Related Experiment Video
Updated: Jan 12, 2026

Competing-Risk Nomogram for Predicting Cancer-Specific Survival in Multiple Primary Colorectal Cancer Patients after Surgery
Published on: September 27, 2024
Real-World Data Analysis on Early Mortality in Chinese Multiple Myeloma Patients: Analysis of 1093 Patients During a
Xue Yang1, Pu Wang1, Tianhong Xu1,2
1Department of Hematology, Zhongshan Hospital, Fudan University, Shanghai, China.
Abstract:
Despite the survival benefit observed in recent decades, early mortality (EM) remains critical in newly diagnosed multiple myeloma (NDMM) patients, with heterogeneous definitions and risk factors across studies. Besides, prognostic models for predicting EM are limited, especially in real-world scenarios. This single-center retrospective study performed comprehensive analyses as well as risk prediction models of 1093 NDMM patients diagnosed between 2007 and 2021 in China, finding the EM rate was 5.9% at 2 months, 8.8% at 6 months and 13.3% at 12 months. Infection and cardiac events were documented as the top two most common causes of early death. Risk factors that were independent predictors of both EM and OS mainly included cardiac amyloidosis, stroke, del (17p) and poor performance status. ISS stage, platelet count, treatment regimen and concomitant cardiac disease influenced OS rather than EM. Multivariate model-based nomograms were constructed to estimate 2-month, 6-month, 12-month mortality, plus OS, enabling individualized outcome prediction, validated via receiver operating characteristic (ROC) and calibration curves. EM is associated with multiple factors in MM patients, and early identification of patients with higher risk of EM may translate into tailored management in clinical practice.
Related Concept Videos
Cancer Survival Analysis
Comparing the Survival Analysis of Two or More Groups
Actuarial Approach
Consider the example of a high-risk surgical procedure with significant early-stage mortality. A two-year clinical study is conducted,...
Kaplan-Meier Approach

