Professor of Radiology and Medicine Vice Chair, Academic Affairs Assistant Dean of Diversity Director, Quantitative Imaging Center (QIC) Boston University School of Medicine, Boston, MA

1. Ali Guermazi, MD, PhD
Professor of Radiology and Medicine
Vice Chair, Academic Affairs
Assistant Dean of Diversity
Director, Quantitative Imaging Center (QIC)
Boston University School of Medicine, Boston, MA
Imaging of Synovitis in OA

2. • Shareholder of BICL, LLC
• Consultant to MerckSerono, AstraZeneca, Pfizer,
Roche, Galapagos, and TissueGene
Disclosure

3. Outline
 Technical aspects – non CE vs CE
 MRI vs Histology
 Synovitis and structure
– Cross-sectional and longitudinal study
 Synovitis and symptoms
– Cross-sectional and longitudinal study
 Clinical trials of synovitis in OA
 New techniques
 Other modalities
 Conclusion

4. MRI Sequences to Assess Synovitis
 T1-weighted: increased volume / thickening of
the synovial membrane
 T2-weighted fat-suppressed (FS) / STIR:
increased water content
 Contrast-enhanced T1-weighted (FS):
vascularity and permeability of the membrane
– Differentiate effusion from true synovial thickening
 Novel non-enhanced MRI sequences to asses
synovitis

5. Assessment of Synovitis using
NCEMRI in OA
Signal changes in Hoffa’s fat pad
(“Hoffa-synovitis”)
Roemer FW et al. AJR 2009
 To date, synovitis in large epidemiological OA studies is assessed
using non contrast-enhanced MRI
Joint effusion
(“Effusion-synovitis”)

6. Roemer et al. AJR 2009
Compare Hoffa signal on NCEMRI with CEMRI in knee OA
 Comparison of sagittal FS PD-w and CE FS T1-w MRI
 50 knee OA patients
 Agreement between unenhanced and CE MRI was fair
to moderate (weighted κ=0.35 and 0.45)
 Hoffa fat pad signal alteration:
– LOW Specificity=10-38%
 Signal in Hoffa's fat pad on NCEMRI do not always
represent synovitis but are a nonspecific albeit
sensitive
 SQ scoring of synovitis in OA ideally should be
performed with T1-weighted CEMRI

7. Guermazi et al.
Ann Rheum Dis 2011
 SQ scoring system for synovitis
 Sagittal and axial CE FS T1-w MRI
 Grades: 0-2 at each of 11 locations:
– medial and lateral parapatellar recess
– suprapatellar, infrapatellar, intercondylar,
medial and lateral perimeniscus
– adjacent to PCL, ACL, loose bodies, Baker cyst
 Moderate to severe synovitis showed
significant association with maximum
WOMAC pain score item compared to
knees with no or equivocal synovitis

8. Literature evidence:
NCEMRI vs. CEMRI
 Is NCEMRI sufficient for imaging assessment
of synovitis in OA trials?
 NO!
 NCEMRI is limited:
 Non-specific
 Cannot differentiate between synovium
and effusion
 Inconsistent association with pain by
different studies

9. Why Gado Is Not Used in OA Studies?
 Time consuming
– Add 5 min to usually 20-30 min exam
 Expensive
– Add $50 to $100 per exam
 Not without risk to the participant
– Very low risk of nephrogenic systemic
fibrosis (NSF)
 Exclude patients with renal insufficiency (GFR)
– Unknown safety issue from Gado deposit

10. Crema et al. OAC 2013
Diagnostic performance of signal changes in HFP
assessed on NCEMRI in detecting synovitis
 393 participants from the MOST
 Hoffa-synovitis scored 0-3 using NCEMRI
 Synovial thickness scored 0-2 on CEMRI in five
parapatellar regions (= reference)
 NCEMRI:
– Sensitivity 71% (infrapatellar) & 88% (intercondylar)
– Specificity 55% (infrapatellar) & 30% (intercondylar)
– No significant association with pain
 CEMRI identifies associations with pain better than
NCEMRI

11. Crema et al.
Continued

12. Crema et al. Osteoarthritis Cartilage 2017
Compare SQ and Q methods for knee synovitis on
NCEMRI and CEMRI
 104 end stage knee OA patients
 NCEMRI used to evaluate effusion-synovitis
and Hoffa-synovitis
 CEMRI used for synovitis assessment
reference standard
 Effusion-synovitis showed superior correlations
and sensitivity than Hoffa-synovitis
– Correlations of effusion-synovitis with synovial
thickness and volume: r = 0.41 and r = 0.43 (P <
.001) r = 0.32 and r = 0.39 (P < .0001)

13. MRI vs. Histology
Loeuille et al. OAC 2011
 30 patients with knee OA
 SQ synovitis assessment using NCEMRI and
CEMRI
– Gold standard = histology of synovial membrane
biopsy
– CEMRI-detected synovitis was correlated with
histologically proven inflammation
– NCEMRI was not associated with histology

14. Loeuille et al. Continued

15. Atukorala et al. Ann Rheum Dis 2014
If synovitis precedes the development of ROA?
 133 knees with ROA, 133 control knees, 4-year
observation
 Persons w/o ROA at baseline
 ORs for occurrence of incidental ROA
associated with the presence of BL effusion-
synovitis: 1.56 – 4.7
 ORs for the occurrence of incidental ROA
associated with the presence of BL Hoffa-
synovitis: 1.80 – 2.47
 Effusion-synovitis and Hoffa-synovitis strongly
predicted the development of incidental ROA

16. Roemer et al. OAC 2010
Anatomical distribution of synovitis and its association
with joint effusion on NCEMRI and CEMRI
 111 patients with knee pain (VIDEO study)
 Anatomical distribution of synovitis and its
association with joint effusion on NCEMRI and
CEMRI
 In a population of mixed ROA severity, the
large majority of knees exhibited definite
synovitis at least in 1 subregion
– Commonest site = posterior to PCL (71.2%)
– Joint effusion as measured on PD FS images does not
only represent effusion but also synovial thickening

17. Roemer et al.
Continued

18. MacFarlane et al. Arthritis Rheum 2019
Association of changes in effusion-synovitis with
progression of cartilage damage
 221 subjects from the MeTeOR study (knee OA +
meniscal tear)
 SQ assessment of effusion synovitis and cartilage
damage
 Patients with extensive effusion-synovitis at baseline had
a relative risk (RR) of progression of cartilage damage
depth of 1.7 (95% CI 1.0-2.7)
 Compared to those with persistently minimal effusion-
synovitis, those with persistently extensive effusion-
synovitis had a significantly increased risk of progression
of cartilage damage depth (RR 2.0 [95% CI 1.1-3.4])

19. Synovitis vs Symptoms
De Lange-Brokaar et al. A&R 2015
 86 subjects with symptomatic knee OA
 SQ assessment of synovitis using CEMRI
 Synovitis at 3 sites was associated with
radiographic severity
 Different patterns of synovitis in knee OA were
observed
 The pattern that included several patellar sites
was associated with pain, whereas other patterns
showed no association, suggesting that pain
perception in patients with knee OA is a localized
response

20. Wallace et al. Arthritis Care Res 2017
Associations between clinical evidence of inflammation
and synovitis in symptomatic knee osteoarthritis
 107 subjects with symptomatic knee OA
 SQ assessment of synovitis using CEMRI
 Significant associations were found between
the number of regions affected by synovitis and
WOMAC pain, effusion, and joint line
tenderness

21. O’Neill et al. Ann Rhem Dis 2016
Synovial tissue volume: a treatment target in knee OA
 120 subjects with painful knee OA, before and
after intra-articular steroid injection
 Synovial tissue volume (STV) measured on post-
contrast MRI
 Synovial tissue volume in knee OA shrinks
following steroid therapy, and rebounds in those
whose pain relapses
=> Synovitis potential treatment target in symptomatic
knee OA

22. Novel Techniques Using NCE-MRI
 Yoo et al. (Radiology, 2017)
 FLAIR-FS sequence with 3T scanner, for peripatellar synovitis
 CE-MRI as gold standard

23. Yoo et al. Continued
 33 patients
 Strong correlations between FLAIR FS and CE
T1W FS in the assessment of peripatellar
synovitis by both readers (correlation coefficient,
0.675–0.973)
 With CE T1W FS as the reference, FLAIR FS
showed relatively good diagnostic performance
for detection of synovitis of any severity (>90%
accuracy)

24. T1-w contrast-enhanced DESS-w subtraction
* *
CE FS T1-w vs. DESS-w Subtraction
McWalter et al. IWOAI 2015

25.  7T MRI now has regulatory clearance to be used
clinically
 Advantages: higher resolution, faster scan time,
specific metabolic imaging (Na+, CEST)
 May have role in further advancing non-enhanced
synovitis assessment
 We could show that CE MRI of synovitis is feasible
 Common assessment approaches feasible at 7T
including SQ, Q, DCE
Roemer et al. Br J radiol, submitted
Multiparametric Synovitis Assessment
at 7T Ultra-High Field MRI

26. Comparison CE vs non CE MRI at 7T DCE at 7T is feasible
Multiparametric Synovitis Assessment
at 7T Ultra-High Field MRI

27. Ultrasound Imaging of Synovitis
 Mostly used in hand OA
 Synovial hypertrophy and
hyperemia (=active synovitis)
 Erosion
 Effusion
 US-detected grey scale synovitis
and Power Doppler signals are
associated with radiographic hand
OA progression after 5 years
Mathiessen A, et al. Ann Rheum Dis 2016
Figure: Hayashi D, et al. Sem Arthritis Rheum 2011

28. PET Imaging of Synovitis
 18FDG PET
 Increased metabolism = active synovitis
 PET-detected synovitis associated with knee pain in OA
 Hybrid PET/CT or PET/MRI for accurate localization
Parsons MA, et al. Nucl Med Commun 2015
Figure: Hayashi D, et al.
PET Clin 2019

29. Summary of synovitis imaging
 Synovitis can be evaluated using NCEMRI
– Hoffa-synovitis and effusion-synovitis (surrogate)
– Novel technical developments promising (e.g. FLAIR-
FS, Diffusion tensor)
 CEMRI enables
– Accurate assessment of synovitis
– Differentiation between synovium and effusion
 CEMRI-detected synovitis is associated with
pain in OA
– Consistent evidence from multiple studies

30.  However, so far, very few high-quality clinical
trials have used synovial inflammation as an
outcome measure and most of them found
no clinical benefits
 Therefore, well-designed trials of the
disease-modifying therapies should consider
synovial inflammation as an important
treatment target in patients with
inflammatory OA phenotypes
Perspective
Wang et al OAC 2018

31. Conclusions (1)
 CEMRI
– Should be recommended if we aim to assess
synovitis accurately in OA (especially in clinical
trials)
– Is a reliable measure of synovitis
– Could be useful in
 clinical trials as a marker of therapeutic response
 clinical practice as a guide to treatment
monitoring

32. Conclusions (2)
 Novel imaging techniques without contrast
administration are promising in showing
synovial inflammation
– Better than Hoffa fat pad surrogate
 Targeting the inflammatory synovium has
great potential to delay or prevent structural
alterations and treat symptoms, especially in
early OA

33. June 26-28, 2019
Charlottetown, Prince Edward Island, Canada
Visit: www.ISOAI.org or www.IWOAI.org

34. What's happening at IWOAI 2019?
• Clinical & Imaging Parameters for DMOAD Trials & Update on DMOAD
Developments
• Pre-workshop on Machine Learning Segmentation Challenge
• Novel Analytics and Computational Approaches
• Linking Imaging with Tissue and Joint Function
• Imaging Early Osteoarthritis
• Phenotypes/Subgroups of Osteoarthritis
• Updates & Insight from APPROACH / OAI / MOST
• Pre-Congress Boat Trip (June 25th) & Post-Congress Bike/Beach Trip (June 29th)
June 26-28, 2019
Charlottetown, Prince Edward Island, Canada
Visit: www.ISOAI.org or www.IWOAI.org