OBJECTIVE Magnetic resonance imaging (MRI) is a standardized method for assisting joint diagnosis. To validate the reliability of different imaging-based grading systems, this study examined (1) the associations between grading systems for osseous change, joint effusion, and the Wilkes classification of temporomandibular joint (TMJ) disorders and (2) the correlation between cytokines in synovial fluid and imaging-based joint scores. METHODS Twenty-seven patients, who routinely received numeric rating scale (NRS) and MRI assessment before TMJ arthrocentesis, were enrolled. Each joint was evaluated through the grading criteria for severity of osseous change and joint effusion by blinded observers using MRI. ImageJ was employed for classifying joint effusion. Joint synovial fluid, collected through arthrocentesis, was examined for cytokine expression by using a Luminex multiplex assay. All data were analyzed using the Pearson correlation analysis. RESULTS The Wilkes classification was strongly correlated with osseous change scores, but not with joint effusion scores. Joint effusion scores significantly correlated with NRS scores, but not with the Wilkes classification and osseous change scores. Compared with osseous change scores, joint effusion scores had a higher correlation with the levels of inflammatory cytokines (interleukin (IL)-8 and soluble IL-6 receptor (sIL-6R)) and with anti-inflammatory cytokines (soluble tumor necrosis factor receptors I and II (sTNF-RI/II)). CONCLUSIONS In patients with TMJ disorders, MRI grades are strongly correlated with NRS scores and levels of cytokines (IL-8, sIL-6R, and sTNF-RI/II) in the synovial fluid. CONCLUSIONS Joint effusion scoring can be a reliable and valid indicator for pathological assessment of TMJ disorders.