Related Experiment Video
Updated: Mar 17, 2026
Establishing a Competing Risk Regression Nomogram Model for Survival Data
Published on: October 23, 2020
Nomogram to predict hemorrhage risk related to anti-tumor therapy in patients with acute leukemia
Xinxin Hu1, Ying He2, Yinghua Xie1
1Department of Hematology, Shanghai Fifth People's Hospital, Fudan University, Shanghai, China.
Introduction:
This study aimed to identify clinical characteristics associated with hemorrhagic events in acute leukemia patients who received anti-tumor therapies and to develop and evaluate a prediction nomogram for hemorrhagic events based on those characteristics.
Method:
This retrospective cohort study included 468 acute leukemia patients, excluding those with acute promyelocytic leukemia, treated at The Shanghai Fifth People's Hospital and Nanyang Municipal Central Hospital between January 2013 and December 2023. The primary endpoint was World Health Organization (WHO) grade 2 or higher hemorrhagic events related to anti-tumor therapy. Patients were randomly divided into training and test groups at a ratio of 7:3. In the training group, univariable logistic analysis and least absolute shrinkage and selection operator (LASSO) regression were performed to identify significant predictors, which were then used to construct a prediction nomogram for hemorrhage risk. Nomogram performance was evaluated by receiver operating characteristic (ROC) curve analysis, calibration curve analysis, and decision curve analysis (DCA). The following five independent variables were identified as predictors of anti-tumor therapy-related hemorrhagic events in acute leukemia patients and used to develop a prediction nomogram: infection status, types of different hemorrhage prevention drugs and blood products administered, platelet (PLT) transfusion, hematocrit, and PLT count.
Result:
On ROC curve analysis, the nomogram exhibited satisfactory performance in both the training group [area under the ROC curve (AUC)=0.741] and test group (AUC=0.718). Calibration plots showed a high degree of consistency between the actual and nomogram-predicted survival rates in both groups, and the nomogram showed good clinical utility on DCA. We successfully developed and validated a nomogram for predicting the risk of anti-tumor therapy-related hemorrhage of WHO grade 2 or higher among patients with acute leukemia.
Conclusion:
This nomogram may provide a practical and user-friendly tool for clinical practice once further validated in perspective large cohort or trials.
