Dynamics of haematological parameters in the prediction of relapse in B-Acute lymphoblastic leukemia.

Abstract

Background: Relapse of disease is the ultimate cause of poor survival in B-Acute lymphoblastic leukemia (B-ALL) and the biggest challenge to treat. With the availability of novel therapies, including targeted therapies, the survival of relapsed patients can be improved if it can be detected at its initial stage. Although many studies have identified risk factors associated with a high probability of relapse, there is no clear indicator that can indicate impending relapse in its initial stage.

Objectives: To investigate the utility of Complete blood cell count (CBC) parameters in predicting impending relapse in pediatric B-ALL patients.

Method: In this retrospective study, we studied 90 pediatric B-ALL patients with age ≤16years, including 30 relapse cases and 60 cases without relapse, registered from January 2015 to June 2015. All these B-ALL cases were diagnosed and treated in Tata Memorial Centre (TMC), Mumbai.

Laboratory parameters were noted from the TMC’s electronic medical records (EMR).

17 Complete blood count (CBC) parameters included were studied at three time points in patients with relapse i.e. 3-6 months before relapse, 1-3 months before relapse and at the time of relapse. While in non-relapse patients, time point studied was at 2 years follow-up from initiation of therapy or last follow-up documented in the EMR.    

Result: Of the 17 CBC parameters studied, platelet count (p=0.026) at 1-3 months before relapse was significantly associated with B-ALL relapse.

Similarly, there was a significant difference between the CBC parameters at the time of relapse with reduced hemoglobin (p=0.015), RBC count (p=0.001), platelet count (p<0.001), absolute neutrophil count (p<0.01), % neutrophil (p=0.037), % monocyte (p<0.01), % eosinophil (p<0.01), % large unstained cells (p<0.01). 

No significant difference was seen between the relapse group (time point 3-6 months before relapse) and the non-relapse group (2 year follow-up / last follow-up).

The ROC - based cut off for Platelet count at the time-point of 1-3 months before relapse and 2yrs follow-up of non-relapsed patients was found to be 340 X 10^9/L (AUC of 62.9%) with 30% sensitivity and 97% specificity.

On the basis of Median platelet count of 244 X 10^9/L at time point of 1-3 months before relapse, the relative risk for platelet count was 1.462 (0.982, 2.174) with p-value 0.0612, which is close to significant value. Thus it showed that risk of getting relapse increases by 46.2% in patients with platelets <244 X 10^9/L as compared to platelets ≥244 X 10^9/L.

Conclusion: Thus, we concluded that platelet count have a high potential to predict impending relapse and provide a basis to investigate further for confirmation with ancillary tests. CBC testing is cheap and available in all laboratories. Hence it can be part of routine close monitoring of B-ALL patients.