Minimal Residual Disease Detection in

Acute Leukemia

11/06/2015

Minimal Residual Disease (MRD) is the presence

of a small number of malignant cells at a level
below the sensitivity of morphologic detection.

Why MRD?

Assess residual burden of the disease

Assess therapeutic response
Prognostic information
Stratification of treatments

Response, MRD, and Relapse

J.J.M van Dongen, et al. Blood. 2015 Jun 25; 125(26): 39964009.

MRD is the strongest predictor for relapse

and poor prognosis.

Childrens Oncology Group (COG) study, Acute lymphoblastic leukemia (ALL):

- End of induction MRD

Borowitz, M, et al. Blood. 2008 Jun 15; 111(12): 54775485.

Day 8 blood MRD:

All ALL

Day29 MRD- ALL

Borowitz, M, et al. Blood. 2008 Jun 15; 111(12): 54775485.

All ALL at end of consolidation

Borowitz, M, et al. Blood. 2008 Jun 15; 111(12): 54775485.

Prognostic significance of MRD in pediatric ALL (Flow cytometry)

Variable

Hazard ratio

Dar 29 MRD>0.01%

4.31

<0.001

NCI risk group

2.25

<0.001

Trisomies 4 and 10

.570

<0.001

Day 8 MRD (PB)>0.01

1.51

.018

TEL-AML1

.778

.151

Day8 M1 marrow

1.034

.789

COG study: AML

MRD at the end of induction I/II

Loken MR, et al. Blood. 2012 Aug 23; 120(8): 15811588.

COG: AML
MRD at the end of therapy

Loken MR, et al. Blood. 2012 Aug 23; 120(8): 15811588.

Impact of Pretransplantation MRD on Outcome of Hematopoietic Cell

Transplantation for AML.

Walter RB, et al. J Clin Oncol. 2011 (29):11901197.

 Assignment into different treatment arms

Treatment intensification
Treatment reduction

Decisions for hematopoietic stem cell transplantation

https://members.childrensoncologygroup.org/_files/meetings/Denver2008/COGALLFrontlineTrials10-22-08Part1.pdf

How to Detect MRD ?

Flow Cytometry is the Standard Method for

MRD Detection Currently in USA.

Flow Cytometry MRD detection

Advantages:

Fast- same day reporting

Sensitive
Applicable for most of the cases
Widely available
Cost effective

Disadvantages:
Not standardized
Subjective interpretation

Pattern Recognition

Two Strategies:
Leukemia-Associated immunophenotype (LAP)
Deviation from normal maturation

Leukemia-Associated immunophenotype (LAP)

based method
- Target the LAP region

- Advantages
- Conceptually simple and objective
- Reduced reagent expense for follow up
- Disadvantages
- Requires pre-treatment sample to define LAP
- Requires immunophenotypic stability
- Any event in pre-defined gate regarded as MRD
- Background noise (nonspecific staining)

Feller et al. Leukemia 2004, 18:1380-1390

6-C B-ALL MRD

Diagnostic BM,

Day 29 BM:

MRD: 0.09%

Immunophenotypic Stability

Immunophenotypic Stability
ALL
- 30 consecutive patients with MRD detectable
- All had some change in immunophenotype
- CD10 and CD34 down-modulation, CD19, CD20 and CD45RA
up-modulation, CD58 stable
- Associated with use of steroid in induction therapy

Gaipa et al (2005) Leukemia 19: 49-56

Immunophenotypic Stability
T-ALL

Roshal, et al (2010) Cytometry 78B:139-146

Dr. Woods ICCS 2015 lecture

Deviation from normal maturation

-Target abnormal population
- Advantages:

Does not require pretreatment sample

Uniform reagent combinations utilized
Improved specificity through population identification
Less sensitive to immunophenotypic instability

- Disadvantages:
Requires detailed immunophenotypic knowledge (expert)
Subjective
More time consuming

Normal B cell Maturation

Wood and Borowitz (2011) Henrys Laboratory Medicine

Normal B cell Maturation

Wood (2004) Methods Cell Biology 75:559-576

ALL Informative Antigens

From Krampera et al (2006) Haematologica 91:1109-1112

Rothenber EV, et al. Nat Rev Immunol. 2008 Jan; 8(1): 9

21.

Normal T cell Maturation

Wood and Borowitz (2011) Henrys Laboratory Medicine

http://www.cytometry.org/newsletters/eICCS-2-1/article3.php

Normal Neutrophil Maturation

Wood and Borowitz (2011) Henrys Laboratory Medicine

Normal Neutrophil Maturation

Normal Blast Maturation

Wood (2004) Methods Cell Biology 75:559-576

AML Informative Antigens

Average 2.3 LAIP per patient

From Kern et al (2005) CRC Rev Onc/Hem 56:283-309

Myeloid leukemia with DS: CD117+CD34min+DR+CD33+CD13+CD71+CD7+CD11b+

End of Induction:

MRD: 0.5%

COG AML MRD Panel

Immunophenotypic Aberrancy Suitable for

MRD Detection

ALL
>95% of pediatric ALL
>90-95% of adult ALL

AML
85% of pediatric AML
80-95% of adult AML

Wood, B, ICCS 2015 meeting

Timing
Induction nadir (day 14)
- Reduced background populations
- Hypoplastic with many apoptotic cells

End of induction
- ALL - Few immature B cells
- AML - Active marrow regeneration, increased precursors

End of consolidation
- ALL - Larger number of immature B cells
- AML - Normal marrow populations

Kinetics of response are important

MRD sample Source

Coustan-Smith, et al. Blood. 2002, 100: 2399-2402

Quantitation

Analytical Sensitivity
ALL: 0.01%, AML: 0.1%
Determinants:
Degree of immunophenotypic aberrancy
Number and immunophenotype of background populations

Identification vs. Enumeratio

Number of events acquired
10~50 events required to identify q population
500,000
Reproducibility (Poisson counting statistics)
CV ~ Sqrt (N)/N
100 events gives CV of 10%
Sensitivity of 0.01% requires 1,000,000 events

Denominator
Total nucleated cells
- Most comparable to morphology
- DNA binding dye often used (Syto16, Draq5, etc.)
Incomplete RBC lysis, platelet aggregates

- Under-representation of NRBCs with lysis and washing

White cells
- CD45 positive cells + neoplasm
- Variable CD45 on early NRBCs
- Overestimation with erythroid hyperplasia

Mononuclear cells
- Exclude granulocytic lineage (high side scatter)
- Most comparable to Ficoll-prepared samples
Early MRD literature used Ficoll

- Reduced variability due to granulocytic degeneration

Shipped samples

Hemodilution
Bone marrow is a semi-solid tissue
- Absolute cell concentration has little meaning

Marrow aspiration is a traumatic procedure

- Variable amount of peripheral blood introduced
- Increased amounts of blood with each subsequent aspiration
1st aspirate should be used for MRD

Not a major problem for many samples

- Problem in hypocellular marrows, high PB WBC count or
poor quality aspirates

No method for accurate correction

- One method for normalization proposed for blast counts

Sources of Variability
Identification (false positive or negative)
- Insufficiently informative reagents
- Improper assay validation
- Immunophenotypic shift
- Inexperienced interpreters
Quantitation
- Too few events acquired
- Denominator effects (2 fold)
- Sample degeneration
- Hemodilution

Reproducibility
Central reference lab system
- East= Dr. Borowitz, John Hopkins University
- West= Dr. Wood, University of Washington

Wood, ICCS meeting 2015

Flow MRD on AALL03B1

80
70
60

% of cases 50
MRD >.01%

40
30
20
10
0

JHU(n=2282)
UW(n=1947)

d29

* day 8 M1 patients excluded

d15*

d8 blood

Unpublished data, courtesy Mike Borowitz

Wood, ICCS meeting 2015

Wood, ICCS meeting 2015

Standardization/COG decentralization
Not standardized
NCI funding mechanism changed
- No longer funds reference lab testing in support of clinical trials
- When: June 2016
Decentralization:
- COG B-ALL Flow MRD Testing Approval Process

COG B-ALL Flow MRD Testing Approval Process

All labs seeking approval will be required to use
the identical standardized protocol developed by
UW/JHU
60 total specimens
15 MRD positive
Day 8 or end of consolidation - may include 4 of these
to get to 60 total
First 5 cases (at least one positive and one negative day 29
case) should be reviewed as the first evaluation step.
45 labs enrolled
15 labs have submitted the first 5 cases
9 passed, including CMH flow lab

Current COG MRD panels on

end-of-induction marrow

Calculation
Tube1/2(MRD/B cells) x Tube3(B cells/Mononuclear cells)

Tube 1 MRD =137, Total B cells 25,711

Tube 2 MRD = 212, Total B cells 39,785
Tube 3 Total B cells 18,345, Mononuclear cells 542,197
MRD:
Tube 1 (137/25,711)*(18,345/542,197)*100
MRD = 0.018% of mononuclear cells
Tube 2 (198/38,785)*(18,345/542,197)*100
MRD = 0.017% of mononuclear cells

Molecular tests
RT-PCR of fusion transcript or mutations

B-ALL: 25~40%, T-ALL: 10 ~15%, AML: 25-40%

Sensitive, 10-4-10-6
Fast
Limited applicability; expression level may change during therapy

RQ-PCR of antigen receptor (TCR/IG) gene Rearrangements

B-ALL: 95%, T-ALL: 95%, AML: 10-15%
Sensitive, 10-4-10-5
Labor intensive, time consuming and expensive (Detection and sequencing of the IGTCR rearrangements at diagnosis and design of the corresponding ASO primers: 3 to 4 weeks;
analysis of follow-up samples: 1 week)

Require specialized laboratories

False negative (oligoclonality, clonal evolution)

RQ-PCR and flow cytometric MRD in childhood ALL

Ryan J., et al. MRD detection in childhood ALL patients at multiple time-points reveals
high levels of concordance between molecular and immunophenotypic approaches. Br
J Haematol 2009 144: 107-115

MRD by Next Gen Sequencing

Determine unique sequence at diagnosis
Deep sequence locus
Count number of unique sequences
Polymorphic loci used for lymphoid neoplasm
IgH, TCR beta, TCR gamma

Highly Sensitive ~10-6

More objective interpretation
Able to standardize
<1 week turnaround time
Not applicable for all pts
Requires pretreatment specimen

Pre- and day 29 post-treatment B lymphoblast frequencies by HTS versus mpFC pretreatment
and day 29 post-treatment, clonal B lymphoblasts were identified by HTS (red) or by mpFC
(blue; N = 91).

David Wu et al. Clin Cancer Res 2014;20:4540-4548

2014 by American Association for Cancer Research

Conclusions
Minimal residual disease (MRD) is strongly associated with poor
outcome
MRD detection is a very important step in current risk-stratified
leukemia treatment
Two sensitive detection methods: Flow cytometry and molecular
Flow cytometry is the standard method for MRD detection in COG
protocols in USA
Flow cytometry has the advantages of fast, cost effective, sensitive
and applicable to wide range of leukemia
Immunophenotype change is common
Accurate FCM detection needs consistent FCM techniques,
informative immunophenotype and knowledge of normal patterns of
antigen expression
NGS, more sensitive MRD detection method is in development