Quantitative MRI Evidence of Synovial Proliferation is Associated with Radiographic Severity of Knee Osteoarthritis

Abstract

Objective

To evaluate the relationships of quantitative and semi-quantitative (SQ) assessments of synovium with knee OA severity by radiographic and 3T MRI findings.

Methods

58 knee OA patients underwent non-fluoroscopic fixed-flexion knee radiographs. Dynamic contrast-enhanced (CE) 3T MRI was performed pre-/post-gadolinium administration to quantify synovial volume (qSV). SQ synovial outcomes were assessed on CE and unenhanced images. Two radiologists scored X-rays using the OARSI atlas; inter-reader agreement was assessed using Kappas and concordance correlation coefficients. Multiple linear and logistic regression analysis was used to assess associations among variables while controlling the effects of age, BMI, gender and meniscal extrusion.

Results

KL grade, diseased compartment joint space width (dcJSW) and diseased compartment joint space narrowing (dcJSN) were significantly associated with synovial proliferation, measured as CE qSV (β = 0.78, p = 0.0001; β = -0.22, p = 0.0003; β = 0.53, p = 0.0001, respectively). Furthermore, qSV strongly correlated with total subchondral BML volume (β = 0.22, p = 0.0003). KL grade, dcJSW, and dcJSN were significantly associated with BLOKS SQ infrapatellar synovitis (OR [95%CI]: 9.05, [1.94,42.3]; 0.75 [0.54,1.03]; 2.22 [1.15,4.31], respectively) and effusion (OR [95%CI]: 5.75, [1.23,26.8]; 0.70, [0.50,0.98]; 1.96, [1.02,3.74], respectively). CE SQ synovitis also significantly associated with KL and dcJSN (β = 0.036, p = 0.0040; β = 0.015, p=0.0266, respectively), and BLOKS synovitis.

Conclusion

Synovitis is a characteristic feature of advancing knee OA stages, and is significantly associated with KL, JSW, JSN, and BMLs. BLOKS synovitis scoring on unenhanced MRI is associated with CE synovitis measures.

Osteoarthritis (OA) is a complex joint disease affecting over 25 million people in the U.S., which is increasing in frequency and severity in the aging population.(1) Pathologically, OA is characterized by progressive loss of articular cartilage and new bone formation, each long appreciated by conventional radiography. However, it is increasingly apparent, based upon pathological, MRI and arthroscopy studies, that progressive OA involves all tissues of the joint and includes bone marrow lesions (BML) and synovial proliferation.(2-4) Ultrasound and MRI studies have highlighted the presence of synovitis in OA, which cannot be detected by radiographs or clinical exam, particularly in early disease.(5-7)

Synovial membrane Volume (SV), measured by contrast-enhanced (CE) MRI, reflects mostly proliferative synovial tissue or synovitis.(8, 9) To our knowledge, no prior studies have used gadolinium-enhanced MRI to quantitatively assess SV and to study its correlation with radiographic disease in knee OA. Non-contrast MRI techniques using semi-quantitative scoring methods may underestimate the degree of synovitis, whereas measurement of quantitative SV (qSV) using gadolinium-enhanced MRI is currently the most sensitive way of detecting synovial hypertrophy in OA.(10-12) Other authors have assessed the correlation between synovitis measures on enhanced and unenhanced MRI as well as the predictive value of such measures with respect to clinical symptoms such as pain.(13-15) However, to the best of our knowledge there have been no prior studies assessing the relationship between both quantitative and semi-quantitative CE measures of synovitis and specific radiographic findings. The objective of the present study was to study the relationship between radiographic assessment of OA in the knee and the degree of synovitis, as measured quantitatively and semiquantitatively on both contrast enhanced and unenhanced MRI.

Patients and Methods

Patient recruitment

Fifty-eight symptomatic knee OA patients were enrolled sequentially as part of a larger ongoing NIH-funded prospective study of 180 subjects evaluating biomarkers in OA. Patients in the substudy presented here were the first 58 enrolled into the larger study, and 53 of them underwent gadolinium-enhanced MR imaging of their signal knee as described in detail below. The first 58 subjects were similar in demographics to the remaining patients in the larger study. Patients were recruited at New York University Hospital for Joint Diseases. To be eligible for the parent study, patients had to be at least 40 years old and respond, on an initial phone questionnaire, “yes” when asked if they had knee pain for most of the last month that was relieved by rest, at least partially. All patients were then evaluated for knee OA by history and physical exam and had to fulfill ACR clinical criteria for the diagnosis of knee OA (knee pain + at least 3 of 6: age >50 yrs, stiffness <30 min, crepitus, bony tenderness, bony enlargement, no palpable warmth) (16). Exclusion criteria were: any other form of arthritis (including rheumatoid arthritis, spondyloarthritis, active crystal arthropathy); BMI ≥33; any disorder requiring the use of systemic corticosteroids within 1 week of screening, history of bilateral knee replacements; major co-morbidities including diabetes mellitus, non-cutaneous cancer within 5 years of screening, chronic hepatic or renal disease, chronic infectious disease, congestive heart failure; and hyaluronan and/or corticosteroid injection to the affected knee within 3 months of screening. The BMI cutoff and other exclusion criteria were those of the parent NIH study of leukocyte gene expression in OA. The Institutional Review Board at NYU Medical Center approved the protocol. Informed consent was obtained from all subjects.

Radiographic Assessments

Knee Radiographs

All patients underwent standardized weight bearing fixed-flexion PA knee radiographs using the SynaFlexer™ X-ray positioning frame (Synarc). Radiographic readings were done separately by 2 musculoskeletal radiologists blinded to patient demographics, clinical information and MRI readings. Disagreements between the two readers were resolved by consensus. X-rays were scored for KL grade (0-4)(17), and medial and lateral JSW were measured at the mid-portion of the joint space via electronic calipers. Diseased compartment joint space width (dcJSW) was defined as the smaller of the two measurements for medial and lateral JSW. The Osteoarthritis Research Society International (OARSI) atlas was used to determine osteophytes, medial and lateral joint space narrowing (JSN), medial tibial/lateral femoral subchondral sclerosis, and medial tibial attrition.(18) Diseased compartment JSN was defined as the JSN score in the compartment designated as more diseased based upon JSW measurements.

Knee MR Imaging Protocol and Analysis

MR imaging was performed on a 3.0T clinical scanner (Magnetom Tim Trio; Siemens Medical Solutions, Erlangen, Germany) using an eight-channel transmit-receive phased-array knee coil (In vivo Corporation, FL, USA). The knee imaging protocol consisted of a sagittal 3D-high resolution T₁-weighed-fast low angle shot (FLASH) sequence with selective water excitation (TR/TE= 25/4ms; flip angle=25; FOV=15×15cm; slice thickness=1.5mm; matrix=512×384; receiver bandwidth=200Hz/pixel) as well as sagittal T₂-weighted fat-saturated spin echo (TR/TE= 4000/75ms; FOV=15×15cm; slice thickness=3mm; matrix=256×128; receiver bandwidth=130Hz/pixel). Synovial membrane was evaluated using dynamic contrast enhanced sagittal 3D-T₁ weighted-FLASH sequence with the following parameters (TR/TE=12/3.9ms, flip angle= 60; FOV=15×15cm, slice thickness=5mm, matrix=256×128, receiver bandwidth=200Hz/pixel, temporal resolution=30sec). This sequence was acquired in contiguous 5 mm sagittal slices throughout the knee before, during and after intravenous bolus administration of double dose contrast agent Gd-DTPA (0.2mg/kg) for assessment of synovial volume. Baseline precontrast static images as well as contrast enhanced dynamic images were acquired after bolus injection.(7, 12) The total acquisition time for the imaging protocol was 24 minutes.

Quantitative SV assessment was performed using MATLAB custom tools for manual segmentation of the entire knee joint based on the contrast enhanced dynamic images (3D-FLASH data set) obtained at 12 minutes of bolus injection with respect to precontrast baseline static images. Signal enhancement patterns were evaluated using regions of interest in the infra-patellar fat pad, the supra-patellar fat pad, the intercondylar notch and along the periphery of joint effusion. All time-enhancement curves were saturated at 10 minutes following the contrast bolus injection. Semiquantitative scoring of synovitis using the standardized BLOKS system was performed by 2 readers (JB, LR) on sagittal T2 weighted fat saturated spin echo images prior to the administration of contrast. (19) Semiquantitative assessment (0-3) of the degree of synovitis in the infrapatellar region as well as binary assessment (0/1) of synovitis in the medial posterior-condylar and lateral posterior-condylar regions was performed. The presence of joint effusion was also evaluated semiquantitatively using the 0-3 BLOKS scoring system. MRI scoring was done independently of radiographic readings, at time points several weeks apart.

A second method of semiquantitative synovitis assessment, similar to that utilized by Roemer et al.(13), was performed on the post-contrast enhanced MR images as follows. Utilizing the contrast enhanced dynamic images (3D-FLASH data set) obtained at 12 minutes after bolus injection, the synovial thickness was measured in 5 regions using the PACS workstation and electronic calipers. Three of the measurements were made on a mid-sagittal image at points 0.5 cm cranial to the superior patellar pole, 1 cm cranial to the femoral osteochondral junction along the femoral margin of the suprapatellar joint space, and 1 cm caudal to the inferior patellar pole along the deep surface of Hoffa's fat pad. Additional measurements were made in the posterior condylar recesses on parasagittal cuts showing these areas to greatest advantage. The measurements in mm were converted to a score for each region according to the scale: 0 = normal; 1 = ≤ 2 mm; 2 = 2-4 mm; 3 = ≥ 4mm. These five scores were the summed and used to arrive at an overall grade as follows: grade 1, score of 0-5; grade 2, score of 6-10; grade 3, score of 11-15.

Figure 1 illustrates examples of the three techniques for scoring synovitis.

Figure 1.

Examples of Synovial Volume Assessments. A and B) Quantitative Synovitis: A) Representative sagittal T1-weighted dynamic contrast enhanced MRI outlining the five different regions summed for total quantitative SV score. B) Corresponding normalized signal intensity curve generated by placing ROIs over enhancing synovium in periphery of these regions demonstrating maximal enhancement at approximately 12 minutes. C) Semi-quantitative Synovitis: Representative midline sagittal post-contrast image obtained 12 minutes after intravenous injection of gadolinium (0.2mmol/kg) at the level of the footprint of the anterior cruciate ligament. Measurements of synovial thickness are made at points 5mm superior to the superior pole of the patella (thick solid white arrow), 1cm superior to the femoral osteochondral junction (dashed white arrow) and 1cm inferior to the inferior patellar pole (thin solid white arrow). D) Semi-quantitative Synovitis: Representative parasagittal post contrast images obtained at the level of the body of the medial and lateral (not shown) menisci. Measurements of synovial thickening were made in the posterior condylar joint recesses (thick white arrow).

Statistical Methods

We explored data graphically and numerically to assess the distributions of measurements and to detect outliers. Side-by-side boxplots were used for graphical displays of data. Multiple linear and logistic regressions were used to assess the relationships among variables while controlling for the effects of age, BMI, gender and meniscal extrusion. Box-Cox transformations were used to normalize quantitative variables prior to applying regression methods. Since KL=2 category contained only 6 subjects, we dichotomized the variable KL grade into KL < 4 and KL = 4 for statistical analysis. Ordinal outcome variables such as infrapatellar synovitis and effusion were dichotomized due to small counts in some of their categories. Spearman's correlation coefficient was also used to assess associations among ordinal variables and variables with deviations from normality. In addition, partial correlation was used to measure the strength of a relationship between two variables, while controlling the effects of other variables, such as age, BMI, gender and meniscal extrusion. Kappa statistic and concordance correlation coefficient were used to assess inter-reader agreement for qualitative and quantitative measurements, respectively.(20, 21) Kappas for KL scores were 0.85 and 0.77 for the right and left knees, respectively. Concordance correlation coefficients for JSW were ≥0.93 for medial and lateral compartments of right and left knees. Kappas for medial and lateral OARSI JSN scores were ≥0.88. For infrapatellar, medial and lateral posterior condylar, Kappas were 1. For effusion, Kappa was 0.92. For contrast-enhanced synovitis, Kappa was 0.81. For meniscal extrusion, Kappa was 1. (Appendix 1). For quantitative variables, averages of the two readers' values were used. Substantial disagreements between readers were resolved by consensus reading.

Results

Demographic, clinical, radiographic, and MRI characteristics of the 58 OA patients included in the study are presented in Table 1. The mean age of patients +/- standard deviation was 62 (+/-10) years, mean BMI was 27.3 (+/- 3.3), and 59% were female. The KL groups were not significantly different with respect to BMI, gender, and meniscal extrusion, but older age was significantly associated with higher KL score (OR = 1.09, 95%CI [1.01, 1.16]).

Table 1. Baseline Characteristics of OA patients (n = 58).

Continuous Variables		Mean (SD)	Median [Min, Max]
Age (years)		62.0 (9.82)	61.0 [42.0, 84.0]
BMI (kg/m2)		27.3 (3.34)	27.0 [19.0, 32.6]
Quantitative		16.5 (8.39)	14.9 [4.10, 40.5]
Synovial Volume (ml)
Contrast-enhanced synovitis		5.75 (1.30)	5 [3, 9]
Total BML Volume (ml)		2.80 (3.80)	1.25 [0, 15.9]
Joint Space Width (mm)
	Medial	3.40 (2.17)	3.50 [0, 8.40]
	Lateral	5.74 (2.39)	6.20 [0, 9.65]
	Diseased Compartment	2.79 (1.96)	3.05 [0, 6.90]
Categorical Variables		0	1	2	3	4
KL Grade		-	-	6 (10%)	34 (59%)	18 (31%)
OARSI Atlas Score		0	1	2	3	NA
Osteophytes	MFC	22 (38%)	23 (40%)	10 (17%)	3 (5%)	-
	MTP	3 (5%)	37 (64%)	16 (28%)	2 (3%)	-
	LFC	20 (34%)	28 (48%)	8 (14%)	2 (3%)	-
	LTP	14 (24%)	31 (53%)	10 (17%)	3 (5%)	-
Joint Space	Medial	6 (10%)	22 (38%)	17 (29%)	13 (22%)	-
Narrowing	Lateral	41 (71%)	9 (16%)	3 (5%)	5 (9%)	-
BLOKS Score		0	1	2	3	NA
Infrapatellar Synovitis		26 (45%)	16 (28%)	7 (12%)	4 (7%)	5 (9%)
Medial condylar Synovitis		15 (26%)	38 (66%)	-	-	5 (9%)
Lateral condylar Synovitis		24 (41%)	29 (50%)	-	-	5 (9%)
Effusion		21 (36%)	17 (29%)	11 (19%)	4 (7%)	5 (9%)
		Absent	Present	NA
Meniscal Extrusion		42 (72%)	14 (24%)	2 (3%)

BMI = body mass index, BML = bone marrow lesion, KL = Kellgren Lawrence grade, LFC = Lateral Femoral Condyle, LTP = Lateral Tibial Plateau, MFC = Medial Femoral Condyle, MTP = Medial Tibial Plateau

Quantitative Synovial Volume (qSV)

The mean +/- standard deviation qSV was 10 +/- 4.6, 13 +/- 6.3 and 23 +/- 8.1 ml for KL2, KL3 and KL4 groups, respectively. KL grade was strongly associated with qSV in the multiple linear regression controlling for age, BMI, gender and meniscal extrusion (β = 0.78, 95%CI [0.42, 1.14], p = 0.0001, Table 2, Figure 2A). As shown in Table 2 and Figure 2B, diseased compartment joint space narrowing (dcJSN), as assessed by the OARSI atlas, was significantly associated with qSV when controlling for the effects of age, BMI, gender, and meniscal extrusion (β = 0.53, 95%CI [0.29, 0.77], p = 0.0001). Diseased compartment joint space width (dcJSW) was negatively correlated with qSV in the multiple linear regression controlling for age, BMI, gender and meniscal extrusion (β = -0.22, 95%CI [-0.34, -0.11], p = 0.0003) (Table 2, Figure 2C). Moreover, the bony changes of osteophyte formation (at the medial tibial and lateral tibial plateaus) both predicted qSV in multiple regression analysis (β = 0.52, 95%CI [0.15, 0.89], p = 0.0079 and β = 0.37, 95%CI [0.09, 0.66], p = 0.0141, respectively) (Table 2). Furthermore, BML volume was significantly associated with qSV (β = 0.22, 95%CI [0.11, 0.33], p = 0.0003) (Table 2, Figure 2D).

Table 2. Multiple linear regressions of Quantitative Synovial Volume: estimated regression coefficients, 95% CI and p-values of covariates (controlling for age, BMI, gender and meniscal extrusion).

Covariate		Estimated Regression Coefficient [95% Confidence Interval]	p-value
KL grade		0.78 [0.42, 1.14]	0.0001
dcJSN		0.53 [0.29, 0.77]	0.0001
Medial JSN		0.31 [0.05, 0.56]	0.0244
Lateral JSN		0.36 [-0.19, 0.91]	0.2074
dcJSW		-0.22 [-0.34, -0.11]	0.0003
Total BML Volume		0.22 [0.11, 0.33]	0.0003
Osteophytes	MTP	0.52 [0.15, 0.89]	0.0079
	MFC	0.27 [0.00, 0.54]	0.0539
	LTP	0.37 [0.09, 0.66]	0.0141
	LFC	0.31 [0.00, 0.61]	0.0532
Semi-quantitative MRI measures by BLOKS	Infrapatellar Synovitis	0.24 [-0.01, 0.49]	0.0692
	Medial Condylar Synovitis	0.65 [0.04, 1.27]	0.0420
	Lateral Condylar Synovitis	-0.12 [-0.64, 0.40]	0.6513
	Effusion	0.60 [0.10, 1.07]	0.0221

BML = bone marrow lesion, dcJSN = diseased compartment joint space narrowing, dcJSW = diseased compartment JSW, KL = Kellgren Lawrence grade, LFC = Lateral Femoral Condyle, LTP = Lateral Tibial Plateau, MFC = Medial Femoral Condyle, MTP = Medial Tibial Plateau

Figure 2.

Side-by-side boxplots showing relationship between quantitative synovial volume (SV) and KL grade (A), diseased compartment joint space narrowing (dcJSN) (B), and diseased compartment joint space width (dcJSW) (C). Lines within each box represent the median values of quantitative SV; boxes represent the interquartile ranges; the whiskers represent the smallest and largest values; and the dots represent the outliers. Numbers represent each individual patient's KL grade. Panel D) shows linear regression of quantitative SV on total summed bone marrow lesion (BML) volume. A Box-Cox transformation with λ = 0.4 was used to normalize BML. Numbers represent each individual patient's KL grade.

Semi-quantitative (SQ) Synovial Volume

Non-contrast SQ evaluation by BLOKS

Infrapatellar synovitis was present in 38% of subjects in KL2-3 group compared to 83% of subjects with KL=4 (Fisher's exact p = 0.0063). Effusion was present in 51% of subjects in KL2-3 group compared to 81% of subjects with KL=4 (Fisher's exact p = 0.0399). KL score, dcJSW and dcJSN were significantly associated with semi-quantitative measurements of infrapatellar synovitis and effusion in multiple logistic regression analyses controlling for age, gender, BMI and meniscal extrusion (Table 3). Furthermore, the average qSV was 13 +/- 6.2 in subjects with zero infrapatellar synovitis and 19 +/- 9.5 in subjects with positive infrapatellar synovitis (Wilcoxon Rank-sum p = 0.0306). The average qSV was 13 +/- 8.3 in subjects with no effusion and 18 +/- 8.3 in subjects with effusion (Wilcoxon rank-sum p = 0.0275). Effusion was also moderately associated with qSV (Table 2).

Table 3. Multiple logistic regressions of BLOKS outcomes of infrapatellar synovitis and effusion: estimated ORs, 95% CI and p-value of covariates (controlling for age, BMI, gender and meniscal extrusion).

Outcome variable	Covariate	Estimated Odds Ratio [95% confidence interval]	p-value
Infrapatellar synovitis (present vs absent)	KL grade1	9.05 [1.94, 42.3]	0.0051
	dcJSW2	0.75 [0.54, 1.03]	0.0727
	dcJSN2	2.22 [1.15, 4.31]	0.0181
Effusion (present vs absent)	KL grade1	5.75 [1.23, 26.8]	0.0258
	dcJSW2	0.70 [0.50, 0.98]	0.0366
	dcJSN2	1.96 [1.02, 3.74]	0.0429

dcJSN = diseased compartment joint space narrowing, dcJSW = diseased compartment joint space width, KL = Kellgren Lawrence grade

¹Covariate KL grade was dichotomized into KL < 4 and KL = 4 (see Methods). The reference group for the odds ratios is KL < 4.

²Covariates dcJSW and dcJSN were treated as continuous variables; therefore, the reported odds ratios are per unit of the covariate.

Contrast enhanced (CE) SQ evaluation

KL grade was significantly associated with CE SV in a multiple linear regression model controlling for age, BMI, gender and meniscal extrusion (β = 0.036, 95%CI [0.013, 0.60], p = 0.0040, Table 4), as was dcJSN (β = 0.015, 95%CI [0.002, 0.028], p = 0.0266). dcJSW, quantitative synovial volume and BML volume were weakly associated with CE SV. Semi-quantitative MRI measures of synovitis by BLOKS were significantly associated with CE SQ SV, including infrapatellar synovitis (β = 0.016, 95% CI [0.003, 0.029], p = 0.0198), medial condylar synovitis (β = 0.052, 95% CI [0.023, 0.082], p = 0.0110) and effusion (β = 0.017, 95% CI [0.005, 0.029], p = 0.0093).

Table 4. Multiple linear regressions of Contrast-Enhanced Semi-Quantitative Synovitis: estimated regression coefficients, 95% CI and p-values of covariates (controlling for age, BMI, gender and meniscal extrusion).

Covariate		Estimated Regression Coefficient [95% Confidence Interval]	p-value
KL grade		0.036 [0.013, 0.060]	0.0040
dcJSN		0.015 [0.002, 0.028]	0.0266
dcJSW		-0.031 [-0.062, -0.000]	0.0557
Quantitative Synovial Volume		0.014 [0.000, 0.029]	0.0558
Total BML Volume		0.006 [-0.001, 0.012]	0.0851
Osteophytes	MTP	-0.004 [-0.026, 0.019]	0.7381
	MFC	0.010 [-0.005, 0.026]	0.1934
	LTP	0.011 [-0.003, 0.026]	0.1346
	LFC	0.012 [-0.006, 0.030]	0.1881
Semi-quantitative MRI measures by BLOKS	Infrapatellar Synovitis	0.016 [0.003, 0.029]	0.0198
	Medial Condylar Synovitis	0.052 [0.023, 0.082]	0.0010
	Lateral Condylar Synovitis	0.011 [-0.015, 0.037]	0.4238
	Effusion	0.017 [0.005, 0.029]	0.0093

dcJSN = diseased compartment joint space narrowing, dcJSW = diseased compartment joint space width, KL = Kellgren Lawrence grade, LFC = Lateral Femoral Condyle, LTP = Lateral Tibial Plateau, MFC = Medial Femoral Condyle, MTP = Medial Tibial Plateau

Discussion

Our study demonstrates that synovial disease is a characteristic feature of progressive OA, which correlates with radiographic assessment of JSW, JSN, and KL grade, especially when analyzed by gadolinium-enhanced MRI. Interestingly, SV was also significantly associated with total BML volume on MRI. While it has been previously reported that synovitis occurs in patients with radiographic OA,(4, 22, 23) this is the first report using Gd-enhanced MR imaging, and exploring relationships between contrast-enhanced quantification of synovial volume and radiographic features of knee OA. We demonstrate that synovial membrane thickening takes place concomitantly with joint space narrowing, osteophytes and subchondral bone marrow lesions. The findings, therefore, further corroborate the concept that OA is a disease of the “whole joint,” raising questions regarding the pathogenic events that drive disease in tandem within the individual compartments. In addition, this is the first study to validate synovitis assessments by the BLOKS scoring system, by comparing these semi-quantitative measures of synovitis and effusion from non-enhanced images to quantitative contrast-enhanced measurements.

The cause of synovitis in osteoarthritis is not well understood, and is probably multifactorial. Synovitis can be induced by the release of cartilage fragments that activate synovial lining cells.(7, 24) Other possible etiologies include synovio-entheseal complex disease secondary to enthesitis, or damage to other structures such as ligaments, tendons etc.(25) There is also evidence that crystals (BCP or CPPD) incite low grade inflammatory changes in chronic osteoarthritis.(26, 27) Finally, it has been reported that cartilage-derived neo-antigens induce infiltration of B cells that are oligoclonal, suggesting that an antigen-driven immune response may play a part in the OA disease process.(28)

Regardless of the causes, our results are consistent with prior reports that synovitis is not restricted to those patients with end stage OA unhdergoing joint replacement surgery, and that it can be seen in the majority of patients with established disease.(7, 12, 29, 30) In fact, Benito et al observed that early OA synovial tissue, taken at arthroscopy, had increased mononuclear cell infiltration, blood vessel formation, and overexpression of inflammatory mediators (TNFα, IL-1β, COX-2, NF-Kb) compared with late OA synovial tissue, taken at knee joint arthroplasty.(29) Studies of late-stage OA have found that OA synovial membranes express similar cytokines to those seen in rheumatoid arthritis (RA) synovium, with the only difference being quantitative(9, 31, 32); it is possible that those differences may be even smaller if early OA synovium was compared to RA specimens. This production of mediators is well known not only to activate chondrocytes, leading them to produce catabolic factors, but also to further stimulate their own production and that of other cytokines, and promote angiogenesis, thus perpetuating the cascade of cartilage destruction.(33, 34) Subclinical chronic synovitis may exist from early stages of knee OA, and it may play an active role in perpetuation of the disease rather than be an innocent bystander. Findings on arthroscopy also support this, as cartilage defects are often seen directly abutting areas of inflammatory synovitis.(4) Furthermore, synovitis in OA has been shown by MRI to have a predilection for the cartilage-pannus junction rather than being randomly distributed,(35, 36) suggesting that there is a predictable geographical pattern to synovitis which makes conditions favorable for disease progression. However, the contribution that the presence of synovitis makes toward progression of disease is still unclear from longitudinal studies to date. (4, 22, 37-39)

There are certain limitations to our study. We excluded markedly obese subjects (BMI ≥33); therefore, while the conclusions cannot be applied to all OA patients, they are applicable to a large subset of patients with knee OA. In addition, the data presented in this manuscript are cross-sectional and no inference regarding disease progression could be made; however, longitudinal data collection from the same cohort is ongoing that would allow the evaluation of the utility of synovitis as a biomarker of disease progression. We also acknowledge that the method by which ‘diseased compartment’ was selected is not ideal. Since the medial joint space is slightly narrower than the lateral in a normal knee, there may be some knees in this study that were classified as medial, which are, in fact, lateral. However, there is no established way to determine diseased compartment and, if misclassification had occurred, this would lead to an underestimation of the strength of the findings. Finally, our study design did not allow for the comparative assessment of synovitis between symptomatic and asymptomatic OA patients since our cohort was entirely symptomatic. This is an important question to answer, particularly with regard to associations of synovitis and pain, and recent publications have addressed this topic.(14, 15) However, we did find that increased levels of synovitis, when measured semiquantitatively on contrast-enhanced MRI, associated with higher WOMAC scores (pain, stiffness, physical function and total). Further studies enrolling larger numbers of patients, particularly asymptomatic patients and healthy controls, who are followed over time, are essential in order to understand the true associations between MRI findings and pathogenesis in knee OA.

In conclusion, we observed that synovial volume, assessed by gadolinium-enhanced 3T MRI, is significantly associated with severity of knee OA by KL grade, JSW and JSN. SV is also significantly associated with radiographic evidence of bone involvement in OA – osteophytes and BMLs. We also demonstrated that BLOKS scoring of infrapatellar synovitis and effusion are associated with contrast-enhanced MRI synovitis. Ongoing longitudinal studies will be necessary to assess the utility of quantitative volumetric assessment of synovium as a biomarker of disease progression or as a surrogate endpoint for DMOAD development.