Development and validation of a predictive model for tumor lysis syndrome in childhood acute lymphoblastic leukemia

Background

Tumor lysis syndrome (TLS) frequently manifests shortly after induction chemotherapy for acute lymphoblastic leukemia (ALL), with the potential for swift progression. This study endeavored to develop a nomogram to predict the risk of TLS, utilizing clinical indicators present at the time of ALL diagnosis.

Methods

We retrospectively gathered data from 2243 patients with ALL, spanning December 2008 to December 2021, utilizing the clinical research big data platform of the National Center for Clinical Research on Children's Health and Diseases. The Least Absolute Shrinkage and Selection Operator (LASSO) method was employed to filter variables and identify predictors, followed by the application of multivariate logistic regression to construct the nomogram.

Results

The LASSO regression identified six critical variables among ALL patients, upon which a nomogram was subsequently constructed. Multifactorial logistic regression revealed that an elevated white blood cell count (WBC), serum phosphorus <2.1 mmol/L, potassium <3.5 mmol/L, aspartate transaminase (AST) ≥50 U/L, uric acid (UA) ≥476μmol/L, and the presence of acute kidney injury (AKI) at the time of initial diagnosis were significant risk factors for the development of TLS in ALL patients (P<0.05). The predictive model achieved an area under the receiver operating characteristic curve (AUC) of 0.824 [95 % CI (0.783, 0.865)], with an internal validation AUC of 0.859 [95 % CI (0.806, 0.912)]. The Hosmer-Lemeshow goodness-of-fit test confirmed the model’s robustness (P=0.687 for the training cohort; P=0.888 for the validation cohort). Decision curve analysis (DCA) indicated that the predictive model provided substantial clinical benefit across threshold probabilities ranging from 10 % to 70 %.

Conclusions

A nomogram incorporating six predictive variables holds significant potential for accurately forecasting TLS in pediatric patients with ALL.