Development and Validation of a Nomogram for Predicting Anti-Tumor Therapy–Related Hemorrhage in Acute Leukemia

The goal of this work was to identify clinical indicators associated with bleeding episodes among individuals with acute leukemia receiving anticancer regimens and to develop and test a predictive nomogram based on those indicators. Encompassing 468 acute leukemia subjects, with acute promyelocytic leukemia diagnoses omitted, this retrospective cohort investigation drew on records from The Shanghai Fifth People’s Hospital and Nanyang Municipal Central Hospital covering January 2013 to December 2023. The primary outcome measure was World Health Organization (WHO) grade 2 or higher hemorrhagic events associated with anticancer treatment. Subjects were randomly split into derivation and validation subsets at a 7:3 ratio. Within the derivation subset, univariable logistic regression and least absolute shrinkage and selection operator (LASSO) regression were applied to identify meaningful predictor variables, which were subsequently used to construct a hemorrhage risk nomogram. The nomogram’s utility was assessed using receiver operating characteristic (ROC) curves, calibration curves, and decision curve analysis (DCA). Five independent variables emerged as correlates of bleeding associated with anticancer treatment in patients with acute leukemia. They were used to assemble the predictive nomogram: presence of infection, classes of distinct anti-hemorrhage prophylactic agents and blood derivatives delivered, platelet (PLT) transfusion, hematocrit value, and PLT count.  ROC curve evaluation revealed that the nomogram possessed sound discriminatory capacity within both the derivation subset [area under the ROC curve (AUC) = 0.741] and the validation subset (AUC = 0.718). Calibration graphs showed strong concordance between observed and nomogram-estimated probabilities across both subsets, and DCA indicated a solid net clinical benefit. A nomogram to estimate the risk of anticancer treatment-attributable hemorrhage of WHO grade 2 or higher in patients with acute leukemia was developed and validated. Pending corroboration in prospective large-scale cohorts or experimental studies, this nomogram may provide a pragmatic, intuitive aid for bedside use.