AbstractBackground and Objectives
Stem cell collection is a standard procedure for the procurement of autologous grafts to rescue myelosuppression induced by high-dose treatments. Accurate prediction of collection yields may contribute to optimize planning and quality control of collection.Materials and Methods
Data of 313 autologous haematopoietic stem cell (AHSC) evaluable collections performed in 208 patients with haematologic and non-haematologic neoplasms from seven centres were prospectively analysed to test the accuracy of yield predictions generated by a formula that required the input of peripheral blood (PB) CD34+ cell precount and desired PB volume to be processed. Data were matched in a standard linear regression, in a zero-point regression analysis and tested for prediction accuracy. Further 165 AHSC collections were analysed on a single-centre basis, using yield predictions as reference standards.Results
Analysis showed high levels of correlation between measured collection yields (my) and predictions (py) (R = 0·85; P = 0·000000) as well as high degree of prediction accuracy (my vs. py at paired t-test: P = 0·114781; median my/py ratio = 1·23). Analysis of additional 165 AHSC collections on a single-centre basis showed that the analysed centres had 70% or more measured yields comprising the 0·6–1·8 interval of the my/py ratio. The observance of the ‘efficiency’ my/py interval assured collection quality control in these centres confirming the reliability of the method.Conclusions
This prediction method generates accurate and immediate yield predictions allowing collection planning and rapid efficiency control. As a consequence of our study, four centres out of seven use the described method to plan both leukapheresis number and single-procedure blood processing volume while the remaining three centres plan leukapheresis number on the basis of our predictions, maintaining a fixed single-procedure 200 ml/kg blood volume processing, according to their centre AHSC collection policy.