Laura BiganzoliOncologia MedicaOspedale di Prato

Istituto Toscano Tumori

Le prospettive di ricerca

Established treatment option for advanced HR+HER2- breast cancer

Finn et al. N Engl J Med 2016 Di Leo et al. ESMO 2017Hortobagyi et al. N Engl J Med 2016

Cristofanilli et al. Lancet Oncol 2016

Sledge et al. J Clin Oncol 2017

PALBOCICLIB RIBOCICLIB ABEMACICLIB

Increased toxicity

Increased costs

Main Challenge

• Identification of the patients to be treated biomarkers of response/resistance

Part 1 PAL + LET(N=34)

LET(N=32)

Number of Events (%) 15 (44) 25 (78)

Median PFS, months(95% CI)

26.1(11.2, NR)

5.7(2.6, 10.5)

Hazard Ratio(95% CI)

0.299(0.156, 0.572)

p-value <0.0001

Part 2 PAL + LET(N=50)

LET(N=49)

Number of Events (%) 26 (52) 34 (69)

Median PFS, months(95% CI)

18.1(13.1, 27.5)

11.1(7.1, 16.4)

Hazard Ratio(95% CI)

0.508(0.303, 0.853)

p-value 0.0046

UNSELECTED (ER+/HER2 neg) CCD1 amplif. and/or p16 loss

Part 1 (N=66)

• Phase II, 1° line• ER+, HER2– BC status

Palbociclib 125 mg QD + Letrozole 2.5 mg QD

Letrozole 2.5 mg QDPart 2 (N=99)

• Same as part 1 but with CCND1 amplification and/or loss of p16

R

PALOMA 1- role of CCD1 and p16

0 4 8 12 16 20 24 28 32 36 40Time (Month)

0

10

20

30

40

50

60

70

80

90

100

Prog

ress

ion

Free

Sur

viva

l Pro

babi

lity

(%)

0 4 8 12 16 20 24 28 32 36 40Time (Month)

0

10

20

30

40

50

60

70

80

90

100

Prog

ress

ion

Free

Sur

viva

l Pro

babi

lity

(%)

Finn R. et al Lancet Oncology 2015; 16: 25–35

PIK3CA WT PIK3CA Mut

PIK3CA status (exon 9 and 20 hotspots) was determined by BEAMING assay on circulating DNA in 395 pts in PALOMA 3

PIK3CA status does not impact the magnitude of benefit from palbociclib

PIK3CA mutation status- PALOMA-3

Cristofanilli M, et al. Lancet Oncol. 2016;17(4):425–439.

Biomarkers of resistance● Genetic loss of RB1 may be a marker of primary resistance to

CDK4/6 inhibitors; uncommon in HR+ subtypes● Functional loss of the Rb pathway can be measured by gene-

expression studies● Gene-expression signatures focusing on inactivation of the Rb

pathway have been developed and characterized in breast cancer patients datasets

● The correlation between these signatures with response to CDK 4/6 inhibitors has not been explored

Thangavel C et al. Endocr Relat Cancer 2011

Rb loss signature in Luminal BC

E2F1 and E2F2 high vs low breast cancers in the TCGA*

RBsig 87 genes

Expression data

Genes correlated with E2F1 and E2F2

expression

Functional analysisAssociation with molecular subtypes*TCGA: The Cancer Genome Atlas, CCLE: Cancer Cell Line Encyclopedia

We have deriveda new signature of Rb

loss-of-function (RBsig) with the specific aim of

testing whether this might help in discriminating between palbociclibresistant vs sensitive

breast cancer cell lines

The RB Sig as a potential tool to predict primary resistance to CDK 4-6 inhibitors

Malorni et al. Oncotarget 2016

RBsig expression in BC subtypes

RBsig levels are higher basal BC and, among Luminal BC, are higher in LumB

Malorni et al. Oncotarget 2016

All tumors

p-value < 7e-32 p-value < 7e-11 p-value < 0.002

Luminal A/B Basal

Rb1 status Rb1 status Rb1 status

RBsig correlates with RB1 status in BC subtypes

RBsig levels are higher in BC samples with loss of Rb, across multiple BC subtypesMalorni et al. Oncotarget 2016

BC gene expression meta-dataset (N=3458)

Recurrence free survival (RFS) of pts with ER+ tumors, untreated or endocrine treated only

Prognostic value in BC patients

ER+, untreatedp= 2.22e-09HR=2.37 (1.8-3.2, p=1.87e-08)

Luminal A, untreatedp= 1.14e-11HR=3.34 (2.3-4.8, p=6.97e-10)

Luminal B, untreatedp= 0.0001HR=2.52 (1.55-4.08, p=0.0003)

200

RBsig

Does RBsig hold prognostic information in ER+ BC?

Malorni et al. Oncotarget 2016

Sensitive/Resistant info*

BC cell lines expression data from RNA-seq

experiment (GSE48213)

discriminate CDK4/6i sensitive vs resistant

BC cell lines

*Finn RS et al. BCR 2009; 11:R77

RBsig

RBsig identifies CDK4/6i resistant vs sensitive cell lines with and Area Under the

Curve (AUC) of 0,7778

Does RBsig predict response to CDK4/6 inhibitors?

Malorni et al. Oncotarget 2016

Possible role of CDK 4/6 inhibitors in HER2-positive BC

Rationale

• 50% of HER2+ BC co-express ER

• CDK4/6 pathway is downstream to both ER and HER2 pathways

• Combination of CDK4/6 inhibitors and endocrine therapies or anti-HER2 agents proved to be synergistic in vitro (Finn RS, et al. Breast Cancer Res 2009)

• CDK4/6 inhibitors have also shown activity in models of acquired resistance to endocrine or HER2-targeted therapies (Witkiewicz AK. et al. Genes & Cancer, 2014)

The NA-PHER2 StudyHPPF x 6 four-weekly cyclesHerceptin+pertuzumab+palbociclib+FLV Surgery

Pts with early/locally advanced HER2+ AND ER+ BC

H= trastuzumab 8mg/kg 6 mg/kg x 6P= pertuzumab 840 mg 420 mg x 6Palbociclib 125 mg/day x 21 days q 28 x 5FLV=fulvestrant 500 mg q 4 wks x 5

ITT population n=30

Gianni et al. SABCS 2016

pCR 30% in breast; 27% in breast & axilla

• De-escalation ie. take maximum benefit from targeted therapy avoiding chemotherapy is an extremely appealing concept in patients with triple positive breast cancer

• Chemotherapy+anti HER2 therapy = standard

BIOMARKERS

Meta-dataset of 10 neoadj. trials of CHT +/- anti HER2 therapy with GEP data RBsig was computed and the correlation with pCR was explored

(ER+ /HER2+ pts N= 211)

ttest: 0.09435

pCR

Non pCR

RBsig in ER+/HER2+ BC

ER+/ HER2+ CT+H (N=117) In ER+/HER2+ pts

treated with CHT+ anti-HER2:

RBsig LOW tumors had lower pCR

rates compared to RBsig HIGH

Risi E. et al SABCS 2016

Signature of RB deficiency have been shown to potentiallypredict response to neoadjuvant chemotherapy

Herschkowitz et al. Breast Cancer Res 2008; Ertel et al. Cell Cycle 2010

Hypothesis

RBsig may help selecting pts with ER+/HER2+ BC who could be spared CHT and treated with ET+ anti-HER2+ CDK4/6 inh

• Pts with RBsig LOW ET+ anti-HER2+ CDK4/6 inh. will be more active than CHT+ anti-HER2

• Pts with RBsig HIGHCHT+ anti-HER2 will be more active than ET+ anti-HER2+ CDK4/6 inh.

TOUCH: trial design

Stratification criteria:•G8 score (>14 vs ≤14)•N1 or T>5cm diameter vs. N0 and T 1-5 cm

Primary objective: To explore the interaction between the RBsig status (HIGH or LOW) and pCRPrimary endpoint: pCR (ypT0/ypTis ypN0)

Conclusions

• CDK4/6 inhibitors represent a new treatment option in ER+/HER2neg MBC

• Biomarkers for selecting patients more likely to benefit from CDK4/6 inhibition would be of great clinical utility to maximize benefit and containing costs

• Biomarkers would be of great utility also to develop new treatment strategies using multiple target blockades

Backup

Statistical assumptions

RBsig Prevalence

A: Palbociclib + letrozole +

trastuzumab

B: Paclitaxel + trastuzumab

Odds Ratio (A:B)

pCR rate

RBsig low 50% 30% 15% 2.429

RBsig high 50% 10% 50% 0.111

unselected 20% 32.5%

Hypothesized pCR rates in subgroups for sample size determination

An assessable sample size of 120 patients with successful RBsig results was determined to provide 86% power for the test of treatment-by-RBsig interaction (two-sided α=0.05).

The enrolled sample size is inflated by 20% to 144 patients to account for non-assessable RBsig status (which is determined after randomization).

RANDOMIZATION

Operable BC ER+/HER2+

Age ≥65 yo

Core biopsy mandatory

Palbociclib 125 mg/day x21 days q28x 4 Letrozole 2.5 mg/day orally x 16 weeksTrastuzumab 8 mg/kg 6 mg/kg q 3 wks x 5 Pertuzumab 840 mg 420 mg q 3 wks x 5

Surgery1:1

TOUCH: trial design

Stratification criteria:•G8 score (>14 vs ≤14)•N1 or T>5cm diameter vs. N0 and T 1-5 cm

Primary objective: To explore the interaction between the RBsig status (HIGH or LOW) and pCRPrimary endpoint: pCR (ypT0/ypTis ypN0)

Paclitaxel 80mg/m2 days 1,8,18 q 28 x4Trastuzumab 8 mg/kg 6 mg/kg q 3 wks x 5 Pertuzumab 840 mg 420 mg q 3 wks x 5

Acknowledgements

(MFAG 14371)

Translational Research Unit, Hospital of Prato

“Sandro Pitigliani” Medical Oncology Unit, Hospital of Prato

Functional Genomics & Bioinformatics Units, Proxenia S.r.l