A new model for predicting the timing of leukapheresis on the basis of CD34⁺ cell and hematopoietic progenitor cell levels

We developed a model (depending on peripheral CD34⁺ cell count and hematopoietic progenitor cell count) to determine the optimal timing of 3-day leukapheresis in patients pretreated with chemotherapy and G-CSF. Marrow potentials were identified on the basis of three patterns of leukapheretic yield. Pattern 1 predicted good marrow potential. The positive predictive value of a first-day leukapheretic yield of >1 × 10⁶ CD34 ⁺ cells/kg (mean 3-day yield = 8.18 × 10⁶ CD34 ⁺ cells/kg, n = 11) was 100%. Pattern 2 predicted poor marrow potential. The negative predictive value of a 3-day leukapheretic yield of >1 × 10⁶ CD34⁺ cells/kg (3-day yield = 0.26 × 10⁶ CD34⁺ cells/kg, n = 1) was 100%. Pattern 3 met neither of the above criteria (mean 3-day yield = 1.37 × 10⁶ CD34 ⁺ cells/kg, n = 19). The marrow potential was borderline and patients could be further divided into two subgroups according to peripheral CD34 ⁺ cell counts when WBC reached >10,000/μ1. The mean yield differed significantly between pattern 1 and 3 (P < 0.001). For patients with good marrow potential, leukapheresis should begin as soon as the WBC count is >5,000/μ1. Patients with borderline marrow potential may benefit from delaying leukapheresis until the WBC level is >10,000/μ1 and leukapheresis extended more than 3 days.