Medical Director AABB Member Orange, California, United States
Background/Case Studies: Hematopoietic Progenitor Cell (HPC) collection requires significant coordination between service lines. Ability to predict CD34 collection yields facilitates these coordination efforts and is valuable for patient counseling.
Study
Design/Methods: HPC collections were performed on validated Spectra Optia (TerumoBCT, Lakewood, CO) apheresis instruments using the continuous mononuclear cell (CNMC) program and fixed 5-hour collection duration by experienced operators. Mobilization regimen was selected by the hematopoietic stem cell transplant team based upon underlying disease, presence of residual disease at time of mobilization, and other factors. Poor mobilization status (ie, pre-CD34/μL of < 10 on day 5 of G-CSF mobilization) would prompt addition of plerixafor where applicable. On each day of collection, a pre-collection peripheral blood CD34/μL (pre-CD34) count and post-collection bag-count would be determined using flow cytometry and standard International Society of Hematotherapy and Graft Engineering (ISHAGE) gating as described elsewhere. The bag count was used to determine the CD34 x 106/kg collection yield. Collection yields were compared to pre-CD34 count using scatterplot and boxplots. Boxplots were analyzed by pre-CD34 count ranges and ranges were progressively collapsed until the interquartile ranges (IQR) no longer overlapped. The final boxplots are represented in Figure A.
Results/Findings: A total of 209 collection events among 95 individual patients for whom pre-CD34 counts and final collection yields were available were analyzed. Data were sorted by pre-CD34 count from smallest to largest and scatterplotted against collection yields. The resultant linear trendline was described by the formula Collection Yield (CD34 x 106/kg) = [(0.0709 x pre CD34/mL) + 0.3415] with an R2=0.8751. After progressive boxplot analysis, the following pre CD34/μL ranges could be determined: 0-9 (n=27/209;13%), 10-19 (47/209;22%), 20-39 (63/209;30%), 40-79 (43/209;21%), and 80-399 (n=25/209,12%). For each of these ranges the median (IQR) collection yield (x106/kg) was 0.6 (0.4-0.7), 1.2 (1.0-1.6), 2.2 (1.7-2.9), 4.4 (2.9-6.3), and 8.9 (5.9-19.1), respectively. Conclusions: The peripheral blood pre-CD34 count drawn the morning of collection can be utilized to predict collection yields. Using a linear trend line, a point estimate of collection yield can be achieved by entering the pre-CD34 count into the formula above. The median (IQR) collection yield can be predicted using boxplot data. Although this data represent a single-center experience using a fixed 5-hour collection, the method presented could be applied by other centers to facilitate collection yield prediction.
