Can we predict it ? Late metastasis Joseph Gligorov MD, PhD ESO Advanced Breast Cancer Task Force APHPHUEPTenon, Paris InsDtut Universitaire de Cancérologie Université Pierre & Marie Curie, Sorbonne Universités
Nov 01, 2014
Can we predict it ?
Late metastasis
Joseph Gligorov MD, PhD ESO Advanced Breast Cancer Task Force
APHP-‐HUEP-‐Tenon, Paris InsDtut Universitaire de Cancérologie
Université Pierre & Marie Curie, Sorbonne Universités
Conflict of interest
• Amgen® • Genomic Health®
• Eisai® • Roche-‐Genentech® • Nanostring ® • Novar;s ®
Clinical trials support, advisory boards, speaker
DEFINITIONS AND IMPORTANCE
Defini;ons & importance • Late:
• Metastasis: . micro mets ? . macro mets ? . the way we detect the mets ?
• Importance: . Is it frequent ?
. Is it life threatening ? . Is breast cancer mortality s;ll important as a main risk of early death aNer a long period of follow-‐up ?
0 2 5 10 15 years ?
Defini;ons & importance • Late:
• Metastasis: . micro mets ? . macro mets ? . Clinical events . Local or distant mets ?
• Importance: . Let’s see the datass it frequent ?
. Is it life threatening ? . Is breast cancer mortality s;ll important as a main risk of early death aNer a long period of follow-‐up ?
0 2 5 10 15 years
FREQUENCY OF LATE METASTATIS
Influence of chemo and pN status
Influence on mortality
Conclusion 1 • According to EBCTCG publica;on
– Overall breast cancer popula;on treated in adjuvant seXng with tamoxifen present an absolute risk of late recurrence aNer 5 years between 10-‐15%
– This risk is higher in pN+ popula;on who do not receive chemotherapy (# 20%) compared to pN0 with or without chemotherapy (7 to 8%)
• The influence of adjuvant tamoxifen – is s;ll clear between 5 to 10 years in ER & PR posi;ve disease – Is less clear aNer 5 years in PR &/or ER poor disease
• Risk of late recurrence clearly impact mortalit
MORE RECENT DATA
Influence of histological type
Pestalozzi et al., J Clin Oncol 26:3006-3014.2008
Prat A, Perou CM. Molecular Oncology 2011
Clinical-‐pathological characterisDcs of the current intrinsic subtypes of breast cancer
Informa;ons from neoadjuvant data
Luminal A Luminal B/HER2-‐ Luminal B/HER2+++
HER2+++ non luminal Triple nega;ve ?
Rates of distant recurrences and breast-‐specific survival in TNBC and other breast cancers.
RFS OS
Dent R et al. Clin Canc Res 2007
HR+ HR-‐
Time to distant recurrence Smoothed hazard rate curves for risk of recurrence
Clearly in:
-‐pN+ populaDon,
-‐ premenopausal women at Dme of tamoxifen
ITT : intent-‐to-‐treat COX : cox regression model IPCW : inverse probability of censoring weighted SCC : Shao, Chang, Chow model
Conclusion 2 • Histological type might influence the risk of late relapse
• Intrinsic subtypes might also influence the risk of late relapse and par;cularly according to the efficacy of systemic treatments – Chemotherapy for TNBC – an;HER2 treatments for HER2 posi;ve BC
• the main popula;on for which the iden;fica;on of a late risk of relapse remains the most important is the HR posi;ve popula;on
• We have possible treatment op;ons to propose to the pa;ents
CAN WE PREDICT LATE METASTASIS ?
New considera;ons
• Target popula;on is HR posi;ve popula;on
• Predic;ng late metastasis: 2 informa;ons – Prognos;c: improving OS
– Predic;on: • defining popula;on s;ll sensi;ve to endocrine treatment
• Trying to find new targets to prevent late relapse
ROR score was calculated using the test variables that include: • Pearson correlaDons with prototypical gene expression profiles for the four intrinsic Subtypes • ProliferaDon score • Pathologic tumor size
OP pN0 pN+
transATAC study: PAM50 vs Oncotype Dx vs IHC4 (RE, RP, HER2,Ki67)
Clinical variables
Sestak I et al, JNCI 2013
Trans ATAC & ABCSG-8 Distant recurrence – post 5 years
San Antonio Breast Cancer – Cancer Therapy and Research Center at UT Health Science Center – December 10-14, 2013
Courtesy of Sestak I et al.
Cha
nge
in L
Rχ2
Sta
tistic
94.1
67.9
0 10 20 30 40 50 60 70 80 90
100
CTS ROR Univariate Multivariate
HR (95% CI ) for IQR
Univariate
CTS Nodal status, grade, tumour size, age, treatment
1.96 (1.73 - 2.21) ROR score (PAM 50) 2.69 (2.12 - 3.43)
61.4
35.3
Multivariate*
1.80 (1.57 - 2.06) 2.07 (1.63 - 2.64)
*When added to other score
10 20 30 40 50 60 70 80 90
100
0
Luminal A vs Luminal B according to PAM50
San Antonio Breast Cancer – Cancer Therapy and Research Center at UT Health Science Center – December 10-14, 2013
HR (95% CI) P-value
Luminal A (N=1530 (71.6%)) - -
Luminal B (N=542 (25.4%)) 2.89 (2.07- 4.02) <0.0001
0 5
10
15
5 6 7 8 9 10 Follow-up time [years]
Luminal B Luminal A
Dis
tant
recu
rren
ce (%
)
4.1%
12.9%
0 5
10
15
Courtesy of Sestak I et al.
E-module in Oncotype Dx is predictive of late reccurence
Among women with tumours most sensitive to oestrogen, with a high E-module score, the recurrence rate more than doubled from 5.7% in the first five years to 13.6% in the subsequent five years. However, if they had a low E-module score, there was little difference in recurrence rates between the first five years and the next five years: 10.3% versus 12.3%.”
Dowsett M et al. EBCC 2014
Breast Cancer Index: • Molecular Grade Index • HOXB13/IL17BR
The predic;ve value of the marker seems depedent of the previous treatment exposure …
Sgroi et al. Lancet Oncol 2013
Endopredict clinicalscore was calculated using: • ProliferaDon associated genes • HR associated genes • Clinical parameters (tumour size & nodal status)
Dubsky P et al. BJC 2013
ER pathway seems to be the “driver”…
Mikempergher L et al. Mol Oncology 2013
Conclusion 3
• Clinical parameters s;ll remains crucial for evalua;ng the risk of late relapse (pN, pT)
• ER pathway ac;va;on seems to be crucial also and maight help to “predict” the benefit of prolonged endocrine treatment in popula;ons at risk of late relapse
• New approaches and signatures might help us to find new tools in pa;ent at risk of late relapse and not candidate for prolonged endocrine treatment
THANKS