Validation of circulating CD34+ cell counts as criteria for initiating peripheral blood stem cell collection in pediatric patients

Abstract

Background & Aim Circulating CD34+ cell (cCD34) counts have been used as a criteria for determining if donors for peripheral blood stem cell harvest are ready for apheresis collection. Different cCD34 counts have been proposed as thresholds for initiation of stem cell collection, but few studies have focused on the validation of these criteria. The aim of this study is to validate an empirical policy of stem cell collection initiation using two independent datasets for statistical training and validation. Methods, Results & Conclusion Retrospective chart reviews were performed to compile patient data in a statistical training dataset that covered the years 2002-2004 and a validation dataset that covered 2005-2018. The training dataset comprised 29 patients and 82 collections. The pre-collection cCD34 counts and the numbers of CD34+ cells collected in this dataset showed a good correlation between the two parameters (Table). cCD34 counts = 9 cells/μl led predominantly to collections with CD34+ cells >=2.0 × 10^6/kg (Good Collections). Basing on this result, a procedural policy was established in which stem cell collections would not proceed if pre-collection cCD34 counts were = 9.0 cells/μl. For cCD34 counts between 4.5 and 9 cell/μl, the decision whether to proceed would be made at the clinician's discretion. For the detection of Poor Collections in the training dataset, the Sensitivity of the criteria was 46.4% (13/28), Specificity was 100% (40/40), Positive Predictive Value (PPV) was 100% (13/13), and Negative Predictive Value (NPV) was 72.7% (40/55). The validity of this policy was then verified using an independent validation dataset comprising 75 patients and 86 collections. The percentage of Poor Collections decreased from 51.2% in the training dataset to 24.4% in the validation dataset, indicating a substantial reduction of inadequate collections after implementation of the policy. For the detection of Poor Collections in the validation dataset, the Sensitivity of the criteria was 35.3% (6/17), Specificity was 100% (63/63), PPV was 100% (6/6), and NPV was 85.1% (63/74). In summary, the use of cCD34 counts to predict if stem cell collections will be Poor Collections is a reliable criteria that can reduce poor collections by almost 50%, thereby saving resources and decreasing unnecessary risks to the donors.