European Research Initiative on CLL (ERIC)

guidelines for the detection of minimal

residual disease in CLL using multi-

parameter (≥4CLR) flow cytometry Andy C. Rawstron; Claudia Fazi; Andreas Agathangelidis; Neus Villamor; Remi Letestu; Josep Nomdedeu; Carlos Palacio; Olga Stehlikova; Karl-Anton Kreuzer; Stuart Liptrot; David O’Brien; Ruth M. de Tute; Martin

Spacek; Johan Dobber; Arnon P. Kater; Peter Gambell; Asha Soosapilla; Gerard Lozanski; Gabriele Brachtl; Ke Lin, Curtis Hanson; Jeffrey L.

Jorgensen; Maryalice Stetler-Stevenson; Constance Yuan; Beth Broome; Fiona Craig; Julio Delgado; Carol Moreno; Francesc Bosch;

Alexander Egle; Michael Doubek; Sarka Pospisilova; Stephen Mulligan; David Westerman; Tait Shanafelt; Andy Pettitt; Thomas J. Kipps; William

G. Wierda; Florence Cymbalista; Michael Hallek; Peter Hillmen;

Emili Montserrat and Paolo Ghia

on behalf of the ERIC consortium

Lessons from the several years of attempting to

harmonise CLL MRD by flow cytometry

• The CLL community needs to progress – Flow MRD can easily achieve 10-5, but do we want this if it

then means there are insufficient cells left to test the novel assay that can achieve 10-6 ?

• The participating labs need flexibility – Variable equipment specification, reagents restricted to

certain suppliers, compatibility with local procedures, locally critical markers (CD160, ROR1)

• Researchers and biotech need to know that their assay meets requirements without re-validating every individual combination – Provide a specification for the required components so that

developers can demonstrate approved standard.

Guidelines on the validation of cell-based

assays • Sensitivity

– Either: “lowest signal detectable above background”

– Or: “true positive / true positive + false negative”

• Limit of Blank (LOB) = highest signal in the absence of measurand, calculated as mean (blank) + 1.645 SD (95% of negative values are below this limit)

• Limit of Detection (LOD) = level at which 95% of samples with low level of measurand are detected above the limit of blank, calculated as LOB + 1.645 SD

• Limit of Quantification (LoQ) = lowest level of measurand that can be reliably detected and whose total error (bias + Imprecision) meets a desired criterion for accuracy (clinical utility)

Cytometry Part B (Clinical Cytometry) 84B: 315-323

Clinically appropriate level of variation:

When BCR-ABL RQ-PCR accepted as a trial

end-point, 95% limit of agreement = ± 5-fold

CML: best 95% LOA +/- 2-fold using synthetic standards

CLL: Target: ±3-fold (0.5log), preferably ±2-fold (0.3 log)

Multiparameter flow CLL MRD Guidelines: aims

of the project

• The primary aim of this project was to identify and

validate the specification of an assay that could provide

a stable core set of markers for MRD detection:

– at the levels required by the IWCLL guidelines or better

– compatible with published outcome data

– independent of instrument/reagent manufacturer

– offer sufficient flexibility for the use and validation of additional

markers.

• The secondary aim of this study was to compare the

assay with High Throughput Sequencing to determine

the most appropriate application of each technology.

4 vs. 6 vs. 8-CLR Component markers

FITC PE PerCP-Cy5.5

PE-Cy7

APC APC-

H7

CD45 CD3 CD19 CD5

CD20 CD38 CD19 CD5

CD81 CD22 CD19 CD5

CD43 CD79b CD19 CD5

CD3 CD38 CD5 CD19 CD79b CD20

CD81 CD22 CD5 CD19 CD43 CD20

4CLR 20-50 CLL events in ≥2 of 3

tubes

6CLR 20-50 CLL events in both tubes

V450 V500 FITC PE PerCP-Cy5.5

PE-Cy7 APC APC-H7

CD5 CD3 CD81 CD79b CD22 CD19 CD43 CD20

0.00001

0.0001

0.001

0.01

0.1

1

10

0.00001 0.0001 0.001 0.01 0.1 1 10

CLL

% le

uco

cyte

s u

sin

g 1

-tu

be

8-C

LR a

ssay

Expected CLL % leucocytes

1-tube 8-CLR dilution analysis

A) 8-CLR 1-tube panel dilution

analysis

V450 V500 FITC PE PerCP-Cy5.5

PE-Cy7 APC APC-H7

CD5 CD3 CD81 CD79b CD22 CD19 CD43 CD20

0.00001

0.0001

0.001

0.01

0.1

1

10

0.00001 0.0001 0.001 0.01 0.1 1 10

CLL

% le

uco

cyte

s u

sin

g 1

-tu

be

8-C

LR a

ssay

Expected CLL % leucocytes

-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

0.0001 0.001 0.01 0.1 1 10 100

Dif

fere

nce

in C

LL %

of

leu

cocy

tes

bet

wee

n 8

-CLR

an

d

core

mar

ker

sin

gle

tub

e as

says

(lo

g tr

ansf

orm

ed)

CLL % of leucocytes average of 8-CLR and core marker single tube assays

Above LOQBelow LOD

6-markers are sufficient for MRD detection

A) 8-CLR 1-tube panel dilution

analysis

B) 6 markers are sufficient for

detection of MRD

V450 V500 FITC PE PerCP-Cy5.5

PE-Cy7 APC APC-H7

CD5 CD3 CD81 CD79b CD22 CD19 CD43 CD20

1-tube 6-CLR analysis limit of detection 10-5

with acceptable inter-operator variation

A) 6-CLR 1-tube core marker panel

dilution analysis: LoD 0.001%

B) Variation tested in 19 students at

ESCCA-sponsored educational

meetings

0.00001

0.0001

0.001

0.01

0.1

1

10

0.00001 0.0001 0.001 0.01 0.1 1 10

CLL

% le

uco

cyte

s u

sin

g M

RD

-co

re p

anel

Expected CLL % leucocytes

Figure 3: A platform-independent specification for MRD-flow

reagents

CLL cells Normal B-cells Normal T-cells

Acceptable Sub-optimal

A platform-independent specification for MRD-flow reagents

Antigen

Typical

Expression

(% pos vs.

control) ‡

Control populations in normal

peripheral blood Minimum

Relative

fluorescence

intensity

(preferred) Positive Negative

CD5 Positive (>20%) CD3+ T-cells CD19+ B-cells >30 (>65)

CD20 Weak CD19+ B-cells CD3+ T-cells >13 (>20)

CD43 Positive (>20%) CD3+ T-cells CD20+ B-cells >15 (>40)

CD79b Weak CD20+ B-cells CD3+ T-cells >15 (>30)

CD81 Weak CD3+ T-cells Granulocytes >12 (>20)

Case 1

Case 2

Case 3

Sequence 1 Sequence 2 Average of S1 & S2

LoD: (≤) 1 in a million

Dilutional analysis (Milan) assessing high-throughput sequencing: analysis of 3 CLL cases diluted into

leucocytes from leucodpletion filters in serial 1:10. Each CLL clone was tagged with 2 sequences, 1

productive and 1 non-productive. The plot shows the CLL sequence as a percentage of nucleated genomes.

For log-transformed data above the LoQ, linearity = 1.15, correlation co-efficient (Pearson R) = 0.980, average

difference = 0.15 log, 95% limit of agreement 0.76 log.

ClonoSEQ HTS in comparison to Sanger

sequencing and MRD-flow in patient samples

• Identifying a clonal marker: HTS detected a dominant clone (median 99.1%, range 7.8-100%) in 66/68 samples vs. 57/68 by Sanger sequencing

– 1 technical issue / 1 LC-MBL

• Clonal IGH CDR3 sequences were identical in 49/57 – 5 cases Sanger-identified clone

detected by HTS (3/5 unproductive, 2/5 not dominant)

– 3 discrepant in LC-MBL (HTS most likely correct)

• Repeat analysis confirms dominant clone in 10/10 CLL/HC-MBL cases

12/13 concordant at 0.01% threshold

(1 flow below LoQ)

3/13 MRD detected by ClonoSEQ

not by flow

REF

MRD-core (CLL

cells % of

leucocytes)

ClonoSEQ

(neoplastic

sequences % of

H24O286M <0.001 0.00035

H003O335M <0.001 0.00054

H235O273M <0.001 0.00061

H78O277M [0.002] 0.0022

H93O275M 0.008 0.0039

H69O280M 0.012 0.0051

H112O101M <0.001 0.0082

H204O278M 0.46 0.46

H104O351M 23 0.55

H104O351M 13 0.55

H67O274M 1.48 0.56

H64O385M 1.59 0.99

H130O285M 0.3 10.5

HTS shows greater variation in quantification above

0.01% but better accuracy below 0.01%

4-CLR vs. ≥6-CLR flow cytometry HTS vs. ≥6-CLR flow cytometry

Comparison of the MRD-core panel with ClonoSEQ high throughput sequencing: analysis of 3 CLL cases

diluted into leucocytes from leucodpletion filters in serial 1:10 dilutions and 13 CLL cases after treatment. The

results are not corrected for LoD/LoQ. For log-transformed data above the LoQ, linearity = 0.905, correlation co-

efficient (Pearson R) = 0.870, average difference = 0.078 log, 95% limit of agreement 1.5 log.

Multiparameter flow CLL MRD guidelines: aims

of the project

• Identify assay specification for CLL MRD analysis to

achieve at least 0.01% limit of quantification

– CD19 (or appropriate gating markers) in combination with CD5,

CD20, CD43, CD79b, CD81 markers meeting defined

specification

– Interoperator variation in analysis: 95% LOA within 2-fold (log-

scale, ±0.3) achieved by operators after reviewing protocol /

test cases.

• Compare ≥4CLR MRD flow with high-throughput

sequencing

– Optimal combination direct quantitative assay to 0.01 -

0.001% (10-4 to 10-5) combined with HTS to exclude disease

below 1 in a million (10-6).

B-cell

enumeration +

clonality

assessment

Relatively inexpensive and simple.

Requires capacity to reflex to full

MRD assay if CLL cells <1%

&/or B-cells polyclonal.

4-colour 4-tube

Published outcome data

Does not require pre-treatment

phenotype for typical CLL

Limit of detection >0.005%

More material required to

achieve higher detection limits

6-colour 2-tube

Harmonised with published

outcome data

Does not require pre-treatment

phenotype for typical CLL

Intermediate LoD/LoQ

Intermediate amount of

materialrequired to achieve

higher detection limits.

6-colour core

panel for ≥6-

colour assays

Flexibility for individual laboratory

requirements

LoD 0.001% (10-5), LoQ 0.0025%

Allows simultaneous analysis of

additional markers

Knowledge of pre-treatment

phenotype preferable.

High throughput

sequencing

LoD 0.0001% (10-6)

Objective analysis, does not

necessarily require expert

interpretation

Further development work on

standardisation of the

quantification

Different ERIC-harmonised approaches for

MRD detecetion

Acknowledgements Johan Dobber, Arnon Kater Academic Medical Center, Amsterdam, The Netherlands.

Remi Letestu, Florence Cymbalista AP-HP, Hôpital Avicenne, Bobigny, France

Martin Spacek Charles University in Prague, Czech Republic

Neus Villamor, Julio Delgado Hospital Clínic, Barcelona, Spain.

Josep Nomdedeu, Carol Moreno Hospital de la Santa Creu i Sant Pau, Barcelona, Spain.

Claudia Fazi, Paolo Ghia Università Vita-Salute San Raffaele, Milan, Italy

Gabriele Brachtl, Alexander Egle Paracelsus Medical University, Salzburg , Austria.

Stuart Liptrot, David O'Brien St. James's Hospital, Dublin, Ireland.

Andy C. Rawstron, Ruth M. de Tute, Peter Hillmen St. James's Institute of Oncology, Leeds Teaching Hospitals, UK.

Olga Stehlikova, Michael Doubek, Sarka Pospisilova University Hospital Brno, Czech Republic.

Karl-Anton Kreuzer, Michael Hallek University of Cologne, Köln, Germany

Carlos Palacio, Francesc Bosch Vall d'Hebron Hospital, Barcelona, Spain.

Peter Gambell, David Westerman Peter MacCallum Cancer Centre, East Melbourne, Australia

Asha Soosapilla, Stephen Mulligan Royal North Shore Hospital, University of Sydney, Australia

Constance Yuan, Maryalice Stetler-Stevenson Center for Cancer Research, National Institutes of Health, Bethesda, MD, USA.

Gerard Lozanski Department of Pathology, The Ohio State University

Curtis Hanson, Tait Shanafelt Mayo Clinic, Rochester, Minnesota, USA.

Jeffrey L. Jorgensen, William G. Wierda MD Anderson Cancer Center, University of Texas

Beth Broome, Thomas J. Kipps Moores Cancer Center, University of California, La Jolla, CA, USA.

Fiona Craig University of Pittsburgh School of Medicine, Pittsburgh, PA, USA

Frans Nauwelaers, Lucia Testolin, Jingyi Chen and Noel Warner. BD

Biosciences provided custom and commercial conjugates for testing

Carol Moreno, EMILI MONTSERRAT and PAOLO GHIA

HTS vs. Sanger: Andreas Agathangelidis

Dilution studies: Claudia Fazi

Ruth de Tute, ANDREW JACK and PETER HILLMEN

Michael Wenger for meeting support

Comparison of flow vs. next

generation sequencing

Which patients can be monitored:

diagnosis of CLL • No universal molecular abnormality

• WHO Definition: CLL cells usually co-express CD5 and CD23

• WHO Immunophenotype: using flow cytometry, the tumour cells express

dim surface IgM/IgD, CD20, CD22, CD5, CD19, CD79a, CD23, CD43 and

CD11c (weak). CD10 is negative and FMC& and CD79b are usually

negative or weakly expressed in typical CLL. Some cases may have an

atypical immunophenotype (e.g. CD5- or CD23-, FMC7+ or CD11c+,

strong sIg, or CD79b+).

• iwCLL/NCI guidelines: CLL cells co-express the T-cell antigen CD5 and

B-cell surface antigens CD19, CD20, and CD23. The levels of surface

immunoglobulin, CD20, and CD79b are characteristically low compared

with those found on normal B cells. Each clone of leukemia cells is

restricted to expression of either kappa or lambda immunoglobulin light

chains. Variations of the intensity of expression of these markers may

exist and do not prevent inclusion of a patient in clinical trials for CLL.

Difficult to validate diagnostic panels – circular argument for the gold standard

CLL Diagnostic Panel

• An invitation to participate in this study was circulated among members of the ERIC and ESCCA scientific groups.

• Responses were received from 158 members of which 154 were actively involved in CLL diagnosis: 100/154 (65%) were from diagnostic laboratories, 14/154 (9.1%) were CLL clinicians and 36/154 (23%) worked in both laboratory and clinic.

• The diagnostic workload was >20 cases per week in 23/154 (15%), 5-20 per week in 82/154 (53%) and <5% in 49/154 (32%).

• Responders were invited to classify 35 antibodies selected from publications on the diagnosis of B-lymphoproliferative disorders as being required, recommended, suggested, uninformative, or of unknown value for the diagnosis of CLL.

>75% of respondents marker is

required for CLL diagnosis:

>50% of respondents marker is

recommended for CLL diagnosis:

Proposal: required (minimum) and

recommended panel for diagnosis • Marker panel required for diagnosis:

– CD19 / CD5 / CD23 / CD20 / Kappa / Lambda

• Additional markers recommended for diagnosis and required for

clinical trials:

– CD43 / CD79b / CD81* / CD22 / CD10

• Frequently recommended but not essential for diagnosis and

monitoring:

– CD45 / CD38 / CD200 / (FMC7)

• Present in current diagnostic criteria but not universally

recommended:

– IgM/D and CD11c:

• * CD81 was considered as a required or recommended by only 40% of participants but this marker is an essential

component for the consensus MRD monitoring panels [REFS] and therefore recommended by the steering committee for

clinical trials.

Figure 3a: A platform-independent specification for MRD-flow

reagents

CLL cells Normal B-cells Normal T-cells

Acceptable Sub-optimal

Antigen

Typical

Expression

(% pos vs.

control) ‡

Are the same

criteria required

for atypical CLL?

Control Population in normal

peripheral blood

Minimum

Relative

fluorescence

intensity

(preferred) Positive Negative

CD19 Positive (>95%) Yes CD20+ B-cells CD3+ T-cells >20†

CD5 Positive (>20%) Yes CD3+ T-cells CD16/56+ NK-

cells >14 (>18)

CD23 Positive (>20%) ? Not required CD20+CD27-

Naïve B-cells

CD20+27+

Memory B-cells >5†

CD20 Weak Not required CD19+ B-cells CD3+ T-cells >5 (>20)

Igκ

Igλ

Weak & restricted

to either Igκ or Igλ

Restricted to either

Igκ or Igλ CD20+ B-cells CD3+ T-cells >10†

CD43 Positive (>20%) Not required CD3+ T-cells CD20+ B-cells >7 (>50)

CD79b Weak Not required CD20+ B-cells CD3+ T-cells >11 (>30)

CD81 Weak Not required CD3+ T-cells Granulocytes >5 (>8)

CD22 Weak Not required CD20+ B-cells CD3+ T-cells >10†

CD10 Negative (<20%) Yes Granulocytes CD20+27+

Memory B-cells >10†

Definition of weak: median fluorescence intensity at least 20% lower than normal peripheral blood B-cells,

range to be determined within each laboratory (ICSH/ISLH/CLIA guidelines for stability require <20%

variation, therefore reduction in fluorescence intensity less than 20% may reflect antigen/sample stability)

† consensus, not specifically validated