Pleural effusion-based nomogram to predict outcomes in unselected patients with multiple myeloma: a large single center experience.

Pleural effusion (PE) is prevalent in unselected "real-life" populations of multiple myeloma (MM). However, its prognostic value on MM is currently elusive. This study aimed to explore the role of PE on MM prognosis and to develop a novel prognostic nomogram for a cohort of Chinese patients with MM. Patients diagnosed with MM form 2000 through 2017 were retrospectively enrolled. PE was evaluated by chest computed tomography (CT) scans. Independent predictors of overall survival (OS) were identified using a multivariable Cox regression model performed on variables selected by the least absolute shrinkage and selection operator (LASSO) algorithm. A nomogram was constructed based on these variables. The concordance index (C-index) and the calibration curve were used to evaluate the predictive performance of the nomogram. Among 861 patients analyzed, 368 patients developed PE. Multivariate cox regression and restricted mean survival time (RMST) analyses revealed that patients with PE experienced worse OS vs. patients without PE. A nomogram predictive of OS was constructed using PE, plasma cell proportion, international staging system (ISS) stage, Charlson comorbidity index (CCI), 1q21 gain, and autologous hematopoietic stem cell transplantation (HSCT). The nomogram showed satisfactory discrimination in the derivation cohort (C-index=0.729) and the validation cohort (C-index=0.684), outperforming the Durie-Salmon (DS) and ISS staging systems. Moreover, the nomogram accurately classified patients into two distinct high- and low-risk groups. PE is frequently encountered in the disease course for MM patients. We derivated and validated a novel nomogram for MM based on PE, outperforming the DS/ISS staging systems.