CD34+ progenitor cells meet metrology

Some flow cytometric (FCM) immunophenotypic analyses of blood cells, like the enumeration of CD4+ T lymphocytes and CD34+ hematopoietic progenitor cells (HPC) can be now fully included in the category of quantitative assays, since a calibration curve can be set and standard reference material is available to determine a result [1]. Comprehensive guidelines for the correct analysis of such measurands are published 2], [3], [4, due to their important clinical impact, and national or international EQA/PT schemes have long been made available to assist the involved laboratorians in the routine practice.

The knowledge of the measurement uncertainty (MU) is commonplace in Clinical Chemistry, with open access tables reporting MU, Biological Variability (BV) and Total Error (TE) parameters for the majority of measurands [5, 6]. The prerequisite for the definition of these crucial performance parameters is that their expected BV is rather stable and predictable in most cases, and the acceptable, desirable and optimal levels of the total allowable imprecision can be therefore calculated using the seminal Fraser’s formulae 7], [8], [9], [10.

The BV of basic peripheral blood lymphocyte subset enumeration by FCM has been also recently estimated, thanks to the large experience on reference range studies and the overall limited quantitative percent and absolute level variations of the different subpopulations also in pathological conditions 11], [12], [13], [14.

The absolute or percent level of CD34+ HPC in peripheral blood, bone marrow and leukapheresis concentrates, on the contrary, is indeed a very peculiar measurand, posing unique methodological problems to laboratorians.

CD34+ cells have long been assumed as ‘blast equivalents’, namely normal or neoplastic cells able to generate a progeny both in vivo and in vitro. Although this assumption proved useful in the laboratory practice, it is however imprecise, since not all blasts express the CD34 antigen and not all CD34+ cells are genuine blasts, like B cell precursors (hematogones) and circulating endothelial cells.

The measurement of CD34+ cell levels by FCM has progressively assumed a great clinical and diagnostic importance, both in laboratory hematology and especially in the setting of autologous and allogeneic HPC transplantation [15].

Healthy subjects in steady-state show typically baseline levels between 0 and 10 CD34+ cells/μL, or less than 0.1 % of leukocytes in peripheral blood [16], and such levels remain remarkably stable in the absence of external stimuli, tissue repair phenomena, bone marrow regeneration or chemotherapy treatments. This measurement has put the basis for high-sensitivity FCM procedures and CD34+ cell count can be considered as the prototype of rare event analyses of clinical relevance [16]. Any evidence of CD34+ cells above 10/μL in peripheral blood may indicate that something wrong is occurring in the bone marrow hemopoietic machinery [17, 18], like myelodysplastic syndrome, myelofibrosis, myeloproliferative disorders or evolution to acute leukemias 19], [20], [21], [22], [23. In this regard, the FCM semi-quantitative assessment of CD34+ blasts in the peripheral blood or bone marrow of patients with acute leukemias remains outside the scope of true quantitative measurements.

The autologous or allogeneic HPC transplantation requires the rapid mobilization of large amounts of CD34+ cells into the peripheral blood of patients or healthy donors, and this can be accomplished thanks to the use of the clinically most relevant mobilizing agents, G-CSF (Filgrastim) and Plerixafor [24]. In good mobilizers, the absolute levels of CD34+ HPCs can raise from baseline levels up to 300 or more/μL in few days, with a great individual variability. The decision to trigger a leukapheresis procedure is made to collect as many CD34+ HPCs as possible, to provide at least 2 million HPCs per kg b.w. of the recipient [15], and this may be typically accomplished as soon as 10–20 to 50 CD34+/μL cells appear in the peripheral blood of mobilized subjects, according to local clinical policies [25].

The resulting leukapheresis bag contains a very concentrated and heterogenous mixture of peripheral blood cells, in which the final collection of CD34+ HPCs is measurable at quite variable but very high concentrations, ranging from some 500 to about 2000 or more/μL.

Not to mention the analysis of HPC in cord blood and in the thawed apheresis bags, presenting further complications, it is evident that with such diverse biological matrices and absolute level dynamics, the attempts to estimate the BV of CD34+ HPCs is a particularly challenging task. Due to the mentioned issues, it is not surprise that the CD34+ HPC level – as a measurand – lacks to date tables reporting MU, BV and TE.

The study by S. Fernández-Luis and Colleagues [26] is at present the first attempt to define the percent allowable total error (%aTE) of FCM absolute CD34+ cell counting, using data from a nationwide Spanish EQA scheme collected over the last twelve years. About 3,700 valid results were obtained from 40 laboratories that received fresh peripheral blood, cord blood or buffy-coat samples, that were analyzed by the ISHAGE method or in-house procedures, mostly by a single-platform technique.

The %aTE was calculated as the state-of-the-art (SOTA) performance of the 80th percentile of the results stratified by three increasing absolute levels of CD34+ cells/μL. As predictable, %aTE tapered from 34.4 % for the lowest level range (<15 cells/μL) to 21.1 % for the highest level (>25 cells/μL). Particular attention was paid to the risk of misclassifying patients due to measurement imprecision: this risk proved quite variable, from 0 to 40 % of cases (average about 13 %), was very dependent on the absolute CD34+ cell level, being most frequent for the lowest cell levels, and did not improve with the passing of time.

Imprecision in the absolute CD34+ cell count may mean misclassification of patients with hematological disorders, may give way to an apparent safe trigger for apheresis procedures, but doomed to the collection of insufficient amounts of CD34+ HPCs, or may give the false impression of having collected enough HPCs for transplantation. The most dangerous error in this field is indeed the overestimation of an actually inadequate HPC collection procedure, that may put patients’ life at risk for the lack of a timely and effective engraftment.

Fernández-Luis and colleagues have also highlighted that the best performing laboratories, in the upper quartile of %aTE data distribution, can show much tighter MU data than the rest of the group, indicating that a path to the continuous improvement of analytical performance up to excellence is possible with procedure standardization, education and practice also in the tricky field of CD34+ HPC enumeration.

The evaluation of %aTE is highly influenced by sample type, logistics, technology, geography and time. The strict adherence to the ISHAGE guidelines and the universal usage of single-platform FCM technique can greatly contribute to performance improvement [27]. It seems therefore likely that the pioneer findings by Fernández-Luis and colleagues may be further improved using highly standardized procedures and stabilized instead of fresh samples, as shown by very preliminary data from UKNEQAS LI exercises included in the forthcoming CLSI H63 guideline on CD34+ cell enumeration [4]. The use of stabilized samples may help to minimize the unavoidable preanalytical decay of fresh samples, surely causing some spread of %aTE calculations, and to focus the studies on MU on the purely analytical variability.

A big missing issue, however, still continues to negatively influence this field, namely the lack of EQA/PT studies using stabilized concentrated apheresis samples. As mentioned, the enumeration of the collected HPCs in leukapheresis suspensions deals with CD34+ cell levels from 50- to 200-fold higher than baseline, and HPCs are admixed to a concentrated mixture of troublesome accompanying cells, like nucleated red cells, large immature myeloid cells, apoptotic or dead cells and a lot of platelets. This final part of CD34+ HPC analysis is crucial to patients’ outcome, but is particularly prone to counting errors due to a number of preanalytical and analytical issues [28]. Unfortunately, concentrated leukapheresis preparations are hardly available for large-scale EQA/PT programs.

To conclude, the absolute enumeration of CD34+ HPC in the peripheral blood of both steady-state and mobilized subjects is a fully mature laboratory technique, for which the SOTA-based %aTE parameters can be evaluated. Ample margins for performance improvement have been demonstrated by studies on EQA/PT programs, indicating that time and practice are still needed in clinical laboratories to implement optimized quality specifications to make CD34+ cell enumeration fully compliant to ISO15189 requirements.