Dynamic nomogram for predicting generalized conversion in adult-onset ocular myasthenia gravis

Abstract

Abstract Purpose To explore the factors and risk mapping model of progression from ocular myasthenia gravis (OMG) to generalized myasthenia gravis (GMG) in adult-onset patients. Methods A retrospective, observational cohort study was performed for 435 OMG patients with onset age older than 14 years old. Multivariate Cox regression was used to identify the independent factors affecting generalized conversions that then were incorporated into the construction of the nomogram. Results Two hundred thirty-seven patients (54.5%) had transformed into GMG after a median of 1.1 years (range 0.1-–9.1 years). The 6-, 12-, and 24-month generalized conversion rates were 31.7%, 49.8%, and 65.4%, respectively. Multivariable analysis showed that the early-onset age, male sex, concomitant autoimmune diseases (AID), positive results of anti-acetylcholine receptor antibodies, repetitive nerve stimulation abnormalities, the presence of thymoma, and prednisone treatment were significantly associated with the generalized conversions (hazard ratio [HR] = 0.598, 0.686, 1.554, 1.541, 2.020, 2.510, and 0.556, respectively). A nomogram was established to predict the possibility of generalization-free survival (GFS) in adult-onset OMG patients, and the model demonstrated good predictive performance with a C-index of 0.736 (95% confidence interval 0.703 ~ 0.769). Moreover, subgroup analyses were performed based on the presence or absence of prednisone therapy, and the results indicated that prednisone therapy has better prevention of generalized conversions in male, non-thymoma patients, and patients without other AID. Conclusion A new predictive nomograph and web-based survival calculator we developed show favorable applicability and accuracy in predicting long-term GFS in adult-onset OMG patients.