with genicular artery embolization (GAE).

Background:

Knee OA is a leading cause of pain and disability worldwide {1}. GAE has emerged as an alternative treatment for symptomatic knee OA when conservative measures fail or if the patient is unsuitable for total knee replacement {1, 2}. Currently the main characteristics being considered are pain and radiographic findings of OA. However, treatment outcomes may vary based on baseline MRI findings and OA severity. Therefore, understanding individualized phenotypes is crucial to improving outcomes.

Clinical Findings/Procedure Details:

GAE’s is thought to alleviate knee pain by treating inflamed synovium and by causing ischemia to pathologic nerves in the knee {4}. Understanding an individual’s baseline presentation, such as synovitis severity, structural and degenerative changes, and OA phenotype is essential. Phenotypes of OA have been proposed for patient stratification based on imaging and molecular biomarkers: i) the inflammatory, ii) the cartilage, iii) the bone, iv) and the atrophic phenotype {5}. Varied MRI imaging features may portend inferior outcomes and include full-thickness cartilage defects, bone marrow lesions, presence of osteophytes, meniscal injury and grade 3 synovitis {6,7}. In this context, a study reported that the serum decrease of NGF levels, an inflammatory biomarker, at 12 months may predict GAE success {8}.  Lastly, reduction in synovitis as assessed by contrast enhanced MRI, seems to be related with pain palliation {9}.

Conclusion and/or Teaching Points:

Individual level differences in synovitis severity and OA phenotype may explain some of the heterogeneity of outcomes after GAE. Patient specific variables such as baseline MRI findings should be considered in patient selection for GAE as they can potentially prognosticate suboptimal response to GAE.