Recurrent hyperactive ESR1 fusion proteins in endocrine therapy-resistant breast cancer

Background

Estrogen receptor-positive (ER-positive) metastatic breast cancer is often intractable due to endocrine therapy resistance. Although ESR1 promoter switching events have been associated with endocrine-therapy resistance, recurrent ESR1 fusion proteins have yet to be identified in advanced breast cancer.

Patients and methods

To identify genomic structural rearrangements (REs) including gene fusions in acquired resistance, we undertook a multimodal sequencing effort in three breast cancer patient cohorts: (i) mate-pair and/or RNAseq in 6 patient-matched primary-metastatic tumors and 51 metastases, (ii) high coverage (>500×) comprehensive genomic profiling of 287-395 cancer-related genes across 9542 solid tumors (5216 from metastatic disease), and (iii) ultra-high coverage (>5000×) genomic profiling of 62 cancer-related genes in 254 ctDNA samples. In addition to traditional gene fusion detection methods (i.e. discordant reads, split reads), ESR1 REs were detected from targeted sequencing data by applying a novel algorithm (copyshift) that identifies major copy number shifts at rearrangement hotspots.

Results

We identify 88 ESR1 REs across 83 unique patients with direct confirmation of 9 ESR1 fusion proteins (including 2 via immunoblot). ESR1 REs are highly enriched in ER-positive, metastatic disease and co-occur with known ESR1 missense alterations, suggestive of polyclonal resistance. Importantly, all fusions result from a breakpoint in or near ESR1 intron 6 and therefore lack an intact ligand binding domain (LBD). In vitro characterization of three fusions reveals ligand-independence and hyperactivity dependent upon the 3' partner gene. Our lower-bound estimate of ESR1 fusions is at least 1% of metastatic solid breast cancers, the prevalence in ctDNA is at least 10× enriched. We postulate this enrichment may represent secondary resistance to more aggressive endocrine therapies applied to patients with ESR1 LBD missense alterations.

Conclusions

Collectively, these data indicate that N-terminal ESR1 fusions involving exons 6-7 are a recurrent driver of endocrine therapy resistance and are impervious to ER-targeted therapies.

Abstract P2-14-01: The impact of local therapy on locoregional recurrence in women with high risk breast cancer in the neoadjuvant I-SPY2 TRIAL

Background: In women with breast cancer receiving neoadjuvant chemotherapy, residual cancer burden (RCB) predicts distant recurrence and survival. In those with high risk tumors, locoregional recurrence (LRR) remains a concern, and has been associated with type of local therapy received. We evaluated the impact of local therapy on LRR in the ISPY-2 TRIAL.

Methods: Data were analyzed in Stata 14.2, using Chi2 test, log rank test, and a Cox proportional hazards model. RCB was considered a categorical variable (0/1 versus 2/3), as described in prior publications. Breast surgery categories were lumpectomy +/- radiotherapy, or mastectomy +/- radiotherapy. Axillary surgery was defined as sentinel lymph node (SLN) surgery (≤6 nodes removed) or axillary dissection (>6 nodes).

Results: Follow up data from the I-SPY2 TRIAL were available for 630 patients (median follow up 2.76 yrs, range 0.4-7.2). Type of local therapy was significantly associated with clinical stage at presentation, with stage III patients most frequently undergoing mastectomy + radiation (p<0.001). Women with higher RCB were more likely to undergo mastectomy than those with lower RCB (61.3% vs 48.8% mastectomy rate, p=0.002), and more likely to receive adjuvant radiotherapy (62.0% vs 53.9%, p=0.048). There was no association between clinical stage, type of surgery, or radiotherapy and LRR (Table). Higher RCB was significantly associated with LRR, with 3 year locoregional recurrence free rate of 95.1% in RCB 0/1 versus 89.9% in RCB 2/3 (p=0.003).

In a Cox model adjusting for clinical stage, tumor subtype, surgical therapy, RCB status, nodal radiation, and age, significant predictors for LRR were tumor subtype and RCB status. Hazard ratio (HR) for LRR in those with RCB 0/1 was 0.39 compared to those with RCB 2/3 (95% CI 0.17-0.87, p=0.021). There was no difference in LRR between breast conservation and mastectomy; within the breast conservation group, those who had lumpectomy alone had higher hazard of LRR compared to those having lumpectomy + radiation (HR 3.1, 95% CI 1.1-9.2, p=0.043).

Conclusions: Extent of surgical therapy was not associated with local tumor control, regardless of advanced tumor stage at presentation. Rather, tumor biology and response to therapy were the best predictors of LRR. These data highlight the opportunity to minimize the morbidity of extensive surgical therapy for patients with excellent response to systemic therapy.

LRR rates by clinical features and treatment status FrequencyLRR RateP valueClinical Stage  0.5I240 (47.5%)5.8% II185 (36.6%)8.7% III80 (15.8%)6.3% Tumor Subtype  0.014ER+PR+Her2-161 (26.4%)3.1% ER+PR-Her2-56 (9.2%)3.6% Her2+176 (28.9%)6.3% Triple negative216 (35.5%)11.1% Local therapy  0.169Lumpectomy85 (13.5%)11.8% Lumpectomy with radiation198 (31.4%)5.6% Mastectomy173 (27.5%)5.2% Mastectomy with radiation174 (27.6%)8.6% Axillary surgery  0.23None5 (0.8%)20% SLN329 (52.2%)5.8% ALND296 (47%)8.5% Axillary radiation  0.535Yes42 (6.7%)9.5% No588 (93.3%)7.0% Axillary management  0.2No surgery or radiation5 (0.8%)20.0% SLN312 (50%)5.3% SLN+Axillary radiation17 (2.7%)8.3% ALND271 (43%)10.3% ALND+Axillary radiation25 (4%)5.4% RCB  0.0020/1293 (50.1%)3.8% 2/3292 (49.9%)10.3% 

Citation Format: Silverstein J, Suleiman L, Yau C, Price ER, Singhrao R, Yee D, DeMichele A, Isaacs C, Albain KS, Chien AJ, Forero-Torres A, Wallace AM, Pusztai L, Ellis ED, Elias AD, Lang JE, Lu J, Han HS, Clark AS, Korde L, Nanda R, Northfelt DW, Khan QJ, Viscusi RK, Euhus DM, Edmiston KK, Chui SY, Kemmer K, Wood WC, Park JW, Liu MC, Olopade O, Leyland-Jones B, Tripathy D, Moulder SL, Rugo HS, Schwab R, Lo S, Helsten T, Beckwith H, I-SPY 2 TRIAL Consortium, Berry DA, Asare SM, Esserman LJ, Boughey JC, Mukhtar RA. The impact of local therapy on locoregional recurrence in women with high risk breast cancer in the neoadjuvant I-SPY2 TRIAL [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P2-14-01.

Abstract P2-03-11: Genetic background determines the algorithm of effectiveness of targeted drugs of RAS and PI3K pathways in TNBC: Testing a combination of MEK 1/2 inhibitor with mTOR kinase inhibitor or AKT inhibitor

Introduction: Most TNBC patients with single-agent targeted therapy in multiple clinical trials develop resistance, leading to disease progression, posing a major challenge in clinical management of mBRCA. In TNBC patients there remains a lack of obvious predictive biomarkers to guide targeted therapy. Although signaling mechanisms of resistance may be either intrinsic or acquired, the role of driver pathways are known be associated with oncogenic evolution of tumor cells to a resistant state. As the PI3K-AKT-mTOR1/2 (PAM) pathway is activated in TNBC (due to alterations in EGFR, 50%; PTEN, 40%; and PIK3CA, 5-10%), the PAM pathway is a candidate for potential molecular targeting of anti-cancer therapeutics. Despite the involvement of PI3K pathway activation in TNBC tumors, single agent PI3K inhibitors show modest clinical activity. Literature references suggest extensive cross-talk exists between the (PAM) and the RAS-RAF-MEK-ERK (MAPK) pathways. Here we tested the efficacy of a combination of MEK 1/2i (AZD6244, GDC0973) with mTOR kinasei (AZD2014, TAK228) or AKTi (AKT5363) in TNBC models. Methods: TNBC cell lines with alterations in KRAS/RAF (MDA-MB231), BRCA-deficient PTEN-loss (SUM149), PTEN-loss (MDA-MB468) and PIK3CA mutation (BT20) were used. Proliferation, apoptosis, cell cycle and 3D clonogenic growth were tested following drug treatment alone or in combination. Signaling events in the respective pathways following drug treatment were interrogated by WB. Results: MEKi led to a brief proliferation inhibition in MDA-MB231 cells. A combination of MEKi and mTORi was additive in decreasing short term proliferation and 10 day 3D growth. AKTi alone had limited effect on 3D growth but in combination with MEKi was profoundly inhibitory. A limited effect of MEKi was observed in BT20 cells; blocking 2D and 3D growth, while a profound effect of MEKi was observed in combination with mTORi (to a lesser extent with AKTi). In MDA-MB468 cells, MEKi did not have an initial appreciable effect (proliferation and G1 arrest), though 3D growth was significantly reduced at 10 days following drug alone or in combination with AKTi. SUM149 cells had G1 arrest with no appreciable change in 3D growth following MEKi. In contrast, single agent mTORi and AKTi reduced 3D growth. MEKi and mTORi doublet slowed proliferation, increased apoptosis and disrupted 3D growth. No single agent treatment with AKTi or mTORi disrupted colony formation however either agent in combination with MEKi blocked 3D growth. Single agents treatments were mostly cytostatic with no increase in apoptosis while a doublet of MEKi plus mTORi induced significant apoptosis.A combination of MEKi with PI3K pathway specific inhibitor significantly blocked phosphrylation of downstream effector molecules. Upregulation of ERK signals following PAM pathway inhibitors were abrogated by prior addition of MEKi. Summary: Combined blockade of the PAM and MAPK pathways (AKTi or mTORi) plus MEK1/2 inhibitor were more effective in attenuating molecular signals. This combination showed enhanced efficacy in TNBC cell models with specific PAM and or MAPK pathway alterations.

Citation Format: Carlson JH, De P, Dey N, Leyland-Jones B. Genetic background determines the algorithm of effectiveness of targeted drugs of RAS and PI3K pathways in TNBC: Testing a combination of MEK 1/2 inhibitor with mTOR kinase inhibitor or AKT inhibitor [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P2-03-11.

Abstract P2-03-06: PTEN expression at the nexus of oncogenic signals in TNBC: Testing combination of p110beta-isoform-specific inhibitor with five PARP inhibitors

Background: TNBC is the most aggressive form of BRCA-associated BC subtype. The loss of PTEN is a common “first event” associated with basal-like subtype (Martins et al., 2012) and this mode of PI3K-pathway activation (deletion/mutation/loss of PTEN) occurs more frequently (35%) than PIK3CA (5-10%) mutations in TNBC (Ellis & Perou, 2013). Several reports suggest upregulation of PI3K/AKT mediated by PTEN loss depends on the PI3Kbeta-isoform. PARP1 is identified as a target of BRCA-defined cancers. Inhibitors of PARP1 have been recently approved by the FDA as targeted agents in BRCA-defined breast cancers. Failure to repair damaged DNA upon PARP inhibition causes accumulation of DNA double-strand breaks and leads to apoptosis of cancer cells. Also, nuclear PTEN controls DNA repair (Bassi et al., 2013). Here we hypothesized that a combination of PARP inhibitor with p110beta-isoform-specific inhibitor would sensitize the effect of PARP inhibitor(s) in PTEN-deprived TNBC model. Methods: We tested five PARP inhibitors, Talazoparib (BMN673, B), Niraparib (N), Olaparib (O), Rucaparib (R) and Veliparib (V) in combination with p110beta specific inhibitor, AZD6482. Four PTEN-null TNBC cell lines (BRCA WT/null), MDA-MB468, HCC70 (p.F90fs*9), BT549 (p.V275fs*1) and SUM149 cell lines were used for the study. Proliferative, apoptotic and PARylation signals following drug combinations were demonstrated by WB in a dose and time-dependent manner. Pro-apoptotic and anti-proliferative effects were verified using complementary 3D ON-TOP assay, real-time proliferation (Incucyte), AnnexinV and cl-caspase3 analyses. As an internal control, we also compared anti-proliferative signals of GDC-0032 (p110 beta sparing inhibitor) and GDC-0941 (pan PI3K inhibitor) with that of AZD6482 at 3/6 hours in MDA-MB468 and SUM149 cells. Mode of apoptosis was tested using triple fluorescence staining in live cells. Results: A dose of 500 nM BMN673 alone was effective in slowing cell proliferation and induced apoptosis. AZD6482 (5-10 uM) did not have anti-proliferative or pro-apoptotic effects at 48 and 72 hours. When compared, combinations of different PARPi (100nM of B, 1uM of N, 10uM of O, 10uM of R and 10uM of V) with AZD6482 abrogated 3D growth in BT549 and MDA-MB468 TNBC cells in a time-dependent manner. The effect of PARP inhibitor was tested by PAR signals which were abrogated either alone or in combination with carboplatin in both BRCA1/2 WT and mutated cells. The most pronounced anti-tumor effect was observed with the combination of B and AZD6482 which was mechanistically explained by the robust increase of AnnexinV positive cells following a single 500nM dose of B at both 48 and 72 hours. Pan PI3K inhibitor, GDC-0941, and AZD6482 blocked the activation of PI3K and its downstream effectors in contrast to p110 alpha-specific inhibitor, GDC-0032 in MDA-MB468 and SUM149 cells. Summary: We demonstrated a remarkable sensitivity of tumor cells to PARP inhibition in PTEN-defined TNBC models and identified PTEN-nullness as a potential predictive biomarker for a possible co-targeting of the PI3K pathway to further sensitize TNBC to PARP inhibitors.

Citation Format: Dey N, Carlson JH, De P, Leyland-Jones B. PTEN expression at the nexus of oncogenic signals in TNBC: Testing combination of p110beta-isoform-specific inhibitor with five PARP inhibitors [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P2-03-06.

Abstract P2-03-08: Preclinical evaluation of the PI3Kα/δ inhibitor, copanlisib in HER2+ breast cancer: A proof of concept study

Purpose: The PI3K-AKT-mTORC1/C2 pathway is frequently activated in HER2+ breast cancer and upregulation of this pathway is a key mechanism of trastuzumab resistance. However, attempts to indirectly target this pathway by using the allosteric mTOR inhibitor everolimus have had limited clinical success. Here, we present the results of a preclinical study of the PI3K α/δ (dominant) inhibitor copanlisib alone and in combination with T-DM1, in HER2 amplified cell lines including a model with acquired resistance to trastuzumab.

Method: Anti-proliferative, apoptotic, cell cycle, and intracellular signaling effects of copanlisib alone and in combination with T-DM1 were evaluated in a panel of HER2 amplified (ER+ or ER-), and HER2 amplified/PIK3CA mutated cell lines, as well as trastuzumab-resistant breast cancer cell lines.

Results: 1) Copanlisib inhibited PI3K, mTOR and their downstream signaling molecules in HER2 amplified/PIK3CA WT or PIK3CA mutated as well as in trastuzumab-resistant cell lines, 2) interestingly, copanlisib also inhibited RAS-MAPK signaling in the earlier time points, 3) copanlisib caused a strong differential growth inhibition in HER2 amplified BC cell lines by 3D-ON-TOP clonogenic assay and real-time monitoring in an IncuCyte Zoom. Inhibition was greater when copanlisib was combined with T-DM1, 4) administration of copanlisib induced cell cycle G0/G1 arrest and resulted in increased apoptosis in a dose-dependent manner, 5) unlike everolimus copanlisib blocked HIF1α accumulation in hypoxic condition and this blocking effect was reversed by prior treatment with the proteasome inhibitor, carfilzomib and 6) copanlisib also attenuated HER2 amplified cell migration, an important phenotypic feature for metastasis.

Conclusions: Copanlisib is highly effective (blocks proliferation, induces apoptosis, and inhibits PI3K and its downstream signaling targets) in HER2 amplified breast cancer cell lines including trastuzumab-resistant and PIK3CA mutated cell lines. The addition of copanlisib to T-DM1 might represent an improved treatment strategy for patients with refractory metastatic HER2+ breast cancer.

Citation Format: De P, Carlson JH, Dey N, Leyland-Jones B. Preclinical evaluation of the PI3Kα/δ inhibitor, copanlisib in HER2+ breast cancer: A proof of concept study [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P2-03-08.

Abstract P1-03-04: Role of PTEN and BRCA1 as determinants of synergy for the combination of vistusertib with carboplatin and olaparib in TNBC

Introduction: Platinum agents are being used in combination with targeted agents in advanced triple-negative breast cancer (TNBC) (See K. Gelmon et al., 2012). Additionally inhibition of PARP is also being considered as a “targeted” therapy for TNBC (Anders CK et al., 2010). PARP inhibitor (i), Lynparza (olaparib,AstraZeneca)met the primary endpoint of a Phase III trial in which Lynparza was compared to physician's choice of a standard of care chemotherapy in patients with HER2-negative metastatic BC harboring germline BRCA1/2 mutations (BRCAm). Based on cBioportal data analyses and experimental studies we and others have reported that more than 30% PTEN loss in TNBC leads to activation/upregulation of the PI3K pathway (Nature. 2012; Ellis and Perou, 2013; Dey et al., 2012; De et al., 2016; Reed and Shokat 2017;). In line with active kinase profiling, genetic and pharmacological data which defined mTOR as an important target in TNBC (Montero JC et al., 2012), we have demonstrated that mTORi has anti-tumor activity in TNBC (De et al., 2014). Aim: These studies focused on exploring the synergy of biology-based targeted drugs PARPi (olaparib, O), mTOR kinase (vistusertib, V), and platinum (C) in TNBC models. Method: TNBC cells of multiple genetic backgrounds were used to test the combination(s) on proliferation and apoptosis by monitoring growth and using real time Annexin V reagent in a microscopy-based assay (Essen IncuCyte Zoom). Flow cytometric analysis of cell cycle progression by PI staining was also used. Long term clonogenic 3D growth was monitored in matrigel. Results: In BRCAm/PTEN null Sum149 and HCC1937 cells and BRCA wild type (wt)/PTEN null MDA-MB-468 the addition of (V) to (O) and (C) enhanced apoptosis induction and further slowed growth. In Sum149 cells, single agent V treatment induced G1 arrest while O plus C or the triple combination increased S phase accumulation. In MDA-MB-468 cells G1 arrest was seen with V alone and in the triplet. In BRCA wt/PTEN null HCC70 cells V decreased cell proliferation and induced G1 arrest. In the HCC70 model, the addition of O plus C did not synergize with V. In BT20, a BRCA wt/PTEN wt but PI3KCA mutant cell line, no effect on proliferation or apoptosis was seen in the O plus C treated arms. V slowed cell proliferation and increased G1 arrest in a dose-dependent manner. As expected in a BRCA wt/PTEN wt, but RAS active mutant cell line MDA-MB-231 this treatment combination was not effective and was used as an internal negative control. Based on ratios of the normalized slopes of proliferation curves (for triplet), the cells were graded in terms of synergy as SUM149>MDA-MB-468>HCC1937 and HCC70>BT20>MDA-MB231. Treatment with the triplet had the largest effect on reducing 3D colony formation and size as compared to control over single or double treatment. Summary: Here, we present the effect of the combination of vistusertib with olaparib plus carboplatin in several TNBC models. Our data demonstrate that increased effectiveness of the triple combination is seen in cells harboring BRCA1 and PTEN-null mutations. The mechanistic role of these two targets on determining this synergy is being worked out and will be presented at the meeting.

Citation Format: Carlson JH, O'Connor MJ, De P, Dey N, Leyland-Jones B. Role of PTEN and BRCA1 as determinants of synergy for the combination of vistusertib with carboplatin and olaparib in TNBC [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr P1-03-04.

Abstract P2-09-02: Blinded molecular subtyping analysis from RNA-Seq of FFPE samples in the GeparQuinto trial reveals predictive value of VEGFA metagene for bevacizumab treatment

Background:

RNA-Seq from total RNA in FFPE tissue can be more challenging due to limited capture of partially degraded RNA. Exome-capture based RNA-Seq may circumvent such problems and allow reproducible complete molecular characterization of low-quality RNA from small clinical samples.

Methods:

HER2 negative patients within the GeparQuinto trial were treated with neoadjuvant anthracycline-taxane-based chemotherapy +/- bevacizumab. Patients with bevacizumab therapy had a significantly higher pCR rate, especially within the triple negative subgroup. We performed exome-capture RNA-Seq on 5µm FFPE sections from pre-therapeutic cores of 400 HER2 negative samples from this trial. In a prospectively planned, blinded study we correlated molecular subtypes and metagenes for proliferation, stroma, MHC2, and VEGFA with clinical and histopathological data. Molecular subtypes were defined using the AIMS methods. Metagenes were calculated as mean values corresponding to previously described gene clusters after platform transfer (Rody et al. 2011 PMID 21978456, Hu et al. 2009 PMID 19291283) and then z-transformed.

Results:

296 samples with RNA-Seq data were classified as either of high (n=226) or of limited quality (n=70). For 22 samples RNA yield was insufficient and 82 did not pass initial QC. 121 (41%), 63 (21%), 34 (11.5%), 46 (15.5%), and 32 (11%) samples were defined as basal-like, HER2-enriched, luminal A, luminal B, and normal-like, respectively. Subtyping was robust with regard to gene filtering, normalization, and sample quality. ER and PR status from local IHC strongly correlated with gene expression (overall correctness 84% and 80% for ER, and 85% and 74% for PR, in samples with high and limited quality, respectively) and luminal subtypes (95% ER positive). Proliferation metagene correlated with histological grade (median -0.73, -0.39, and 0.53 in G1, G2, and G3, respectively; P<0.001) and MHC2 metagene correlated strongly with TIL counts (Rho=0.53, P<0.001). Among the high quality samples response rates (49.3% pCR overall) differed significantly by subtype, with higher pCR rates in basal-like (68.9%) and HER2-enriched (45.5%) than in luminal B (35.7%), luminal A (17.9%), and normal-like (20.0%). MHC2- (OR 1.60, 95%CI 1.21-2.12, P=0.001), proliferation- (OR 2.88, 95%CI 2.00-4.16, P<0.001), and VEGFA-metagenes (OR 1.92, 95%CI 1.41-2.60, P<0.001) were significant predictors for pCR. In a multivariate logistic regression (adjusted for bevacizumab treatment and hormone receptor status) both VEGFA metagene (OR 2.59, 95%CI 1.40-4.77, P=0.002) and the interaction between the VEGFA-metagene and bevacizumab treatment arm (P=0.023) significantly predicted pCR.

Conclusions:

Exome-capture RNA-Seq allows robust genomic characterization of clinical samples with limited FFPE material from core biopsies, and molecular subtypes and immune metagenes are predictive for pCR. The VEGFA metagene is a specific predictor for response to neoadjuvant bevacizumab treatment.

Citation Format: Karn T, Meissner T, Weber K, Sinn B, Denkert C, Budczies J, Nekljudova V, Fasching PA, Holtrich U, Schem C, Solbach C, Hartmann A, Röcken C, Untch M, Young BM, Willis S, Leyland-Jones B, von Minckwitz G, Loibl S. Blinded molecular subtyping analysis from RNA-Seq of FFPE samples in the GeparQuinto trial reveals predictive value of VEGFA metagene for bevacizumab treatment [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr P2-09-02.

Abstract PD8-10: APOBEC mutation signature in breast cancer correlates with tumor mutation burden and poor responses to therapy

Introduction

Mutational processes can be characterized by unique combinations of mutation types in the form of mutational signatures and have been associated with age, known mutagenic exposures, defects in DNA maintenance, or the APOBEC family of cytidine deaminases. We asked whether mutation signatures could be extracted from DNA sequence information in a targeted 434 gene panel covering 297 breast cancer specimens.

Materials and Methods

Targeted whole exome sequencing (Illumina, 2x50bp) of a 434 gene panel was performed on a set of 297 primary and metastatic breast tumor samples. Tissue of origin included breast (56%), liver (15%), lymph node (10%), lung (3%) and others (16%). Alignment was done with BWA against the human reference hg19 and variant calling was performed using VarDict. Germline variants were filtered based on allele frequencies, cohort specific population frequencies, as well as using 1000 Genomes and ExAC population frequencies. For somatic signature inference, only single nucleotide variants were retained. Panel specific trinucleotide frequencies were computed and normalized towards whole genome frequencies and somatic signatures were inferred using deconstructSigs method.

Results

We identified a total of 26 signatures from the set of 30 known signatures in our patient samples. Due to the small panel size, there was only a limited number of mutations available per patient to infer somatic signatures. On average, we identified two somatic signatures per sample. Most common mutation signatures identified were: Signature 1 (90.8%) - result of an endogenous mutational process initiated by spontaneous deamination of 5-methylcytosine; Signature 6 (21.8%) - defective DNA mismatch repair; Signature 15 (15.6%) - defective DNA mismatch repair; Signatue 7 (9.9%) - ultraviolet light exposure; and Signature 10 (6.5%) - altered activity of POLE. An APOBEC specific signature was identified in 20 (7%) samples. APOBEC positive samples showed significantly higher tumor mutational burden (10.7 vs. 5.7 mutations/mb) as compared to APOBEC negative samples (p<=0.001). PIK3CA was found to be mutated in 80% of APOBEC positive samples, compared to 36% of APOBEC negative samples. In addition, we found higher rates of mutations in TP53 (70% vs. 50%), MLL3 (50% vs. 19%) and MLL2 (25% vs 14%) of APOBEC positive patients. Response rates of APOBEC positive patients were significantly worse than of APOBEC negative patients, with 50 percent of patients having progressive disease compared to 25 percent of APOBEC negative patients(p=0.07, borderline).

Conclusions

We demonstrate the feasibility of a targeted sequencing approach to extract somatic mutation signatures from breast tumor samples, and we highlight the potential of using the APOBEC signature to predict therapeutic responses.

Citation Format: Meissner T, Amallraja A, Willis S, Harris R, Leyland-Jones B, Williams C. APOBEC mutation signature in breast cancer correlates with tumor mutation burden and poor responses to therapy [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr PD8-10.

Abstract P1-06-08: Independent validation of EarlyR gene signature in E2197: A randomized clinical trial comparing doxorubicin plus docetaxel to doxorubicin plus cyclophosphamide as adjuvant chemotherapy in breast cancer

Background: EarlyR is a prognostic gene signature score in ER+ breast cancer (BC) computed from the expression values of ESPL1, SPAG5, MKI67, PLK1 and PGR using a nonlinear mathematical formula. EarlyR has been validated in multiple cohorts profiled on Affymetrix and Illumina microarrays and by RNA-seq. This study sought to assess the prognostic features of EarlyR in a cohort of E2197.

Patients and Methods: Illumina DASL assay was used to measure gene expression in FFPE tissue of primary BC from a case-cohort sampling subset of women in E2197 treated with doxorubicin plus docetaxel (AT) or doxorubicin plus cyclophosphamide (AC). ER+ patients received hormone therapy at physician's discretion. After 79.5 months median follow-up, disease-free survival was 85% in both treatment arms. Among patients centrally reviewed with sufficient RNA material for the DASL assay, 319 with ER+ status and assessed for EarlyR are included in the analytic cohort. EarlyR scores and pre-specified risk strata (≤25=low, 26-75=intermediate, >75=high) were computed, while blinded to clinical data. The analysis endpoint was disease-free survival (DFS), defined as the time from randomization to date of invasive BC recurrence or death from any cause within 8 years. Weighted Cox proportional hazards models were used to associate EarlyR score or risk strata with DFS. Variances of the estimated coefficients were adjusted to account for the case-cohort design.

Results: The distribution of the EarlyR risk groups was 59% low, 11% intermediate and 30% high risk in this ER+ cohort. The continuous EarlyR score was significantly prognostic of DFS up to 8 years after randomization (p = 0.02). Patients with low EarlyR score (≤ 25) had significantly lower risk of BC recurrence within 8 years (p = 0.031, univariate HR=0.562, 95%CI: 0.334-0.948) compared to those with high EarlyR score (> 75). Analysis within the AC arm showed that patients with low EarlyR score had significantly lower risk of 8-year BC recurrence (p = 0.023, univariate HR=0.392, 95%CI: 0.175-0.878) compared to those with high EarlyR score. Within the AT arm there was no significant difference in 8-year DFS prognosis between any of the EarlyR risk groups.

Conclusions: This study confirmed the prognostic significance of EarlyR using FFPE tissue in a cohort of patients treated with AC chemotherapy from E2197. Patients with high EarlyR score who were treated with AC had significantly higher risk of recurrence than low EarlyR score patients treated with AC. On the other hand, prognosis of high EarlyR score AT-treated patients was not significantly lower than the prognosis of low EarlyR score AT-treated patients. Further study in a larger cohort is needed to assess the relative benefits of AC versus AT within the EarlyR high risk group and the EarlyR low risk group.

Citation Format: Badve S, Wang V, Willis S, Leyland-Jones B, Gokmen-Polar Y, Shulman L, Martino S, Sparano J, Davidson N, Goldstein L, Buechler S. Independent validation of EarlyR gene signature in E2197: A randomized clinical trial comparing doxorubicin plus docetaxel to doxorubicin plus cyclophosphamide as adjuvant chemotherapy in breast cancer [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr P1-06-08.

Abstract P1-14-01: Final analysis of overall survival (OS) for the epoetin alfa (EPO) phase 3 study, EPO-ANE-3010, of EPO plus standard supportive care (SOC) versus SOC in anemic patients with metastatic breast cancer (MBC) receiving standard chemotherapy

Background: In the interim analysis of study EPO-ANE-3010, for the primary endpoint of progression-free survival (PFS), the non-inferiority objective in ruling out a 15% increased risk in progressive death (PD) or death per investigator-determined PD was not achieved (JCO 34:1197-1207, 2016). PFS, based on independent review committee (IRC)-determined PD, however, met the non-inferiority criteria. At the interim analysis, OS with 1,337 deaths was reported; we now report the final analysis at 1,653 deaths and the updated PFS.

Methods: This multinational (19 countries and 132 participating sites), phase 3, randomized, open-label noninferiority study included anemic (≤11.0 g/dL hemoglobin) women receiving first- or second-line standard chemotherapy for MBC (Eastern Cooperative Oncology Group performance status of 0 or 1). Subjects were randomized (1:1) to receive either standard SOC for treatment of anemia plus EPO (40,000 IU subcutaneous) weekly up to 4 weeks after the last dose of cytotoxic chemotherapy, or SOC alone. The primary endpoint was PFS (using Cox's regression model). Secondary endpoints included OS, time to tumor progression (TPP), overall response rate (ORR) and safety assessments.

Results: A total of 2,098 subjects were enrolled (EPO plus SOC: n=1,050; SOC alone: n=1048). Demographic and baseline characteristics were well-balanced across the groups; median age was 52 years, most were white (67.5%) or Asian (30.5%) and median BMI was 26.0 kg/m2. Primary efficacy analysis (based on investigator-determined PD) showed a median PFS of 7.4 months for both groups (hazard ratio [HR], 1.094; 95% CI: 0.996, 1.201);upper bound exceeded prespecified noninferiority margin of 1.15. A 9% increased risk for PD/death in the EPO plus SOC group was observed and did not statistically rule out a 15% increased risk. Median PFS per IRC-determined PD was 7.6 months in both groups (HR, 1.028; 95% CI: 0.922, 1.146), this met pre-defined non-inferiority margin of 1.15 with a 3% risk increase in PD/death in EPO plus SOC group. At the final analysis for OS, median OS was 17.8 months in the EPO plus SOC group and 18.0 months in the SOC group; HR: 1.073 (95% CI: 0.974, 1.182); median TPP was 7.5 months in both groups (HR, 1.099; 95% CI, 0.998 to 1.210), and ORR was 50% in the EPO plus SOC group and 51% in the SOC group (odds ratio, 0.939; 95% CI, 0.789, 1.117). Red blood cell (RBC) transfusions were 5.8% versus 11.5% (P<0.001), and thrombotic vascular events were 2.8% versus 1.4% (P=0.038), respectively, in EPO plus SOC group and SOC group. The incidence of death due to PD were similar in both groups (EPO plus SOC: 93%; SOC: 91%).

Conclusion:The primary endpoint, PFS based on investigator-determined PD, did not meet noninferiority criteria but for PFS based on IRC-determined PD, noninferiority criteria was met. Overall, this study did not statistically rule-out a 15% increased risk in PD/death. The final analysis did not show statistically different OS in the EPO plus SOC group versus the SOC group. No new safety signals were noted with EPO treatment and the results are consistent with the known safety profile of EPO.

Citation Format: Leyland-Jones B, Bondarenko I, Nemsadze G, Smirnov V, Litvin I, Kokhreidze I, Abshilava L, Janjalia M, Li R, Lakshmaiah KC, Samkharadze B, Tarasova O, Shparyk Y, Polenkov S, Vladimirov V, Han J, Safonov I, Appiani C, Leitz G. Final analysis of overall survival (OS) for the epoetin alfa (EPO) phase 3 study, EPO-ANE-3010, of EPO plus standard supportive care (SOC) versus SOC in anemic patients with metastatic breast cancer (MBC) receiving standard chemotherapy [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr P1-14-01.

Abstract PD4-15: A tale of two pathways: Mutations in PI3K pathway in TNBC patients matter for the oncogenic cooperation with DNA damage repair pathway

Background: Mutations guide targeted therapy in the personalized medicine. In the opening chapter of our recently edited book (Dey et al., 2016), Prof. L. Cantley elegantly elucidated the basic signaling of the PI3K pathway in cancers. Mutations in the PI3K pathway are not only common and subtype-specific in BC but are also contextual. Alterations in DNA damage repair (DDR) pathway involving HRD (Homologous Recombination Defect) genes are one of the important contextual events of the upregulation of the PI3K pathway (De et al., 2016). Aim: Here we interrogated the contextuality of alterations of the PI3K and DDR pathway genes in our Avera patients. The mechanism of contextual cooperation between the pathwayswas experimentally validated. Methods: We examined mutation profile (FoundationOne) of our patients (Avera Cancer Institute) and patients fromthe TCGA data (cBioPortal). We validated the cooperation of the two pathwaysexperimentally by the synergy model of mutation-specific drugs; PI3K-PTEN-mTOR pathway inhibitor(s) and PARP inhibitor(s) using TNBC model.Results: We analyzed alterations of 17 and 12 genes of the PI3K and DDR pathways respectively in subtypes of BC. In luminal A and HER2-enriched (TCGA, Nature 2012), the alteration of PIK3CA reached 49 % and 47% as compared to 37 % in luminal B and 25% in basal-like.In the basal-like/TNBC subtype (cBioPortal) 12 DDR pathway genes (CHEK1/2, RAD51, BRCA1/2, MLH1, MSH2, ATM, ATR, MDC1, PARP1, FANCF) were altered in 90.1 % of cases, and 17 PI3K pathway genes were altered in 88.9 % of cases.Our ER+ve patients presented a diverse variety of PIK3CA mutations (E545K,  E545A, E545G, E542K, E453K, E762K, E365K, N345K, C420R, E81K, Q546R, C420R, E726K, E81K, E970K, H1047R, H1047L P104L, P539R, G106R,G1049R, R93Q,N345T, V105_E109>E, L113del, K111del) as compared to a less diverse type of PIK3CA mutations (Amplification, E542K, H1047R) in our TNBC patients. In our TNBC patients, the predominant type of mutation in PI3K pathway genes was found in PTEN consisting of Y68C, Y180*, loss, loss exons 1-5, and deletion exon1. The other most common mutation found in TNBC patients was in TP53 (>80%) and somatic BRCA1/2 (˜15%) genes. The interaction between the two pathways was evaluated using the mostly altered oncogenes and tumor suppressor genes (PTEN, AKT1/2, TSC1/2, mTOR, RICTOR, RHEB, BRCA1/2, ATM, ATR, FANCF) applying STRING10 to test the association at the highest 0.900 confidence views. Finally, we experimentally validated the contextual synergy of 2 pathways by demonstrating that a node-specific inhibition of the PI3K-mTOR pathway by GDC-0980 in the presence of carboplatin resulted in (1) an enhanced impairment of DSB repair and (2) a subsequent sensitization to PARPi (i). This effect occurred simultaneously with the inhibition of classic PI3K-mTOR survival signal(s) which induced a robust antiproliferative/proapoptotic effect even in BRCA-competent TNBC cells. The absence of PTEN, on the other hand, sensitized TNBC cells to PARPi in the presence of carboplatin, an effect more pronounced in BRCA-loss. Conclusion: Our data showed that the PI3K pathway cooperates with the DDR pathway in the breast oncogenesis especially basal-like and TNBC.

Citation Format: Dey N, Williams C, Krie A, Klein J, Williams K, Carlson JH, De PK, Leyland-Jones B. A tale of two pathways: Mutations in PI3K pathway in TNBC patients matter for the oncogenic cooperation with DNA damage repair pathway [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr PD4-15.

Hybrid capture-based genomic profiling of circulating tumor DNA from patients with estrogen receptor-positive metastatic breast cancer

BACKGROUND

Genomic changes that occur in breast cancer during the course of disease have been informed by sequencing of primary and metastatic tumor tissue. For patients with relapsed and metastatic disease, evolution of the breast cancer genome highlights the importance of using a recent sample for genomic profiling to guide clinical decision-making. Obtaining a metastatic tissue biopsy can be challenging, and analysis of circulating tumor DNA (ctDNA) from blood may provide a minimally invasive alternative.

PATIENTS AND METHODS

Hybrid capture-based genomic profiling was carried out on ctDNA from 254 female patients with estrogen receptor-positive breast cancer. Peripheral blood samples were submitted by clinicians in the course of routine clinical care between May 2016 and March 2017. Sequencing of 62 genes was carried out to a median unique coverage depth of 7503×. Genomic alterations (GAs) in ctDNA were evaluated and compared with matched tissue samples and genomic datasets of tissue from breast cancer.

RESULTS

At least 1 GA was reported in 78% of samples. Frequently altered genes were TP53 (38%), ESR1 (31%) and PIK3CA (31%). Temporally matched ctDNA and tissue samples were available for 14 patients; 89% of mutations detected in tissue were also detected in ctDNA. Diverse ESR1 GAs including mutation, rearrangement and amplification, were observed. Multiple concurrent ESR1 GAs were observed in 40% of ESR1-altered cases, suggesting polyclonal origin; ESR1 compound mutations were also observed in two cases. ESR1-altered cases harbored co-occurring GAs in PIK3CA (35%), FGFR1 (16%), ERBB2 (8%), BRCA1/2 (5%), and AKT1 (4%).

CONCLUSIONS

GAs relevant to relapsed/metastatic breast cancer management were identified, including diverse ESR1 GAs. Genomic profiling of ctDNA demonstrated sensitive detection of mutations found in tissue. Detection of amplifications was associated with ctDNA fraction. Genomic profiling of ctDNA may provide a complementary and possibly alternative approach to tissue-based genomic testing for patients with estrogen receptor-positive metastatic breast cancer.

Abstract P6-08-07: Gain and amplification of RAC1 GTP-ase in BC: Explaining alterations in patients by experiments using TNBC model

INTRODUCTION: RAC1-GTPase which transduces signals from cell surface integrins, have been implicated in metastasis. We reported that Wnt-beta-catenin pathway (WP) that signals metastasis (BMC Cancer, 2013), is one of the salient genetic features of Triple-Negative Breast Cancer (TNBC) (PlosOne, 2013).AIM: We demonstrated that TNBC cells acquire integrin-directed metastasis-associated (ID-MA) phenotypes following an upregulation of the WP (Oncotarget, In Press). Here we examined how WP signals are transduced in the context of ID-MA phenotypes in TNBC.METHOD: We documented gain and amplification of RAC1 gene in Breast Invasive Carcinoma subtypes from cBioPortal. The outcome for RFS was studied in the Hungarian ER-ve BC cohort.Mechanistically, we studied fibronectin-directed (1) migration, (2) matrigel- invasion, (3) RAC1 activation, (4) actin dynamics (confocal microscopy) and (5) podia-parameters using pharmacological agents (sulindac sulfide), genetic tools (beta-catenin siRNA), WP modulators (Wnt-C59, XAV939), RAC1 inhibitors (NSC23766, W56) and WP stimulations (LWnt3ACM, Wnt3A recombinant) in a panel of 6-7 TNBC cell lines, RESULTS: The collective percentage of gain and amplification of RAC1 were (1) 31% of total 1105 breast invasive carcinoma samples, (2) 29% of total 594 ER+ve samples, (3) 39% of total 174 ER-ve samples, (4) 38% of total 120 HER2+ve samples and (4)35% of total 82 TNBC samples (brca/tcga/pub2015; Cell 2015).In invasive ductal BC subtypes, gain and amplification of RAC1 were (1) 32% of total 201 Luminal A samples, (2) 37% of total 122 PAM50 Luminal B samples, (3) 47% of total 51 PAM50 Her2-enriched samples and (4) 33% of total 107 PAM50 Basal-like samples. In invasive lobular cancers, gain and amplification of RAC1 were 24% of total 127 samples.Involvement of WP in different TNBC cells was tested following stimulation by LWnt3ACM and Wnt3Arecombinant protein and different inhibitors of WP by both qRT-PCR and WB for beta-catenin, active beta-catenin, cMYC, cyclin D1and WP specific several stem cell markers. The WP attenuation, which (a) decreased cellular levels of beta-catenin, as well as its nuclear active-form, (b) decreased fibronectin-induced migration & invasion, (c) altered actin dynamics and (d) decreased podia-parameters was successful in blocking fibronectin-mediated RAC1/Cdc42 activity. Both Wnt-antagonists and RAC1 inhibitors blocked fibronectin-induced RAC1 activation and inhibited fibronectin-induced ID-MA phenotypes following WP stimulation by LWnt3ACM and Wnt3Arecombinant protein. High expression of RAC1 was associated with poor outcome for RFS with HR=1.48 [CI: 1.15-1.9] p=0.0019 in the Hungarian ER-veBC cohort.CONCLUSION:In TNBC model, the activation of RAC1 signals downstream of WP mediated ID-MA phenotypes. The identification of the functional relationship between RAC1 signaling and the WP activation in the control of ID-MA mechanistically explains how the activation of WP in TNBC is associated with the high metastatic incidences and a dismal outcome.

Citation Format: Dey N, Carlson JH, Jepperson T, Willis S, De P, Leyland-Jones B. Gain and amplification of RAC1 GTP-ase in BC: Explaining alterations in patients by experiments using TNBC model [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P6-08-07.

Abstract P6-08-04: Preclinical efficacy of dasatinib in combination with PARP inhibitor plus standard cytotoxic agent in triple-negative breast cancer xenograft model

Background: Dasatinib is an orally-active ATP-competitive small molecule kinase inhibitor that potently inhibits Abl kinase, Src family kinases and other kinases (Lombardo et al., 2004). Src, one of the key targets of dasatinib is involved in the regulation of cell proliferation, survival and apoptotic ability of cancer cells (Tryfonopoulos et al., 2011; Pusztai et al., 2014). Dasatinib has shown its anti-proliferative and anti-metastatic effectiveness against triple-negative breast cancer (TNBC) in both preclinical and clinical studies (Finn et al., 2011). Several molecular targets including poly ADP ribose polymerase (PARP) are under clinical investigation for the treatment of TNBC. Recently, PARP inhibitors in combination with chemotherapy have shown promising results in this disease in clinical and preclinical studies (Tutt et al., 2010; Kim et al., 2013; De et al., 2014). Here, we hypothesize that dasatinib in combination with PARP inhibitor (ABT888) plus standard cytotoxic agent (carboplatin) will attenuate the growth of both TNBC cell lines and xenograft tumors. Methodology:We have used BT-20 (PIK3CA mutated, H1047R), HCC70 (PTEN null), HCC1937 (BRCA1 mutated, PTEN null), MDA-MB-231 (KRAS/BRAF mutated), MDA-MB-468 (PTEN null) and SUM149PT (BRCA1 mutated, PTEN null) cells for in vitro study. Survival/proliferation, colony formation and apoptosis were examined by using 2D proliferative/growth assay, 3D-ON-TOP assays, and annexinV staining respectively. We next studied the activation status of Src and its downstream signaling. We also have evaluated the effects on tumor growth inhibition of dasatinib/ABT888/carboplatin as a single agent or in combination by using mouse xenograft model. Results: We observed that 1) Dasatinib inhibited Src activation in all tested lines, induced dephosphorylation of ERK1/2 and S6 RP; 2) level of Cyclin D1 was decreased by dasatinib treatment; 3) high anti-proliferative activities were observed following the treatment of dasatinib along with ABT888 plus carboplatin in both 2D proliferation assay and 3D-ON TOP colony formation assay; 4) dasatinib in combination with ABT888 plus carboplatin inducing early stage apoptosis was seen by Annexin V staining in all tested cell lines; 5) dasatinib alone or combined with ABT888 or carboplatin or in triple combination inhibited tumor growth in TNBC xenograft models, the best tumor inhibition result was induced by triple combination (comparing to no treatment control, the mean tumor volume was decreased ~ 87% ). Conclusion: Our in vitro and in vivo studies suggest that dasatinib may enhance the antitumor activity of PARP inhibitor plus standard cytotoxic agent in TNBC. Mechanistic studies of xenograft tumor samples are ongoing, the results of which will be presented in the meeting.

Citation Format: Sun Y, Lin X, Carlson JH, De P, Dey N, Jepperson T, R & D NCI, Williams C, Leyland-Jones B. Preclinical efficacy of dasatinib in combination with PARP inhibitor plus standard cytotoxic agent in triple-negative breast cancer xenograft model [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P6-08-04.

Abstract P1-05-23: Utilities and challenges of RNA-Seq based expression and variant calling in a clinical setting

Introduction

Variant calling based on DNA samples has been the gold standard of clinical testing since the advent of Sanger sequencing. The use of DNA variants has proved a great value to guide treatment in cancer patients. However, DNA based analysis will not inform about expression status of the gene harboring a particular variant. RNA has long been used to monitor expression. To this point RNA assays and analysis are confined to the research laboratory and rarely used clinically except in specifically defined gene signatures such as PAM50 and OncoType Dx. Beyond expression, RNA has the ability to confirm expression of DNA variants and identify fusion events. We hypothesize that the combination of DNA and RNA based data will allow the determination of variant specific expression status and improve clinical diagnostics. It has been previously shown that RNA sequencing (RNA-Seq) based variant calls are highly accurate and confirm DNA based variant calls. In this study we investigated the utility of RNA-Seq as a diagnostic assay integrated with DNA based sequencing data.

Materials and Methods

Targeted DNA sequencing of 321 genes was performed on 37 patient samples (FFPE), including 22 breast cancer samples by a commercial vendor. RNA-Seq on the same patient samples was performed using 100ng of total RNA. Libraries were run on the Illumina NextSeq 500 with a minimum of 75M paired 75bp reads. To evaluate RNA-seq expression reproducibility, replicates of 6 normal ovarian tissue samples (min. 50M reads) were run in sets of triplicates. STAR was used for alignment (hg19) and gene expression quantification (RefSeq). RNA-Seq based variant calling was performed using the SNPiR pipeline. Based on the results of the commercial assay, DNA based variants were examined for expression of the corresponding genes and ability to confirm variants in the RNA-Seq data.

Results

RNA expression data showed no corresponding gene expression for at least one single nucleotide variant (SNV) in 9/37 patients analyzed (24.3%). In 18/37 patients (48.6%) SNV corresponding expression was in the lowest quartile of expression values. Variant calls could be confirmed by RNA-Seq for 95/455 SNVs, with adequate coverage in 263 of the remaining 360 variant locations (median coverage: 34). Of these, a homozygous reference call was made in 166/263 SNVs. Concordance for RNA-Seq gene level expression data between replicates was > 0.995.

Conclusions

These findings suggest that RNA-Seq based data can provide clinical value when using gene expression values in combination with DNA based variant calls. We found gene level expression to be highly reproducible and will further investigate the use of spike in controls to determine clinically usable expression ranges and lower limit of expression values. To our knowledge, it has not been shown that RNA-Seq based variant calls are reproducible which is the focus of our current research as this will be one requirement for usage in a regulated environment. While our use of RNA Seq is currently limited to gene expression level data, we have demonstrated a clinically relevant benefit to using RNA Seq data as an additive feature to the current standard of DNA variant calling.

Citation Format: Young B, Mark A, Meissner T, Amallraja A, Andrews A, Connolly C, Williams C, Leyland-Jones B. Utilities and challenges of RNA-Seq based expression and variant calling in a clinical setting [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P1-05-23.

Abstract P3-07-02: Are we missing actionable targets in breast cancer? Novel insights into recurrent Ret alterations

Background: Recurrent gene fusions in breast cancer have been rarely reported suggesting that they either are not present or are not easily detected by standard sequencing methods. Comprehensive genomic profiling (CGP) by hybrid capture-based, high depth next-generation sequencing approaches, can be used to detect recurrent rearrangements and other genomic alterations involving target genes. We found that CGP can identify recurrent alterations involving RET, a known oncogenic tyrosine receptor kinase, in a subset of breast cancer.

Methods: CGP using FoundationOne platform was performed interrogating the entire coding region for up to 315 cancer-related genes and introns of up to 28 genes involved in rearrangements at a depth of 500-1000X in formalin-fixed, paraffin embedded tumor tissue (Foundation Medicine, MA). Engineered representative RET fusion vectors were synthesized and expressed in non-tumorigenic cell lines (breast MCF10A and mouse 3T3 fibroblasts), and cells were evaluated for RET kinase signaling, drug response, and tumorigenicity. Patient-derived xenografts (PDX) generated from two triple negative breast cancers (TNBCs) were used in an ex vivo assay (Response3DXTM, Molecular Response LLC, San Diego, CA).

Results: Twenty-two RET rearrangements were identified in 8119 (0.27%) breast cancer cases. Of these, 5 rearrangements were activating RET fusions including CCDC6-RET (n=4) and NCOA4-RET (n=1), that have been described in other cancer types. Five other cases had clear evidence of genomic rearrangement involving RET, but the 5' partners could not be definitively identified. The remaining twelve cases had complex rearrangements of RET including internal duplications. RET amplification was also observed, both in TNBC and in a HER2+ breast cancer at onset of resistance to HER2-targeted therapy.

Both NCOA4-RET and a novel RASGEF1A-RET fusion were characterized in vitro. Non-tumorigenic cells engineered to stably overexpress either RET fusions demonstrated transformed phenotypes. The fusions were constitutively active, as shown by endogenous phosphorylation of the kinase domain, and drove activation of downstream signaling as shown by increased phosphorylation of ERK and AKT. Cells transformed by RET-fusions were exquisitely sensitive to treatment with RET inhibitors. Interestingly, a PDX model of RET-amplified TNBC was sensitive to treatment with a PIK3CA inhibitor. An index case of ER+/PR-/HER2+, metastatic breast cancer that had radiographic evidence of disease progression while on trastuzumab, pertuzumab, and anastrazole was found to have a NCOA4-RET fusion by CGP. Subsequent treatment with with cabozantinib plus anastrazole led a rapid clinical and radiographic response.

Conclusions: CGP can identify recurrent RET rearrangements in breast cancer that act as primary oncogenic drivers and can be therapeutically targeted. RET alterations may also play a role in acquired resistance to HER2-targeted therapies, suggesting a role for combined RET and HER2-targeted therapy in this setting. Our data demonstrate that RET alterations can be identified by clinical-grade CGP and are promising candidates as therapeutic targets in selected breast cancer patients.

Citation Format: Hirshfield KM, Paratala BS, Hindoyan A, Dolfi SC, Yilmazel B, Schrock A, Gay L, Ali SM, Ross JS, Williams CB, Nair P, Ganesan S, Leyland-Jones B. Are we missing actionable targets in breast cancer? Novel insights into recurrent Ret alterations [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P3-07-02.

Abstract P6-08-09: Cancer stem cells define 3D clonogenic growth response to rational combinations of PI3K-isoform specific inhibitors in TNBC

This abstract was not presented at the symposium.

Abstract P4-12-01: Independent validation of EarlyR gene signature in BIG 1-98: A randomized, double-blind, phase III trial comparing letrozole and tamoxifen as adjuvant endocrine therapy for postmenopausal women with hormone receptor-positive, early breast cancer

Background: EarlyR is a prognostic gene signature score in ER+ breast cancer (BC) computed from the expression values of ESPL1, SPAG5, MKI67, PLK1 and PGR using a novel algorithm. EarlyR has been validated in multiple cohorts profiled on Affymetrix and Illumina microarrays. This study sought to verify prognostic features of EarlyR in a cohort of BIG 1-98.

Patients and Methods: Illumina DASL assay was used to measure gene expression in FFPE tissue of primary BC from a case-cohort sampling subset of postmenopausal women in BIG 1-98 treated with adjuvant endocrine therapy (letrozole or tamoxifen). Chemotherapy treatment was at the discretion of individual physicians and patients. Among the 1218 patients centrally reviewed with sufficient RNA material for the DASL assay, 1174 with ER+ status and assessed for EarlyR are included in the analytic cohort. EarlyR scores and pre-specified risk strata (≤25=low, 26-75=intermediate, >75=high) were computed, while blinded to clinical data. The analysis endpoints included distant recurrence free interval (DRFI) defined as time from randomization to BC recurrence at a distant site within 8 years and BC free-interval (BCFI) defined as time from randomization to first invasive BC recurrence at a local, regional or distant site or invasive contralateral BC within 8 years. Weighted proportional hazards models (univariate and multivariate, stratified by treatment assignment) were used to adjust for Kaplan-Meier, hazard ratio estimates and Wald test statistics to obtain unbiased analyses and to give consistent estimates.

Results: The distribution of the EarlyR risk groups was 67% low, 19% intermediate and 14% high risk in this ER+ cohort. EarlyR was prognostic for 8-year DRFI (P-trend=0.008). Patients with high EarlyR risk score (>75) had significantly increased risk of distant recurrence within 8 years (univariate HR=1.73, 95%CI: 1.14-2.64) compared to low EarlyR risk group (≤25). The estimated 8-year DRFI (95%CI) is 84%(80%-88%) for high risk vs. 91%( 89%-92%) for low risk, corresponding to an absolute DRFI risk reduction of 7% (low vs high). EarlyR is also prognostic of 8-year BCFI in ER+ (P-trend=0.002) with the estimated 8-year BCFI (95%CI) 79%(75%-84%) for high risk vs. 88%( 86%-89%) for low risk. Consistent results were observed in ER+, HER2- (P-trend=0.01 for DRFI, P-trend=0.004 for BCFI), in ER+, LN- (P-trend=0.05 for DRFI, P-trend=0.03 for BCFI) and ER+, LN+ (P-trend=0.08 for DRFI, P-trend=0.03 for BCFI) subsets.

Conclusions: This study confirmed the prognostic significance of EarlyR using FFPE tissue from the BIG 1-98 trial. In analyses of all ER+ patients and subsets LN-, LN+ and HER2-, EarlyR classifies 65%-70% of patients as low risk, 11-16% as high risk, and < 20% as intermediate risk. In these subsets, the size of the low risk group is larger and the size of the intermediate risk group is smaller than those reported for commercially available signatures. EarlyR identifies a set of high-risk patients with relatively poor prognosis who may be considered for additional treatment. The clinical utility of EarlyR requires further study.

Citation Format: Buechler S, Gray KP, Gökmen-Polar Y, Willis S, Thürlimann B, Kammler R, Leyland-Jones B, Badve SS, Regan MM. Independent validation of EarlyR gene signature in BIG 1-98: A randomized, double-blind, phase III trial comparing letrozole and tamoxifen as adjuvant endocrine therapy for postmenopausal women with hormone receptor-positive, early breast cancer [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P4-12-01.

Abstract P1-05-22: The value of RNA-Seq for the detection of clinically actionable targets in breast cancer - A small cohort analysis

Introduction

Next generation sequencing has facilitated the understanding of pathogenesis and molecular heterogeneity of breast cancer (BC) as well as accelerated the path towards precision medicine. DNA sequencing (DNA-Seq) based assays for the detection of mutations and alterations in solid and hematologic cancers are finding their way into clinical practice and are readily available as clinical products. RNA sequencing (RNA-Seq), so far being vastly applied in the research context, promises to expand the diagnostic, prognostic and therapeutic use of this technology in cancer. Beyond mutational status, RNA-Seq enables the detection of fusions, quantification of gene expression level, detection of differentially expressed genes, molecular based subtyping, and risk-stratification. In this study we analyzed RNA-Seq and copy number data from BC patients that had undergone DNA-Seq based diagnostics through commercial providers with the goal to detect additional actionable targets.

Materials and Methods

We included 18 BC patients (5/18 triple negative) that had previously undergone DNA-based targeted (321 genes) sequencing. RNA-Seq to a minimum of 75M reads (75pb) was performed using 100 ng of total RNA on the Illumina NextSeq 500 platform. STAR was used for alignment (hg19) and gene expression quantification (RefSeq). Fusions were detected using STAR-Fusion. DESeq2 was utilized to identify patient specific differentially expressed genes by analyzing samples individually against a set of 13 controls from healthy breast tissue generated in-house. Copy number variations (CNVs) were detected using the Nanostring CNV Cancer panel (89 genes) on the Nanostring nCounter platform. Differentially upregulated or amplified genes were queried against DGIdb and Gene Drug Knowledge database for suitable drug matches, limiting the queries to clinically actionable antineoplastic drugs.

Results

Analyzing the cohort of 18 BC patients, we detected on average 26 BC relevant genes (526 total, log2 FC > 2) to be upregulated per patient. Querying the upregulated genes against DGIdb, we found a total of 18 genes that had drug matches and fulfilled the criteria of being actionable antineoplastic drugs, with 17/18 samples having a minimum of two gene targets (avg: 4). Most frequent upregulated genes were TOP2A (83%), AURKA (61%), AURKB (56%), RET (39%)and FGFR3 (28%). In the case of CNVs, 12/18 patients showed at least one gene target with clinically actionable drugs associated. This was observed across 12 gene targets that were amplified (avg: 3) and 4 gene targets that underwent deletions (avg: 1). Most frequent CNVs included MYC (14%) and CCND1 (12%). 4/7 patients having an AURKA overexpression also showed an AURKA amplification on the CNV assay. 10/18 patients had fusions events, with an average of three fusions per patient, including GAB2-WNT11, PAK1-TENM4 and FGFR2-CEP55 fusions.

Conclusions

We show that RNA-Seq and copy number assays provide additional clinical value by detecting suitable drug targets beyond traditional DNA-based approaches. We are conducting further analysis on how these additionally derived drug targets could improve the current treatment schedule of those patients.

Citation Format: Meissner T, Amallraja A, Mark A, Andrews A, Connolly C, Young B, De P, Williams C, Leyland-Jones B. The value of RNA-Seq for the detection of clinically actionable targets in breast cancer - A small cohort analysis [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P1-05-22.

Abstract P2-03-02: Differential mutation pattern between neoadjuvant and metastatic settings in breast cancer patients

Background: Dysregulated signaling pathways occur in human cancers including breast cancer, making it a rational target for novel genome guided combinatorial personalized therapies. The aim of the present study was to investigate the different genetic mutation pattern between neoadjuvant and metastatic settings in breast cancer patients to guide research and clinical treatment.

Material and Methods: 150 breast cancer patients were involved in this study. 38 patients were receiving neoadjuvant treatment and 112 patients were in the metastatic setting. Tumor specimens obtained from the 150 patients were subjected to genetic mutation testing by FoundationOne. Genetic alterations detected by FoundationOne test were collected and analyzed.

Results: 96 and 149 different genes where reported by FoundationOne in neoadjuvant and metastatic setting respectively. The average number of non-synonymous mutation was five per case in the neoadjuvant setting and six per case in the metastatic setting. TP53 (58%), MYC (32%), PIK3CA (29%), PTEN (16%), CDH1 (13%), CCND1 (11%), EMSY (11%), LYN (11%) and ZNF703 (11%) were the most seen mutations in neoadjuvant setting. TP53 (40%), PIK3CA (39%), MYC (22%), CCND1 (21%), FGF19 (21%), FGF4 (21%), CDH1 (20%), FGF3 (19%), ERBB2 (17%), ESR1 (14%), FGFR1 (14%), ZNF703 (14%), GATA3 (13%), MYST3 (11%), PTEN (11%), EMSY (10%), NF1 (10%) and ZNF217 (10%) were the most seen mutations in metastatic setting. ESR1 and GATA3, which are seen in 14% and 13% of metastatic breast cancer patients, were not reported in neoadjuvant breast patients. Moreover, among the 16 metastatic breast cancer patients who has ESR1 mutation, 9 (56%) of them presented with PIK3CA or other genetic mutations which are directly involved in the phosphoinositide 3-kinase (PI3K)/AKT pathway.

Conclusion: A significantly more mutation in Receptor Tyrosine Kinases (RTKs)/ Growth Factor Signaling ( especially ERBB and FGFR pathways) was reported in the metastatic setting compare to the neoadjuvant setting, suggesting a critical role of the RTKs in metastatic breast cancer patients. The coexisting of ESR1 and PI3K/AKT pathway alteration and the absence of ESR1 in neoadjuvant setting also suggested that in early stage breast cancer patients who have a PI3K pathway alterations; there is a higher chance to develop ESR1 mutation with disease progression.

Citation Format: Xu B, Williams C, De P, Dey N, Klein J, Williams K, McMillan A, Leyland-Jones B. Differential mutation pattern between neoadjuvant and metastatic settings in breast cancer patients [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P2-03-02.

Abstract P6-03-01: A combination of dual inhibition in HER2-network by T-DM1 and GDC-0980 provides maximal antitumor efficacy in preclinical model of HER2+ breast cancer

Background: PIK3CA mutation is associated with a lower pCR rate in primary HER2+ breast cancer (BC) treated with trastuzumab and lapatinib in addition to chemotherapy (from five clinical trials, PMID: 27177864). The BOLERO-1 study showed that the efficacy of a combination of mTOR inhibitor (everolimus) plus trastuzumab (T) and paclitaxel was not very efficacious with HER2+ advanced BC patients except for HER2+/ER- BC patients (PMID: 26092818). In the same line, BOLERO 3 trial data showed the same combination (T + everolimus + vinorelbine) is also not efficacious in T-resistant, HER2+ advanced breast cancer women (median PFS 7 months with everolimus and 5.78 months with placebo, HR: 0.78) (PMID: 24742739). T-DM1 does not have typical adverse events of chemotherapy. Therefore, there has been interest in combining it with other targeted agent. Here we tested the efficacy of a combination of T-DM1 plus GDC-0980 (a dual PI3K/mTOR inhibitor) in HER2+/T-resistant BC cell lines in vitro and in vivo. Methodology: Here we have studied the in vitro and in vivo effects of GDC-0980 along with T-DM1 in HER2+/T-sensitive (BT474), HER2+/T-resistant (BT474HerR), and HER2+/PIK3CA (HCC1954, MDA-MB453) mutated models. We assessed in vitro anti-proliferative, pro-apoptotic and activation status of the PI3K-AKT-mTOR signaling pathway following the combination of GDC-0980 plus T-DM1 in HER2+ BC cell lines. We next evaluated the impact of GDC-0980 plus T-DM1 on tumor growth and angiogenesis using xenograft models. Results: 1) GDC-0980 inhibited downstream activation of the PI3K-mTOR signaling pathway effectors, p-AKT (Ser473, The308), p-P70S6K, p-S6RP and p-4EBP1, and this inhibition was more pronounced when GDC-0980 was combined with T-DM1, 2) similarly the anti-proliferative activity of a combination of GDC-0980 plus T-DM1 was significantly higher by 3D-ON-TOP clonogenic assay following heregulin stimulation, 3) consistent with anti-proliferative effects of GDC-0980, the proportion of cells in the G1 phase of the cell cycle increased in HER2+ cell lines with a concomitant decrease in the S phase of their treatment with GDC-0980, 4) the initiation of apoptotic activity (annexin V) of GDC-0980 was significantly superior to that of an allosteric inhibitor of mTOR, RAD001. GDC-0980 also induced apoptotic markers like cleaved CASPASE3, cleaved PARP1, BIM in HER2+ BC cells and 5) a combination of GDC-0980 plus T-DM1 significantly blocked tumor growth to tumor regression in the HER2+/T-sensitive, HER2+/T-resistant and HER2+/PIK3CA mutated BC xenograft models. Along with its anti-tumor effect, this combination effectively decreased tumor angiogenesis (tumor micro-vessel density via CD31 staining). Conclusions: A combination of GDC-0980 plus T-DM1 significantly blocked in vitro and in vivo HER2+ breast tumor cells growth irrespective of PIK3CA mutation status. This strategy warrants further clinical investigation.

Citation Format: De PK, Carlson JH, Sun Y, Lin X, Friedman L, Dey N, Leyland-Jones B. A combination of dual inhibition in HER2-network by T-DM1 and GDC-0980 provides maximal antitumor efficacy in preclinical model of HER2+ breast cancer [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P6-03-01.

Abstract P3-07-36: Immune related gene expression signatures predict benefit of letrozole over tamoxifen in BIG 1-98

Background: The prognostic significance of increased levels of CD8+ tumor infiltrating lymphocytes(TILs) in ER- breast cancer has been described. We sought to identify possible immune-related biomarkers for predicting benefit from letrozol(LET) or tamoxifen(TAM) for recurrence in ER+ breast cancer.

Patient and Methods: We used Illumina DASL Assay to measure gene expression in FFPE primary breast cancers from a subset of postmenopausal patients enrolled in the BIG 1-98 randomized phase 3 trial comparing 5 years LET (n=344) vs TAM (n=381) as adjuvant endocrine therapy. Gene sets (n=1910) that represent cell states and perturbations within the immune system from the Human Immunology Project Consortium were used in an exploratory analysis to identify possible predictive signatures.

Results: We identified five distinct gene signatures from previously reported laboratory experiments associated with immune cell differentiation that are highly predictive of benefit (reduced breast cancer recurrence risk) of LET over TAM, each with gene signature p-values<1E-5 and signature-by-treatment interaction p<1E-6. The signatures predict a similar pattern that patients at low-risk score benefit from LET and patients with high-risk score appear to have an advantage with TAM after 5 years. The gene signatures originate as a result of being differentially expressed in the following previously reported experiments. [RAP2A EEF2K TRAF3IP2 GPR37L1 DDX54] down regulated comparing TLR3 and TLR9 agonists in dendritic cells. [RPA1 DUSP4 NUDT18 ZFYVE28] up regulated in comparison of T follicular helper versus Th17 cells. [MAPK15 CCR4 SORCS2 RAMP1 SH3PXD2A] up regulated in regulatory T cell versus CD4+ T cells. [NDUFA6 GIMAP1 CPNE3 ST3GAL6 CCDC88A] down regulated in comparison of untreated CD8+ dendritic cells versus treated with IFNG. [GPN1 COX17 CUL2 CDSA] down regulated in naïve vs stimulated CD8 T cells after 48 hours. We further investigated the signatures using Hungarian Academy of Sciences (HAS) cohort (Gyorffy B 2010), which is a collection of smaller published affymetrix cohorts combined into a larger ER+, TAM treated cohort (n=700). One signature was not tested because two genes were not present in the affymetrix cohort. Three of the remaining four signatures gave informative prognostic results in the HAS cohort, and the signature associated with differentiation of CD8+ dendritic cells was highly prognostic with HR=0.36 (0.26-0.49) p=1E-11.

Conclusion: The role of selective estrogen receptor modulators on immune response has been well described, where TAM has been shown to prevent differentiation and activation of dendritic cells (Naibandian 2005). Similarly, it has been shown that MET inhibitors negatively regulate neutrophils suggesting that anti-MET drugs in cancer could impact immune response (Finisquerra 2015). These findings suggest that if TAM is a negative regulator of immune response why in the ATAC clinical trial, the combination therapy of anastrozole plus TAM were not significantly different from TAM alone were anastrozole was superior. With the increasing importance of understanding the role of immune response on outcome and the use of combination therapies the assessment of TILs in the neoadjuvant setting will be critical for guiding therapy.

Citation Format: Willis S, Gray KP, Regan MM, Rae JM, Kammler R, Young B, Ditzel HJ, Lyng MB, Colleoni M, Viale G, Leyland-Jones B. Immune related gene expression signatures predict benefit of letrozole over tamoxifen in BIG 1-98. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P3-07-36.

Abstract P1-14-03: The evaluation of trebananib plus standard neoadjuvant therapy in high-risk breast cancer: Results from the I-SPY 2 TRIAL

Background: I-SPY 2 is a multicenter phase 2 trial using response-adaptive randomization within biomarker subtypes to evaluate a series of novel agents when added to standard neoadjuvant therapy for women with high-risk stage II/III breast cancer. The primary endpoint is pathologic complete response (pCR). The goal is to identify/graduate regimens with ≥85% Bayesian predictive probability of success (statistical significance) in a 300-patient phase 3 neoadjuvant trial defined by hormone-receptor (HR), HER2 status & MammaPrint (MP). Regimens may also leave the trial for futility (< 10% probability of success) or following accrual of maximum sample size (10%< probability of success <85%). We report the results for trebananib, an angiopoietin-1/2-neutralizing peptibody that inhibits interaction with the Tie2 receptor.

Methods: Women with tumors ≥2.5cm were eligible for screening. MP low/HR+/HER2- tumors were ineligible for randomization. Serial MRI scans (baseline, 2 during treatment and pre-surgery) were used in a longitudinal model to improve the efficiency of adaptive randomization. Participants are categorized into 8 subtypes based on: HR status, HER2 status and MP High 1 (MP1) or High 2 (MP2). MP1 and MP2 are determined by a predefined median cut-point of I-SPY 1 participants who fit the eligibility criteria for I-SPY 2. Trebananib was initially assigned to HER2- patients only; once safety data with trastuzumab (H) were obtained, it was also assigned to HER2+ patients. Analysis was intent to treat -- patients who switched to non-protocol therapy were designated non-pCRs.

Results: Trebananib +/-H did not meet the criteria for graduation in any of the 10 signatures tested. When the maximum sample size was reached, accrual ceased. We report probabilities of trebananib +/-H being superior to control and Bayesian predictive probabilities of success in a 1:1 randomized neoadjuvant phase 3 trial for the 10 biomarker signatures, using the final pCR data from all patients.

SignatureEstimated pCR Rate (95% probability interval)Probability Trebananib Is Superior to ControlPredictive Probability of Success in Phase 3Trebananib (n=134)Control (n=133)ALL0.259(0.16 -0.36)0.158(0.09-0.23)0.9860.564HR+0.157(0.05-0.26)0.115(0.03- 0.20)0.8050.281HR-0.378(0.22-0.53)0.207(0.11- 0.31)0.9910.784HER2+0.279(0.07-0.49)0.17(0.04-0.30)0.8790.553HER2-0.254(0.15-0.36)0.155(0.08-0.23)0.9810.555MP20.342 (0.19-0.49)0.177(0.07-0.28)0.9910.786HR-/HER2-0.368 (0.21-0.53)0.201(0.10-0.30)0.9880.771HR-/HER2+0.444(0.15-0.74)0.244(0.07-0.42)0.9260.739HR+/HER2+0.201(0.01-0.39)0.135(0.01-0.26)0.7750.41HR+/HER2-0.143(0.04-0.24)0.11(0.03-0.19)0.7580.248

Citation Format: Albain KS, Leyland-Jones B, Symmans F, Paoloni M, van 't Veer L, DeMichele A, Buxton M, Hylton N, Yee D, Lyandres Clennell J, Yau C, Sanil A, I-SPY 2 Trial Investigators, Berry D, Esserman L. The evaluation of trebananib plus standard neoadjuvant therapy in high-risk breast cancer: Results from the I-SPY 2 TRIAL. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P1-14-03.

Abstract P4-08-04: Navigating genomic landscape to find a PI3K-signaling algorithm for a rational combinatin in precision medicine

Background: Treatment of BC is conventionally based on the presence/absence of ER/PR or HER2 status of the primary tumor. We have enriched this approach by including major genetic and proteomic changes in tumors of individual patients in order to develop a better treatment-rationale based on an alteration driven signaling algorithm. Methods: Genomic and proteomic data from 75 BC patients seen in our center were retrospectively analyzed. Patients were re-biopsied after consultation and samples were characterized (IHC for ER, PR, and HER2; FFPE samples for genomic [Foundation Medicine] and proteomic analyses [Theranostics]). In vivo studies were conducted using xenograft models. Results: Although alterations of PIK3CA, PIK3R1, AKT, PTEN, MDM2, MDM4, TSC1, mTOR and RICTOR are most frequently observed in our patients, there is a distinct pattern of alteration(s) of the PI3K pathway genes in different subtypes of BC. A total of 76 genes were altered in 48 ER+BC patients. In 79% of ER+BC patients the above mentioned PI3K pathway genes were altered. Analyzing the set of alterations of genes in individual patients, we observed that within these 48 patients 25% exhibited alterations in more than one node of the pathway; the most common combination (alterations) being the amplification/mutation of PIK3CA with the amplification of MDM2/4 genes. The percentage of patients belonging to HER2+ & TNBC exhibiting similar alterations in the PI3K pathway genes were significantly lower (∼40%). Our previous in vivo studies demonstrated that GDC-0980 and BEZ235 enhanced the antitumor activity of ABT888 plus carboplatin in TNBC or trastuzumab in HER2+ BC respectively and blocked the growth of established xenograft tumors by 80% to 90% with a concomitant decrease in tumor Ki67, pS6RP and CD31. Mechanistically the action of the PI3K-mTOR pathway targeted drug(s) was tested using cell line based models of BC subtypes pertaining to their respective genomic alterations. A combination of a pan-PI3K pathway inhibitor, GDC-0941 or isoform-specific inhibitors along with AI, trastuzumab, or HRD inhibitors (PARP) blocked proliferative signals and enhanced apoptosis (cleaved caspase3) in ER+/PIK3CA mutated, HER2+/PIK3CA mutated or PTEN-null TNBC cells respectively as demonstrated by WB, flow cytometry, cell proliferation, viability and cytotoxicity assays. A recent study demonstrated that exposure to chemotherapy induced a phenotypic shift or cell state transition towards a transient CD44Hi/CD24Hi chemotherapy-tolerant state, leading to the activation of downstream non-receptor tyrosine kinase signaling towards an emerging adaptive resistance (Goldman et al., Nature Comm. 2015). Hence drug combination(s) are being tested for their effect on CD44/CD24 expression levels, results of which will be presented in the meeting. Conclusion: Plotting the genetic alterations from the patient on the signaling landscape will be useful in cracking the code leading to improved treatment options. Patient specific in-depth plotting of genetic alterations of the PI3K-mTOR pathway and the relevance of these alterations in the context of (1) mechanisms of PI3K-mTOR pathway targeted drugs and (2) cell signaling are critical in determining choice of drugs in BC subtypes.

Citation Format: Carlson JH, Krie A, Williams C, Sun Y, Lin X, Williams K, Klein J, Friedman L, De P, Dey N, Leyland-Jones B. Navigating genomic landscape to find a PI3K-signaling algorithm for a rational combinatin in precision medicine. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P4-08-04.

Abstract P5-16-01: Implementation of routine genomic and proteomic profiling of metastatic breast cancer patients in a community cancer center

Background: The optimal treatment strategy for patients with metastatic breast cancer (MBC) is currently unknown. Resistance to standard therapies, including anthracyclines and taxanes, limit the number of treatment options in many patients to a small number of non-cross resistant regimens. Rational combination approaches that are selected based upon genomic and proteomic analysis represents a possible advance that warrants extensive exploration.

Methods: Single center analysis of 77 consecutive metastatic breast cancer patients seen over a 12 month period (June 2014 through May 2015). All patients were referred for sequencing and the metastatic disease was rebiopsied. All samples were sent for standard pathologic, genomic (FoundationOne), and proteomic (TheraLink) analysis.

Results: Genomic and proteomic analysis yielded actionable targets in a majority of cases (89%). The most common pathways involved were the following: PI3K/Akt/mTOR (73%), MAPK (46%), ErbB (36%), FGFR (25%), and Jak/STAT (11%). Over 100 unique molecular aberrations were identified in 40 evaluable patients. Current outcomes are summarized in Table 1. The overall response rate was 45%, with another 43% of patients with stable disease. Average number of prior therapies was over 4, with a range of 1-11.

Table 1ER+/HER2-CR = 3PR = 8SD = 15PD = 2NE = 20ER+/HER2+CR = 2PR = 2SD = 1PD = 0NE = 9ER-/HER2+CR = 0PR = 1SD = 1PD = 1NE = 2Triple NegativeCR = 0PR = 2SD = 0PD = 2NE = 6Total evaluable patients = 40Overall CR = 13%Overall PR = 33%Overall SD = 43%Overall PD = 13%Total Not Evaluable = 37 pts (48%)Overall Response Rate (ORR) = 45%     CR = complete response PR = partial response SD = stable disease PD = progressive disease NE = not evaluable

Conclusion: Since current literature suggests that an overall response rate of approximately 10% or less is expected for patients that have received greater than 4 previous lines of therapy, the ORR seen in this analysis is quite remarkable. Most patients in this analysis were treated with FDA approved drugs off label, which provided additional challenges and was the primary reason that many patients were not evaluable. Patients were only evaluable if they received the recommended therapy and were measured for outcome. Our initial data provides growing evidence that it is critical to incorporate genomic and proteomic analysis (preferably as early as possible in the disease course) to allow for the best chance of disease response.

Citation Format: Williams CB, Krie A, De P, Dey N, Klein J, Williams KA, Hoogeveen M, Solomon B, Leyland-Jones B. Implementation of routine genomic and proteomic profiling of metastatic breast cancer patients in a community cancer center. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P5-16-01.

Abstract P6-09-02: Analysis of the International tamoxifen pharmacogenomics consortium (ITPC) dataset shows that genotyping DNA derived from tumor does not introduce CYP2D6 genotyping error or mask an association with tamoxifen efficacy

Background: The anti-estrogen tamoxifen is metabolized into the more potent anti-estrogen, endoxifen, primarily by polymorphic CYP2D6. An association between CYP2D6 genotype and tamoxifen efficacy was assessed in a meta-analysis of 4,973 breast cancer patients by the ITPC. A subgroup analysis in 1,996 estrogen receptor (ER)-positive postmenopausal patients receiving 20[thinsp]mg/day tamoxifen for 5 years (criterion 1) found CYP2D6 poor metabolizer genotype was associated with worse invasive disease free survival (IDFS; hazard ratio (HR) 1.25, 95% confidence interval (CI) 1.06-1.47, P=0.009). This meta-analysis did not include data from two prospective trials (Anastrozole, Tamoxifen, Alone or In Combination (ATAC) and Breast Intergroup (BIG) 1-98) both of which meet the patient criteria but failed to replicate the ITPC findings. Some ITPC investigators criticized ATAC and BIG 1-98 for genotyping CYP2D6 from tumor-derived DNA, positing this leads to genotyping errors detected by Hardy-Weinberg Equilibrium (HWE). However, ITPC analyses did not exclude tumor-derived CYP2D6 genotypes or report HWE. Therefore, we re-analyzed the ITPC data to investigate claims that tumor-derived genotyping causes HWE departure and masks the association between CYP2D6 genotype and tamoxifen efficacy.

Methods: The ITPC dataset was filtered to patients fulfilling criterion 1. HWE for CYP2D6*4 was analyzed in Caucasian patients (n=1,619) by study, by DNA source (tumor or blood), and in the entire subgroup. The ITPC meta-analysis was rerun stratified by DNA source using the same specifications (patients, phenotype, genotype, statistical model, etc.) as in ITPC.

Results: Significant HWE deviation for CYP2D6*4, the most common variant (MAF =0.2), was not observed in any study genotyped from tumor but was observed in one study genotyped from blood. Combining studies led to significant HWE deviations in studies genotyped from both blood and tumor, and for the entire subcohort (Table 1). Associations between CYP2D6 genotype and IDFS stratified by DNA source yielded similar, non-statistically significant, results (blood: n=933, HR=1.19, 95% CI 0.91-1.57, P=0.21; tumor: n=997, HR=1.19, 95% CI 0.99-1.44, P=0.07).

Conclusions: HWE deviations for CYP2D6*4 are not uniformly or exclusively found in studies using tumor DNA, but can occur as a statistical consequence of combining genotypes from heterogeneous populations like in the multi-institution BIG 1-98 and ATAC studies. Re-analysis of the ITPC dataset stratified by DNA source refutes the hypothesis that genotyping tumor DNA masks a pharmacogenetic association. These findings reaffirm the validity of the BIG 1-98 and ATAC analyses and support inclusion of these studies in the ITPC meta-analysis to rigorously assess the association between CYP2D6 genotype and tamoxifen efficacy.

CYP2D6*4 HWE Test in Caucasian Patients from Each ITPC Study, by DNA Source, and CombinedDNA SourceStudy NumberNHWEBlood2700.68 4530.03 6130.28 84640.12 930.56 102220.0002 Total8250.0004Tumor51970.41 62170.04 83800.04 Total7940.0037CombinedTotal16190.000006

Citation Format: Kidwell KM, Hertz DL, Leyland-Jones B, Regan MM, Dowsett M, Rae JM. Analysis of the International tamoxifen pharmacogenomics consortium (ITPC) dataset shows that genotyping DNA derived from tumor does not introduce CYP2D6 genotyping error or mask an association with tamoxifen efficacy. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P6-09-02.

Abstract P3-04-02: Absence of PTEN facilitates the anti-tumor efficacy of GDC-0980 in combination with ABT888 plus carboplatin in BRCA1-competent triple negative breast cancer

Introduction: PI3K pathway, in addition to its pro-proliferative and anti-apoptotic effects on tumor cells, is known to contribute to DNA-damage repair (DDR). We hypothesized that GDC-0980, a dual PI3K-mTOR inhibitor, will induce an efficient anti-tumor effect in BRCA-competent PTEN-null TNBC cells when combined with PARP inhibitor, ABT888 and carboplatin. We propose that in PTEN-null BRCA-competent TNBC model, the growth of TNBC tumor will be blocked due to the inhibition of (1) HR and NHEJ and (2) PI3K-mTOR pathway mediated survival signals following treatment with GDC-0980, when combined with PARP inhibitor (impaired DNA-SSB-repair) and carboplatin (increased DNA-DSB). Purpose: Here we tested the efficacy of a combination of GDC-0980 with ABT888 plus carboplatin in BRCA-competent PTEN-null model of TNBC. Methods: Athymic mice bearing PTEN-null TNBC xenograft tumors were treated with GDC-0980 alone or in combination with ABT888 and carboplatin. Results: Dual inhibition of PI3K and mTOR by GDC-0980 alone as well as in the presence of carboplatin plus ABT888 changed the state of the repair of DNA-damage in BRCA-competent PTEN null TNBC cells, which led to increased cellular apoptotic signals in addition to decreased survival/proliferative signals. GDC-0980 treatment led to DNA damage (increased pgH2AX), gain in PAR and a subsequent sensitization of BRCA-competent PTEN-null MDA-MB468 TNBC cells to ABT888 plus carboplatin with a time-dependent (1) decrease in proliferation signals (pAKT T308/S473, pP70S6K, pS6RP), PAR/PARP ratios, PAR/pgH2AX ratios, live/dead cell ratios, cell-cycle progression and clonogenic 3D growth and (2) increase in apoptosis markers (cleaved-caspase 3, 9, BIM, cleaved-PARP and annexinV positivity). These effects are more pronounced in MDA-MB468 than in RAS/RAF mutated MDA-MB231 cells. GDC-0980 alone and in combination with ABT888 plus carboplatin inhibited cell cycle progression, increased apoptosis, and decreased live/dead cell ratios in BRCA-competent PTEN null TNBC cells. GDC-0980 alone and in combination with ABT888 plus carboplatin attenuated anchorage -dependent and -independent clonogenic 3D growth comparatively more in BRCA-competent PTEN-null cells TNBC cells than MDA-MB231 cells. GDC-0980 in combination with ABT888 plus carboplatin blocked the growth of established PTEN-null TNBC tumors as compared to vehicle control(s) with a concomitant decrease in tumor Ki67 and CD31 IHC-stains. Conclusion: This is the first mechanism-based study to demonstrate that in BRCA-competent PTEN-null TNBC model, GDC-0980 enhanced antitumor activity of ABT888, in the presence of carboplatin by inhibiting DDR system in conjunction with the inhibition of PI3K-mTOR pathway-mediated proliferative, and anti-apoptotic signals. Considering (1) the importance of PARP as the target in TNBC, (2) the existence of a large percentage of BRCA-competent TN and/or basal type BC patients and (3) the high frequency of PTEN-null-ness in this subset of BC, this combination merits further investigation.

Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P3-04-02.

Abstract P5-06-01: The PI3K inhibitor GDC-0941 combines with trastuzumab for superior anti-tumor efficacy in HER2+ breast cancer models

Background: PI3K-AKT-mTOR pathway signaling is important for the oncogenic function of HER2. Activating alterations of this pathway are frequently observed in HER2-enriched breast cancer and generally herald a poor response and resistance to trastuzumab (T). Purpose: Targeting the PI3K-AKT-mTOR pathway is an attractive strategy in HER2+ breast cancer that is refractory to trastuzumab. The hypothesis is that the suppression of this pathway by pan-PI3K inhibitor, GDC-0941 may lead to overcome trastuzumab-resistance. Experimental Design: The antiproliferative and HER2-mediated cellular signaling (pAKT, pP70S6K, pS6RP, p4EBP1 and p-ERK) effects of GDC-0941 alone and in combination with T were evaluated in HER2 amplified T-sensitive (BT474), T-resistant (BT474HR), and HER2 amplified/PIK3CA mutated (HCC1954, UACC893) BT cell lines by MTT assay and Western blots. Clonogenic growth was tested by 3D ON-TOP assay and apoptosis markers were also tested. Athymic mice bearing BT474 and BT474HR xenograft tumors were treated with GDC-0941 and T (alone and in combination). Results: (1) GDC-0941 exhibited in vitro cell killing activity in MTT assay with IC50's ranging from 0.35 μM to 1 μm and potency was augmented by the addition of T, (2) inhibition of phosphorylation of AKT(S473, T308), P70S6K, S6RP, and 4EBP1(T37/46, T70) was observed following GDC-0941 treatment, and the combination of GDC-0941 and T more effectively blocked the PI3K-AKT-mTOR pathway, (3) GDC-0941 treatment increased apoptosis markers (CL-CASPASE3 and annexinV positivity), (4) GDC-0941 dose-dependently blocked 3D-ON-TOP clonogenic growth of HER2+ cells. This effect was potentiated in the presence of T and (5) in vivo, the combination of GDC-0941 and T significantly reduced established tumor growth in both sensitive (82%) and resistant (79%) models with concomitant decrease of different PD markers. Conclusions: Our data suggest that 1) therapeutic targeting of the PI3K-AKT-mTOR signaling should be effective in abrogating resistance to T therapy in HER2+ BT, and 2) targeting both the HER2 and the PI3K signaling pathways is an attractive strategy to enhance the clinical efficacy of T therapy, as well as to prevent or delay the development of resistance.

Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr P5-06-01.

Abstract S6-05: High levels of APOBEC3B, a DNA deaminase and an enzymatic source of C-to-T transitions, are a validated marker of poor outcome in estrogen receptor-positive breast cancer

Two recent observations have connected the innate immune DNA cytosine deaminase APOBEC3B to the genetic evolution of breast cancer. First, APOBEC3B was shown to be up-regulated in the majority of breast cancers, and, in breast cancer cell lines, its activity was causally linked to a doubling of the number of C-to-T transitions over time and to a delay in cell cycle progression (1). Second, sequencing of the complete genome of 21 breast cancers independently suggested that APOBEC deaminase activity could be responsible for 2 of 5 mutational imprints identified, which involved clustered (also called kataegis) and dispersed C-to-T transition mutations in the context of 5’TC dinucleotide motifs (2).

In the current study, we addressed a possible association of APOBEC3B expression with outcome in clinical breast cancer. For this we measured using real-time RT-PCR APOBEC3B mRNA levels in 1,491 primary invasive breast cancers and correlated these levels with disease-free survival (DFS), metastasis-free survival (MFS) and overall survival (OS) using univariate and multivariable Cox regression analysis. In addition, we independently validated our findings in available gene expression datasets with appropriate follow-up.

In univariate analyses including all patients, increasing levels of APOBEC3B mRNA analyzed as a continuous variable were significantly associated with shorter DFS, MFS and OS (Hazard Ratio [HR] = 1.29, 1.31 and 1.36, respectively, all P<0.001). To determine the relation of APOBEC3B mRNA expression with the natural course of the disease without the potential confounding effects of systemic adjuvant therapy, we restricted our next analyses to MFS in 829 patients with lymph node-negative disease who had not received any (neo)adjuvant systemic therapy. This analysis showed that APOBEC3B mRNA expression was, in univariate, and in multivariable analysis, including the traditional prognostic factors (age, menopausal status, tumor size, grade and steroid hormone receptors), a marker of pure prognosis specifically in patients with estrogen receptor-positive (ER+) disease (univariate HR = 1.30; P = 0.003; multivariate HR = 1.22, P = 0.042).

To substantiate and validate our findings, we analysed 4 independent available datasets containing in total 5,760 breast cancer cases in which APOBEC3B mRNA expression was measured by probes on microarrays and found that higher APOBEC3B mRNA expression (dichotomised by mean) was significantly associated with poor outcome in all 4 cohorts ([Metabric, 1,491 ER+ cases, HR = 1.82; P<0.001], [Affymetrix compiled dataset-1, 2,407 cases, HR = 2.22; P = 0.001], and [BIG 1-98; 1,207 cases, HR = 2.13; P<0.001 of late recurrence>5 years], and [Affymetrix dataset-2, 643 ER+ cases, HR = 2.04; P = 0.001]).

Altogether, our analyses show that APOBEC3B mRNA - and as a result likely DNA deamination – is a validated predictor of poor outcome in breast cancer, supporting the notion that APOBEC3B is a potentially interesting clinical target for therapeutic intervention to prevent breast cancer progression and metastasis, particularly in ER+ disease.

1. Burns, M.B. et al. Nature 494, 366-70 (2013); 2. Nik-Zainal, S. et al. Cell 149, 979-93 (2012).

Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr S6-05.

Abstract OT3-4-04: InVite: an observational pilot study evaluating the feasibility of genome-wide association studies using self-reported data from patients with metastatic breast cancer treated with bevacizumab

Background: Personalized healthcare tailors treatments to patients and their disease characteristics through the use of genetics and other biomarkers. Genetic differences among individuals may explain variations in drug treatment response, including side effects. With such information physicians could make more informed decisions about drugs and dosing for a given individual, thereby improving patient care. Although there has been some success, to a large extent genetic variation related to drug response remains unexplained.

Bevacizumab, a humanized monoclonal antibody against the angiogenic factor VEGF, has demonstrated activity in patients with metastatic breast cancer (MBC). The InVite study will evaluate the feasibility of performing genomewide association studies using self-reported information collected via an online platform from patients with MBC who have been treated with bevacizumab. Using novel methodology in a convenient, user-friendly, and scientifically rigorous format, InVite ultimately aims to identify potential pharmacogenetic associations in this patient population.

Trial design: InVite is a pilot, non-interventional, observational, web-based, prospective cohort study designed to collect patient-reported safety, efficacy, and genetic data from patients with MBC treated with bevacizumab. Data on demographics, breast cancer disease status, cancer treatment history, bevacizumab-related outcomes, and certain safety events will be collected directly from patients entirely via online surveys. Patients will be asked to complete surveys at the time of enrollment and then every 3 months for 1 year after enrollment. A saliva sample for DNA collection will be gathered using an at-home kit. Evaluations of data quality and collection feasibility will be conducted intermittently. There will be an optional substudy to allow for blood sample collection for DNA analysis.

Eligibility criteria: ≥18 years of age, residing in the US, locally recurrent breast cancer or MBC, currently being or having been treated with bevacizumab starting on or before Dec 31, 2011, fluent in English, and access to a computer with an internet connection.

Specific aims: The primary objective is to assess the feasibility of recruiting subjects and collecting biospecimens and self-reported data online. The secondary objective is to characterize the patient population. Exploratory objectives include analyzing potential associations between genetic polymorphisms and 1) bevacizumab-induced hypertension, the most common bevacizumab-related adverse event, and 2) patient-reported time-to-progression.

Statistical methods: Baseline demographics, clinical and treatment characteristics of enrolled patients will be summarized. Each polymorphism genotyped will be tested for association with the defined endpoint using appropriate statistical modeling.

Present and target accrual: Accrual as of May 23, 2012 is 82 patients. Target accrual is 1000 patients.

Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr OT3-4-04.

Abstract S4-4: Independent validation of Genomic Grade in the BIG 1-98 study

Background: We have reported that a 97-gene signature, the Genomic Grade Index (GGI) can separate histological grade (HG) 2 breast tumors into low vs high GGI tumors with different outcomes. We have also developed a quantitative reverse transcriptase polymerase chain reaction (qRT-PCR) tool including 6 reporter and 3 control genes that can reliably evaluate Genomic Grade (GG) in paraffin embedded tumors.

Methods: We evaluated the qRT-PCR GG in women treated with either tamoxifen or letrozole monotherapy in the Breast International Group (BIG) 1–98 study with 8.1 year median follow-up. The association between continuous GG and distant recurrence-free interval (DRFI) was evaluated in Cox regression models stratified for the 2 vs 4 arm randomization option, chemotherapy and hormonotherapy use, with and without adjustment for clinicopathological characteristics. Estrogen, progesterone receptor (ER, PR), Ki67 and HG were centrally reviewed. The clinicopathological model included age and log2 tumor size, ER and PR as continuous variables, nodal status (N0 vs 1–3 vs >3), HG (1 vs 2 vs 3) and HER2 (negative vs positive). Similar analyses were performed for log2 Ki67 as continuous variable. Added prognostic value was assessed using the likelihood ratio statistic.

Results: We obtained GG in 883 (62%) of 1428 samples. Non evaluable results were due to pre-analytical issues or technical failure. Among the 883 patients, 521 (59%) had N0 disease, 435 (49%) were treated with tamoxifen and 84 (10%) had distant recurrences. The GG classified 502 patients with HG2 tumors as either GG1 (N = 202, 40%), equivocal (N = 220, 44%) or GG3 (N = 80, 16%). In univariate analysis, increasing GG and Ki67 were significantly associated with lower DRFI (Table). When the 773 N0/1-3 patients were analyzed based on HG, the Kaplan-Meier estimate for 10-year DRFI was 98% (95% confidence intervals 96–100%) for the 123 HG1, 92% (88–95%) for the 456 HG2 and 84% (78–90%) for the 194 HG3 patients. In the N0/1-3 population, 10 year DRFI based on GG was 95% (96–100%) for the 290 GG1, 92% (89–95%) for the 309 GG equivocal and 84% (77–90%) for the 178 GG3 patients. Interestingly, when the 456 HG2 and N0/1-3 patients were analyzed based on GG, the 10-year DRFI was 95% (92–100%) for the 185 HG2/GG1 patients, 92% (88–96%) for the 202 HG2/GG equivocal patients, 86% (76–96%) for the 69 HG2/GG3 patients. In all patients either GG or Ki67 as continuous variable significantly improved the clinicopathological model in predicting distant recurrence (Table).

Conclusion: Either Genomic Grade or centrally reviewed Ki67 provides independent prognostic information for risk of distant recurrence beyond clinicopathological characteristics in patients treated with endocrine therapy.

Abstract P5-19-03: Olaparib plus carboplatin in combination with vandetanib inhibited the growth of BRCA-wt triple negative breast tumors in mice: Outside BRCA-box

Introduction: PARP inhibition has emerged as one of the most exciting and promising ‘targeted’ therapeutic strategies in the treatment of advanced triple negative breast cancer (TNBC) – the intended ‘target,’ being DNA repair (Anders CK, 2010). Although olaparib is known to have antitumor activity in BRCA-related TNBC cells, a limited number of preclinical and clinical studies have shown antitumor efficacy of olaparib in non-BRCA-related BC (Shimo T, 2012). Understanding the biology of TNBC cells has identified molecular targets including RTK(s), such as EGFR.

Purpose: Here we tested the combination of a PARP inhibitor, olaparib (O) (AstraZeneca) plus carboplatin (C) with the EGFR/VEGFR inhibitor, vandetanib (V) (AstraZeneca) in a BRCA-wt TNBC model.

Methods: Athymic mice bearing TNBC (BRCA-wt VEGFR expressing MDA-MB231& MDA-MB468 [EGFR amplified/overexpressed]) xenograft tumors (200 mm3) were treated with O plus C in combination with V (Arms: vehicle control 1, vehicle control 2, C+O, V, C+O+V). In vitro effects of V in combination with O plus C (or temozolomide) on clonogenic growth (3D on-TOP assay), proliferation (MTT and CelltiterGLO), apoptosis (Apoptosis Array), cell signaling marker(s) (Western Blot), and tumor cell phenotypes (fibronectin-directed migration, matrigel-invasion, and vascular mimicry) were investigated in a panel of five BRCA-wt and BRCA-mutated TNBC cell lines. The effects of V were tested on (a) cell signaling marker(s), (b) angiogenesis marker (HIF-1alpha), and (c) angiogenesis related phenotypes (vitronectin-directed migration, and cord formation) in HUVEC.

Results: (1) O plus C in combination with V caused a regression of the in vivo growth of established tumors by 50% which was evident in both the BRCA-wt TNBC models tested. Interestingly, a marked suppression of the progression of tumor-growth was observed in the O plus C arm. (2) In vitro, V alone (10 µM) inhibited baseline as well as EGF-induced phosphorylation of AKT (S473/T308), S6RP, 4EBP1 and ERK. (3) TNBC cells exhibited higher sensitivity to V in clonogenic assays when combined with a 10 µM fixed dose of O and C/temozolomide. (4) A combination of V with O plus C increased cleaved caspase-3, PARP cleavage, and pro-apoptotic signals, while inhibiting vascular mimicry, migration, and invasion in MDA-MB231, MDA-MB468, and SUM149 cells. (5) Treatment with V blocked cord formation, migration, EGF-induced HIF-1α accumulation, and phosphorylation of AKT, 4EBP1, and ERK in HUVEC.

Conclusion: A profound anti-tumor efficacy of O plus C in combination with V in BRCA-wt TNBC model can be explained by a significant anti-proliferative/pro-apoptotic and anti-migratory/anti-invasive actions of the drugs (alone or in combination), as observed both in tumor cells as well as in endothelial compartments. The combination of O plus C plus V merits further investigation in TNBC.

Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr P5-19-03.

Abstract P5-18-02: Selective Crossover in Randomized Trials of Adjuvant Trastuzumab for Breast Cancer: Coping with Success

BACKGROUND: Disease-free survival (DFS) is often a primary endpoint of randomized trials of adjuvant therapies for breast cancer, but long-term follow-up of DFS and especially overall survival (OS) remain important. When the primary DFS results favor the experimental arm, patients (pts) assigned to the control group may select the option to crossover to receive the experimental treatment via protocol amendment. Such “selective crossover” disturbs the integrity of the randomized comparison for any efficacy endpoints that rely on further follow-up. Selective crossover, which is motivated by positive results having been observed in the current trial, is distinct from so-called “unplanned crossover,” which refers to non-adherence to protocol. In this abstract, we discuss the consequences of selective crossover for trials evaluating adjuvant trastuzumab, using the HERA (HERceptin Adjuvant) trial as an example, and present a variety of alternative analysis approaches.

METHODS: HERA enrolled 5102 women with HER2-positive early breast cancer who had completed all surgery and (neo)adjuvant chemotherapy to compare 1 or 2 years of trastuzumab treatment vs observation. After a positive first interim analysis at 1y median follow-up (MFU) showed that 1 year of trastuzumab significantly improved DFS vs observation [MJ Piccart-Gebhart et al; NEJM 2005], event-free patients in the observation group were offered crossover to receive trastuzumab. 885 (52%) of the 1698 pts in the observation group selectively crossed over to trastuzumab.

RESULTS: Previously reported intention-to-treat (ITT) analysis of HERA at 4y MFU showed a decreasing effectiveness of trastuzumab with respect to DFS compared with those at 2y MFU [L Gianni et al, Lancet Oncol 2011; I Smith et al, Lancet 2007]. In addition, the ITT analysis of OS at 4y MFU showed little effect of trastuzumab, while the analysis artificially censoring follow-up in the observation group at the time of selective crossover showed a substantial OS advantage for trastuzumab.

Abstract PD10-03: Predictive value of a proliferation score (MS) in postmenopausal women with endocrine-responsive breast cancer: results from International Breast Cancer Study Group (IBCSG) Trial IX

Background: While representing the largest fraction of women diagnosed with primary breast cancer, older postmenopausal women with ER+, HER2− tumors are less responsive to chemoendocrine therapy than younger women and have been underrepresented in molecular profiling of randomized trials. IBCSG Trial IX, a randomized controlled trial in postmenopausal women, median age 61y, with node negative disease, failed to demonstrate the benefit of preceding tamoxifen (T) by 3 cycles of CMF for ER+ tumors. We sought to determine if MS, a proliferation score, could identify a subset of women who differentially benefit from addition of chemotherapy to T in this trial.

Methods: From 1988–1999, 1669 eligible patients (1040 with ER+, HER2− tumors) were randomized to CMF→T vs T. Disease-free survival (DFS) was the primary trial endpoint; breast cancer-free interval (BCFI) which excludes second (non-breast) malignancies and censors deaths without prior cancer event was also evaluated. Analysis was limited to the first 7 years of follow-up. From 671 (ER+, HER2−) available subjects, 568 were successfully profiled by RT-PCR. The mRNA expression levels of 14 equally-weighted proliferation genes and 3 normalization genes were used to generate MS; predetermined binary categorization of MS was used. Analysis of this post hoc, pre-specified study used results from centralized laboratory IHC and Cox models to assess the predictive value of MS on DFS and BCFI, adjusting for traditional risk factors of local treatment, age, ER, PR, Ki67, tumor size and grade.

Results: Subgroups of MS (low, 169 samples (30%) and high, 399 samples (70%)) were identified. MS by treatment interaction was significant for DFS and BCFI (each p ≤ 0.004). Among patients with low MS, CMF→T improved DFS (HR 0.19, 95% CI 0.06–0.59) and BCFI (HR 0.19, 95% CI 0.05–0.72) vs T; 7y DFS was 95% vs 83% with CMF→T vs T. Among patients with high MS, CMF→T did not improve DFS (HR 1.27, 95% CI 0.79–2.05) or BCFI (HR 1.37, 95% CI 0.80–2.33) and 7y DFS of 81% for CMF→T and T. Continuous MS was moderately correlated with log Ki67 (r = 0.47) but not correlated with ER or PR. The MS by treatment interaction remained significant with Ki67 in the model.

Conclusions: Low MS was associated with differential benefit favoring those women receiving CMF→T vs T alone for both DFS and BCFI in the first 7 years. The effect was independent of traditional risk factors including Ki67. Hence this study, which is unconfounded by chemotherapy-induced ovarian ablation in younger women, identifies a subset of postmenopausal women with ER+, HER2− tumors that benefit from CMF chemotherapy. This seemingly incongruous observation is consistent with a) the prior observation that only the low-proliferation subgroup by PAM50 11-gene signature benefits from the addition of weekly paclitaxel to adjuvant FEC (GEICAM/9906), b) the ability of MS to identify a subset of women with tumors with disseminated luminal progenitor cells activated through the agonistic activity of tamoxifen, and c) the repetitive dosing of cyclophosphamide and taxol being hypothesized to act via tumor stroma/anti-angiogenesis. The relative contribution of these factors is under investigation.

Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr PD10-03.

Abstract OT2-3-11: Tivozanib in combination with paclitaxel vs placebo with paclitaxel in patients with locally advanced or metastatic triple-negative breast cancer

Background: Triple-negative breast cancer (TNBC) is an aggressive cancer with inferior survival outcomes. Although weekly paclitaxel (WP) is effective in the treatment (tx) of metastatic breast cancer (MBC), optimization of therapies for patients (pts) with TNBC is essential. Angiogenesis is a hallmark of advanced cancer, with subset analyses suggesting activity of angiogenesis inhibitors in TNBC. Tivozanib (TIVO) is a potent and selective inhibitor of vascular endothelial growth factor receptors (VEGFR) 1, 2, and 3 with a promising role in metastatic renal cell carcinoma, and established safety in Phase I combination with WP in MBC.

Purpose: This Phase II trial will assess the efficacy and safety of TIVO + WP in the first-line setting for pts with advanced or metastatic TNBC and evaluate the performance of candidate angiogenesis biomarkers.

Objectives: The primary objective of this study is to compare progression-free survival (PFS) of pts treated with TIVO + WP vs pts treated with placebo (PB) + WP. Secondary objectives include objective response rate (ORR), overall survival (OS), safety and tolerability, quality of life, and correlative candidate biomarker endpoints. The pharmacokinetics of TIVO + WP also will be characterized.

Study Design and Methods: This multicenter, randomized, PB-controlled, two-arm study will enroll pts with metastatic or unresectable TNBC (evaluable per RECIST) and no prior systemic therapy. Pts must have confirmed available archival tumor tissue. Pts will be stratified by ECOG performance score and number of metastatic sites, then randomized to receive either oral TIVO 1.5 mg once daily for 3 weeks (wks) on/1 wk off and intravenous WP 90 mg/m2 for 3 wks on/1 wk off, or PB + WP. One cycle will be 4 wks; tx will continue until disease progression or unacceptable toxicity. Archival tumor tissue and blood samples will be evaluated for response biomarkers, including a hypoxia sensitivity gene signature, a myeloid resistance gene signature, and angiogenic ligands. All pts will be followed for survival until death. Adverse events will be monitored throughout the study. Pharmacokinetic samples will be collected during cycles 1 and 2. PAM-50–defined intrinsic molecular subtype populations also will be evaluated retrospectively.

Recruitment of 130 patients is planned, with an interim analysis after 80 pts to measure ORR (130 pts with a total of 82 investigator-assessed PFS events provides 80% power to detect statistically significant PFS differences between tx arms). Endpoint analyses will use the intent-to-treat population. The primary efficacy analysis will use investigator assessments of response and a two-sided 95% confidence interval for the hazard ratio produced using Cox proportional hazards regression models. OS will be compared using the log-rank test. Analyses of candidate biomarkers and determination of an optimal predictive cutoff for response also are planned. Trial enrollment will commence in fall 2012.

Conclusion: This study will determine whether TIVO, a selective and potent VEGFR inhibitor, combined with WP improves clinical outcomes in pts with TNBC, and whether clinical activity is associated with candidate angiogenesis biomarkers.

Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr OT2-3-11.