To further investigate if MRI findings or the microscopic characteristics of synovitis could be used as criteria for the indications of knee replacement.

Thirty-one patients with end-stage knee osteoarthritis (OA) who met the fourth and fifth grade of the Ahlback classification and excluded other diseases were selected and received the following examinations: (1) noninvasive MRI evaluation for grading the synovial thickness of the investigated sites (i.e., medial and lateral parapatellar recesses, and medial and lateral suprapatellar pouches; synovitis was categorized into grade 0~3 on the basis of the thickness); (2) hematoxylin and eosin (HE) staining and pathological score grading of samples from the synovial membrane at each observed site; (3) Pearson’s linear test or Spearman rank test which was expressed as Spearman rank correlation (rs) was used to analyze the correlation between synovial thickness and pathological scores.

Statistical analysis revealed no significant difference in MRI synovial thickness grade across the different regions of the knee. The mean pathological scores of the different regions showed no significant differences, neither did the mean pathological parameter scores (F=2.146, P>0.05). The grade of MRI-determined thickness of synovitis in patients with end-stage knee OA was positively correlated with the pathological scores (rs=0.81, 0.56, 0.59, 0.52 respectively; P <0.01) and the three pathological parameters (rs=0.4, 0.47, 0.36; P <0.05).

There is positively correlation between the grades of MRI-determined thickness and pathological scores of synovial membrane in patients with end-stage knee OA combining synovitis, which can be used as drug selection reference of synovitis-sensitive OA patients and the criteria of therapeutic efficacy for knee joint replacement.