A Nomogram for the Prediction of Progression and Overall Survival in Childhood Acute Lymphoblastic Leukemia

Background: Advances in treatment and supportive care have significantly improved the overall survival (OS) of pediatric patients with acute lymphoblastic leukemia (ALL). However, there is a large number of these patients who continue to relapse after receiving standard treatment. Accurate identification of patients at high risk of relapse and targeted therapy may significantly improve their prognosis. Therefore, the aim of this study was to identify significant prognostic factors for pediatric ALL and establish a novel nomogram for the prediction of survival. Methods: The ALL clinical data of Phases I and II of the Therapeutic Applicable Research to Generate Effective Treatments (TARGET) project were merged and randomly divided into training and validation groups. The LASSO regression model was used to select the specific factors related to the OS of the training group and generate prognostic nomograms according to the selected characteristics. The predictive accuracy of the nomogram for OS was verified using the concordance index of the training and validation groups, the area under the receiver operating characteristic curve for prognostic diagnosis, and the calibration curve. Results: A total of 1,000 children with ALL were included in the TARGET project. Of those, 489 patients had complete follow-up data for further analysis. The data were randomly divided into the training group (n = 345) and the validation group (n = 144). Seven clinical characteristics, namely age at diagnosis, peripheral white blood cells, bone marrow and CNS site of relapse, ETV6/RUNX1 fusion, TCF3/PBX1, and BCR/ABL1 status, were selected to construct the nomogram. The concordance indices of the training and validation groups were 0.809 (95% confidence interval: 0.766–0.852) and 0.826 (95% confidence interval: 0.767–0.885), respectively. The areas under the receiver operating characteristic curve of the 3-year, 5-year, and 10-year OS in the training group were 0.804, 0.848, and 0.885, respectively, while that of the validation group were 0.777, 0.825, and 0.863, respectively. Moreover, the calibration curves demonstrated a favorable consistency between the predicted and actual survival probabilities. Conclusions: Independent predictors of OS in children with ALL included age at diagnosis, white blood cells, bone marrow site of relapse, CNS site of relapse, ETV6/RUNX1 fusion, TCF3/PBX1, and BCR/ABL1 status. The nomograms developed using these high-risk factors can more simply, accurately, and quantitatively predict the survival of children, and improve treatment and prognosis.