Predictors of remission and relapse in retroperitoneal fibrosis

Objectives: In retroperitoneal fibrosis (RPF), glucocorticoids (GC), alone or in combination with immunosuppressive agents, induce remission in 80%–90% of patients but up to two thirds of them relapse. There is limited knowledge on outcome predictors in RPF. We aimed to identify clinical, laboratory and imaging predictors of remission and relapse in RPF. Methods: We included consecutive RPF patients treated with 6–9-month courses of GC with/without immunosuppressive agents. Baseline and post-treatment computed tomography, magnetic resonance imaging and 18F-fluorodeoxyglucose (18F-FDG) positron emission tomography (PET) were assessed. The potential predictive value of the examined parameters as predictors of remission and time-to-relapse was analysed using logistic and Cox regression models. Results: Of 152 patients screened, 115 were included. Of them, 101 (87.8%) achieved remission a median of 4 months (interquartile range 3–5) after starting treatment. At multivariable analysis, smoking (odds ratio [OR] 0.34, 95% confidence interval [CI] 0.11–0.99) and atypical RPF localization (e.g., pelvic) (OR 0.11, 95% CI 0.02–0.52) were negatively associated with remission, whereas pre-treatment 18F-FDG-PET activity was positively associated (OR 11.51, 95% CI 1.35–98.20). A median of 33 months (17–57) after treatment initiation, 42% patients relapsed (median time from remission to relapse, 14 months [8–26]). Thoracic vessel involvement and positive 18F-FDG-PET at the end of treatment independently predicted relapse (hazard ratio [HR] 2.61, 95% CI 1.19–5.68 and HR 3.47, 95% CI 1.54–7.82, respectively). Conclusions: Metabolic activity of RPF at 18F-FDG-PET is an important predictor of remission and relapse. Smoking and atypical localization are negatively associated with remission, whereas thoracic aorta involvement is associated with relapse risk.