Minimal Residual Disease assessment inMultiple Myeloma by Next-Generation Sequencing

:Prof Hervé AVET-LOISEAU, MD, PhDInstitut Universitaire du Cancer

Toulouse, France

MRD Meta-Analysis in Myeloma Nikhil Munshi, MD

Lipper Myeloma CenterDana-Farber Cancer Institute

Why to assess MRD in Myeloma?The main rationale is the correlation response/outcome

CR, n=278

nCR, n=124PR, n=280

PD, n=25

0 12 24 36 48 60 72 84 96

Months from diagnosis

0,1

0,2

0,3

0,4

0,5

0,6

0,7

0,8

0,9

1,0

Cum

ulat

ive

Prop

ortio

n Su

rviv

ing

RC vs TBRP RC vs RP TBRC vs RP

P=0.01P<10-6

P=0.04

OS

Lahuerta et al. JCO 2008

CR nCR PR PDMedians EFS, months 61 40 34 13Medians OS, months NR NR 61 15

Why to assess MRD in Myeloma?Treatment advances have increased the likelihood of achieving CR.

Why to assess MRD in Myeloma?However, a large majority of pts with CR eventually relapse, suggesting that undetectable, but clinically meaningful MRD may be present.

Why to assess MRD in Myeloma?

Diagnosis 1012

CR 1010 80% HD trials

MRD: What are the techniques?

(Old) Gold Standard: Multiparametric Flow Cytometry

Myeloma cells present a specific phenotype / normal PC

This phenotype is stable during evolution

clonal PCs

normal PCs

PCA1

PCA

2

clonal PCs

normal PCs

MRD: What are the techniques?

Next-Generation Sequencing

Myeloma cells present unique Ig gene rearrangements

These clonal rearrangements are stable during evolution

Locus IGK 2p11

Mix J

Locus IGH 14q32

Mix B Mix CMix E

Mix J

Allèle IgH fonctionnelhypermuté

Functional Allele

Non-Functional allele

Mix D

Mix J

NGS: Technical principles

Non B cell LeukocytesNormal B cellsMyeloma cells

FREQUENCY OF MYELOMA CLONE

AMONG B CELLS = SL / (SL + SB)NUMBER OF MYELOMA MOLECULESPER LEUKOCYTE = SL x (NR/SR) / NTOT

NGS: Technical principles

Is MRD clinically pertinent?

BLOOD, 19 JANUARY 2012 - VOLUME 119, NUMBER 3

BLOOD, 15 MAY 2014 - VOLUME 123, NUMBER 20

Significant Impact of Minimal Residual Disease (MRD) Status On Survival Outcomes In pts (pts) With

Multiple Myeloma (MM) Who Achieve Complete Response (CR): A Meta-Analysis

• A total of 405 published articles with MRD– 25 articles recently published articles

• Of these, 21 reported overall survival (OS) or progression-free survival (PFS) results, as well as MRD status

• Overall, 2,208 pts were evaluated for MRD• Nine publications reported conventional CR at the time of

MRD measurement. Six represented unique data sets.

Munshi N et al., JAMA Oncol, in press

MM, multiple myeloma; MRD, minimal residual disease; SCT, stem cell transplant.

Refining the literature search

Exclusion criteria• Publications were excluded if they:

– Only included patients with relapsed and/or refractory MM

– Included patients who had undergone allogenic SCT– Assessed MRD in apheresis product – Reported on the same study population used in an

already-included trial• Analysis was restricted to techniques with a

detection limit of≤ 0.01%

496

496

1,329

1,515

0 500 1,000 1,500 2,000

Patientsassessed

for OS

Patientsassessedfor PFS

Number of patients assessed

Any response-achievingpatients (n = 13 studies)

CR-achieving patients(n = 4 studies)

21 articles retrieved in total

Number of patients with PFS and OS data allowing for analysis

1. Paiva B, et al. J Clin Oncol. 2011;29:1627-33.2. Paiva B, et al. Blood. 2012;119:687-91.

3. Rawstron AC, et al. J Clin Oncol. 2013;31:2540-7.4. Swedin A, et al. Br J Haematol. 1998;103:1145-51.OS, overall survival; PFS, progression-free survival.

Munshi N et al., JAMA Oncol, in press

The effect of MRD status on PFS and OS (All patients)

Munshi N et al., JAMA Oncol, in press

The effect of MRD status on PFS and OS(All patients)

Munshi N et al., JAMA Oncol, in press

The effect of MRD status on PFS and OS (CR patients)

Munshi N et al., JAMA Oncol, in press

18

The effect of MRD status on PFS (CR patients)

Data are adjusted for different proportions of patients being MRD-positive and MRD-negative by study.

100

80

60

40

20

00 2 4 6 8 10 12

Time (years)

PFS

(%)

2 (adjusted) = 35.85;P < 0.0001

MRD-negative (n=389)MRD-positive (n=155)

Number at risk by year:389 359 301 211 155 96 65 35 16 12 1155 129 86 51 33 19 12 7 7 5 0

• 3-year PFS: 70% (MRD–) vs. 46% (MRD+)

• 5-year PFS: 48% (MRD–) vs. 27% (MRD+)

• Majority of MRD-positive patients progressed by 6 years; nearly 50% of MRD-negative patients progression free

CR patients Median PFSMRD-negative 56 monthsMRD-positive 34 months

CR-achieving patients

Munshi N et al., JAMA Oncol, in press

Munshi N, et al. IMW 2015, abstract 0394.

Data are adjusted for different proportions of patients being MRD-positive and MRD-negative by study.

The effect of MRD status on OS (CR patients)

CR, complete response; MRD, minimal residual disease; NR, not reached; OS, overall survival.

100

80

60

40

20

00 2 4 6 8 10 12

Time (years)

OS

(%)

2 (adjusted) = 15.06;P < 0.0001

MRD-negative (n=362)MRD-positive (n=134)

Number at risk by year:362 359 331 274 218 138 76 34 8 3 1134 131 111 81 55 35 20 10 5 5 2

• Median OS was not reached for MRD-negative pts versus 82 months for MRD-positive pts)

• OS @ 3-years, 94% versus 80% OS @ 7-years, 67% versus 47%• OS @ 5-years, 80% versus 61%

CR patients Median OSMRD-negative 112 monthsMRD-positive 82 months

CR-achieving patients

19

Conclusions of the meta-analysis

MRD is definitely predictive of both longer PFS and OS

Most of the available results are from MFC

MFC has a quite low sensitivity (10-4)

Following question:

Would a higher sensitivity have a better predictivity?

Higher sensitivity: NGS

Martinez-Lopez J et al., Blood 2014;123:3073

IFM DFCI 2009 Trial700 patients < 66y,

Newly diagnosed symptomatic MM

3 RVD

5 RVD MEL200 + ASCT

2 RVD

12 months Lenalidomide maintenance

MRD*

MRD*

MRD*

MRD** Primary objective = 7-color Flow, Secondary objective = Molecular

54% Conventional CR

289 patients analyzed by NGS

Applicability of NGS: 92% (8% clone ID failure)

Median follow-up: 44 months

IFM 2009 trial

P-value (trend) : p<0.0001

<10-6

[10-6;10-5[

[10-5;10-4[

>=10-4

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Patie

nts

with

out p

rogr

essi

on (%

)

79 79(0) 70(9) 59(11) 50(9) 38(11) 28(6) 6(9) 0(3)[10-4;10-3[49 49(0) 47(2) 45(2) 43(2) 34(7) 22(4) 8(6) 2(0)[10-5;10-4[31 31(0) 30(1) 28(2) 27(0) 22(4) 17(1) 8(2) 4(1)[10-6;10-5[87 87(0) 87(0) 85(2) 83(2) 74(6) 54(4) 31(3) 8(0)<10-6

N at risk(events)

06

1218

2430

3642

48

Months since randomization

MRD at pre-maintenance

IFM 2009 trial

P-value (trend) : p<0.0001

<10-6

[10-6;10-5[[10-5;10-4[

>=10-4

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Patie

nts

with

out p

rogr

essi

on (%

)

40 40(0) 40(0) 40(0) 33(6) 23(9) 15(6) 4(4) 2(1)[10-4;10-3[23 23(0) 23(0) 23(0) 22(1) 19(3) 14(2) 3(5) 2(0)[10-5;10-4[29 29(0) 29(0) 29(0) 28(0) 22(5) 16(3) 4(4) 1(1)[10-6;10-5[86 86(0) 86(0) 86(0) 86(0) 77(5) 61(3) 36(5) 10(0)<10-6

N at risk(events)

06

1218

2430

3642

48

Months since randomization

MRD at post-maintenance

IFM 2009 trial

P-value : p<0.0001

Negative (<10-6)

Positive

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Patie

nts

with

out p

rogr

essi

on (%

)

159 159(0) 147(12) 132(15) 120(11) 94(22) 67(11) 22(17) 6(4)MRD positive87 87(0) 87(0) 85(2) 83(2) 74(6) 54(4) 31(3) 8(0)MRD neg (<10-6)

N at risk(events)

06

1218

2430

3642

48

Months since randomization

MRD at pre-maintenance

IFM 2009 trialFCM Negative Patients

P-value : p<0.0001

Negative (<10-6)

Positive

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Patie

nts

with

out p

rogr

essi

on (%

)

92 92(0) 92(0) 92(0) 83(7) 64(17) 45(11) 11(13) 5(1)MRD positive86 86(0) 86(0) 86(0) 86(0) 77(5) 61(3) 36(5) 10(0)MRD neg (<10-6)

N at risk(events)

06

1218

2430

3642

48

Months since randomization

MRD at post-maintenance

IFM 2009 trialFCM Negative Patients

MRD by NGS vs MFC

NGS• Highly sensitive (> 10-6)• Standardized• Informative in 92% of pts• Frozen samples

FCM• Less sensitive (10-5)• Many different panels• Informative in 100% of pts• Fresh samples

Conclusions

The highest sensitivity is the most discriminant 10-6 is required

NGS is probably the technique of choice

MFC should be used in case of clone ID failure

MRD could (should?) be the main objective of future trials

MRD could identify cured patients

Conclusions

Limitations: MRD evaluates MM in one BM region

What about other regions?

Solutions:

• cfDNA sequencing analyses? ongoing• Imaging techniques: PET-TDM+++

Bartel et al. Blood 2009;114:2068-2076

Zamagni et al, Blood 2011;118:5989-5995

IFM 2009 trialPET‐CT normalisation before maintenance

Impact on PFS (62% normalised)

p < 0.001

69%

51.6%

Conclusions

New concepts in assessment of response in MM

IMWG consensus (Kumar S et al., Lancet Oncol, in press)

- MRD in BM with the most sensitive technique

- PET-TDM