Establishment And Clinical Application Of A Novel Prognostic Score System For Acute Lymphoblastic Leukemia

Background and purposeAcute lymphocytic leukemia(ALL)is a hematological malignant disease.The cure rate of adolescents and adults ALL is still poor.Accurately identifying high-risk patients and deciding personalized strategy can significantly improve such patients’prognosis.There is currently no study to establish a prognostic model for predicting the prognosis of adolescents and adult ALL patients.This study aims to establish a simple and effective prognostic model for≥14-year-old ALL patients based on the large sample of data from a single-center to decide clinical treatment decisions.MethodsWe retrospectively analyzed the clinical data of 321 primary diagnosed ALL patients of the Department of Hematology,Fujian Medical University Union Hospital from January 2017 to June 2020.These patients were randomly divided into the training set and the validation set with a ratio of 2:1.Cox regression model was applied to analyze the independent factors affecting the prognosis,and the nomogram was used to construct a prognostic model.ROC curve,decision curve analysis,and time-dependent ROC was used to evaluate the predictive accuracy,clinical applicability,and stability,respectively.ResultsA total of 321 ALL patients were included,of which 214 were in the training-set and 107 were in the validation-set.The median follow-up time was 28(1-48)months.There was no significant statistical difference between the training-set and validation-set In the clinical data(all P>0.05).Multivariate Cox regression analysis of the training-set showed that age>50 years,WBC count>27.83×10~9/L,no response of induced treatment,and relapse were independent risk factors(all P<0.05).Platelet count>131×10~9/L was the independent protective factor(P<0.05).The nomogram was established according to these independent prognostic factors in the training set,and the corresponding total score can be calculated according to each clinical variable.Age≤50years old scored 0 points,and age>50 years old scored 10 points;Patients with negative response of induced treatment scored 8 points,and patients with positive response of induced treatment scored 0 points;Patients without relapse scored 0 points,and patients with relapse scored 7.375 points;PLT≤39×10~9/L scored 9.125 points,PLT between39×10~9/L and 131×10~9/L scored 4.5 points,and PLT>131×10~9/L scored 0 points;WBC≤27.83×10~9/L scored 0 points,and WBC>27.83×10~9/L scored 4.75 points.The total score≤15.9375 points was defined as the low-risk group,and the total score>15.9375 points was defined as the high-risk group.The calibration curve of both two sets showed the predicted probability was consistent with the actual probability.AUC value was 0.811(95%CI:0.753-0.868).The time-dependent ROC curve confirmed that the prognostic model was well stable,and decision curve analysis showed the nomogram has good clinical applicability.The survival curves of either overall survival(OS)or progression-free survival(PFS)showed that the prognosis of low-risk patients was significantly better than that of high-risk patients(P<0.001).Further analysis showed that in both the training-set and validation-set,the prognosis of high-risk patients was significantly worse than that of low-risk patients(all P<0.001).The survival curve of bone marrow relapse revealed that the OS of patients with relapse was significantly worse than that of without relapse(P<0.001),and the 3-year OS rate was10%(95%CI:6%-14%)and 57%(95%CI:53%-61%),respectively.The survival curves showed for both OS and PFS,the prognosis of ALL with transplantation was significantly better than that of without transplantation(all P<0.001).Further stratified analysis showed that in the high-risk patients,the OS and PFS of patients with transplantation were significantly better than that of without transplantation(all P<0.05).While in the low-risk patients,compared with non-transplantation patients,receiving transplantation treatment can only significantly prolong the PFS of ALL patients(P<0.05),but not benefiting the OS of ALL patients(P>0.05).Conclusion We have established a simple and effective prognostic prediction model for≥14-year-old ALL patients that can provide the accurate risk stratification and decide the clinical strategy based on the evidence-based medicine.