imeneo study

International MEta-analysis

of circulating tumor cell detection

in early breast cancer pts

treated by NEOadjuvant chemotherapy (IMENEO study)

FC Bidard*, S Michiels, V Mueller, S Riethdorf, LJ Esserman, A Lucci, B Naume, J Horiguchi, R Gisbert-Criado, S Sleijfer, M Toi, JA Garcia-Saenz, A Hartkopf, D Generali, F Rothé, J Smerage, L Muinelo, J Stebbing, P Viens, M Magbanua, CS Hall, O Engebraaten, D Takata, J Vidal-Martínez, W Onstenk, N Fujisawa, E Diaz-Rubio, FA Taran, MR Cappelletti, M Ignatiadis, C Proudhon, D Wolf, J Bowman Bauldry, E Borgen, R Nagaoka, V Carañana, J Kraan, M Maestro, SY Brucker, K Weber, F Reyal, D Amara, MG Karhade, RR Mathiesen, H Tokiniwa, A Llombart-Cussac, K d'Hollander, P Cottu, JW Park, S Loibl, JY Pierga, K Pantel

* Medical Oncology, Institut Curie, Paris, France

This presentation is the intellectual property of the presenter. contact [email protected] for permission to reprint/ distribute.

San Antonio Breast Cancer Symposium – December 6-10, 2016



FC BidardResearch grant from Janssen Diagnostics to Institut Curie

K d'HollanderIDDI employee

Disclosures



Circulating Tumors Cells (CTCs)CTCs are the “seeds” of distant metastases

In metastatic BC patients- prognostic factor before & during therapy- dynamic changes of CTCs >> serum markersCristofanilli et al., N Engl J Med 2004; Bidard et al., Lancet Oncol 2014

In non-metastatic BC patients- prognostic factor before adjuvant therapyLucci et al., Lancet Oncol 2012; Rack et al., J Natl Cancer Inst 2014

CTCs can be counted with the FDA-cleared CellSearch® system



CTCs in neoadjuvant settingFirst study with CellSearch®, N=118 pts (REMAGUS02)

Numerous studies initiated worldwidevery few published / heterogeneous results

No significant association between CTC detection and pCR

Prognostic impact on:

Distant-Metastasis-Free SurvivalPierga et al., Clin Cancer Res 2008

Overall SurvivalBidard et al., Ann Oncol 2010



ObjectivesStudy objectives were pre-specified in a written protocol

Iry objective: Overall Survival (OS)

IIry objectives: - Association with baseline characteristics (includ. T4d vs non T4d)- Distant Disease-Free Survival (DDFS)- Locoregional Relapse-Free Interval (LRFI)- Association with pCR- Exploration of CTC-positivity thresholds- To measure the added value of CTC count to clinically-optimized prognostic models



Methods- literature & abstracts search up to Dec 2014- direct contact with all centers deemed to have eligible data:

CTC count by CellSearch® Early BC pts treated with neoadjuvant chemotherapy (NCT) Survival (published or not)

Non-overlapping CTC time points:○ [-5;0] weeks before NCT = baseline○ [1;8] weeks after start of NCT○ [-5;0] weeks before the surgery○ [1;52] weeks after surgery

StatisticsCox regression models (stratified by study) & landmark method Overfitting bias of multivariate prognostic models (used to report average increases in likelihood ratio) was limited by resampling procedures



Data collection

#2,000 potentially eligible pts from 18 centers

2 centersoff study

Letter of intent

call for data

2,239 pts data received

data cleaning 83 excluded

2,156 individual patients 21 studies 16 centers



N= 1,574 pts N= 290 pts

N= 1,200 ptsN= 285 pts

[-5;0]w before start of NCT[1;8w]w after start of NCT[-5;0]w before surgery[1;52]w after surgery

CTC data collected:

7.5ml of blood (1 tube) analyzedfor >95% of CTC points

Data collection

#2,000 potentially eligible pts from 18 centers

2 centersoff study

Letter of intent

call for data

2,239 pts data received

data cleaning 83 excluded

2,156 individual patients 21 studies 16 centers



N patients ≥1 CTC ≥2 CTC ≥5 CTC1574 25.2% 12.6% 5.9%Before NCT

CTC detection



N patients ≥1 CTC ≥2 CTC ≥5 CTC1574 25.2% 12.6% 5.9%

1200 15.1% 5.3% 1.0%

Before NCT

CTC detection

Before surgery

Decrease during NCT: p<.0001



N patients ≥1 CTC ≥2 CTC ≥5 CTC continuous1574 25.2% 12.6% 5.9%

cT size p<.0001 p<.0001 p<.0001 p<.0001cT1 122 (7.9%) 18.9% 8.2% 3.3%cT2 770 (49.8%) 22.3% 10.3% 3.5%cT3 343 (22.2%) 24.2% 12.2% 6.1%cT4a-c 108 (7.0%) 28.7% 16.7% 8.3%cT4d 204 (13.2%) 41.2% 24.5% 15.7%

cN status p=.051 p=.021 p=.009 p=.024cN0 656 (41.9%) 22.7% 10.4% 4.1%cN+ 911 (58.1%) 27.1% 14.4% 7.2%

Subgroup p=.23 p=.028 p=0.54 p=.12

HER2+ 365 (23.2%) 24.1% 11.0% 4.7%HR+ 800 (51.0%) 24.1% 11.5% 5.3%Triple Neg. 405 (25.8%) 28.4% 16.5% 8.4%

CTC detection & baseline characteristics

Before NCT



N patients ≥1 CTC ≥2 CTC ≥5 CTC continuous1370 22.9% 10.8% 4.4%

cT size p=.31 p=.15 p=.049 p=.21

cT1 122 (9.1%) 18.9% 8.2% 3.3%cT2 770 (56.9%) 22.3% 10.3% 3.5%cT3 343 (25.5%) 24.2% 12.2% 6.1%cT4a-c 108 (8.0%) 28.7% 16.7% 8.3%cT4d EXCLUDED

cN status p=.22 p=.19 p=.24 p=.17

cN0 604 (44.3%) 21.4% 9.6% 3.6%cN+ 760 (55.7%) 24.2% 12.0% 5.1%

Subgroup p=.44 p=.36 p=.13 p=.37

HER2+ 292 (21.4%) 20.2% 8.9% 2.4%HR+ 738 (54.0%) 23.2% 11.0% 5.1%Triple Neg. 336 (24.6%) 24.4% 12.5% 4.8%

CTC detection & baseline characteristics, T4d excluded

Before NCT



CTC detection: association with pCR

pCR was defined as ypT0/isN0 in 92.5% of patients (N=1916/2072)

pCR was observed in 24.3% of patients (N=503/2072)

N patients ≥1 CTC ≥2 CTC ≥5 CTC continuousp=.076 p=.65 p=.90 p=.10

pCR 374 (24.0%) 21.7% 12.0% 6.1%No pCR 1183 (76.0%) 26.3% 13.0% 5.9%

p=.45 p=.13 p=.53 p=.52pCR 300 (26.3%) 13.7% 7.0% 1.3%No pCR 841 (73.7%) 15.7% 4.6% 1.0%

CTC before NCT

CTC before surgery



N pts0 CTC 11751 CTC 1992 CTC 59

3-4 CTC 47≥ 5 CTC 93

CTC before NCT & Overall Survival



N pts % Events Hazard Ratio0 CTC 1175 9.8% 11 CTC 199 10.6% 1.09 [0.65-1.69]2 CTC 59 23.7% 2.63 [1.42-4.54]

3-4 CTC 47 29.8% 3.84 [2.08-6.66]≥ 5 CTC 93 46.2% 6.25 [4.34-9.09]

Stratified p value <.0001

months

CTC before NCT & Overall Survival

Same HR observed without T4d tumors No interaction found with tumor subtype



CTC before NCT & Distant Disease-Free Survival


3-4 CTC 47 38.3% 3.44 [1.96-5.55]≥ 5 CTC 93 58.1% 5.00 [3.57-7.14]


monthsSame HR observed without T4d tumors No interaction found with tumor subtype



CTC before NCT & Locoregional Relapse-Free Interval


3-4 CTC 47 4.3% 1.23 [0.20-4.00]≥ 5 CTC 93 22.6% 4.16 [2.32-6.66]

months


Same HR observed without T4d tumors No interaction found with tumor subtype



Overall clinical validity (threshold ≥2 CTC)Multivariate analysesTime point OS

HR p

CTC at baseline(landmark analysis)

4.19 [2.97-5.88]

<.0001

cT T3-T4T4d

1.49 2.94

.0023

cN cN1 1.65 .0045Subgroup HER2+

Triple Neg1.69 5.24

<.0001

pCR No 5.88 <.0001



Overall clinical validity (threshold ≥2 CTC)Multivariate analysesTime point OS DDFS LRFI

HR p HR p HR p

CTC at baseline(landmark analysis)

4.19 [2.97-5.88]

<.0001 3.79 [2.84-5.03]

<.0001 3.20[1.93-5.19]

<.0001

CTC [-5;0]w before surgery(landmark analysis)

2.56[1.45-4.23]

.0020 2.69[1.67-4.12]

<.0001 1.05[0.32-2.55]

.92



Changes during therapy; ≥2 CTC

Distant Disease-Free SurvivalLandmark analysis

Overall SurvivalLandmark analysis

N

<2 / <2 CTC 609

<2 / >2 CTC 21

>2 / <2 CTC 103

>2 / >2 CTC 11

N

<2 / <2 CTC 606

<2 / >2 CTC 21

>2 / <2 CTC 103

>2 / >2 CTC 10



Changes during therapy; ≥2 CTC

Distant Disease-Free SurvivalLandmark analysisStratified p value p<.0001

Overall SurvivalLandmark analysisStratified p value p<.0001

N HR

<2 / <2 CTC 609 1

<2 / >2 CTC 21 3.57>2 / <2 CTC 103 4.29>2 / >2 CTC 11 9.68

N HR

<2 / <2 CTC 606 1

<2 / >2 CTC 21 3.89>2 / <2 CTC 103 4.18>2 / >2 CTC 10 4.94



Added value to comprehensive prognostic models

Model (1) Model (2) Endpoint Average increase in Chi2 [95% CI] P value

Baseline

(resampling procedure)

Baseline + CTC baseline

OS

DDFS

LRFI

Baseline+ CTC baseline+ pCR(resampling & landmark)

Baseline+ CTC baseline+ pCR+ CTC before surgery

OS

DDFS

LRFI

Likelihood ratio tests



Added value to comprehensive prognostic models

Model (1) Model (2) Endpoint Average increase in Chi2 [95% CI] P value

Baseline

(resampling procedure)

Baseline + CTC baseline

OS 28.9 [13.3-44.1] <.0001

DDFS 38.7 [21.9-58.5] <.0001

LRFI 8.7 [2.4-19.2] .003

Baseline+ CTC baseline+ pCR(resampling & landmark)

Baseline+ CTC baseline+ pCR+ CTC before surgery

OS 3.1 [0.03-11.3] .07

DDFS 3.1 [0.02-9.7] .07

LRFI 0.9 [0.00-3.5] .34

Likelihood ratio tests



Conclusion

CTC number-dependent impact on OS, DDFS and LRFI

significant above ≥2 CTC/7.5ml

Post-neoadjuvant survival does not exclusively rely on breast cancer

characteristics & pCR

CTC complements (but not dupplicates) usual prognostic factors

Next steps: - clinical utility trials (e.g. post-neoadjuvant therapy)

- further biological characterization


S MichielsJY PiergaP CottuC ProudhonF ReyalP Viens

B NaumeE BorgenO EngebraatenRR Mathiesen

S SleijferJ KraanW Onstenk

D GeneraliMR Cappelletti

Thank you

K d'HollanderE Coart

K PantelA HartkopfS LoiblV MuellerS RiethdorfSY BruckerFA TaranK Weber

J HoriguchiM ToiN FujisawaR NagaokaD TakataH Tokiniwa

JA Garcia-SaenzR Gisbert-CriadoL MuineloV CarañanaE Diaz-RubioA Llombart-CussacM MaestroJ Vidal-Martínez

LJ EssermanA LucciJ SmerageD AmaraJ Bowman BauldryCS HallD HayesMG KarhadeM MagbanuaJW ParkD Wolf

M IgnatiadisF Rothé

J Stebbing

imeneo study

Health & Medicine