validation and quality control of immunophenotyping in ... · marilyn a. owens , horacio g. vall ,...

Journal of Immunological Methods 243 (2000) 33–50www.elsevier.nl / locate / jim

Validation and quality control of immunophenotyping in clinicalflow cytometry

a , a b c*Marilyn A. Owens , Horacio G. Vall , Anne A. Hurley , Susan B. WormsleyaIMPATH, Inc., 5230 Pacific Concourse Drive, Los Angeles, CA 90045, USA

bComprehensive Cytometric Consulting, 62 Bayberry Lane, Hanover, MA 02339, USAcPharMingen (BD Biosciences), 10975 Torreyana Rd, San Diego, CA 92121, USA

Abstract

Clinical flow cytometry has evolved from two-parameter quantitative assessment of peripheral blood lymphocytes tosix-parameter qualitative evaluation of bone marrow for hematopathology. Leukemia and lymphoma immunophenotypingrepresent an extremely important complement to morphology in the diagnosis and monitoring of hematopoietic malignancies.The complexity of five- and six-parameter analyses and the interpretation of the data rely on standardization and validationof the instrument, the reagents and the procedure. In addition, flow cytometry laboratories in the U.S. are required todocument proficiency testing, sample preparation, method accuracy, specificity, sensitivity and precision. NCCLS and theU.S.–Canadian Consensus Conference have provided recommendations, but each laboratory is ultimately responsible forvalidating its own qualitative and quantitative procedures. This paper reviews procedures for validation and quality control ofall aspects of the operation of a clinical flow cytometry service. 2000 Elsevier Science B.V. All rights reserved.

Keywords: Immunophenotyping; Flow cytometry; Phenotypic analysis; Leukemia diagnosis; Lymphoma diagnosis

1. Introduction unusual in the clinical laboratory, allowing thesimultaneous measurement of five or six different

Flow cytometry is a dynamic technology which parameters, respectively (Nicholson et al., 1996). Ahas allowed the multi-parameter analysis of recent consensus conference recommended five-pa-heterogeneous cell populations to develop as a rameter immunophenotyping to be the minimalclinical service (Owens and Loken, 1995). Complex standard for hematological malignancies, forwardanalyses are able to combine immunophenotyping of and right angle light scatter and three colors ofboth surface and cytoplasmic antigens, DNA analysis fluorescence (Stelzer et al., 1997). With so manyand functional evaluations. Subsets of cells can be variables in these analyses, standardization and vali-identified and characterized by patterns of maturation dation of instrumentation and methodology is essen-antigens and staining intensity which can assist in tial to ensure the technical quality of the resultsdiagnostic and prognostic interpretations as well as (Brando and Sommaruga, 1993; Carter et al., 1992;the detection of minimal residual disease. Cavelli et al., 1993; National Committee for Clinical

Three- and four-color immunophenotyping are not Laboratory Standards, 1992; Hurley, 1988, 1997a,c).Reagents must be well characterized for specificity

*Corresponding author. and performance with different fluorochromes. All

0022-1759/00/$ – see front matter 2000 Elsevier Science B.V. All rights reserved.PI I : S0022-1759( 00 )00226-X

34 M.A. Owens et al. / Journal of Immunological Methods 243 (2000) 33 –50

monoclonal-fluorochrome combinations must be 2. Responsibilities of the clinical servicecritically evaluated for staining intensity, spectraloverlap, and instrument compensation (Hurley, Flow cytometry falls under the Centers for Dis-1997b; McCoy et al., 1990; Muirhead, 1993). Fur- ease Control and Prevention (CDC) category ofthermore, fluorescence patterns must be character- high complexity laboratory testing. Documentationized for diagnostic combinations of antigens and of staff qualifications and training as well as ana-diagnostic interpretation (McCoy and Overton, lytical accuracy, sensitivity, precision and QC are1996). required. In some states, such as California, per-

In attempts to assist with standardization, the sonnel responsible for generating flow cytometricU.S.–Canadian Consensus Conference in 1997 and results must be licensed medical technologists whoNational Committee for Clinical Laboratory Stan- are required to have 12 h of continuing educationdards (NCCLS) guidelines in 1998 provided recom- yearly. Other states may not have the same person-mendations for clinical flow cytometry in hemato- nel qualifications. At the very least, training andpathology (Borowitz et al., 1997; Davis et al., 1997; proficiency in the technology must be documented.National Committee for Clinical Laboratory Stan- Laboratories should develop SOPs for training,dards, 1997a; Stelzer et al., 1997; Stewart et al., with supervisor verification of proficiencies in in-1997; Braylan et al., 1997a,b). On the regulatory cremental responsibilities for each staff member.side, the Clinical Laboratory Improvement Act Instrument manufacturers and scientific societies(CLIA 88) (Department of Health and Human also conduct training courses and provide certifi-Services, 1992) has influenced laboratory staffing, cates of training. Additionally, the American Socie-training, validation and documentation. As we enter ty of Clinical Pathology (ASCP) offers a specialtythe 21st century, the laboratory’s responsibilities exam for national certification in flow cytometrycontinue to increase. A recent publication by Hurley (QCYM).and Zito (1998b) offers approaches for CLIA com-pliance in the clinical flow cytometry laboratory andincludes templates for appropriate forms to satisfy 2.1. Proficiency testingdocumentation.

The goal of this article is to discuss how regula- All high-complexity laboratories must enroll in atory oversight influences the laboratory’s validation proficiency testing (PT) program that meets CLIAand quality control documentation, particularly in regulations (Subpart I) (Department of Health andhematopathology. Although this article covers reg- Human Services, 1992) and is approved by Healthulatory issues as they apply in the U.S., all lab- and Human Services (HHS). Regulations require thatoratories world-wide are responsible for maintaining the laboratory inform HHS which programs it willperformance standards. Good Laboratory Practices use, list the tests performed for each program,include Standard Operating Procedures (SOPs), participate in a program for at least a year beforeQuality Control (QC) and Quality Assurance (QA) choosing a different program, and must notify HHSas integral to providing good patient care (Hurley before changing. The laboratory must also authorizeand Zito, 1998a). Regardless of requirements placed the PT program to release data to HHS.on manufacturers to provide analyte-specific reagents Proficiency Testing samples must be treated the(ASR) or in vitro diagnostic (IVD) reagents, each same as patient specimens, using the same personnellaboratory should validate their own panels for and work processes (Dorsey, 1975). Results must notsensitivity, specificity, and correlation with morphol- be discussed with other laboratories and PT samplesogy and clinical findings. The following sections will may not be outsourced under any circumstances.consider laboratory responsibilities, reagent and in- Documentation for handling, preparation, processing,strument validations, QC, QA, and troubleshooting analysis and interpretation of PT samples must beas a model for the integration of new technologies kept (Grannis et al., 1972). Final reports must beinto diagnostic hematopathology. retained for a minimum of 2 years.

M.A. Owens et al. / Journal of Immunological Methods 243 (2000) 33 –50 35

2.2. Sample handling that all clinical testing be characterized for accuracy,specificity, sensitivity and precision (National Com-

All laboratories must establish specimen require- mittee for Clinical Laboratory Standards, 1997b).ments, recommended transport conditions and However, for leukemia and lymphoma immuno-criteria for acceptability. Procedures for handling, phenotyping by flow cytometry, there are no consen-packaging, labeling and transporting potentially in- sus standards or recommendations for these assess-fectious biological specimens have been published ments; thus each laboratory must establish its own(Nicholson et al., 1994; National Committee for performance criteria (National Committee for Clini-Clinical Laboratory Standards, 1990, 1997a). The cal Laboratory Standards, 1997a).U.S.–Canadian Consensus strongly recommendedthat every attempt be made to derive useful in-formation from any specimen submitted for leukemia 2.4. Accuracyor lymphoma immunophenotyping analysis (Stelzeret al., 1997). Requesting physicians should always be Analytical accuracy compares the test result to ainformed about sample quality issues; any compro- reference, or ‘gold’ standard. In hematology, normalmised specimen, whether evaluated or not, must be cell populations can be counted microscopically withdescribed in the test interpretation. a hemocytometer or with automated equipment and

Clinical specimens must be appropriately iden- results compared for accuracy (Wooten and King,tified — the minimum information is a unique 1953). In abnormal populations, especially when thepatient identifier (for cytogenetic analyses, two iden- characterization is based on multi-parameter data, thetifiers may be required), test ordered and date of comparison becomes complex. For hematopathology,sampling. Other information helpful for interpreta- the gold standard is morphology. The flow cyto-tion include presumptive diagnosis, age, gender, date metric assessment of accuracy must therefore com-and time of specimen collection, source of specimen, pare to morphology. With rare events, however,name of physician, and recent treatment or medica- morphologic diagnosis is difficult. Flow cytometrytions. Confidentiality must be assured and docu- allows characterization of cell populations withmentation tracking the specimen’s handling is essen- complex phenotypes and should also assess accuracytial. Tests may only be performed if requested by an by comparison to previously characterized cellsauthorized person. Verbal requests must be followed (Stewart and Stewart, 1997b). Sources of well-char-by written authorization. All requests must be re- acterized cell populations are cryopreserved pedi-tained at least 2 years. greed specimens from another validated laboratory or

relapse specimens from previously characterizedcases. Another source of specimens for assessing

2.3. Sample preparation accuracy can be cases diagnosed by cytogenetics andmolecular biology. Documentation comparing at

Sample preparation for flow cytometric analysis least 20 leukemia and lymphoma cases with com-must consider the type of specimen submitted and plete histopathology and clinical findings should bethe number of cells available for analysis (Stewart on file to support analytical accuracy.and Stewart, 1994). Peripheral blood, bone marrow, Another aspect of accuracy that should not beor tissue specimens should be processed to contain a ignored is the qualitative staining pattern used tosuspension of the cells of interest, eliminating eryth- identify cell lineages. Descriptions of dim, moderate,rocytes, at a concentration optimal for monoclonal and bright staining patterns can be diagnostic or

7staining (0.5–1310 /ml; 0.1 ml per tube). All prognostic and should be well characterized andprocessing procedures must be written, approved and documented. For example, when multiple antibodiesdaily records maintained. Test records must be conjugated to fluorochromes are used in combina-retained for at least 2 years. tion, accuracy should be assessed for each antibody

Good laboratory practices and CLIA 88 require separately and the frequency and intensity results


compared to those obtained when using the com- pedigreed malignant specimens or well-characterizedbined reagents. transfected cell lines used as positive controls.

2.6. Sensitivity2.5. Specificity

Reagent sensitivity describes the ability to detectSpecificity of monoclonal reagents is defined by the minimum staining intensity above non-specific or

how well the antibody recognizes the correct an- negative staining (Steen, 1991). The sensitivity oftigenic target. Manufacturers are responsible for detection is dependent on the titration of monoclonalreagents having the correct specificity listed on their reagents (Stewart and Stewart, 1997a), the properlabels. However, specificity in a ‘home-brew’ diag- instrument setup and calibration, the number of cellsnostic test has a broader meaning. For leukemia and counted and the flow rate of the instrument (Wittruplymphoma testing, the interpretation of the flow et al., 1994; Zucker et al., 1991). Documentation ofpanel should be compared with morphology and reagent titrations and parallel testing of each new lotclinical presentation to assess the ‘specificity’ of the of antibody is required. Instrument calibration andtesting (Zagursky et al., 1995). Each laboratory documentation is also required.should establish its own expected rate of discrepancybetween flow and morphology, most likely ,5%. 2.7. PrecisionThe laboratory should then assess, on a case-by-casebasis, the reason for a discrepancy, document the Precision is a standard analytical parameter whichdiscrepancy as a QA assessment, and monitor trends measures the reproducibility of a single samplequarterly. stained and analyzed in duplicate at least 10 times.

Specificity of flow cytometric reagents can be NCCLS Guidelines recommend 20 replicates. Nor-assessed by consensus workshops, publications or by mal peripheral blood, cell lines, blood standards orin-house validation. Most notable are the publica- CD Chex may be used. CD Chex are preserved whitetions from the various Human Leukocyte Differentia- cell controls manufactured by Streck Laboratories.tion Antigen Workshops, the latest being Leucocyte Quantitative mean and standard deviation for eachTyping VI (Kishimoto et al., 1997), which aptly and monoclonal antibody should be documented and a 2succinctly summarize the vast body of testing results S.D. range developed. Determination of instrumenton most if not all of the monoclonal antibodies precision is accomplished by running the samecommonly used in leukemia and lymphoma immuno- stained sample at least three times, with resultsphenotyping. Helpful data summaries, reviews, and within 2 S.D.citations pertaining to these reagents can be found onthe HLDA Web page at http: / /www.mh-han- 2.8. Analyte-specific reagentsnover.de /projekte /hlda7/hldabase /select.htm. Alsorefer to the PROW database (Protein Reviews on the In the United States, clinical laboratories, reg-Web). Other sources of specificity data are clinical ulated under CLIA 88 and performing physician-texts and journal articles. ordered flow cytometric testing for leukemia and

However, flow cytometry monoclonal reagents are lymphoma immunophenotyping, are required to beconsidered ‘home-brew’ reagents and each labora- qualified to perform high-complexity testing andtory must document staining performance, or ‘diag- must use reagents that are manufactured and labelednostic specificity’ on both normal and abnormal cell as Analyte Specific Reagents (ASRs).populations, identifying both positive and negative Manufacturers of ASR must register their facility,staining (Stewart and Stewart, 1995). Clinical hema- list their products, follow current Good Manufactur-topoietic specimens will almost always contain nor- ing Practices (cGMP) under the new Quality Systemmal along with any abnormal populations and thus Regulations (QSRs), and comply with medical de-will allow both evaluations in one specimen. Diag- vice reporting requirements. Manufacturers are re-nostic immunophenotyping must be validated with stricted from providing any statement regarding


analytical or clinical performance, number of tests ance (Belk and Sunderman, 1947; Sax et al., 1967;provided, or instructions for use. Advertising and Shuey and Cebel, 1949; Paxton et al., 1989).promotional material for an ASR product mustinclude the identity and purity (including source and 3.1. Instrument validationmethod of acquisition) of the analyte-specific reagentand the identity of the analyte [Code of Federal Validation of instrument performance falls intoRegulations, CFR 809.30(d), Federal Register, two areas: (a) instrument setup and daily qualifica-1997]. tion of both light scatter and fluorescence measure-

The identity of the analyte for monoclonal anti- ments, and (b) cross-instrument performance usingbody reagents is provided by the antigen distribution relevant clinical specimens.and supporting references. The HLDA references forwhen the antibody clone was first clustered are 3.2. Instrument setup and daily qualificationprovided for the identity of the ASR. The labelingfor analyte-specific reagents must include the reagent NCCLS recommends that the setup of a flowname, concentration, purity and quality, statement of cytometer is comprehensive enough to assure properwarnings or precautions for users, date of manufac- optical alignment for adequate light scattering andture, lot number, expiration date, storage instructions, fluorescence sensitivity and resolution as well as thenet quantity of contents, name and place of business, proper degree of compensation to correct for spectraland the following statement: ‘‘Analyte Specific overlap when multiple fluorochromes are used. WhileReagent. Analytical and performance characteristics hardware configuration may not allow the laboratoryare not established’’ [CFR 809.10(e)]. to align optics, a check of performance must be

It is the responsibility of the testing laboratory made and documented. In addition, since the hard-using these reagents to validate their performance in ware and optics are different for flow cytometers‘home-brew’ clinical assays. In addition, any labora- made by different manufacturers (the two major onestory reports using Class I ASRs must contain the being Beckman-Coulter and Becton Dickinson, nowfollowing disclaimer: ‘‘This test was developed and known as BD Biosciences), the recommended pro-its performance characteristics determined by hlab- cedures for instrument setup and performance qual-oratory namej. It has not been cleared or approved ification are instrument-specific.by the U.S. Food & Drug Administration’’ [CFR809.30(e)]. Further clarifying comments may also be 3.3. Light scatter sensitivity and resolutionincluded.

Research institutions and other non-clinical lab- Optical alignment for optimal sensitivity andoratories that use ASRs to make tests for purposes resolution of both forward (FSc) and side (SSc)other than providing diagnostic information to pa- scatter can be assessed by running uniformly sizedtients and practitioner are not restricted under the beads that fall within the light scatter ranges ob-ASR regulation [CFR 809.30(g)]. served with most clinical samples, at a constant PMT

voltage on a daily basis. The mean FSc and SScchannel numbers and percent coefficient of variation(% C.V.) should be recorded. The acceptable ranges

3. Validation of Immunophenotyping for each parameter can be established by firstrunning the beads 20 times over a 5 day time period

How does a laboratory validate flow cytometric at the same PMT setting. Levy–Jennings graphs areimmunophenotyping? The basic components are the then used to plot the values obtained daily and anvalidation of the instrument, the individual reagents, action plan is established for what to do when anyand the staining protocols used to create a final parameter falls outside of the expected range (Alleninterpretation (Whitehurst et al., 1975). Each step in et al., 1969; Henry, 1959; Levy and Jennings, 1950).the testing process must be specified and quality However, since beads and cells do not alwayscontrol measurements included to monitor perform- behave similarly on a flow cytometer, it is also


recommended that the instrument be set up for ticular staining protocol, the compensation betweenrunning clinical samples using biological material fluorescence signals must be properly set to preventand that the proper resolution of different cell types spill-over of one fluorescence signal into anotherbe determined in daily instrument qualification pro- (Bagwell and Adams, 1993). This is usually donecedures (Henry and Segalove, 1952; Ladenson, automatically using hard dyed beads as seen in Fig. 11975). Normal peripheral blood leukocytes pro- for four-color compensation of FITC, PE, PerCP andcessed in a similar manner as the clinical samples APC on a FACSCalibur (BD Biosciences).can be used for this purpose. For example, the However, the use of a biological compensationlymphocytes can be set to fall within a specified control is also recommended. When performing four-range based on FSc. The PMT for SSc can then be color immunophenotyping using a single laser at 488adjusted so that the granulocytes fall within a nm on a Coulter XL (Beckman-Coulter), and aspecified range, with the final FSc vs. SSc cytogram combination of monoclonal antibodies directlyrequired to demonstrate good separation of lympho- conjugated to FITC, PE, PE-Cy5 and ECD, ancytes, monocytes and granulocytes. appropriate biological compensation control would

consist of normal donor peripheral blood lympho-3.4. Fluorescence sensitivity and resolution cytes stained with CD4–FITC, CD8–PE, CD2–PE/

Cy5 and CD45–ECD, as seen in Table 1.Acceptance values for monitoring fluorescence Hypothetically, in this example, there are 12

sensitivity and resolution can be established by two different combinations of fluorescence signals thatmethods: (a) recording the channel number and C.V. need to be checked for compensation but, due toof calibration beads with a pre-determined, fixed instrument design, the optical filters used, and thelaser power, filters, PMT voltages and gains, or (b) inherent spectral emissions of each fluorochrome, notrecording the high voltage and gain to position the all combinations can or need be considered. Usingbeads in the same channel each time. For either the Cytosetting Matrix Display on System II 3.0approach, acceptance ranges can be established by Expo software in the Coulter XL, each of the fourusing at least 20 replicate data sets collected over at fluorescence PMTs are first adjusted to the properleast a 5 day period; new ranges must be developed settings based on the single color fluorescenceeach time a new lot of beads is phased into service. histogram and compensated with the compensationRegardless of the method used, procedures for what matrix. For a complete discussion of four-colorto do if any of the measured and plotted parameters compensation, see Stewart and Stewart (1999).do not fall within the acceptable range must also beestablished, followed and documented by the clinicallaboratory.Values for the acceptable ranges displayed 4. Monitoring instrument performanceon the Levy–Jennings graphs for each parameter aretypically updated based on each successive 20 day In the setup and qualification of a flow cytometercycle of data collection. However, QC specifications for daily use, NCCLS recommends a check offor instrument performance must be established to optical alignment, fluorescence resolution and in-allow for detection of significant trends or drifts over tensity. Using the Coulter XL, a calibrator suspen-time which require corrective action. sion of latex beads with a known concentration of

fluorescence (DNA Check) can be used to verify an3.5. Fluorescence compensation appropriate cell-wide laminar flow, a proper align-

ment between the laser beam and the cell at theCompensation should also be evaluated at the time interrogation point and a proper adjustment of laser

of initial instrument setup and then monitored daily, power and/or PMT voltage. The ranges for thepreferably using a biological compensation control in particles must be established, running 20 timesaddition to hard dyed beads provided by the instru- during 5 days and keeping the PMT voltage constant.ment manufacturer. Depending on the number, type The placement of the beads must meet the estab-and combination of fluorochromes used in a par- lished ranges and be recorded in the QC log. Any


Fig. 1. Column A: fluorescent beads run without compensation on FACSCalibur (four color) FITC, PE, PerCP, APC. Column B: beadsshowing compensation for FL12%FL251.2 and FL22%FL1523.3, FL22%FL350, FL32%FL452.5, FL42%FL355.0.


Table 1Signal volts and gain showing matrix compensation for four-color phenotyping on a single laser Coulter XL MCL instrument

Signal Compensation

Signal Volts Gain Total Signal out5Y2%Xgain

Y X→ FL1 FL2 FL3 FL4

FS 25 5.0 5.38 FL1 0.9 0.0 0.0SS 402 20.0 44.12FL1 687 1.0 FL2 11.2 12.8 4.7FL2 873 1.0FL3 718 1.0 FL3 0.0 31.1 2.1FL4 909 1.0AUX 69 5.0 6.04 FL4 0.0 0.0 19.1

values out of range must be investigated, corrective unstable. Monitoring linearity is a check of PMTaction taken and documented. voltages. Using the same PMT settings used for

For the BD FacsCalibur, the fluidics and cali- clinical specimens, a set mixture of four to fivebration check is performed by running CaliBrite multi-level fluorescence beads, each with a knownBeads and allowing the instrument software to make relative fluorescence intensity level, should be run.adjustments in setup based on lot-specific perform- Products are available from several manufacturers,ance expectations. Documentation of acceptable for example Flow Cytometry Standards, orcalibration is provided for record keeping. These Spherotech. Acceptable mean fluorescence intensityrecords can be used for trend analysis of the laser (MFI) ranges for each bead in the mixture should bepower and voltage settings to indicate potential established by 20 replicate runs over a 5 day timetechnical problems. period. These same bead mixtures should then be run

NCCLS recommends daily performance monitor- once a month and the linearity, or the relativeing of fluorescence intensity, color compensation and difference in MFI between each of the beads, shouldverification of system performance using a QC remain constant for each fluorescence PMT (Fig. 2).(normal) specimen. Each laboratory must have QC Graphs of linearity should be drawn followingprocedures to monitor instrument performance. For manufacturers’ recommendations to accommodatehematopoietic immunophenotyping, color compensa- Log 0.tion and verification of performance can be evaluatedwith each specimen panel. Although every patient isdifferent, the CD4/CD8/CD45/CD2 combination 4.2. Cross-instrument performancecan be used to check compensation. Settings willrarely need to be changed. Laboratories that are performing the same clinical

Optics should not be changed by a clinical labora- immunophenotyping protocols on more than onetory without specific validated procedures. It is more instrument should include a semi-yearly cross-instru-appropriate to have the manufacturer check and ment comparison of at least five different andadjust optics during routine maintenance and to representative clinical samples stained with each ofprovide documentation for the laboratory. the standard protocols used in the laboratory. The

results obtained for each sample should not differ4.1. Fluorescence linearity between instruments by more than a pre-defined

acceptable variance (Vogt et al., 1991, 1994). ThisLinearity of fluorescence detection should be cross-instrument sample testing process should be

checked on a monthly basis or as recommended by documented and corrective action plans must bethe instrument manufacturer (Vogt et al., 1989). established and followed when cross-comparisonLinearity should be constant unless the laser is results fail to meet performance specifications.


Fig. 2. Beads with five different fluorescence intensities for each fluorochrome. Setup regions A to E will provide the log intensity channelvalue. A plot of log intensity vs. bead number should provide a linear distribution.

5. Preanalytical QC be documented before transfer to the testing labora-tory.

Laboratories must not only establish acceptancecriteria for immunophenotyping specimens, but musthave procedures to assess acceptability and docu-mentation to assist in troubleshooting and the inter- 6. Reagent and method validationpretation of results. NCCLS recommends a visualanalysis on receipt to confirm specimen quality. An immunophenotyping procedure first requiresHemolysis, partial draw (especially in ACD tubes), the selection of monoclonal reagents and fluoro-temperature extremes and improper labeling should chromes which are to be used in the multi-parameter


analysis and interpretation of results. The hemato- reagents are combined in testing panels is left up topathology testing panel must be designed to clearly the testing laboratory, but should be based on adistinguish normal and abnormal immunoreactivity target-oriented strategy (Table 2).patterns based on differences in light scatter and/or All the combinations have to be validated in eachfluorescence intensity. These pattern comparisons are laboratory and the analysts must be familiar with theintegral to the interpretation of results. Therefore, the patterns associated with each combination. Differentpanel must include all relevant markers and reflect clones (under the same CD) may perform differently.the visual expectations of the diagnostic interpreters. Similarly, the same CD labeled with different fluoro-

The U.S.–Canadian Consensus Conference chromes will show a different level of intensity andstopped short of recommending a diagnostic panel possible color overlap. Fluorochrome intensity andfor leukemia and lymphoma immunophenotyping, antigen expression are both important in establishinghowever there was agreement on 42 determinants the best antibody combination to evaluate antigenwhich can provide diagnostic information in these density (Schwartz et al., 1990, 1996a,b). Someprocedures (Stewart et al., 1997). Each specimen tandem dye combinations can help to minimize thesemust be considered a unique case and fully evaluated problems.to minimize missing any abnormality. The selection Different combinations of antibodies can be usedof panel reagents needs to balance the economic to provide the same clinical interpretation and noneed for a minimal number of monoclonal antibodies single panel will accommodate all leukemia orto identify abnormal populations with the scientific lymphoma phenotypes. To compound the difficulty,need to detect abnormal antigen expression. The new markers and reagents, when determined to becommittee concluded that it is more important that clinically valuable, must be validated by each labora-each laboratory have adequate experience with the tory. Any panel of monoclonal reagents used by abinding characteristics of their testing reagents than laboratory must be chosen not only for their techni-specifying what those reagents should be. How cal performance, but also to satisfy the experience

Table 2aSome common tube combinations

Population Three-color Four-colorto identify combination combination

B cells Kappa/Lambda/CD20 K/L/CD45/CD20CD10/CD5/CD19 CD10/CD5/CD45/CD19CD5/CD23/CD20 CD5/CD23/CD20/CD19

Plasma cells CD38/cytoK/CD45CD38/cytoL/CD45

T cells CD7/CD3/CD45 CD7/CD3/CD45/CD56CD4/CD8/CD3 CD4/CD8/CD45/CD2

Myeloid CD33/CD13/CD45 CD13/CD14/CD45/CD33CD34/CD117/CD45 CD15/CD117/CD45/CD34CD11b/CD16/HLA-DR CD11b/CD16/CD45/HLA-DR

Myeloid wmonocytic diff CD64/CD13/CD45 CD64/CD14/CD45/CD33

Myeloid w CD34/CD41/CD61 CD41/CD61/CD45/CD34megakariocytic diff CD42b/CD61/CD45 CD42b/CD61/CD41/CD34

Red cell precursors CD71/Glycophorin A/CD45a These combinations are for examples only and do not represent guidelines or consensus recommendations.


and expectations of the technical and medical per- proach is to set the PMT with the single reagent andsonnel who must differentiate abnormal populations then check compensation with the final cocktail.from normal cells.

The proper combination of monoclonal antibodieswithin a cocktail must consider antigen expression 7. Sample preparationon normal and abnormal cells as well as the fluoro-chrome combinations to minimize interference be- Integral to the analytical procedure, the laboratorytween reagents. One example of potential interfering must decide on and validate a sample preparationmarkers is bright CD15/CD16/CD45/CD13 on procedure. Erythrocyte lysis is the most commonlygranulocytes. For multiple fluorochrome combina- used method for preparing peripheral blood or bonetions, the laboratory must not only validate that each marrow specimens for clinical immunophenotyping.monoclonal performs as expected in the procedure, Lysis can be performed either before staining withbut must also prove that the performance is not monoclonal antibodies or after. Whichever methodaffected by co-blocking of epitopes, fluorescence for erythrocyte lysis is used must be validated priorquenching, or energy transfer (Pollice et al., 1992). to reporting clinical results. Lysis is optically moni-

Antibody cocktails may be purchased from the tored by observing a cloudy red suspension of cellsmanufacturer or prepared by the laboratory. Whether change into a clear, translucent solution in 5 to 10reagents are mixed immediately before each use or min. If red cells are still present after centrifugationstored as a premix solution, performance of each and resuspension, the lysing process should becocktail must be validated and expiration dates repeated.documented. Tissue specimens should be disaggregated and

Each monoclonal antibody in each cocktail must filtered to create a single cell suspension prior tobe titrated individually for optimal signal-to-noise staining. This laboratory procedure must also beseparation. However, potential interference between written and validated. Cytological slides should becertain clones and/or fluorochromes necessitates also prepared and reviewed with the H&E slides of theperforming checkerboard titrations to select the final tissue to assess the loss of fragile cells, such asoptimal working dilutions once the reagents are used Reed–Sternberg cells. Some large cells in fibroticin combination. For each new multi-color combina- tissues may be difficult to remove and cell loss willtion added to the panel, it is necessary to first occur during the processing. Other sampling errorsvalidate that the performance (mean intensity fluores- are due to focal lymphoma in the lymph node, notcence and percent positive) of each antibody when seen in the isolated cells or in the fixed tissue. Cellused in combination is comparable (within 2 S.D.) to fragility may be minimized by gentle tissue process-the performance of each antibody used alone at the ing and teasing the tissue apart with a needle.same concentration on the same target cell popula- Reproducible monoclonal antibody staining reliestion. If reagents are stored in a pre-mixed cocktail on the proper proportion of antibodies to cells. Forformat, it is also important to institute sufficient QC hematopoietic immunophenotyping, cell countsprocedures, using appropriate control cells, to verify should be performed on all specimens to ensurestability of performance of the cocktailed reagents correct proportions of cells to monoclonal antibodiesover time. Stable cocktails will provide results and thus accurate staining intensity patterns. NCCLSwherein the mean channel and percent of gated cells recommends the following adjustments based onwill not differ more than 2 S.D. between cocktailed WBC count of peripheral blood. (There are noand single reagents. Documentation must be summa- similar recommendations for bone marrow and eachrized and maintained in the laboratory and should laboratory should establish and validate its owninclude instrument photomultiplier tube (PMT) set- procedures.)tings and compensation. No more than one combina-tion of PMT and compensation may be used within a (a) WBC ,1000 cells /ml: use 200 ml whole bloodpanel (Stewart and Stewart, 1999). The best ap- and adjust amount of lysing reagent appropriately;


(b) WBC 1000–10,000 cells /ml: use 100 ml 30%, a gating discrimination strategy should be usedwhole blood and standard volume of lysing to limit the analysis to live cells, minimizing non-reagent; specific binding from dead cells which could lead to(c) WBC 10,000–20,000 cells /ml: use 50 ml misdiagnoses. Several different methodologies havewhole blood and adjust amount of lysing reagent been used to report viability, trypan blue, 7-amino-appropriately; actinomycin D (7-AAD), or propidium iodide (PI).(d) WBC .20,000: dilute whole blood with PBS The most common is trypan blue dye exclusionto achieve WBC concentration 1000–10,000, then visualized microscopically with a hemocytometer.use 100 ml and standard volume of lysing reagent. Dead cells stain blue because their membranes are

broken, allowing the dye to enter the cell. Viability isLight scatter patterns must be appropriate after reported as a percentage of the total cell population.

lysing and staining. The exclusion of dead cells is now being common-ly performed on the flow cytometer using dyes such

7.1. Viability as 7-AAD (Schmid et al., 1992) or PI (Sasaki et al.,1987) (see Fig. 3). Several others have been de-

Assessment of viability is crucial for evaluating scribed and, for most common clinical instruments, aleukemia and lymphoma specimens by flow cytom- relatively simple dead cell discrimination can beetry because cell membrane integrity is required for performed simultaneously with immunophenotyping.antigen expression.Viability at the time of collection, For example, using 7-AAD, the most commonthe influence of transportation environment and the three-color application in single laser instruments istime before testing can all affect sample viability and FITC in FL1, PE in FL2 and 7-AAD in FL3. In thiscompromise tumor cell detection. setting, 7-AAD (E 5 655 nm) is preferred over PImax

By recent consensus recommendations, flow cy- (E 5 625 nm) because it has less spectral overlapmax

tometry performed on leukemias and lymphomas with PE (E 5 578 nm) In our experience withmax

should be held to less stringent viability restrictions four-color immunophenotyping with a dual laserthan quantitative immunophenotyping (Stelzer et al., system (488 nm/635 nm on the FACSCalibur) the1997). When viability is low, for example less than most powerful combination for excluding dead cells

Fig. 3. Dead cell discrimination on two lymphoma tissues. (A) 7-AAD vs. Forward Scatter with 18.4% dead cells. (B) CD451PI vs. FSCwith 29.1% dead cells.


while evaluating four surface markers is FITC in be influenced by the donor, the target cell type, andFL1, PE in FL2, PerCP1PI in FL3, and APC in the particular staining method used. For cytoplasmicFL4. staining, the use of isotype controls is recommended

due to the significant influence of cell size on the7.2. Staining process QC degree of non-specific staining (Jacobberger and

Bauer, 2000). It is important that each CD mono-Once the instrument setup and performance vali- clonal antibody has the corresponding fluorochrome

dation is complete, a normal blood is run to check and isotype reagent to monitor non-specific bindingthe staining process. For this purpose, there are some performance.commercially available preserved cell preparations The presence of non-specific binding of an isotype[e.g., CD Chex (Streck Labs), CD Chex Plus (Streck can affect the interpretation of the fluorescenceLabs), Immunotrol (Beckman-Coulter)] which con- intensity patterns and, as such, is critically importanttain most of the cellular markers evaluated in in the analysis of leukemia and lymphoma immuno-leukemia and lymphoma immunophenotyping. Stain- phenotyping.ing each new lot of stabilized cells in triplicate willprovide a range to be used for quality control. Most 7.4. Procedure controlimportantly, each leukemia or lymphoma specimenwill contain a normal cell population which can be Fluorescence staining intensity patterns are moreused to document appropriate staining patterns. easily interpreted if there is a common factor in each

tube. For example, CD45 can be used as the ‘anchor’7.3. Isotype controls in each tube of a multi-color cocktail, providing a

reference population for comparisons to other an-The appropriate use of isotype controls is still tigenic expression patterns. The most important

controversial in the clinical flow cytometry com- parameter, however, is experience-based. The analystmunity (Borowitz et al., 1997; National Committee must understand the markers and the patterns used tofor Clinical Laboratory Standards, 1997a). For quan- identify an abnormal population. Running a normaltitative immunophenotyping, the isotype control was control will provide a baseline for all the markers.included to provide a negative cell reference and to Any deviation from the normal pattern should beset up the axes for assessing the numbers of cells in reported and confirmed with other markers or tech-each quadrant. However, immunophenotyping of nologies, such as cytogenetic or molecular tech-leukemias and lymphomas is a qualitative assessment niques.and does not usually require quadrant analysis.Furthermore, since all leukemias and lymphomascontain a mixture of cells, those cells that do not 8. Data analysis and interpretationexpress the antigen will be the negative control forthe ones that do. Immunophenotyping by flow cytometry can be

Most of the commercially available antibodies used to distinguish abnormal cell populations fromhave been selected by screening for the IgG class of normal cells and, in the process, identify the profileimmunoglobulin. For IgG class antibodies, however, of antigens in the abnormal population (Almasri etit is important to recognize that some IgG subclasses al., 1992; Borowitz et al., 1993a; Braylan et al.,will be more problematic due to their increased 1993; Cheson et al., 1996; Hertler et al., 1988;binding to Fc receptors present on various cell types. Jamieson et al., 1993; Roth and Schmitz, 1996;In general, the order of ‘stickiness’ is IgG2b. Tomer and Harker, 1989). Various patterns of im-IgG2a.IgG1. As a result of their carbohydrate munoreactivity with monoclonal antibodies are com-structure, some antigenic targets (e.g., CD15 and bined with light scatter to characterize populations ofCD57) will invariably result in generation of only interest (Davis et al., 1994; Kilo and Dorfman, 1996;IgM class antibodies. High non-specific binding of Loeffler et al., 1992; Segal et al., 1991; Wormsley etIgM class antibodies is commonly observed and can al., 1990; Davis and Bigelow, 1990; Shimenti et al.,


1992). The specific reagent combinations can be populations should only be performed if all cells ofchosen to identify abnormal antigen expression and interest are contained within the analysis gate. Strate-to characterize cells at various stages of maturation gies for specific diseases may include gating on B(Civin and Loken, 1992; Ford et al., 1995; Hurwitz cells to determine clonality or identification ofet al., 1988; Macedo et al., 1995; Seivers et al., leukemic blasts in a CD45 and right angle light1996). scatter display.

The complement of monoclonal antibodies neces- In Fig. 4, gating by FSc vs. SSc is shown in thesary to identify hematopoietic malignancies is not left panel and gating by CD45 vs. SSc for the samestandardized. In fact, strategies for immuno- case is shown on the right. The better gating strategyphenotyping must consider antibody specificity, uses CD45 to differentiate three populations (imma-fluorochrome, laboratory workload and, most im- ture myeloid, immature monocytes and normal lym-portantly, appropriate multi-parameter combinations. phocytes) (Borowitz et al., 1993b).Most laboratories use a comprehensive screening Data analysis of abnormal populations requirespanel and add confirmatory monoclonal antibodies if that fluorescence intensity be measured. The calcula-needed (Hassett and Parker, 1995). Reimbursement tion of percent positive cells does not aid in thefor laboratory testing is a major factor, thus lab- interpretation, while presence of inappropriate deter-oratories strive to make diagnoses with a minimal minants or inappropriate intensity expression ofnumber of antibodies. The U.S.–Canadian Consen- antigens can be diagnostic (Bene et al., 1995; Presssus Conference estimated that the average North et al., 1989). Resolving dimly positive populationsAmerican laboratory uses 19 antibodies to diagnose a from negative populations is crucial in the accurateleukemia and 16 for a lymphoma. European diag- assessment of fluorescence intensity (Givan et al.,noses often involve a larger number of monoclonal 1991). There are no hard and fast rules for definingantibodies (Stewart et al., 1997). these populations; in fact, reagent combinations are

Gating is one of the most important parameters in generally optimized to provide discrimination be-multi-parameter data analysis. In a diagnostic case, tween these populations based on the subjectiveall cells should be evaluated. The U.S.–Canadian review of the pathologist (Campana and Pui, 1993;consensus recommended that the initial analysis Rameshwar et al., 1994). For critical markers, suchinclude all viable cells. Further analyses on gated as immunoglobulin light chains, in the diagnosis of

Fig. 4. Comparison of gating strategies: FSc vs. SSc on left; CD45 vs. SSc on right.


Borowitz, M.J., Guenther, K.L., Shults, K.E., Stelzer, G.T., 1993b.non-Hodgkin’s lymphoma, the redundant use ofImmunophenotyping of acute leukemia by flow cytometricantibodies to light chains using alternative fluoro-analysis: use of CD45 and right-angle light scatter to gate onchromes, antibody pairs from different vendors, orleukemic blasts in three-color analysis. Am. J. Clin. Pathol.

monoclonal versus polyclonal reagents, can signifi- 100, 534–540.cantly enhance confidence in interpretation of mono- Borowitz, M.J., Bray, R., Gascoyne, R., Melnick, S., Parker, J.W.,

Picker, L., Stetler-Stevenson, M., 1997. U.S.–Canadian con-clonality in the face of very low level expression.sensus recommendations on the immunophenotypic analysis ofDiagnostic criteria for immunophenotyping werehematologic neoplasia by flow cytometry: data analysis andsummarized in the U.S.–Canadian Consensus byinterpretation. Cytometry 30, 236–244.

Borowitz et al. (1997).Brando, B., Sommaruga, E., 1993. Nationwide quality control trial

The interpretation of the flow cytometric data on lymphocyte immunophenotyping and flow cytometer per-relies on an experienced diagnostic interpreter, usual- formance in Italy. Cytometry 14, 294–306.

Braylan, R.C., Benson, N.S., Iturraspe, J., 1993. Analysis ofly a pathologist, reviewing all data, including mor-lymphomas by flow cytometry: current and emerging strate-phology (Banks et al., 1992; Ryan et al., 1991).gies. Ann. New York Acad. Sci. 677, 364–378.Drawing conclusions from either microscopy or flow

Braylan, R.C., Borowitz, M.J., Davis, B.H., Stelzer, G.T., Stewart,cytometry is an experience-based skill and the final C.C., 1997a. U.S.–Canadian consensus recommendations oninterpretation should reflect a synthesis of all avail- the immunophenotypic analysis of hematologic neoplasia byable information about the case (Davis et al., 1997; flow cytometry. Cytometry 30, 213.

Braylan, R.C., Atwater, S.K., Diamond, L., Hassett, J.M., John-Timm et al., 1995; Geisler et al., 1991). All reportsson, M., Kidd, P.G., Leith, C., Nguyen, D., 1997b. U.S.–must be sent to the requesting party and must also beCanadian consensus recommendations on the immuno-retained for a minimum of 2 years.phenotypic analysis of hematologic neoplasia by flow cytom-etry: data reporting. Cytometry 30, 245–248.

Campana, D., Pui, C.H., 1993. Detection of minimal residualdisease in acute leukemia: methodologic advances and clinical

References significance. Blood 85, 1416–1434.Carter, P.H., Resto-Ruiz, S., Washington, G.C., Ethridge, S.,

Allen, J.R., Earp, R., Farrell, C.E., Gruemer, H.D., 1969. Ana- Palini, A., Vogt, R., Waxdal, M., Fleisher, T., Noguchi, P.D.,lytical bias in a quality control scheme. Clin. Chem. 15, 1039. Marti, G.E., 1992. Flow cytometric analysis of whole blood

Almasri, N., Duque, R., Iturraspe, J., Everett, E., Braylan, R., lysis, three anticoagulants, and five cell preparations. Cytom-1992. Reduced expression of CD20 antigen as a characteristic etry 13, 68–74.marker for chronic lymphocytic leukemia. Am. J. Hematol. 40, Cavelli, T., Denny, T., Paxton, H., Gelman, R., Kagan, J., 1993.259. Guideline for flow cytometric immunophenotyping: a report

Bagwell, C.B., Adams, E.G., 1993. Fluorescence spectral overlap from the National Institute of Allergy and Infectious Diseases,compensation for any number of flow cytometric parameters. Division of AIDS. Cytometry 14, 702–715.In: Landay, A., Ault, K.A., Bauer, K.D., Rabinovitch, P. Cheson, B., Bennett, J., Grover, M., Kay, N., Keating, M.,(Eds.), Clinical Flow Cytometry. Ann. New York Acad. Sci. O’Brien, S., Rai, K., 1996. National Cancer Institute-spon-677(94), 167–184. sored working group guidelines for chronic lymphocytic

Banks, P.M., Chan, J., Cleary, M.L., Delsol, G., De Wolf-Peeters, leukemia: revised guidelines for diagnosis and treatment.C., Gatter, K., Grogan, T.M., Harris, N.L., Isaacson, P.G., Blood 87, 4990–4997.Jaffe, E.S. et al., 1992. Mantle cell lymphoma: a proposal for Civin, C., Loken, M., 1992. Cell surface antigens on humanunification of morphologic, immunologic, and molecular data. marrow cells: dissection of hematopoietic development usingAm. J. Surg. Pathol. 16, 637–640. monoclonal antibodies and multiparameter flow cytometry. In:

Belk, W.P., Sunderman, F.W., 1947. A survey on the accuracy of Murphy, M. (Ed.), Concise Reviews in Clinical and Ex-chemical analysis in clinical laboratories. Am. J. Clin. Pathol. perimental Hematology. AlphaMed Press, Dayton, OH.17, 854. Davis, B.H., Bigelow, N.C., 1990. Clinical flow cytometric

Bene, M.C., Castoldi, G., Knapp, W., Ludwig, W.D., Matutes, E., reticulocyte analysis. Pathobiology 58, 99–106.Orfao, A., vant Veer, M.B., 1995. Proposals for the immuno- Davis, R.E., Longacre, T.A., Cornbleet, P.J., 1994. Hematogoneslogical classification of acute leukemias: European Group for in the bone marrow of adults. Immunophenotypic features,the Immunological Characterization of Leukemias (EGIL). clinical settings, and differential diagnosis. Am. J. Clin.Leukemia 9, 1783–1786. Pathol. 102, 202–211.

Borowitz, M., Bach, B.A., Bauer, K.D., Duque, R.E., Horton, Davis, B.H., Foucar, K., Szczarkowski, W., Ball, E., Witzig, T.,A.F., Johnson, R.L., Marti, G., Muirhead, K., Peiper, S., Foon, K.A., Wells, D., Kotylo, P., Johnson, R., Hanson, C.,Rickman, W., 1993a. Immunophenotyping of Leukemic Cells: Bessman, D., 1997. U.S.–Canadian consensus recommenda-Proposed Guidelines. NCCLS, Villanova, PA. tions on the immunophenotypic analysis of hematologic


neoplasia by flow cytometry: medical indications. Cytometry in B lineage acute lymphoblastic leukemia. Blood 72, 299–30, 249–263. 307.

Department of Health and Human Services, 1992. Health Care Jacobberger, X., Bauer, K., personal communication.Financing Administration, Public Health Service 42 CFR Part Jamieson, G.P., Brocklebank, A.M., Snook, M.B., Sawyer, W.H.,405, Clinical Laboratory Improvement Act 1988. Buolamwini, J.K., Paterson, A.R.P., Wiley, J.S., 1993. Flow

Dorsey, D.B., 1975. The evaluation of proficiency testing in the cytometric quantitation of nucleoside transporter sites onUSA. In: Proceedings of the Second National Conference on human leukemic cells. Cytometry 14, 32–38.Proficiency Testing, Bethesda, MD, Information Services. Kilo, M., Dorfman, D., 1996. The utility of flow cytometric

Federal Register, 1997. Medical devices. Classification / reclassifi- immunophenotypic analysis in the distinction of smallcation; restricted devices; analyte specific reagents. Health and lymphocytic leukemia from mantle cell lymphoma. Am. J.Human Services, Federal Register 61, Docket No. 96N-0082. Clin. Pathol. 105, 451–457.

Ford, R.A., Tamayo, P., Martin, B., Niu, K., Claypool, K., Kishimoto, T., Kikutani, H., von dem Borne, A.E.G., Goyert,Cabanillas, F., Ambrus, J., 1995. Identification of B-cell S.M., Mason, D.Y., Miyasaka, M., Moretta, L., Okumura, K.,growth factors (interleukin-14; high molecular weight-B-cell Shaw, S., Springer, T.A., Sugamura, K., Zola, H. (Eds.), 1997.growth factors) in effusion fluids from patients with aggressive White Cell Differentiation Antigens. Leucocyte Typing, Vol.B-cell lymphomas. Blood 86, 283–293. VI. Garland, New York.

Geisler, C., Larsen, J., Hansen, N., Hansen, M., Christensen, B., Ladenson, J.H., 1975. Patients as their own controls: use of theLund, B., Nielsen, H., Piesner, T., Thorling, K., Andersen, A., computer to identify laboratory error. Clin. Chem. 21, 1648.1991. Prognostic importance of flow cytometric immuno- Levy, S., Jennings, E.R., 1950. The use of control charts in thephenotyping of 540 consecutive patients with B-cell CLL. clinical laboratory. Am. J. Clin. Pathol. 20, 1059.Blood 78, 1795–1802. Loeffler, D.A., Kuruga, M.A., Juneau, P.L., Heppner, G.H., 1992.

Givan, A.L., Calvert, J.E., Shenton, B.K., 1991. The effect of Analysis of distribution of tumor- and preneoplasia-infiltratingimprovements in cytometer sensitivity on the detection of lymphocytes using simultaneous Hoechst 33342 labeling andCD5-positive B cells with dim fluorescence. Cytometry 12, immunophenotyping. Cytometry 13, 169–174.360–365. Macedo, A., Orfao, A., Vidrailes, M., Lopez-Berges, M., Valverde,

Grannis, G.F., Gruemer, H.D., Lott, J.A., Edison, J.A., McCabe, B., Gonzalez, M., Caballero, M., Ramos, F., Martinez, M.,W.C., 1972. Proficiency evaluation of clinical chemistry Fernandes-Calvo, J., 1995. Characterization of aberrantlaboratories. Clin. Chem. 18, 222. phenotypes in acute myeloblastic leukemia. Ann. Hematol. 70,

Hassett, J., Parker, J., 1995. Laboratory practices in reporting flow 189–194.cytometry phenotyping results for leukemia / lymphoma speci- McCoy, Jr. J.P., Overton, W.R., 1996. A survey of currentmens: results of a survey. Cytometry 22, 264–281. practices in clinical flow cytometry. Am. J. Clin. Pathol. 106,

Henry, R.J., 1959. Use of the control chart in clinical chemistry. 82–86.Clin. Chem. 5, 309. McCoy, J., Carey, J., Kause, J., 1990. Quality control in flow

Henry, R.J., Segalove, M., 1952. The running of standards in cytometry for diagnostic pathology. I. Cell surface immuno-clinical chemistry and the use of the control chart. J. Clin. phenotyping and general laboratory procedures. Am. J. Clin.Pathol. 5, 305. Pathol., S27–S37.

Hertler, A.A., Schlossman, D.M., Borowitz, M.J., Laurent, G., Muirhead, K., 1993. Quality control for clinical flow cytometry.Jansen, F.K., Schmidt, C., Frankel, A.E., 1988. A phase I In: Bauer, K., Duque, R., Shankey, T. (Eds.), Clinical Flowstudy of T101-ricin A chain immunotoxin in refractory chronic Cytometry: Principles and Applications. Williams & Wilkins,lymphocytic leukemia. J. Biol. Response Modifiers 140, 2322– Baltimore, pp. 177–199.2329. National Committee for Clinical Laboratory Standards, 1990.

Hurley, A.A., 1988. Quality control in flow cytometry. Cytometry Procedures for the handling and processing of blood speci-(Suppl. 3). mens: approved guideline. NCCLS document H18-A. NCCLS,

Hurley, A.A., 1997a. Principles of quality control. In: Current Villanova, PA.Protocols in Cytometry. Wiley, New York. National Committee for Clinical Laboratory Standards, 1992.

Hurley, A.A., 1997b. Components of quality control. In: Current Clinical applications of flow cytometry. Quality assurance andProtocols in Cytometry. Wiley, New York. immunophenotyping of peripheral blood lymphocytes. Instru-

Hurley, A.A., 1997c. Quality control in phenotypic analysis by mentation Section, National Committee for Clinical Labora-flow cytometry. In: Current Protocols in Cytometry. Wiley, tory Standards. Approved Guideline. NCCLS document H42-New York. T. NCCLS, Villanova, PA.

Hurley, A.A., Zito, D.R., 1998a. Regulatory affairs. Cytometry 34 National Committee for Clinical Laboratory Standards, 1997a.(5). Clinical applications of flow cytometry. Immunophenotyping

Hurley, A.A., Zito, D.R., 1998b. The ISAC Cytometry CLIA of leukemic cells. Approved Guideline. NCCLS documentCompliance Manual. International Society for Analytical H43-A. NCCLS, Villanova, PA.Cytology/Wiley Liss. National Committee for Clinical Laboratory Standards, 1997b.

Hurwitz, C., Loken, M., Graham, M., Karp, J., Borowitz, M., Clinical applications of flow cytometry: quality assurance andPullen, D., Civin, C., 1988. Asynchronous antigen expression immunophenotyping of peripheral blood lymphocytes. Ap-


proved Guideline. NCCLS document H42-A. NCCLS, Vil- dardized fluorescence calibration plot using matching spectrallanova, PA. calibrators. Cytometry (Commun. Clin. Cytometry) 26, 22–31.

Nicholson, J., Kidd, P., Mandy, F., Livnat, D., Kagan, J., 1996. Schwartz, A., Fernandez Repollett, E., Vogt, R., Gratama, J.,Three-color supplement to the NIAID DAIDS guideline for 1996b. Standardizing flow cytometry: construction of a stan-flow cytometric immunophenotyping. Cytometry (Commun. dardized fluorescence calibration plot using matching spectralClin. Cytometry) 26, 227–230. calibrators. Cytometry 26.

Nicholson, J., McDougal, J., Hearn, T., Collins, C., Valleca, W., Segal, G.H., Edinger, M.G., Owens, M., McNealis, M., Lopez, P.,Jones, W., Gerger, A., 1994. CDC (1994) revised guidelines Perkins, A., Linden, M.D., Fishleder, A.J., Stoler, M.H.,for the performance of CD41 T-cell determinations in persons Tubbs, R.R., 1991. Concomitant delineation of surface Ig,with human immunodeficiency virus (HIV) infections. J.

B-cell differentiation antigens, and HLA-DR on lymphoidMMWR 43, 1–21.

proliferations using three-color immunocytometry. CytometryOwens, M., Loken, M., 1995. Flow Cytometry Principles For

12, 350–359.Clinical Laboratory Practice. Quality Assurance For Quantita-Seivers, E., Lange, B., Buckley, J., Smith, F., Wells, D., Daig-tive Immunophenotyping. Wiley-Liss, New York.

neault-Creech, C., Shults, K., Bernstein, I., Loken, M., 1996.Paxton, H., Kidd, P., Landay, A., Giorgi, J., Flomenberg, N.,Prediction of relapse of pediatric acute myeloid leukemia byWalker, E., Valentine, F., Fahey, J., Gelman, R., 1989. Resultsuse of multidimensional flow cytometry. J. Natl. Cancer Inst.of the Flow Cytometry ACTG Quality Control Program:88, 1483–1488.analysis and findings. Clin. Immunol. Immunopathol. 52, 69–

Shimenti, K.J., Lacerna, K., Wamble, A. et al., 1992. Reticulocyte74.quantification by flow cytometry, image analysis, and manualPollice, A.A., McCoy, Jr. J.P., Shackney, S.E., Smith, C.A.,counting. Cytometry 13, 853–862.Agarwal, J., Barholt, D.R., Janocko, L.E., Homicek, F.J.,

Shuey, H.E., Cebel, J., 1949. Bull. US Army Med. Dept. 9, 799.Singh, S.G., Hartsock, R.J., 1992. Sequential paraformal-Steen, H.B., 1991. Noise, sensitivity, and resolution of flowdehyde and methanol fixation for simultaneous flow cyto-

cytometers. Cytometry 13, 822–830.metric analysis of DNA, cell surface proteins, and intracellularStelzer, G.T., Marti, G., Hurley, A., McCoy, Jr. P., Lovett, E.J.,proteins. Cytometry 13, 432–444.

Press, O.W., Farr, A.G., Borroz, K.I., Anderson, S.K., Martin, P.J., Schwartz, A., 1997. U.S.–Canadian consensus recommenda-1989. Endocytosis and degradation of monoclonal antibodies tions on the immunophenotypic analysis of hematologictargeting human B-cell malignancies. Cancer Res. 49, 4906– neoplasia by flow cytometry: standardization and validation of4912. laboratory procedures. Cytometry 30, 214–230.

Rameshwar, P., Denny, T.N., Stein, D., Gasc, P., 1994. Monocyte Stewart, C.C., Stewart, S.J., 1994. Cell preparation for theadhesion in patients with bone marrow fibrosis is required for identification of leukocytes. In: Darzynkiewicz, A., Robinson,the production of fibrogenic cytokines. Potential role for J., Crissman, H. (Eds.). Methods in Cell Biology, Vol. 41.interleukin-1 and TGF-B. J. Immunol. 153, 2819–2830. Academic Press, New York, pp. 39–60.

Roth, G., Schmitz, G., 1996. Consensus protocol for the flow Stewart, C.C., Stewart, S.J., 1995. The use of directly andcytometric immunophenotyping of hematopoietic malignan- indirectly labeled monoclonal antibodies in flow cytometry. In:cies: Working Group on Flow Cytometry and Image Analysis. Davis, W.C. (Ed.). Methods in Molecular Biology, Vol. 45.Leukemia 10, 877–895. Humana Press, pp. 129–147.

Ryan, D.H., Nuccie, B.L., Abboud, C.N., Winslow, J.M., 1991. Stewart, C.C., Stewart, S.J., 1997a. Titering antibodies. In:Vascular cell adhesion molecule-1 and the integrin VLA-4 Robinson, J.P., Darzynkiewicz, Z., Dean, P., Dressler, L.,mediate adhesion of human B cell precursors to cultured bone Rabinovitch, P., Stewart, C., Tanke, H., Wheeless, I. (Eds.),marrow adherent cells. J. Clin. Invest. 88, 99–1004. Current Protocols in Cytometry. Wiley, New York, pp. 4.1.1–

Sasaki, D., Bumas, S., Engleman, E., 1987. Discrimination of 4.1.13.viable and nonviable cells using propidium iodide in two color Stewart, C.C., Stewart, S.J., 1997b. Phenotypic analysis. In:immunofluorescence. Cytometry 8, 413–420. Robinson, J.P., Darzynkiewicz, Z., Dean, P., Dressler, L.,

Sax, S.M., Dorman, L., Lebenson, D.D., Moore, J.J., 1967. Design Rabinovitch, P., Stewart, C., Tanke, H., Wheeless, L. (Eds.),and operation of an expanded system of quality control. Clin. Current Protocols in Cytometry. Wiley, New York, pp. 6.0.1–Chem. 13, 825. 6.2.18.

Schmid, I., Krall, W., Uittenbogaart, C., Braun, J., Giorgi, J., 1992. Stewart, C.C., Stewart, S.J., 1999. Four color compensation.Dead cell discrimination with 7-amino-actinomycin D in Commun. Clin. Cytometry 38, 161–175.combination with dual color immunofluorescence in single Stewart, C.C., Behm, F.G., Carey, J.L., Cornbleet, J., Duque, R.E.,laser flow cytometry. Cytometry 13, 204–208. Hudnall, S.D., Hurtubise, P.E., Loken, M., Tubbs, R.R.,

Schwartz, A., Fernandez-Repollet, E., Velilla, M., Ocasio, B., Wormsley, S., 1997. U.S.–Canadian consensus recommenda-1990. Calibration of flow cytometers to directly measure tions on the immunophenotypic analysis of hematologicnumbers of antibody binding sites on cells. Abstracts for the neoplasia by flow cytometry: selection of antibody combina-XIV International Meeting of the Society for Analytical tions. Cytometry 30, 231–235.Cytology, Ashville, NC. Cytometry (Suppl. 4), 463B. Timm, E.A., Podniesinski, E., Duckett, I., Cardott, J., Steward,

Schwartz, A., Fernandez-Repollet, E., Vogt, R., Gratama, J., C.C., 1995. Amplification and detection of a-chromosome1996a. Standardizing flow cytometry: construction of a stan- DNA sequence by fluorescence in situ polymerase chain


reaction and flow cytometry using cells in suspension. Cytom- Whitehurst, P., DiSilviio, T.V., Boadjian, G., 1975. Evaluation ofetry 22, 250–255. discrepancies in patients results — An aspect of computer-

Tomer, A., Harker, L.A., 1989. Quantitative cytometric measure- assisted quality control. Clin. Chem. 21, 87.ment of normal human megakaryocytes. Nouv. Rev. Fr. Wittrup, K.D., Westerman, R.J., Dessai, R., 1994. FluorescenceHematol. 31, 233. array detector for large-field quantitative fluorescence. Cytom-

Vogt, R.F., Cross, G.D., Henderson, L.O., Phillips, D.L., 1989. etry 16, 206–213.Model system evaluating fluorescein-labeled microbeads as Wooten, I.D.P., King, E.J., 1953. Normal values for bloodinternal standards to calibrate fluorescence intensity of flow constituents: inter-hospital differences. Lancet 1, 470.cytometers. Cytometry 10, 294–302. Wormsley, S.B., Baird, S.M., Gadol, N., Rai, K.R., Sobol, R.E.,

Vogt, R.F., Cross, G.D., Phillips, D.L., Henderson, L.O., Hannon, 1990. Characteristics of CD11c1CD51 chronic B-cellW.H., 1991. Interlaboratory study of cellular fluorescence leukemias and the identification of novel peripheral bloodintensity measurements with fluorescein-labeled microbead B-cells subsets with chronic lymphoid leukemia immuno-standards. Cytometry 12, 525–536. phenotypes. Blood 76, 123–130.

Vogt, R., Cross, G., Phillips, D., Henderson, I., Hannon, W., 1994. Zagursky, R.J., Sharp, D., Solomon, K., Schwartz, A., 1995.Interlaboratory study of cellular fluorescence intensity mea- Quantitation of cellular receptors by a new immunocytochemi-surements with fluorescence intensity for clinical and research cal flow cytometry technique. BioTechniques 18 (3).applications of immunophenotyping by flow cytometry. In: Zucker, R.M., Elstein, K.H., Gershey, E.L., Massaro, E.J., 1991.Tyler, H. (Ed.). Reviews in Biotechnology and Bioengineer- Increasing sensitivity of the ortho analytical cytofluorographing, Vol. 1. Ablex, Norwood, NJ. by modifying the fluid system. Cytometry 11, 848–851.

validation and quality control of immunophenotyping in ... · marilyn a. owens , horacio g. vall ,...

Documents