SIR 2024
Pain Management/MSK
Stephen Worrell, BSc
Medical Student
University of Pittsburgh School of Medicine
Financial relationships: Full list of relationships is listed on the CME information page.
Will S. Lindquester, MD
Independent IR Resident
Emory University School of Medicine
Disclosure information not submitted.
Anish Ghodadra, M.D.
Interventional Radiology Attending Physician
University of Pittsburgh Medical Center
Disclosure information not submitted.
To identify radiomic features of the infrapatellar fat pad (IFP) on MRI (magnetic resonance imaging) that correlate with objective clinical outcomes following geniculate artery embolization (GAE).
Materials and Methods:
All patients with Kellegren-Lawrence grade < 4 knee osteoarthritis (KOA) and pre-procedural MRI who underwent GAE between April 2022-April 2023 were included in the study. For each patient, the IFP on pre-procedure MRI was manually segmented, and 3D Slicer’s SlicerRadiomics extension was used to compute first order and gray level correlation matrix (GLCM) texture features. KL score, number of arteries embolized, baseline, 1, 3, and 6-month post procedural KOOS score and 30 second chair stand test were also recorded. Texture features were analyzed for correlation with the change in 30 second chair stand test from baseline to most recent post-procedural follow up.
Results:
Ten patients who underwent GAE for KOA using Teleflex GelBead 100 – 300 µm microspheres were included in the study. Mean age and BMI of patients were 53.5 years (SD: 11.1) and 35 (SD: 7), respectively and 60% of patients were female. 8 patients had KL grade 2 KOA and 2 patients had KL grade 3 KOA. A median of 2 genicular arteries were embolized among the study population. 7 patients had 6 month follow-up and 3 patients had 3 month follow-up. Median (IQR) KOOS Pain score was 40.5 (35.3, 49.8) at baseline, and 75 (65.5, 91.3) at most recent follow up, with pair-wise Wilcoxon-Signed Rank test showing a mean increase of 32.7 (p = 0.004). Median 30 second chair stand was 11 (9, 15) at baseline, and 15 (11.8, 27.3) at most recent follow up, with Wilcoxon-Signed Rank test showing a mean increase of 6.4 (p = 0.006). Change in 30 second chair stands was not correlated with change in KOOS Pain (r² = 0.03, p = 0.62). The GLCM texture feature, correlation, on pre-procedural MRI was significantly inversely correlated with the post procedural change in 30 second chair stands (r² = 0.55, p = 0.015).
Conclusion:
Our results suggest GAE objectively improves knee function as measured by the 30 second chair stand test, with 80% of patients meeting minimal clinically important difference (MCID) criterion. The change in 30 second chair stands was independent of change in KOOS Pain score. Our results also demonstrate that MRI texture analysis of the IFP, a highly vascularized and innervated structure of the knee joint implicated in KOA pathogenesis, correlates with the change in 30 second chair stands following GAE. Overall, pre-operative MRI and GLCM texture analysis of the IFP may be a useful biomarker in determining objective patient responsiveness to GAE.