Osteoarthritis (OA) is a debilitating disease that affects 27 million
 Americans, causing pain, stiffness and loss of mobility¹. Magnetic resonance imaging (MRI) provides an excellent opportunity to non-invasively study and understand complex disease processes involved in OA². However, characterization of bone pathology on MRI is limited to qualitative assessment of osteophytes and marrow signal changes³ and is poorly understood. ¹⁸F-fluoride is a long recognized bone-seeking agent that is able to probe bony remodeling⁴. This study aims to correlate and characterize bone pathology observed on MRI with ¹⁸F-fluoride uptake on PET.Both knees of nine subjects with idiopathic or traumatic radiographic knee OA (Kelgren-Lawrence grade 1-3) were imaged on a 3T whole-body PET-MR hybrid system (GE Healthcare, Milwaukee, WI) following injection of between 2.5 and 5 mCi of ¹⁸F-fluoride under an approved IRB protocol. Two experienced radiologists identified bone pathology on MR images. This included areas of high T₂ signal referred to as “bone marrow lesions” (BMLs)⁵, osteophytes identified on ultra-short echo time (uTE) MR images and subchondral sclerosis (areas of low signal intensity of subchondral bone on both fat-saturated T2-weighted FSE images and Dixon fat images). Additionally, BMLs and osteophytes were scored according to MRI Osteoarthritis Knee Score (MOAKS)³. The maximum pixel standardized uptake values (SUV_max) from volumes of interest (VOI) on PET images corresponding to bone pathology identified on MRI were compared between different bone pathologies as well as between individual pathology lesion grade. VOIs with SUV greater than 4 times the natural bone background (mean SUV of background bone) were identified (VOI_High) on PET SUV maps. A Wilcoxon signed-rank test was used to compare SUV_max from different bone pathology to each other as well as within BML and osteophyte grade. Additionally, SUV_max from VOI_High that didn’t correlate with MRI findings were identified and were compared to MR findings of bone pathology.Figure 1 displays example PET (SUV) and MRI images of two patients with traumatic OA showing correlation between BML (blue arrow), osteophytes (red arrows), and sclerosis (green arrows) on MRI with high ¹⁸F-fluoride uptake on PET. High ¹⁸F-fluoride uptake in subchondral bone did not always correspond to structural damage detected on MRI (purple arrows) (23 of the 103 areas of VOI_High) (Figure 2a). Additionally, many of the small osteophytes used to identify early signs of OA on radiographs did not show focal uptake of ¹⁸F-fluoride (orange arrows) (Figure 2b). Bone marrow lesions (BMLs) observed on T2-weighted FSE images acquired with fat saturation consistently correlated with ¹⁸F-fluoride PET VOI_High (25 of 26 lesions) (Figure 3b). Furthermore, SUV_max associated with BMLs was significantly higher than that of osteophytes or subchondral sclerosis (p<0.001) (Figure 3). Additionally, significant correlations were observed between SUV_max and BML grade based off MOAKS scoring (p<0.05) (Figure 4a). Association between ¹⁸F-fluoride VOI_High and MRI findings of osteophytes (49 of 70) and subchondral sclerosis (6 of 10) was less consistent. However, similar to BMLs, there were significant correlations between SUV_max and osteophyte grade (p<0.05) (Figure 4b). The sensitivity of ¹⁸F-fluoride PET to bone remodeling make it an optimal tool to better understand subchondral bone pathology on MRI. Subchondral bone is a region that is associated with the development of pain as well as cartilage degeneration. Early results suggest that BMLs are significantly more metabolically active than osteophytes and sclerosis. Furthermore, significant differences were observed in ¹⁸F-fluoride PET uptake between osteophyte and BML grade identified with MR. However, there were large standard deviations observed within each type of bone pathology and within lesion grade. The sensitivity of ¹⁸F-fluoride PET/MR to bone remodeling may help us to better understand and characterize these lesions in subchondral bone. Additionally, a lack of metabolic activity may signal that certain pathology play a reduced role in OA progression. Lastly, high ¹⁸F-fluoride uptake that does not correlate with MR findings suggests that metabolic abnormalities in the bone may occur prior to when structural changes are seen on MRI.This work demonstrates that simultaneous ¹⁸F-fluoride PET/MR may provide additional metabolic information regarding bone pathology compared with conventional MR. This will allow for a better understanding of the role of bone degeneration in OA disease processes. Additionally, ¹⁸F-fluoride PET/MR may detect knee abnormalities unseen on MRI alone and is a promising tool for detection of early metabolic changes in OA.