The prediction of haemophilic arthropathy progression based on MRI findings and clinical characteristics.

objectiveTo identify magnetic resonance imaging (MRI) and clinical characteristics that are closely associated with the progression of haemophilic arthropathy (HA) after different therapies and to establish a prediction model for HA progression using Cox proportional hazards regression, thus facilitating the development of personalized clinical replacement therapy plans. MATERIALS AND

methodsRetrospective clinical and imaging data were collected from HA patients registered at the Henan Provincial Registration Management Center of Haemophilia from December 2010 to May 2023. The inclusion criteria were joints with a history of haemorrhage and initial/posttreatment reevaluation with X-ray and MRI. Joints with severe damage (i.e., a Pettersson score > 6) were excluded. Joint disease progression was defined as a > 1-point increase in the Pettersson score. Progression-free survival (PFS) was the primary outcome. MRI observations revealed joint effusion, synovial hypertrophy, haemosiderin deposition, bone destruction or cystic degeneration at the joint margins, and cartilage destruction. Age, body mass index (BMI), factor VIII (FVIII) activity, activated partial thromboplastin time (APTT), prothrombin time (PT), therapy type, annual joint bleeding rate (AJBR), and the Haemophilia Joint Health Score (HJHS) were also assessed. Subsequently, univariate and multivariate Cox proportional hazards regression models were employed to analyse the clinical and imaging characteristics influencing HA progression. Factors with a P < 0.15 in univariate analysis were subsequently included in the multivariate analysis. The impact of various imaging and clinical characteristics on PFS was assessed via Kaplan‒Meier (K-M) survival curves.

resultsThis study included 98 joints across 65 patients. During the follow-up period, 63 joints exhibited progression. Both univariate and multivariate Cox analyses revealed that MRI-detected synovial hypertrophy (MRI-SP) was an independent risk factor for HA progression. Incorporating BMI into the model improved its predictive performance (Model 1: c-index = 0.671, P < 0.01). Spearman's correlation analysis revealed strong correlations between baseline MRI-SP and detected haemosiderin deposition (r = 0.73) as well as AJBRs (r = 0.66). K-M survival curves indicated that patients receiving prophylactic treatment and those with less severe MRI-SP had better progression-free survival.

conclusionMRI-detected synovial hypertrophy is an independent risk factor for HA progression. The predictive model, which includes BMI as a covariate for assessing the risk of HA progression, can serve as an auxiliary tool for developing personalized treatment plans for HA patients.