Nomogram Model for Dynamic and Individual Prediction of Cardiac Response and Survival for Light Chain Amyloidosis in 737 Patients With Cardiac Involvement

ObjectiveLight chain amyloidosis (AL) with cardiac involvement is associated with poor prognosis. The existing prognostic assessment system does not consider treatment-related factors, and there is currently no effective system for predicting the response. The purpose of this study was to build an individualized, dynamic assessment model for cardiac response and overall survival (OS) for AL patients with cardiac involvement. MethodsThe records of 737 AL patients with cardiac involvement were collected through cooperation with 18 hospitals in the Chinese Registration Network for Light-chain Amyloidosis (CRENLA). We used univariate and multivariate analyses to evaluate the prognostic factors for OS and cardiac response. Then, two nomogram models were developed to predict OS and cardiac response in AL patients with cardiac involvement. ResultsA nomogram including four independent factors from the multivariate Cox proportional hazards analysis-Mayo staging, courses of treatment, hematologic response, and cardiac response-was constructed to calculate the possibility of achieving survival by adding all the points associated with four variables. The higher the score, the more likely death would occur. The other nomogram model included the courses of treatment, hematological response, and different treatment regimens, and was correlated with cardiac response. The higher the score, the more likely a cardiac response would occur. ConclusionIn conclusion, based on the large Chinese cohort of patients with AL and cardiac involvement, we identified nomogram models to predict cardiac response and OS. These models are more individualized and dynamic, and therefore, they have important clinical application value.

Automated summary

Based on a large Chinese cohort of AL patients with cardiac involvement, nomogram models were identified to predict cardiac response and OS. These models, which consider treatment-related factors, are individualized and dynamic, and have important clinical application value.