The Value of Preprocedural MR Imaging in Genicular Artery Embolization for Patients with Osteoarthritic Knee Pain
What is the value in preprocedural MRI in genicular artery embolization (GAE) for patients with osteoarthritic knee pain?
Take away point
Large bone marrow lesions and severe meniscal injury on MRI as well as high Kellgren-Lawrence grade on plain radiograph indicated poor response to GAE.
Choi JW, Ro DH, Chae HD, Kim DH, Lee M, Hur S, Kim H, Jae HJ, and Chung JW. The Value od Preprocedural MR Imaging in Genicular Artery Embolization for Patients with Osteoarthritic Knee Pain. J Vasc Interv Radiol. 2020;31(12):2043-2050. doi.org/10.1016/j.jvir.2020.08.012
Click here for abstract
Single-institution retrospective study of 28 knees in 18 patients undergoing GAE for osteoarthritic knee pain.
National Research Foundation of Korea (grant no. NRF-2018R1D1A1A02086183).
Academic hospital, Seoul National University Hospital, South Korea.
Figure. Numbers of knees that showed > 30% reduction in pain scores at 1 and 3 month follow up following GAE based on (a) KL grades on plain radiographs, (b) BML grades on MRI, and (c) meniscal injury grades on MRI.
Geniculate artery embolization (GAE) is a relatively new treatment for osteoarthritic knee pain with little known about the ideal indications for the procedure. Given that the genicular arteries supply soft tissue structures of the knee, the authors hypothesized that MR imaging could add value to preprocedural assessment of indications for GAE in addition to the known Kellgren-Lawrence (KL) grading system based on plain radiography. The authors performed a retrospective study of 28 knees in 18 patients treated with GAE with 1 and 3 month follow up data.
Patients presenting with osteoarthritic knee pain of over 6 moths duration despite conservative therapy were assessed for GAE. Knees without an MRI within 1 month of treatment or with diagnoses other than osteoarthritis were excluded. Additionally, one patient with knee replacement surgery within the follow up period and one patient who did not follow up were also excluded.
Prior to GAE, patients ranked their pain for each affected knee based on a 100-mm visual analog scale (VAS). Prior to the procedure, knees were evaluated by an orthopedic surgeon and pain was localized via physical exam and denoted with radiopaque markers. During the procedure, the genicular arteries that most closely supplied the marked areas were embolized. If pain was diffuse or unable to be localized, the genicular arteries exhibiting abnormal hypervascular staining on digital subtraction angiography (DSA) were embolized. At 1 and 3 month follow up visits, patients again ranked their pain in each affected knee on the same 100-mm VAS.
KL grades were assigned based on preprocedural radiographs reviewed by a musculoskeletal radiologist. Preprocedural MRIs were reviewed twice (at 3 months apart) by a single, blinded, musculoskeletal radiologist with specific grading of 6 parameters including: cartilage defects (scored 0-3 based on deepest loss among lesions), osteophytes (scored 0-3 based on length of extrusion from the cortex), subchondral cysts (scored 0-3 based on largest diameter), bone marrow lesions (BML; scored 0-3 based on largest diameter), meniscal injury (scored 0-4 based on tear morphology and severity), and joint effusion (scored 0-3 based on number of distended synovial recesses). Intra-observer disagreements were reviewed by a second musculoskeletal radiologist.
A total of 28 knees in 18 patients met criteria with preprocedural VAS pain scores ranging from 30-100 (mean 84.3; SD 15.5). Three of the graded imaging parameters were significantly related to GAE response (defined as a reduction in VAS pain score by > 30%): KL grade (p<.001), BML grade (p<.001), and meniscal injury grade (p=.003). With a cutoff of ≤ 2 for each significant parameter, sensitivity and specificity in predicting GAE response were as follows: 100% and 68.2% for KL grade, 83.3% and 77.3% for BML grade, and 83.3% and 72.7% for meniscal injury grade.
The authors investigate the value of preprocedural MRI in a relatively small cohort of patients undergoing GAE for osteoarthritic knee pain. The authors discuss that the value in preprocedural MRI likely lies in identifying potential GAE nonresponders rather than patients who will benefit from treatment. The significant parameters of KL, BML, and meniscal injury grade are hypothesized as indicators of bone-centered pain, which is less likely to be adequately treated with GAE. Of note, the authors mention that although the menisci are soft tissue structures, they are not supplied by nerves, and therefore meniscal injury noted on MRI was felt to represent bone-centered pain via degenerative joint-space narrowing. Similarly, cartilaginous defects, although not statistically significant, were also associated with poor GAE response. There are several limitations with this study; most notable are its small sample size and lack of control group. Patients with prior surgical treatments were included in the study, and although the most recent surgery was 9 months prior to GAE, the clinical significance of prior surgery remains unclear without a comparison group. Additionally, the applicability of the findings is limited given the time-consuming grading of parameters on each knee MRI, which are not currently standard reported findings. Overall, this study is a useful initial exploration into the value of preprocedural imaging for GAE, which remains a relatively novel therapy and merits future investigation.
Catherine (Rin) Panick, MD
Resident Physician, Integrated Interventional Radiology
Dotter Interventional Institute
Oregon Health & Science University