Cooled radiofrequency ablation (cRFA) of the knee geniculate nerves is a promising treatment option for knee osteoarthritis, but has variable efficacy. Fluoroscopy is the typical guidance of choice, but uses bony landmarks to target these nerves which have variable positions around the knee. Alternatively, ultrasound allows for targeting based on visualization of bony landmarks, the nerves, and their accompanying vascular structures, and thus may offer greater accuracy. However, there are few studies which directly compare the clinical efficacy of these two guidance options. This study will compare these two options.

Background: Cooled radiofrequency ablation (cRFA) of the knee geniculate nerves is a promising treatment option for knee osteoarthritis, but has variable efficacy. Fluoroscopy is the typical guidance of choice, but uses bony landmarks to target these nerves which have variable positions around the knee. Alternatively, ultrasound allows for targeting based on visualization of bony landmarks, the nerves, and their accompanying vascular structures, and thus may offer greater accuracy. However, there are few studies which directly compare the clinical efficacy of these two guidance options. Objective: To compare the effectiveness of geniculate nerve cRFA on knee osteoarthritis outcomes when using either ultrasound or fluoroscopic guidance. The primary outcome measures will be patient pain levels and patient-reported function. The secondary outcome measures will be objective performance-based functional outcomes. The exploratory outcome measures will be blood biomarkers of inflammation, extracellular matrix turnover, and cartilage degradation. Methods: Knee osteoarthritis patients (Kellegren-Lawrence Scale 2-4) that have been refractory to conservative treatments and report at least 80% pain relief with diagnostic geniculate nerve blocks will be enrolled and randomized to fluoroscopic (N=45) or ultrasound (N=45) cRFA treatment arms. Baseline assessments will include 1) Numeric Pain Rating Scale, Western Ontario and McMaster Universities Osteoarthritis Index, and PROMIS depression and pain interference measures; 2) the Osteoarthritis Research Society International recommended physical performance assessments including the 30-Second Chair Stand Test, Stair Climb Test, 40m Fast Paced Walk Test, Timed Up and Go Test, and Six Minute Walk Test; and 3) in a subset of patients (N=24), blood biomarkers including MMP-3, IL-1β, TNF-α, COMP, and CTX-II. These measures will be re-assessed at one month (pain and patient-reported function only, via phone contact), three months, and 6 months after the cRFA procedure. Study endpoints will be assessed using a mixed model repeated measures analysis of variance using an alpha level of 0.05. Statement of Relevance: Knee osteoarthritis patients are commonly treated by sports medicine providers, who are typically well-trained in the use of ultrasound. Should ultrasound be demonstrated as effective for cRFA, sports medicine providers would be well-positioned to use relatively low-cost and accessible imaging technology to significantly improve patient pain and function using cRFA.

Inclusion Criteria:

• 1) patients of either gender with primary knee OA of one or both knees fulfilling the diagnostic criteria for knee OA by the American College of Rheumatology,
• 2) Kellgren-Lawrence score of two to four,
• 3) reported NPRS pain intensity of at least four on most or all days of the past week
• 4) pain resistant to conventional treatments which may include but is not limited to medications (e.g. acetaminophen, oral or topical non-steroidal anti-inflammatory drugs, opioids), physical therapy, and intra-articular injections (e.g. corticosteroids, hyaluronic acid, platelet-rich-plasma) for at least 3 months. In the case of bilateral knee OA, the most symptomatic knee will be treated and studied.

Exclusion Criteria:

• 1) age <35 years,
• 2) non-English speaking patients,
• 3) body mass index greater than 40,
• 4) previous radiofrequency ablation procedure for the knee,
• 5) active systemic or local infections at the site of needle/cRFA probe placement,
• 6) previous knee joint replacement surgery,
• 7) autoimmune or inflammatory cause of knee arthritis such as rheumatoid or psoriatic arthritis,
• 8) non-ambulatory patients,
• 9) patients who are unable to provide their own consent (e.g. dementia),
• 10) unstable medical or psychiatric illness,
• 11) patients with pacemakers, spinal cord stimulators, deep brain stimulators, or similar devices,
• 12) patients seeking care as a part of workman's compensation or have litigation pending
• 13) a negative response to diagnostic geniculate nerve lidocaine injections.

A random number table will be used for group allocation and the values recorded separately into opaque, sealed envelopes. This will be performed by a clinical coordinator who is not involved in study to ensure allocation concealment. Outcomes for the primary, secondary, and exploratory aims will be completed by research staff who are blinded to the subject treatment allocation. Data analyses will also be completed in a blinded fashion by the study statistical consultant.