Improving umbilical cord blood processing to increase total nucleated cell count yield and reduce cord input wastage by managing the consequences of input variation

Abstract

This work could pave the way for cord blood banks to improve UCB processing and increase efficiency through higher yields and lower costs.

Challenges to the Sustainability of the U.S. Public Cord Blood System

“…[121][122][123][124][125][126] Not surprisingly, the likelihood of both CBU use and transplant success is greatest among high-TNC count units. 81,121,[127][128][129] Among all transplanted CBUs from NMDP data from 2007 to 2016, only 11 percent had a TNC count below 1.25×10 9 (see Figure 7.1). The share of transplanted CBUs, however, was 28 percent and 59 percent for CBUs with TNC counts of 1.25-1.75×10 9 and 1.75×10 9 or more, respectively.…”

“…Based on an analysis of 1,839 CBUs collected by the Anthony Nolan Cell Therapy Centre (ANCTC), the population-average TNC count is estimated to be about 1.10×10 9 , with a standard deviation of 0.65×10 9 ; the distribution is skewed toward lower TNC counts. 121 Thus, assuming a lognormal distribution implies that 40 percent of collected CBUs would not be banked if the TNC threshold matched the HRSA threshold of 9.0×10 9 . A TNC threshold of 1.50×10 9 would yield a discard rate of 72 percent.…”

“…These concerns can be mitigated by ensuring that individuals collecting the CBUs are highly trained (training is correlated with collection of higher-TNC count units 131,132 ), or endeavoring to improve processing and reduction in waste of the already collected CBUs. 121 The current structure of the NCBI program complicates this decision for banks because the program implicitly incentivizes the collection of lower-TNC count units by allowing CBBs to collect subsidies for banked units with TNC counts as low as 0.9×10 9 .…”

“…Figure 6.2. 121 We assumed that the distribution of TNC counts is log-normally distributed and estimate its mean and standard deviation. While Naing et al do not provide information on individual data points, they provide information on the mean (1.086×10 9 ) and standard deviation (0.651×10 9 ) of their data.…”

“…While Naing et al do not provide information on individual data points, they provide information on the mean (1.086×10 9 ) and standard deviation (0.651×10 9 ) of their data. 121 This allowed us to estimate the log-normal preprocessing TNC-count distribution. First, we took 1,000 random draws from a normal distribution with mean 1.086×10 9 and standard deviation 0.651×10 9 .…”