Many clinical flow cytometry laboratories are examining the pros and cons of switching to higher color analyses.  Whether it’s switching from a four or five color assay to a new 8 (or more) color instrument, or adding one more color, implementation of a higher color assay may seem a daunting task.  Actual experience from users provides valuable, unbiased, input that can help answer questions related to this major investment in money, time, and effort.   In order to provide this actual user feedback, we have gathered information from four members of the ICCS Education Committee who have completed or are completing a switch to a higher color flow cytometry platform for clinical analysis. Three other users outside the education committee also shared their experiences by answering a questionnaire. It is hoped that this summary will help those who are making the decision about the purchase of these systems and provide guidance in validating the new system in their laboratory.

Our seven respondents reported the following -

Prior Laboratory Set-up: The person leading the change to the new higher color assays was either the medical director, technical director or analytical specialist in the lab, with at least 5 years, and in most cases over 10 years, of experience in that job. All respondent labs perform leukemia/ lymphoma immunophenotyping (LLI), but the volumes vary from less than 1000 to greater than 200,000 per year. Minimal residual disease detection is offered by all, but the proportion of these cases ranges from less than 1% of total LLI cases to over 100%.  Most labs perform PNH assay and a majority of labs offer CD4 counts and lymphocyte subset analysis. Other assays offered include stem cell counts, platelet function and immunodeficiency panels, and HLA typing. In the past, most respondents had gained similar experience while implementing a change from three color to a higher color assay, most often from FACScan to FACSCaliber, but also from a BD instrument to a Coulter FC-500 in some cases, or both.  Prior to the switch to a new generation, high color instrument, most of our respondents were using either FACCaliber or FC500, but one respondent set up a new flow cytometry laboratory from the ground up. Users of BD instruments employed the vendor’s software for instrument set up, and all but one lab used vendor software for analysis, while the users of Coulter instruments selected a third party software for data analysis even though they used the vendor software for set up.

High Color Set Up Implemented:  Four labs switched to FACSCanto II, two to LSR-II and one changed from a 3 color panel to a 5 color panel on the same instrument (FC500) operational in the lab. Those moving to FACSCanto II are using the instrument for 6 or 8 color assays and LSR-II users for 10 or 11 color assays.   While all but one laboratory is using the vendor’s software for instrument set up and acquision, all but two are using third party software for data analysis.  The goals of the majority of labs in making the switch to a higher color system were to meet, or exceed, industry standards and to provide more accurate results, and a sizeable minority were also attracted by the potential of providing more complete analysis of  small samples.  Other reasons given by some users for making the switch include ability to manage increased workload, faster throughput, aging previous instrument, or to match experience gained in a previous lab. 

Design of High Color Assays:  The overall strategy used to implement the high color assay was different in different labs- one lab migrated their existing low color assay to the new instrument (largely driven by the fact that the old instrument was failing and needed to be replaced soon) and then validated he 8 color panels. Four respondents validated the high color panels first and then migrated the test to the new instrument. Only two labs migrated all assays at once, one being a lab where the same instrument was being used but with higher color assays  and one where the director had previous experience of setting up the same high color instrument.  Most labs had decided to migrate the lymphoma analysis first, followed by acute myeloid leukemia analysis, and finally more specialized assays including cytoplasmic staining.   The approach to the design of antibody panels was different among the respondents.   Four labs designed the panels from the ground up without consideration to legacy panels while three respondents made efforts to maintain as many antibody combinations from previous panels in the higher color panel.  One respondent with the longest experience (already published independently) based panels on linage specific screening. Highly expressed antigens were coupled with dim fluorochromes and dimly expressed antigens coupled with bright fluorochromes. They made all efforts to incorporate their 4 color antibodies with known performance and intensity from the 4-color assay taking into consideration fluorochrome availability of antibodies and their intensities. Interestingly one respondent noted that while there are guidelines about how to select antibody-fluorochrome pairs, they found it difficult to predict what different combinations will look like until trying them.  As a result, they had to try different antibody clones, but often started off with the reagents from a familiar vendor, with clones that had been previously used in the lab, and then went to additional antibodies and/or fluorochromes if the results were not optimal. Effort was made by four labs to have as many tubes in which all possible colors were used, while one respondent gave a low priority to that aspect of panel design.  Several labs tested different antibody clones (from same or different vendors) for a particular antibody and different fluorochromes for the same antibody.  The latter was especially noted for fluorochromes excited by the violet laser.

Assay Validation: All except one lab titrated individual antibodies and all validated the new combinations on normal and abnormal specimens before putting them in use. Additional validation steps were taken by some labs to test the stability of antibody “cocktails”. Labs also tested the antibody combinations in “fluorochrome – 1”  iterations,  which were noted by one respondent to be particularly useful when struggling with compensation issues.  However, when that lab uses the huge matrix with antibody specific tandem compensation it has generated previously, “fluorescence – 1” experiments rarely identify a problem (although they are still used routinely as an upfront check).  One lab employed cell lines representative of all anticipated cell types to be encountered in leukemia/lymphoma analysis, both individually and in “spiking” experiments into normal blood or marrow for MRD simulation.  One respondent mentioned that they did not have any previous experience with this degree of testing of antibodies, but after the 8 color validation, it has become a routine.   

Some Antibody Problems:  Some respondents mentioned that the instrument and software switch went relatively smoothly when they migrated their four color panels to the new instrument while others found the analysis software incompatible with the newer computer operating system and/ or cumbersome to use and had to switch to third party software.  More problems were mentioned with panel set up and validation. Problems  encountered with specific fluorochromes  included,  non specific staining with CD5 Pe-Cy7 (resolved by switching to  CD5-PC7 from Beckman Coulter  which has a blocking agent already in it), pacific blue fluorescence tagged to CD19 and CD4 antibodies was dim even on cells expected to have high expression of the antigen, making it impossible to discriminate normal and abnormal expression level (solved by changing to different fluorochrome),  Pacific Orange was found to be too dim to be useful and its substitute, AmCyan, showed excessive spillover in FITC channel requiring an unacceptable level of compensation. An alternative fluorochrome for this slot was made available more recently, but proved adequately bright only in combination with CD45. Similarly AlexaFluor 700 was found to be too dim with the red laser and needed to be replaced by the brighter APC-AlexaFluor700 tandem. In general, there is still room for improvement with fluorochromes. One respondent mentioned that changing to 8-color combinations was difficult and is still a work in progress, and continues wryly that they have certainly learnt a lot about flow cytometry during the process, which frankly they did not need to know before.  Initial difficulty in setting compensation while using tandems was a recurring theme and more experienced operators noted the variability of the violet laser and the need for prolonged set up when any component is adjusted.  The stability of the tandem fluorochromes in cocktails was low in some, but not all, labs.  Complement in the plasma was found to have interactions with some antibodies.

Advantages of the New System:   Respondent are pleased with their new set up because it ‘s faster, more stable, performs automatic compensation, and provides walk away convenience.  The digital electronics may underlie the stability and accuracy, which provides better correlation of atypical populations with their markers.  Fewer tubes are required to perform the entire leukemia and lymphoma work up. For example, following panel has been found to work well in one lab (coupled to FITC, PE, Per CP-Cy 5.5, PE-Cy7, APC, APC-CH7, H450, and H500, in each tube) – tube 1: 71/123/8/33/38/14/56/45;  tube 2: 64/117/4/33/34/HLA-DR/7/45;  Tube 3: 61/13/19/33/41/16/11c/45;  Tube 4: Simultest  K/L/19/5/10/20/-/45; tube 5: 52/26/4/8/1A/3/7/4; and  Tube 5: 103/22/19/5/23/25/11c/45. This enthusiasm is mirrored by another respondent that having CD45, CD19, CD20, CD10, CD5, kappa, lambda, CD38 or CD45, CD2, CD3, CD4, CD7, CD8, CD56, CD16&CD57 all in the same tube is great.  The system is more sensitive and it’s been easier to find small abnormal populations. However, the respondent  also cautioned that it is easy to get fooled by some normal T-cell subsets that we have not recognized before.  Overall, the higher color assays were felt to provide better analysis capabilities. The use of the biexponential scale on the histogram axes makes the delineation of negative population more accurate and facilitates identification of really DIM populations.   The anticipated gains in productivity and accuracy and the ability to more completely analyze small specimens (FNA, CSF) or specimens for MRD detection is noted.

Disadvantages of the New System:  On the practical side, respondents have mentioned the higher cost of maintaining the system, without realizing significant savings in labor etc.  Analysis has become more complex and time consuming.  The need for highly trained personnel to analyze these complex data is emphasized.  On the technical side, the more experienced operators have noted higher backgrounds because of fluorescent crosstalk.  Fiber optic connectivity between laser and flow cell has proven to be troublesome and there is loss of power.  The scatter on these instruments causes the monocytes to be often lost in the lymphs and many beads cannot be resolved by forward scatter.

Final Thoughts:   The respondents reported widely different times taken for full implementation. In an ideal situation, when the director was already intimately familiar with the instrument, had been involved in the validation of the same multicolor panel before, and had planned the set up at the new laboratory for 2 months, the assay could go live in 4 months after the instrument was delivered.  Slightly longer duration of about 9 months was reported by an experienced technical director, who was able to devote his full attention to the validation. On the other hand, others reported a longer time line of over 2 years for full implementation.  In these cases, the availability of suitable fluorochromes for the violet laser has been a major drag.  Even experienced mavens of the art and science of flow cytometry were hampered by the shortcomings of the available fluorochrome combinations.  If our respondents had to do this all over again, they may have chosen to migrate their current four color assays directly to the new instrument, allowing the technical staff to get acquainted with the new instrument and software.  Being thoroughly familiar with many technical aspects of flow cytometry and having a dedicated person to do the development and validation are other suggestions given in hindsight. The design of antibody panels based on basic principles and not consolidating all antibodies for a particular lineage in the same tube are also mentioned as possible alterations from past experience. However, there is general agreement that this is a great learning experience.

Based on a questionnaire answered by members of the Education Committee of International Clinical Cytometry Society (ICCS) and other eminent flow cytometrists. 



Summary prepared by
Anand S. Lagoo, MD, PhD
Director, Clinical Flow Cytometry Laboratory
Associate Professor of Pathology
Duke University Medical Center, Durham, NC 27710.
anand.lagoo@duke.edu

The input from the following flow cytometry experts is gratefully acknowledged (alphabetically) – Fiona Craig, Katy Dougherty, David  Grier, Steve Kussick, Anand Lagoo, Horacio Vall, and Paul Wallace.