Prognostic value of CEC count in HER2-negative metastatic breast cancer patients treated with bevacizumab and chemotherapy: a prospective validation study (UCBG COMET)

BACKGROUND

Proof of concept studies has reported that circulating endothelial cell (CEC) count may be associated with the outcome of HER2-negative metastatic breast cancer (mBC) patients treated by chemotherapy and the anti-VEGF antibody bevacizumab. We report the results obtained in an independent prospective validation cohort (COMET study, NCT01745757).

METHODS

The main baseline criteria were HER2-negative mBC, performance status 0-2 and no prior chemotherapy for metastatic disease. CECs were detected by CellSearch® from 4 ml of blood at baseline and after 4 weeks of weekly paclitaxel and bevacizumab therapy. CEC counts (considered both as a continuous variable and using the previously described 20 CEC/4 ml cutoff) were associated with clinical characteristics and progression-free survival (PFS).

RESULTS

CEC count was obtained in 251 patients at baseline and in 207 patients at 4 weeks. Median baseline CEC count was 22 CEC/4 ml (range 0-2231). Baseline CEC counts were associated with performance status (p = 0.02). No statistically significant change in CEC counts was observed between baseline and 4 weeks of therapy. High baseline CEC count was associated with shorter PFS in univariate and multivariate analyses (continuous: p < 0.001; dichotomized: HR 1.52, 95% CI [1.15-2.02], p = 0.004). CEC counts at 4 weeks had no prognostic impact.

CONCLUSION

This study confirms that CEC count may be associated with the outcome of mBC patients treated with chemotherapy and bevacizumab. However, discrepancies with previous reports in terms of both the timing of CEC count and the direction of the prognostic impact warrant further clinical investigation.

Impact of breast cancer molecular subtypes on the incidence, kinetics and prognosis of central nervous system metastases in a large multicentre real-life cohort

BACKGROUND

Metastatic breast cancer (MBC) behaviour differs depending on hormone receptors (HR) and human epidermal growth factor receptor (HER2) statuses.

METHODS

The kinetics of central nervous system (CNS) metastases (CNS metastasis-free survival, CNSM-FS) and subsequent patient's prognosis (overall survival, OS) according to the molecular subtype were retrospectively assessed in 16703 MBC patients of the ESME nationwide multicentre MBC database (Kaplan-Meier method).

RESULTS

CNS metastases occurred in 4118 patients (24.6%) (7.2% at MBC diagnosis and 17.5% later during follow-up). Tumours were HER2-/HR+ (45.3%), HER2+/HR+ (14.5%), HER2+/HR- (14.9%) and triple negative (25.4%). Median age at CNS metastasis diagnosis was 58.1 years (range: 22.8-92.0). The median CNSM-FS was 10.8 months (95% CI: 16.5-17.9) among patients who developed CNS metastases. Molecular subtype was independently associated with CNSM-FS (HR = 3.45, 95% CI: 3.18-3.75, triple-negative and HER2-/HR+ tumours). After a 30-month follow-up, median OS after CNS metastasis diagnosis was 7.9 months (95% CI: 7.2-8.4). OS was independently associated with subtypes: median OS was 18.9 months (HR = 0.57, 95% CI: 0.50-0.64) for HER2+/HR+ , 13.1 months (HR = 0.72, 95% CI: 0.65-0.81) for HER2+/HR-, 4.4 months (HR = 1.55, 95% CI: 1.42-1.69) for triple-negative and 7.1 months for HER2-/HR+ patients (p <0.0001).

CONCLUSIONS

Tumour molecular subtypes strongly impact incidence, kinetics and prognosis of CNS metastases in MBC patients.

CLINICAL TRIAL REGISTRATION

NCT03275311.

UCBG 2-04: Long-term results of the PACS 04 trial evaluating adjuvant epirubicin plus docetaxel in node-positive breast cancer and trastuzumab in the human epidermal growth factor receptor 2-positive subgroup

PURPOSE

We conducted a double-randomised phase III trial to evaluate a concomitant taxane-anthracycline regimen in node-positive breast cancer and the efficacy of trastuzumab in the human epidermal growth factor receptor 2 (HER2)-positive subpopulation.

METHODS

A total of 3010 patients with node-positive breast cancer were randomly assigned to receive 6 cycles of 500 mg/m² of fluorouracil, 100 mg/m² of epirubicin and 500 mg/m² of cyclophosphamide (FEC) or 75 mg/m² of epirubicin and 75 mg/m² of docetaxel (ED). Patients with HER2-positive tumours were secondary randomly assigned to either trastuzumab or observation. The primary end-point was disease-free survival (DFS) in the two chemotherapy arms.

RESULTS

After a 115-month median follow-up, DFS was not significantly better in the ED arm (DFS: 70%, 95% confidence interval [CI]: 67-72) than in the FEC arm (DFS: 68%, 95% CI: 65-70; hazard ratio [HR] = 0.88, 95% CI: 0.77-1.01; p = 0.064). The OS was not different between FEC (OS: 80%, 95% CI: 78-83) and ED (OS: 81%, 95% CI: 79-83); HR = 0.97, 95% CI: 0.81-1.16; p = 0.729). ED appeared more toxic. In the 528 HER2-positive subset, there was trend for a higher DFS, in the intention-to-treat population, in the trastuzumab arm (DFS: 68%, 95% CI: 61-74) than in the observation arm (DFS: 60%, 95% CI: 54-66; HR = 0.77, 95% CI: 0.57-1.03; p = 0.079). In the per-protocol population, DFS was significantly higher in the trastuzumab arm (DFS: 70%, 95% CI: 63-76) than in the observation arm (DFS: 59%, 95% CI: 53-65; HR = 0.69, 95% CI: 0.51-0.94; p = 0.0156). The OS was not different between these 2 arms.

CONCLUSION

This study did not show superiority of the concomitant anthracycline-taxane arm which was more toxic in high-risk node-positive breast cancer patients. Long-term results of the HER2-positive subpopulation are in line with those of the other adjuvant trastuzumab trials but quantitatively less pronounced mostly because of lack of power.

[E-health and "Cancer outside the hospital walls", Big Data and artificial intelligence]

To heal otherwise in oncology has become an imperative of Public Health and an economic imperative in France. Patients can therefore receive live most of their care outside of hospital with more ambulatory care. This ambulatory shift will benefit from the digital revolution and the development of digital health or e-health. Cancer research will also benefit with Big Data and artificial intelligence, which gather and analyze a huge amount of data. In this synthesis, we describe the different e-health tools and their potential impacts in oncology, at the levels of education and information of patients and caregivers, prevention, screening and diagnosis, treatment, follow-up, and research. A few randomized studies have already demonstrated clinical benefits. Large Big Data projects such as ConSoRe and Health Data Hub have been launched in France. We also discuss the issues and limitations of "cancer outside the hospital walls and e-health" from the point of view of patients, health care professionals, health facilities and government. This new organization will have to provide remote support "outside the walls" with care and follow-up of quality, continuous and prolonged in total safety and equity. Ongoing and future randomized clinical trials will need to definitively demonstrate areas of interest, advantages and drawbacks not only for patients, but also for caregivers, health facilities and governments.

First Measurement of Near-Threshold J/ψ Exclusive Photoproduction off the Proton

We report on the measurement of the γp→J/ψp cross section from E_{γ}=11.8  GeV down to the threshold at 8.2 GeV using a tagged photon beam with the GlueX experiment. We find that the total cross section falls toward the threshold less steeply than expected from two-gluon exchange models. The differential cross section dσ/dt has an exponential slope of 1.67±0.39  GeV^{-2} at 10.7 GeV average energy. The LHCb pentaquark candidates P_{c}^{+} can be produced in the s channel of this reaction. We see no evidence for them and set model-dependent upper limits on their branching fractions B(P_{c}^{+}→J/ψp) and cross sections σ(γp→P_{c}^{+})×B(P_{c}^{+}→J/ψp).

Abstract CT140: Circulating tumor cells as a tool to guide first line therapy in metastatic breast cancer: subgroup analyses of the STIC CTC Phase III utility trial

Background: Endocrine therapy (ET) is the preferred 1st line therapy in ER+ HER2- metastatic breast cancer (MBC) patients, but not all benefit from such strategy for which predictive biomarkers are lacking. Prior studies showed that circulating tumor cells (CTC) count is the strongest prognostic factor beyond PS in that population. The STIC CTC trial (NCT01710605) assessed whether CTC count could drive the choice between frontline ET or chemotherapy (CT, with or without ET maintenance).

Methods: For this multicenter phase 3 non-inferiority trial, main inclusion criteria were: ER+ HER2- MBC with no prior therapy, PS≤2, no contra-indication to ET or CT, informed consent. Patients were randomized 1:1 between clinically-driven choice (single agent ET if clinicallow, CT if clinicalhigh, as decided by physician based on clinical factors, blinded to CTC results) or a CTC-driven choice [ET if CTClow (&lt;5 CTCs/7.5mL), CT if CTChigh (≥5 CTCs/7.5mL)]. The primary objective was treatment efficacy ([PFS hazard ratio (HR)], non-inferiority being established if the upper bound of the PFS HR 2-sided 90%CI is &lt;1.25; secondary objectives included subgroup analyses (CTC status, patient characteristics) and OS.

Results: In this trial, 778 patients were randomized between both strategies. Of all patients, 71.0%, 26.6% and 2.4% were considered as endocrine-sensitive, with secondary or with primary endocrine resistance, respectively. In both arms (clinically- and CTC-driven), prognostic evaluation by physician/CTCs and allocated treatments were as follows: (i) clinicallow/CTClow 46.5% and 48.6% of patients, all treated with ET. (ii) clinicalhigh/CTChigh 14.0% and 13.0%, all treated with CT. (iii) clinicallow/CTChigh 26.1% and 24.2%, treated with ET or CT respectively. (iv) clinicalhigh/CTClow 13.4% and 14.0%, treated with CT or ET respectively. CTC-driven strategy met the primary endpoint, yielding a non-inferior PFS (median: 17 months; 95%CI [15.4-20.3] vs 18 months 95%CI[13.9-23.3]). In the 2 discordant subgroups, preplanned analyses showed the following results: in the clinicallow/CTChigh subgroup, patients treated with CT had a significantly longer PFS than those treated with ET (15.6 months, 95%CI [12.2-22.7] vs 10.5 months, 95%CI [7.3-15.4], log rank p=0.002). In the clinicalhigh/CTClow subgroup, patients treated with CT had a non-significant PFS advantage over patients treated with ET. Pooling these two subgroups of patients (N=292) with discordant treatment recommendations (depending on clinician or CTC count standpoint), an exploratory analysis showed that patients treated with CT frontline had significantly longer PFS (HR=0.66; 95%CI [0.51-0.85]) and OS (HR=0.65; 95%CI [0.43-0.98]); 2-year OS were 82.9% versus 74.7% in patients treated with CT (±maintenance ET) or ET, respectively.

Conclusion: This trial demonstrates the utility of CTC count to select 1st line therapy in ER+ HER2- MBC patients, especially in those for whom single agent ET is the recommended therapy based on clinical factors. With this modern prognostic biomarker, the STIC CTC trial is the first one to identify potential ER+ HER2- MBC patients who might derive more benefit from frontline CT followed by maintenance ET than from frontline ET, challenging current standards.

Funding: French Ministry of Health (PSTIC 2011); Menarini SB; Institut Curie.

Citation Format: Francois-Clement Bidard, William Jacot, Sylvain Dureau, Etienne Brain, Thomas Bachelot, Hugues Bourgeois, Anthony Goncalves, Sylvain Ladoire, Herve Naman, Florence Dalenc, Joseph Gligorov, Marc Espie, Christelle Levy, Jean-Marc Ferrero, Delphine Loirat, Paul Cottu, Veronique Dieras, Marie-Emmanuelle Legrier, Frederique Berger, Catherine Alix-Panabieres, Jean-Yves Pierga. Circulating tumor cells as a tool to guide first line therapy in metastatic breast cancer: subgroup analyses of the STIC CTC Phase III utility trial [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2019; 2019 Mar 29-Apr 3; Atlanta, GA. Philadelphia (PA): AACR; Cancer Res 2019;79(13 Suppl):Abstract nr CT140.

Real-life activity of eribulin mesylate among metastatic breast cancer patients in the multicenter national observational ESME program

Eribulin mesylate (EM) was recently approved for metastatic breast cancer (MBC) chemotherapy (CT) in late lines by the FDA, with debated results in second line. We evaluated outcomes in breast cancer patients receiving EM as second, third and fourth line in a national real-life cohort of 16,703 consecutive MBC patients initiating their first metastatic therapeutic line between 2008 and 2014. Primary and secondary objectives were overall survival (OS) and progression-free survival (PFS). An imbalance was seen for HER2+ tumors and concomitant anti-HER2 targeted therapies use, we thus performed a subanalysis in HER2- patients. PFS and OS were significantly better in EM patients in third and fourth lines, compared to "Other chemotherapies" patients (PFS: 4.14 vs. 3.02 months, p = 0.0010; 3.61 vs. 2.53 months, p = 0.0102, third and fourth-line; OS: 11.27 vs. 7.65 months, p = 0.0001; 10.91 vs. 5.95 months, p < 0.0001, third and fourth-line). No significant difference was reported in second-line (PFS: 5.06 vs. 4.14 months, p = 0.1171; OS: 13.99 vs. 11.66 months, p = 0.151). Among HER2- patients, a significant difference was seen for all lines, including 2nd-line (PFS: 4.57 vs. 3.91 months, p = 0.0379; OS: 14.98 vs. 10.51 months, p = 0.0113). In this large real-world database, HER2-negative MBC patients receiving EM in second or later CT line presented significantly better PFS and OS. This difference disappeared in second line in the overall population, probably because of the imbalance in HER2-targeted treatments use. Our results mirror those of the published randomized trials. The effect of anti-HER2 therapies addition in this setting still needs to be defined.

Express study: A trial in progress exploring the association between low level of genomic alteration and exceptional and unexpected response to targeted therapies in patients with solid tumors

TPS3159

Background: Molecular targeted agents (MTA) resulted in breakthroughs in selected niches. It is often assumed that tumor regression is consecutive to an oncogenic de-addiction effect. An emerging hypothesis suggests that genomic instability may be associated with poor response to MTA. Indeed, the accumulation of defects in multiple oncogenes or tumor suppressor genes may result in the activation of multiple oncogenic pathways. These multiple signaling would mechanically result in a limitation of the oncogenic de-addiction process. Another hypothesis, suggests that tumor heterogeneity could also be associated with poor outcome under MTA. Such heterogeneity could also result from the genomic instability, and be appraised by bioinformatic and functional approaches. In this study, we thought to investigate whether molecular profiles reflecting a low level of genomic alterations in genes causally implicated in oncogenesis could be associated with an exceptional response (ER) to MTA. Methods: This is an exploratory, multicenter, multicohort, prospective trial conducted in 264 adult patients, with advanced breast, lung, colorectal, ovarian, kidney cancers and melanoma, having presented an ER to an approved MTA. ER is defined using the definition chosen by the NCI which combines the three criteria: - complete or partial response, - lasting > 6 months, - and not expected in > 10% of the patients in this drug – organ situation. The primary objective is to assess whether ER can be associated with a low level of genomic instability in the tumor. Low genomic instability is defined by the presence of less than the 5th quantile of genomic alterations (mutations, amplifications, deletions) to be expected in the given tumor type as per TCGA database. For each tumor type, the null hypothesis H0: π = 0.05 will be tested, against the one-sided alternative hypothesis π > 0.05. For each of the 6 cohorts, a sample size of 44 patients is necessary to achieve 80% power at π = 15 with a one-sided level 5% test. Patients presenting an ER will be identified retrospectively, in a nationwide manner, then monthly reviewed and validated for inclusion by a panel of pathology experts. As of February 2019, 75 patients have been included. The identification of molecular traits associated with ER might serve the development of predictive classifiers for precision medicine. This study also represents a unique opportunity to better understand cancer biology. Clinical trial information: NCT02701907.

TAKTIC: A prospective, multicenter, uncontrolled, phase IB/II study of LY2780301 (LY) in combination with weekly paclitaxel (wP) in HER2-negative locally advanced (LA) or metastatic breast cancer (MBC) patients

1091

Background: Phosphatidylinositol-3-kinase (PI3K)/AKT/mammalian target of rapamycin (mTOR)-pathway is frequently activated in HER2-negative breast cancer and may play a role in taxane-resistance. LY is a dual inhibitor of p70 S6 kinase and AKT. TAKTIC study aimed to determine the recommended phase II dose (RP2D) of the combination of LY with wP (phase Ib) and to estimate overall response rate (ORR) of this regimen (phase II) in HER2-negative LA or MBC patients, both in the overall patient population and in patients with activation of PI3K/AKT pathway (PI3KAKT+). Methods: HER2-negative inoperable LA or MBC patients (pts), with (phase Ib) or without (phase II) previous cytotoxic treatment for advanced disease were eligible. Oral LY (400 or 500mg) was administered daily in combination with intravenous wP (70 or 80mg/m2). A modified CRM using an adaptive Bayesian model guided the dose escalation of both agents. PI3KAKT+ was defined as activating mutation of PIK3CA and/or AKT by targeted NGS or homozygous loss of PTEN by array comparative genomic hybridization (aCGH) or loss of PTEN by immunohistochemistry, as evaluated on available fresh tumor tissue. Results: A total of 12 and 35 patients (pts) were included in the phase Ib and II, respectively. In phase Ib, only 1 dose-limiting toxicity (confusion) was observed at the last dose level (LY, 500 + wP, 80), which was determined as RP2D. Main drug-related adverse events (AE) in phase Ib were skin toxicity (92% of pts, G3-4 in 33%), and paresthesia (50% of pts, G3-4 in 8%). In the phase II study, ORR was 62.9 % [44.9,78.5] including 1 CR and 21 PR in the overall population and 55.6 % [30.8,78.5] in PI3KAKT+ pts (10 PR in 18 pts). Median progression-free survival was 12.4 months [7.9,17.9] and 6-month clinical benefit rate was 82.9% [66.4,93.4]. AEs in phase II were similar to phase I part, except that 17% of pts experienced pneumonia (G3-4 in 9%). Conclusions: Combining LY and wP in HER2-negative LA or MBC was feasible with preliminary evidences of efficacy, independently of PI3K/AKT activation. Clinical trial information: NCT01980277.

Outcomes of talazoparib (TALA) versus physician's choice of chemotherapy (PCT) in patients (pts) with advanced breast cancer (ABC) and a germline BRCA (gBRCA) mutation by line of chemotherapy (CT) in the EMBRACA trial

1071

Background: The PARP inhibitor TALA was approved in the US for treatment of g BRCA-mutated ABC based in part on the EMBRACA study. Understanding the outcomes of EMBRACA pts relative to prior CT is a current unmet need. Methods: EMBRACA was a randomized Phase 3 trial comparing TALA 1 mg daily vs PCT (capecitabine, eribulin, gemcitabine, vinorelbine) in g BRCA-mutated ABC. Clinical outcomes were assessed by line of prior CT for ABC in intent-to-treat (ITT), triple-negative breast cancer (TNBC), and hormone receptor-positive (HR+) breast cancer cohorts. Results: 431 pts were randomized (ITT; TALA 287; PCT: 144). TALA was generally more effective than PCT across efficacy endpoints regardless of line of CT (Table). For the ITT population, TALA improved progression-free survival (PFS) and objective response rate (ORR) vs PCT for each line of CT assessed. Other prespecified subgroups (TNBC and HR+) will be presented. Conclusions: In pts with g BRCA-mutated ABC, TALA demonstrated improvements in clinical outcomes compared with PCT regardless of prior lines of CT. Clinical trial information: NCT01945775. [Table: see text]

Multimodality liquid biopsy for early monitoring and outcome prediction in first-line metastatic HER2-negative breast cancer: Final results of the prospective cohort from the French Breast Cancer InterGroup Unicancer (UCBG)— COMET study

3019

Background: Circulating Tumor Cells (CTC) are independent markers of progression-free survival (PFS) and overall survival (OS) in patients (pts) with metastatic breast cancer (MBC). Monitoring circulating tumor DNA (ctDNA) can detect mutation associated with resistance to treatment and its variations reflect changes in tumor burden. We prospectively monitored CTC, Circulating Endothelial Cells (CEC), serum markers and ctDNA during first line chemotherapy for MBC. Methods: The French cohort COMET is a prospective study including first line HER2 negative pts receiving weekly paclitaxel and bevacizumab . Blood samples were obtained at baseline (BL) and before the second cycle of chemotherapy (C2).We present here the final planned analysis. Results: From 09/2012 to 11/2014, 286 patients were included: 198 for ctDNA, 251 for CEC and 283 for CTC. Median age was 56 years and 23% of pts had triple negative BC. At baseline, 71% of pts had ≥1 detectable CTC per 7.5 ml of blood (median 4 CTC, range 1- 30,000). With a threshold of ≥5 CTC, 49% of pts were positive at baseline and 22% at C2. For ctDNA, out of the first 196 pts analyzed, 147 had at least one somatic mutation (SNV) detected in plasma (75%). The average number of mutations per pt was 2.4 (range 1 to 9). Most commonly mutated genes were TP53 and GATA3. ESR1 was mutated in 10.6% of the pts and restricted to the ER+ subgroup. PIK3CA was mutated in 23.2% of the pts. Median Allelic Frequency was 9.1% . Only 68 pts (36%) had detectable ctDNA at C2. At baseline, CTC and ctDNA levels were correlated (r = 0.40, p < 0.0001). Despite no complete overlap, 24 pts (12%) had no CTC nor ctDNA detected at baseline. Median follow-up was 53 months and median OS was 32 months. Detectable CTC and ctDNA at baseline and at C2 were significantly associated with decreased PFS and OS. CEC and serum markers level had no prognostic value. At multivariate analysis, triple negative status, detectable ctDNA at C2, CTC ≥5 at C2 and grade 3 on primary tumor were independent prognostic factors. Conclusions: This is the largest prospective cohort assessing the respective prognostic values of early CTC and ctDNA changes in homogenously treated first line MBC pts. Early decrease of CTC and ctDNA after one cycle of chemotherapy are independent predictive markers of favorable outcome, with a stronger value for ctDNA compared to CTC. Clinical utility of early ctDNA variations monitoring and changes in mutation profile remain to be demonstrated. Clinical trial information: NCT01745757.

RUBY: A phase II study testing rucaparib in germline (g) BRCA wild-type patients presenting metastatic breast cancer (mBC) with homologous recombination deficiency (HRD)

1092

Background: PARP-inhibitors improve PFS in mBC patients (pts) harboring a gBRCA mutation (mut). However, unlike ovarian cancer, there is no evidence until now that this class of agents has efficacy in gBRCA wild-type (WT) pts. In RUBY, we evaluated rucaparib in gBRCA WT mBC pts, and whose tumor present with HRD as assessed by genome-wide Loss of Heterozygosity (LOH) score. Methods: 713 gBRCA WT women with HER2- mBC, initiating a first metastatic chemo, were screened for high (≥18%) genomic tumor LOH generated from a SNP array on metastatic sample. Eligible pts with a high LOH score or somatic (s) BRCA mut and ≥1 prior chemo regimen were proposed to enter RUBY and receive oral rucaparib 600 mg BID continuously in 28-day cycles until disease progression. The primary endpoint was clinical benefit rate (CBR), defined by complete (CR) and partial response (PR) or stable disease (SD) ≥16 weeks. We used a Simon’s two-stage design (p0=20%; p1=40%), responses in ≥4/17 pts were expected to move to second step, and ≥11/37 pts to be considered of clinical interest (α=10% and power of 90%). Whole genome sequencing (WGS) was performed retrospectively to further assess potential biomarkers of PARP inhibitor response. Results: Tumors from 221 (31%) pts were LOH high. 41 pts were enrolled, including 4 pts with sBRCA mut. Median prior metastatic chemo lines was 2 (1-5), 17 pts had TNBC at diagnosis. As of 14 Jan 2019, 16 pts were alive, 5 are still on treatment. The median number of cycles was 2 (1 -20), and 37/40 patients were evaluable for CBR. 5 pts (13.5%) demonstrated clinical benefit (1 CR [LOH high], 3 PR [2 LOH high, 1 sBRCA2] and 1 SD>31 weeks [sBRCA1]). 19 pts had grade 3-4 toxicities. 3 pts discontinued due to toxicity. 4/5 responders pts had their tumor profiled by WGS: preliminary analyses showed that 4 pts presented high large scale state transitions, and 3 presented a somatic biallelic loss of function in HR-related genes. The fifth responder harbored a mut on gPALB2 and sBRCA2 at inclusion. Conclusions: In this study, rucaparib demonstrates antitumor activity in a subset of gBRCA WT mBC pts whose tumor has high LOH. Final analyses of WGS will provide insights about HRD signatures and drivers alterations associated with response. Clinical trial information: NCT02505048.

A Comparison of DNA Mutation and Copy Number Profiles of Primary Breast Cancers and Paired Brain Metastases for Identifying Clinically Relevant Genetic Alterations in Brain Metastases

Improving the systemic treatment of brain metastases (BM) in primary breast cancer (PBC) is impaired by the lack of genomic characterization of BM. To estimate the concordance of DNA copy-number-alterations (CNAs), mutations, and actionable genetic alterations (AGAs) between paired samples, we performed whole-genome array-comparative-genomic-hybridization, and targeted-next-generation-sequencing on 14 clinical PBC–BM pairs. We found more CNAs, more mutations, and higher tumor mutational burden, and more AGAs in BM than in PBC; 92% of the pairs harbored at least one AGA in the BM not observed in the paired PBC. This concerned various therapeutic classes, including tyrosine-kinase-receptor-inhibitors, phosphatidylinositol 3-kinase/AKT/ mammalian Target of Rapamycin (PI3K/AKT/MTOR)-inhibitors, poly ADP ribose polymerase (PARP)-inhibitors, or cyclin-dependent kinase (CDK)-inhibitors. With regards to the PARP-inhibitors, the homologous recombination defect score was positive in 79% of BM, compared to 43% of PBC, discordant in 7 out of 14 pairs, and positive in the BM in 5 out of 14 cases. CDK-inhibitors were associated with the largest percentage of discordant AGA appearing in the BM. When considering the AGA with the highest clinical-evidence level, for each sample, 50% of the pairs harbored an AGA in the BM not detected or not retained from the analysis of the paired PBC. Thus, the profiling of BM provided a more reliable opportunity, than that of PBC, for diagnostic decision-making based on genomic analysis. Patients with BM deserve an investigation of several targeted therapies.

Abstract P2-06-04: Enhancing abiraterone acetate efficacy in androgen receptor-positive triple negative breast cancer: Chk1 as a potential target

Purpose: Our aim was to identify predictive factors of abiraterone acetate (AA) efficacy and putative new druggable targets in androgen receptor (AR)-positive triple-negative breast cancer (TNBC) treated in the UCBG 2012-1 trial.

Material and methods: We defined AA response as either complete or partial response, or stable disease at 6 months. Using Ampliseq, we sequenced 91 general and breast cancerassociated genes from the tumor DNA samples. We analyzed transcriptomes from the extracted RNA samples on a Nanostring platform and performed immunohistochemistry (IHC) on tumor samples using tissue microarrays. We assessed AA and CHK1 inhibitors (GDC-0575 and AZD7762) efficacies, either alone or in combination, on cell lines grown in vitro and in vivo.

Results: Classical IHC apocrine markers, including AR, FOXA1, GGT1 and GCDFP15, allowed identifying AA responders and non-responders. All responders have clear apocrine features. Transcriptome analysis revealed that 31 genes were differentially expressed in the two subgroups, 9 of them being linked to proliferation and DNA damage repair. One of the most significant differences was the overexpression in non-responders of CHEK1, a gene encoding Chk1, a protein kinase that can be blocked by specific inhibitors. In vitro, AA and Chk1 inhibitor combination showed additive or slightly synergistic effect on cell viability, cell cycle, apoptosis and accumulation of DNA damages. In vivo, orthotopic xenograft experiments confirmed the efficacy of this combination therapy.

Conclusions: This study suggests that apocrine features can be helpful in the identification of AA-responders. We identified Chk1 as a putative drug target in AR-positive TNBCs.

Citation Format: Grellety T, Callens C, Richard E, Pulido M, Goncalves A, Gestraud P, MacGrogan G, Bonnefoi H, Cardinaud B. Enhancing abiraterone acetate efficacy in androgen receptor-positive triple negative breast cancer: Chk1 as a potential target [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 P2-06-04.

Abstract GS3-07: Clinical utility of circulating tumor cell count as a tool to chose between first line hormone therapy and chemotherapy for ER+ HER2- metastatic breast cancer: Results of the phase III STIC CTC trial

Background: In ER+ HER2- metastatic breast cancer (MBC) patients, the clinical choice between 1st line hormone therapy (HT, the recommended option) or chemotherapy (CT) is based on the absence of “visceral crisis” or adverse prognostic factors, with no proven/objective criteria. In that context, STIC CTC (NCT01710605) was set up as a strategy trial to test whether circulating tumor cells (CTC) count could help customize the choice between 1st line HT or CT.

Methods: For this multicenter phase 3 non-inferiority trial, the main inclusion criteria were: ER+ HER2- MBC with no prior therapy, PS≤2, no contra-indication to HT or CT and informed consent. The a priori treatment choice (HT or CT) and CTC count (CellSearch®) were obtained in all patients prior to randomization. Patients were randomized 1:1 between clinically-driven choice (CTC count not disclosed, HT or CT administered as decided a priori), or a CTC-driven choice (HT if &lt;5 CTC/7.5ml, CT if ≥5 CTC/7.5ml). The primary objective was treatment efficacy (PFS hazard ratio), non-inferiority being established if the upper bound of the PFS HR 2-sided 90%CI is ≤1.25; secondary objectives included subgroup analyses (CTC status, patient characteristics) and OS.

Results: 761 MBC patients were randomized between 02/2012 and 08/2016. Baseline characteristics: 7.8% of patients had a PS=2, 24.1% had a de novo metastatic disease; 63.3% received prior adjuvant HT and 49.9% prior adjuvant CT; 31.3% had ≥3 metastatic sites. A priori treatments (HT or CT) and CTC count (&lt; or ≥5 CTC/7.5ml) were well balanced between the two arms. After randomization, in the clinically-driven arm, N=267 (72.4%) patients received HT and N=102 (27.6%) CT (as decided a priori). In the CTC-driven arm: (1) the a priori choice of HT was confirmed by a low CTC count in N=181 (67.5%) of patients, while N=87 (32.5%) were switched to CT due to a high CTC count; (2) the a priori choice of CT was confirmed by high CTC count in only N=48 (48%) patients, while N=52 (52%) were switched to HT. The primary endpoint was met, PFS being not inferior in the CTC-driven arm (HR=0.98, 90%CI=[0.84–1.13]). Analyses focusing on discordant subgroups showed that for patients with a priori choice of HT but with high CTC count (leading to a switch to CT in the CTC-arm), PFS was significantly longer in the CTC-driven arm than in the standard arm (HR=0.67, 95%CI=[0.49–0.92]; p=0.01), with a non-significant trend toward longer OS (HR=0.68, 95%CI=[0.44–1.07]; p=0.09). Inversely, for patients with a priori choice of CT but with low CTC count (i.e. de-escalation to HT in the CTC arm), PFS was not statistically different between the two arms.

Conclusion: This trial demonstrates the clinical utility of CTC count as an objective decision tool when considering 1st line therapy in ER+ HER2- MBC. In most patients, CTC count did confirm the a priori clinical choice; however, trial results show that in discrepant cases, CTC count may be trusted for either escalating (i.e. considering CT in patients if high CTC count) or de-escalating (i.e. considering HT in patients if low CTC count) 1st line therapy.

Funding: French National Cancer Institute; Menarini Silicon Biosystems.

Citation Format: Bidard F-C, Jacot W, Dureau S, Brain E, Bachelot T, Bourgeois H, Goncalves A, Ladoire S, Naman H, Dalenc F, Gligorov J, Espie M, Levy C, Ferrero J-M, Loirat D, Cottu P, Dieras V, Simondi C, Berger F, Alix-Panabieres C, Pierga J-Y. Clinical utility of circulating tumor cell count as a tool to chose between first line hormone therapy and chemotherapy for ER+ HER2- metastatic breast cancer: Results of the phase III STIC CTC 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 GS3-07.

Abstract PD2-03: Circulating tumor DNA (ctDNA) and circulating tumor cells (CTC) predictive value in HER2 negative metastatic breast cancer patients treated with first line weekly paclitaxel and bevacizumab: Results of a prospective cohort from the French Breast Cancer InterGroup Unicancer (UCBG): COMET study

Background: Increased levels of CTC and a persistent elevated level after just one cycle of chemotherapy are very strong and independent markers of worse progression-free survival (PFS) and overall survival (OS) in patients (pts) with metastatic breast cancer (MBC) (Bidard et al, Lancet Oncol 2014). ctDNA can be used to detect mutation associated with resistance to treatment. It has also been shown that dynamic changes in ctDNA levels closely reflect changes in tumor burden. We prospectively monitored CTC and ctDNA early variations during first line chemotherapy for MBC.

Patients & methods: The French cohort COMET is a prospective study including first line HER2 negative patients (pts) receiving weekly paclitaxel and bevacizumab according to EMA approved combination. The aim of this cohort is to evaluate clinical, biological and radiological parameters associated with pts outcome (CTC, serum markers, ctDNA, pharmacogenomic polymorphisms, metabolomic parameters, visceral fat, serum estradiol level and quality of life). We present here the first planned analysis on pts evaluated for CTC (CellSearch) and ctDNA using targeted sequencing (Roche SeqCap technology) of a panel of 46 genes and 8 promoters, using unique molecular identifiers to increase ctDNA detection sensitivity. Blood samples were obtained at baseline (BL) and before the second cycle of chemotherapy (C2).

Results: From 09/2012 to 5/2014, 218 pts were included in this substudy. Median age was 55 years and 22% of pts had triple negative BC. At BL, 70% of pts had ≥1 detectable CTC per 7.5 ml of blood (median 4 CTC, range 1- 30,000) and 37% at C2. With a threshold of ≥5 CTC, 47% of pts were positive at BL and 22% at C2. For ctDNA, out of the first 141 pts analyzed, 105 had at least one somatic mutation detected in plasma (74%). The average number of mutations per pt was 2.7 and most commonly mutated genes were TP53 and PIK3CA. ESR1 was found mutated in 9% of all cases and restricted to the ER+ subgroup. Median Allelic Frequency was 10% (range 0.6-83%). Only 33% of pts had detectable ctDNA at C2. At BL, CTC and ctDNA levels were correlated (r=0.46, p&lt;0.0001). Despite no complete overlap, 11% of pts had no CTC nor ctDNA detected. Median follow-up was 53 months and median OS was 32 months. Increased level of CTC and ctDNA were significantly associated with decreased PFS and OS. At C2, ≥5 CTC or still detectable ctDNA were strong markers of reduced OS: HR 4.6 (CI95 3.1-7) and HR 3.2 (CI95 1.8 – 5.5), respectively (both p&lt; 0.0001). At multivariate analysis for PFS, detectable ctDNA at C2 and triple negative status were the only significant prognostic factors. None of serum marker level at BL or their early variations had prognostic value.

Conclusion: This is the largest prospective cohort assessing the respective prognostic values of early CTC and ctDNA changes in homogenously treated first line MBC patients. Analysis of mutations profile variations and comparison with primary tumor and metastasis biopsies are ongoing and may reveal early mechanisms of resistance.

Citation Format: Pierga J-Y, Silveira A, Lorgis V, Tanguy M-L, Tredan O, Dubot C, Jacot W, Goncalves A, Debled M, Levy C, Ferrero J-M, Jouannaud C, Luporsi E, Mouret-Reynier M-A, Dalenc F, Lemonnier J, Berger F, Proudon C, Bidard F-C. Circulating tumor DNA (ctDNA) and circulating tumor cells (CTC) predictive value in HER2 negative metastatic breast cancer patients treated with first line weekly paclitaxel and bevacizumab: Results of a prospective cohort from the French Breast Cancer InterGroup Unicancer (UCBG): COMET 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 PD2-03.

Abstract P6-21-01: Xentuzumab (BI 836845), an insulin-like growth factor (IGF)-neutralizing antibody (Ab), combined with exemestane and everolimus in hormone receptor-positive (HR+) locally advanced/metastatic breast cancer (LA/mBC): Randomized phase 2 results

Background:

Xentuzumab (Xen), an IGF-1/-2-neutralizing Ab, binds IGF-1 and IGF-2, inhibits their growth-promoting signaling, and suppresses AKT activation by everolimus (Ev). This Phase 1b/2 trial evaluates Xen in combination with Ev and exemestane (Ex) in HR+/HER2− LA/mBC.

Methods:

The two-arm, open-label, randomized Phase 2 part enrolled female patients (pts) with HR+/HER2− LA/mBC not amenable to curative therapy and refractory to nonsteroidal aromatase inhibitors. Pts were randomized (1:1) to: oral Ev (10 mg/d) + Ex (25 mg/d); or Xen (1000 mg/wk iv) + Ev (10 mg/d) + Ex (25 mg/d). Randomization was stratified by visceral metastases (VM; Y vs N). Treatment continued in 28-day cycles until progression, intolerable adverse events (AEs) or other reasons for discontinuation. Primary endpoint was progression-free survival (PFS), with an interim futility analysis incorporated in the study design.

Results:

Following the results of the interim analysis, the Data Monitoring Committee (DMC) advised early termination of the trial and discontinuation of Xen treatment. Thus, Xen treatment exposure time and time-to-event data for the Xen+Ev+Ex arm are limited. Of the 139 women treated (Xen+Ev+Ex 70; Ev+Ex 69), 77% had VM. Median PFS was not significantly different between arms (Xen+Ev+Ex vs Ev+Ex, 7.3 vs 5.6 months; HR [95% CI] 0.97 [0.57–1.65]; p=0.91). In a pre-specified subgroup of pts without VM, Xen+Ev+Ex showed favorable PFS vs Ev+Ex (HR 0.21 [0.05–0.98]; Pint=0.0141). Pint values &lt;0.05 were also observed for ad hoc subgroups: measurable disease at baseline; bone-only metastases. Rates of total AEs/grade ≥3 AEs/drug-related AEs were similar between arms (Xen+Ev+Ex, 100/60/96%; Ev+Ex, 99/58/96%). The most common AEs overall were diarrhea (44 vs 33%), mucosal inflammation (39 vs 32%), rash (34 vs 33%) and stomatitis (34 vs 38%); most were grade 1/2. 6% of pts in the Xen+Ev+Ex arm discontinued Xen due to AEs. Ev/Ex discontinuations (Xen+Ev+Ex vs Ev+Ex) occurred in 13/6% vs 23/6%; 1 pt each in the Xen+Ev+Ex arm died from pneumonitis and liver injury and 1 pt each in the Ev+Ex arm died from Burkitt's lymphoma, acute kidney injury and metastases to the peritoneum.

Conclusion:

In the overall population, PFS did not improve with the addition of Xen to Ev+Ex and the trial was therefore discontinued early. Nevertheless, a favorable signal was observed in the pre-specified subgroup of pts without VM when treated with Xen+Ev+Ex, which warrants additional investigation. The safety profile was comparable between arms.

Citation Format: Crown J, Sablin M-P, Cortés J, Bergh J, Im S-A, Lu Y-S, Martínez N, Neven P, Lee KS, Morales S, Pérez-Fidalgo JA, Adamson D, Goncalves A, Prat A, Jerusalem G, Schlieker L, Espadero R-M, Bogenrieder T, Chin-Lun Huang D, Schmid P. Xentuzumab (BI 836845), an insulin-like growth factor (IGF)-neutralizing antibody (Ab), combined with exemestane and everolimus in hormone receptor-positive (HR+) locally advanced/metastatic breast cancer (LA/mBC): Randomized phase 2 results [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 P6-21-01.

Evaluation of the Incidence of Hematologic Malignant Neoplasms Among Breast Cancer Survivors in France

IMPORTANCE

Breast cancer survivors are at an increased risk of developing certain types of hematologic malignant neoplasm after diagnosis.

OBJECTIVE

To estimate the incidence of various types of hematologic malignant neoplasm in breast cancer survivors, both in absolute terms and in association with the general population.

DESIGN, SETTING, AND PARTICIPANTS

This nationwide cohort study conducted in France used data from the French National Health Data System, a database that contains all of French residents' health-related expenses. All French women aged 20 to 85 years with an incident breast cancer diagnosis between July 1, 2006, and December 31, 2015, were included (n = 439 704) and followed up until hematologic malignant neoplasm occurrence, death, loss of follow-up, or December 31, 2016, whichever came first. Comparisons were made with all French women in the general population who were registered in the French general health insurance program each year from January 1, 2007, and December 31, 2016. Data analysis was performed from January 23, 2018, to May 25, 2018.

MAIN OUTCOMES AND MEASURES

Main outcomes were incident hematologic malignant neoplasm cases occurring at least 6 months after breast cancer diagnosis. The various types of hematologic malignant neoplasm considered were acute myeloid leukemia (AML), myelodysplastic syndrome (MDS), myeloproliferative neoplasms (MPNs), multiple myeloma (MM), Hodgkin lymphoma or non-Hodgkin lymphoma (HL/NHL), and acute lymphoblastic leukemia or lymphocytic lymphoma (ALL/LL). Incidence of these various types was estimated among breast cancer survivors and compared with the incidence in women in the general population.

RESULTS

The 439 704 women in the study had a median (interquartile range [IQR]) age of 59 (50-69) years and were followed up for a median (IQR) duration of 5 (2.8-7.5) years. Overall, 3046 cases of hematologic malignant neoplasm occurred: 509 cases (16.7%) of AML (crude incidence rate [CIR] per 100 000 person-years, 24.5; 95% CI, 22.4-26.8), 832 cases (27.3%) of MDS (CIR, 40.1; 95% CI, 37.4-42.9), and 267 cases (8.8%) of MPN (CIR, 12.8; 95% CI, 11.4-14.5). Lymphoid neoplasm cases included 420 cases (13.8%) of MM (CIR, 20.3; 95% CI, 18.4-22.3), 912 cases (29.9%) of HL/NHL (CIR, 44.4; 95% CI, 41.1-50.0), and 106 cases (3.5%) of ALL/LL (CIR, 5.1; 95% CI, 4.2-6.2). Compared with the general population, breast cancer survivors had statistically significantly higher incidence of AML (standardized incidence rate ratio [SIRR], 2.8; 95% CI, 2.5-3.2) and MDS (SIRR, 5.0; 95% CI, 4.4-5.7) and, to a lesser extent, MM (SIRR, 1.5; 95% CI, 1.3-1.7]) and ALL/LL (SIRR, 2.0; 95% CI, 1.3-3.0).

CONCLUSIONS AND RELEVANCE

The finding that AML and MDS still occur among breast cancer survivors today, and that ALL/LL and MM may also be of concern, merits the continuous monitoring of hematologic malignant neoplasms and the thorough investigations into their underlying mechanisms.

Prognostic and predictive factors for angiosarcoma patients receiving paclitaxel once weekly plus or minus bevacizumab: an ancillary study derived from a randomized clinical trial

BACKGROUND

We report here a correlation analysis conducted along with a phase II trial assessing bevacizumab in combination with weekly paclitaxel.

METHODS

Circulating pro/anti-angiogenic factors were assessed on day 1 (D1) and day 8 (D8). The prognostic value for progression-free survival (PFS) was evaluated using a Cox model with biomarkers as continuous variables.

RESULTS

Among the 51 patients enrolled and treated in this trial, biomarker analysis was performed for 42: 18 in Arm A (single-agent) and 24 in Arm B (combination). With a median follow-up of 46 months, PFS was 5.5 versus 5.7 months, respectively (p = 0.75). According to univariate analysis, factors associated with a poor PFS were as follows: visceral angiosarcoma, de novo angiosarcoma, and high PlGF and low VEGF-C baseline values. In multivariate analysis, de novo angiosarcoma (HR = 2.5; p = 0.024) and baseline VEGF-C value (HR = 0.7; p = 0.003) were significant prognostic factors. We observed a significant increase in circulating PlGF (< 0.001) and a decrease in VEGF (< 0.001) during bevacizumab treatment. An increase in FGF was associated with a poor outcome.

CONCLUSIONS

De novo angiosarcoma and a low baseline level of VEGF-C were found to be associated with a poor prognosis. Addition of bevacizumab induces major changes in circulating biomarkers (VEGF and PlGF) in a short timeframe without impacting PFS.

TRIAL REGISTRATION

Retrospectively registered on EudraCT N° 2009-017020-59 and NCT01303497 (February 24, 2011).

Time trends of overall survival among metastatic breast cancer patients in the real-life ESME cohort

AIM

Real-life analysis of overall survival (OS) trends among metastatic breast cancer (MBC) patients may help define medical needs and evaluate the impact of public health investments. The present study aimed to evaluate the independent impact of the year of MBC diagnosis on OS in the Epidemio-Strategy-Medical-Economical (ESME)-MBC cohort.

METHODS

ESME-MBC (NCT03275311) is a French, national, multicentre, observational cohort including 16,702 consecutive newly diagnosed MBC patients (01 January 2008-31 December 2014). Of 16,680 eligible patients, 15,085 had full immunohistochemistry data, allowing classification as hormone receptor-positive and HER2-negative (HR+/HER2-, N = 9907), HER2-positive (HER2+, N = 2861) or triple-negative (HR-/HER2-, N = 2317) subcohorts. Multivariate analyses of OS were conducted among the full ESME cohort and subcohorts.

RESULTS

Median OS of the whole cohort was 37.22 months (95% confidence interval [CI], 36.3-38.04). Year of diagnosis was an independent predictor of OS (hazard ratio 0.98 [95% CI, 0.97-1.00], P = .01) together with age, subtype, disease-free interval, visceral metastases and number of organs involved. Median OS of HR+/HER2-, HER2+ and HR-/HER2- subcohorts was, respectively, 42.12 (95% CI, 40.90-43.10), 44.91 (95% CI, 42.51-47.90) and 14.52 (95% CI, 13.70-15.24) months. Year of diagnosis was a strong independent predictor of OS in HER2+ subcohort (hazard ratio 0.91 [95% CI, 0.88-0.94], P < .001), but not in HR+/HER2- nor HR-/HER2- subcohorts (hazard ratio 1.00 [95% CI, 0.98-1.01], P = .80 and 1.00 [95% CI, 0.97-1.02], P = .90, respectively).

CONCLUSIONS

The OS of MBC patients has slightly improved over the past decade. However, this effect is confined to HER2+ cases, highlighting the need of new strategies in the other subtypes.

Circulating vascular endothelial growth factor (VEGF) as predictive factor of progression-free survival in patients with advanced chordoma receiving sorafenib: an analysis from a phase II trial of the french sarcoma group (GSF/GETO)

Background

Patients with advanced chordoma are often treated with tyrosine kinase inhibitors without any predictive factor to guide decision. We report herein an ancillary analysis of the the Angionext phase II trial (NCT 00874874).

Results

From May 2011 to January 2014, 26 were sampled. The 9-month PFS rate was 72.9% (95%-CI: 45.9-87.9). During sorafenib treatment, a significant increase in PlGF (18.4 vs 43.8 pg/mL, p<0.001) was noted along with a non-significant increase in VEGF (0.7 vs 1.0 ng/mL, p=0.07). VEGF at D1 >1.04 ng/mL (HR=12.5, 95%-CI: 1.37-114, p=0.025) and VEGF at D7 >1.36 ng/mL (HR=10.7, 95%-CI: 1.16-98, p=0.037) were associated with shorter PFS. The 9-month PFS rate was 92.3% (95%-CI: 56.6-98.9) when VEGF at D1 was ≤1.04 ng/mL versus 23.3% (95%-CI: 1.0-63.2) when >1.04 ng/mL.

Patients and Methods

Chordoma patients were treated with sorafenib 800 mg/day for 9 months, unless earlier occurrence of progression or toxicities. Six biomarkers (sE-Selectin, VEGF, VEGF-C, placental growth factor (PlGF), Thrombospondin, Stem Cell Factor (SCF)) were measured at baseline (day 1: D1) and day 7 (D7).

Conclusion

High levels of VEGF was associated with poor outcome.

Abstract P2-01-02: Heterogeneity and variability of human epidermal growth factor receptor 2 (HER2) expression on circulating tumor cells (CTC) in HER2 negative metastatic breast cancer patients treated with first line weekly paclitaxel and bevacizumab in a prospective cohort from the French Breast Cancer InterGroup Unicancer (UCBG): COMET study

Background: It has been reported in women with advanced estrogen-receptor (ER)-positive/(HER2)-negative breast cancer the acquisition of a HER2-positive CTC subpopulation during therapy (Jordan NV Nature 2016). The clinical significance of acquired HER2 heterogeneity during the evolution of metastatic breast cancer is unknown. We report here the analysis of HER2 status of CTC before and after one cycle of treatment in HER2 negative metastatic breast cancer patients treated with first line weekly paclitaxel and bevacizumab.

Patients & methods:The French cohort COMET is a prospective study including first line HER2 negative patients (pts) receiving weekly paclitaxel and bevacizumab according to EMA approved combination. The aim of this cohort is to evaluate clinical, biological and radiological parameters associated with pts outcome. We confirmed previously the outcome of patients with high CTC count at base line and after one cycle of treatment (Bidard et al, Lancet Oncol 2014). We present here the analysis on 203 pts evaluated for the expression of HER2 on CTC using the FDA cleared CellSearch method. The HER2 expression of CTCs (CB11 clone) was categorized (class 0–3) as described by Riethdorf et al.CCR2010.

Results: At base line, 144 out of 203 pts had at least one detectable CTC (71%), (median 4, and range 1- 30,000). Among them, 104 (72%) had one or more HER2 positive CTC (1-21,484). In 25 patients with HER2 2+ primary tumor with FISH or CISH non amplified, the incidence of CTC HER2+ cases (13/25, 52%) was similar than in pts with HER2 0 or HER2 1+ (51%) primary tumor. In each case, 3 to 100% of detectable CTC could be HER2+ stained (median 50% of CTC). Only 12 cases (8% of all CTC cases) had 2+ HER2 staining score on CTC and none 3+. After one cycle of treatment, the number of pts with detectable CTC dropped to 64, including 42 with HER2+ CTC (65%). Out of these cases, 14 were 3+ or 2+ HER2 score (22% of CTC+ cases). This was a significant increase compared to baseline (8%) (p&lt;0.001), including 6 cases with 100% of HER2+ CTC. To note, 7 patients without HER2+ CTC at baseline, had detectable HER2+ CTC after one cycle of treatment. With a median follow-up of 2 years, correlation of CTC variations with pts outcome is planned.

Conclusion: HER2 staining on CTC was heterogeneous with HER2 positive and negative subpopulations in the same patient with primary HER2 negative breast cancer. We observe a variability of HER2 CTC status with an increased intensity or appearance of immunostaining in few cases during treatment. We hypothesize that these phenotypes changes within patient-derived circulating tumor cells could contribute to progression of breast cancer and acquisition of drug resistance.

Citation Format: Pierga J-Y, Proudon C, Tredan O, Decraene C, Dubot C, Lorgis V, Jacot W, Goncalves A, Debled M, Levy C, Ferrero J-M, Jouannaud C, Luporsi E, Mouret-Reynier M-A, Dalenc F, Lemonnier J, Berger F, Bidard F-C. Heterogeneity and variability of human epidermal growth factor receptor 2 (HER2) expression on circulating tumor cells (CTC) in HER2 negative metastatic breast cancer patients treated with first line weekly paclitaxel and bevacizumab in a prospective cohort from the French Breast Cancer InterGroup Unicancer (UCBG): COMET study [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 P2-01-02.

Abstract P1-07-07: Overtime distribution and predictors of local recurrences (LRs) in patients with hormone receptor positive (HR+) and node positive (N+) breast cancers (BCs): 10 -year follow-up analysis of UNICANCER-PACS 01 and PACS04 trials

Purpose:Incidence of LRs in patients (pts) treated for HR+ HER2- localized BC and distribution overtime have not been described in recent years after introduction of new generation of adjuvant therapies and more extensive use of radiotherapy. We evaluated the incidence and distribution overtime of LRs in pts with HR+ HER2- N+ BCs who entered PACS 01 and PACS04 trials.

Patients and Methods: Data were analyzed from 2909 pts with HR+/HER2- BC out of 5008 included in both trials. Pts underwent mastectomy or lumpectomy plus axillary dissection for a localized N+ BC and, according to study design, were randomized to: 6 cycles of FE100C (standard arm) versus FE100C x 3 cycles followed by docetaxel 100 mg/m2 x 3 cycles (FEC-D) (PACS01) or 6 cycles of Epirubicin 75mg/m2 and Docetaxel 75 mg/m2 (ED75)(PACS04). Loco-regional radiotherapy was mandatory after lumpectomy and recommended in other cases. All pts received 5 years of hormone therapy (HT). A competing risk multivariate analysis was conduct using Fine and Gray model to identify risk factors associated to isolated LRs. Competing events were nodal recurrence, contralateral BC, distant metastasis and death. Cumulative incidence associated to each event was estimated by a Kablfleish-Prentice estimator.

Results: Pts' median age was 50 (22-65); 67.2% underwent lumpectomy, 32.8% mastectomy; 67.6% had 1 to 3 N+, 32.4% more than 3 N+; 45.7% had lymphovascular invasion; 49.5% received FE100C, 35.8% ET75, 14.7% had FEC-D; while radiotherapy was given to 97.3% and HT to 92.2%, of whom 90.5% received tamoxifen. At a median follow-up of 9.1 years, 60 pts (2.1%) experienced LR as first event. The 5-year and 10-year cumulative incidence of LRs were 1.04% and 2.53%, respectively. The cumulative incidence of LRs increased from the 5th year, and the annual risk tended to remain constant over time. Multivariate analysis of competing risk showed that younger age, conservative surgery and omission of HT (not prescribed or non-adherence) were independently associated with risk of developing LRs.

Table 1. Multivariate analysis on competing risk of predictors of LRsVariablesHR 95%CIP valueAge at entry (&lt;35 years, ≥ 35)*0.95 [0.92; 0.99]0.009Mastectomy, lumpectomy0.39 [0.17; 0.86]0.020&gt; 20mm, ≤20 mm0.68 [0.37; 1.24]0.203N+ &gt;3, 1-31.73 [0.99; 3.02]0.055Grade II/III, I1.06 [0.50; 2.24]0.885PR+,PR-1.78 [0.70; 4.53]0.223Type of chemotherapy 3FEC-3D, 6FEC/6ET1.32 [0.65; 2.69]0.446Number of cycles 6, &lt;60.71 [0.17; 0.75]0.630Hormone therapy Yes,No0.36 [0.17; 0.75]0.006*treated as continuous variable

Conclusion: Our analysis showed that incidence of LRs in pts with HR+ N+ BCs treated within PACS trials were considerably lower as compared to earlier studies. These findings may reflect differences in treatment era, as the more extensive use of radiotherapy and new generation of adjuvant chemotherapy. Despite current adjuvant strategies, young age at diagnosis and omission of HT remain independent risk factors of LRs.

Citation Format: Pistilli B, Filleron T, Mazouni C, Zingarello A, Lacroix-Triki M, Rivera S, Coudert B, Serin D, Canon J-L, Campone M, Bachelot T, Goncalves A, Levy C, Cottu P, Petit T, Eymard J-C, Tunon De Lara C, Roché H, Roca L, Lemonnier J, Delaloge S. Overtime distribution and predictors of local recurrences (LRs) in patients with hormone receptor positive (HR+) and node positive (N+) breast cancers (BCs): 10 -year follow-up analysis of UNICANCER-PACS 01 and PACS04 trials [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 P1-07-07.

Abstract P6-14-02: Real-life activity of eribulin among metastatic breast cancer patients in the multicenter national observational ESME program

Background: In 2014, UNICANCER (composed of 18 French Comprehensive Cancer Centers) launched the Epidemiological Strategy and Medical Economics (ESME) program to investigate real-world data in solid tumors. Real-world data give the opportunity to assess for the activity of specific drugs outside clinical trials. Eribulin is approved for pre-treated metastatic breast cancer (MBC). Marketing authorization has been granted in France in July 2012. However few data are available regarding its efficacy in real life. We evaluated eribulin use as second and third line of chemotherapy in MBC patients from the ESME database.

Methods: Data from all newly diagnosed MBC patients having initiated at least one treatment between Jan. 2008 and Dec. 2014 are included in the ESME database. Data were collected retrospectively using a clinical trial-like methodology. Primary endpoint was overall survival (OS), defined from the starting date of second or third line chemotherapy (eribulin versus other chemotherapy). Progression-free survival (PFS) was calculated as a secondary endpoint.

Results: Of 16,703 MBC patients included in the ESME database, 7,412 received at least 2 lines of chemotherapy: eribulin/other chemotherapy, total 1,966/5,446, second line 363/5,446, third line 654/2,669. Depending on second or third line chemotherapy use classification, median age was 59 years (range 20-97) and 58 year (range 21 – 94), triple negative tumors accounted for 20% and 19% of cases, and median follow-up reached 26 months and 22 months respectively.

Table reports median OS and PFS, according to lines and type of chemotherapy.

 OS eribulin (months)OS other chemotherapy (months)pPFS Eribulin (months)PFS other chemotherapy (months)pSecond line12.4 (11.3-15.1)11.8 (11.3-12.3)0.4654.1 (3.7-4.9)4.1 (4.0-4.3)0.9225Third line10.3 (9.3-11.5)7.7 (7.3-8.0)&lt;.00013.6 (3.2-3.9)3.0 (2.9-3.2)0.0058

Supportive analyses (using a propensity score for adjustment and as a matching factor for nested case–control analyses) and sensitivity analyses will be available for full presentation at the meeting.

Conclusion: In this large-scale real-life setting, MBC patients treated with third line eribulin showed an improved OS and PFS compared with those receiving another chemotherapy. The difference was not statistically significant for second line treatment.

Citation Format: Jacot W, Heudel P-E, Fraisse J, Gourgou S, Guiu S, Dalenc F, Pistilli B, Campone M, Levy C, Debled M, Leheurteur M, Chaix M, Lefeuvre C, Goncalves A, Uwer L, Ferrero J-M, Eymard J-C, Petit T, Mouret-Reynier M-A, Courtinard C, Cottu P, Robain M, Mailliez A. Real-life activity of eribulin among metastatic breast cancer patients in the multicenter national observational ESME program [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 P6-14-02.

Abstract PD7-06: MAAT: Menses after adjuvant treatment. Prediction of menses recovery after chemotherapy for early breast cancer (BC) by using a nomogram model in UNICANCER PACS04 and PACS05 trials

Purpose:The likelihood of menses recovery (MR) is largely variable in premenopausal patients (pts) receiving adjuvant chemotherapy for BC. Quantifying this probability for each single patient could impact discussion of chemotherapy side effects and better individualize fertility counseling.We performed a pooled analysis from PACS04 and PACS05 randomized trials aiming to develop a nomogram to estimate the probability of menses recovery at 6 and 18 months (mos) after the end of adjuvant chemotherapy (CT) for premenopausal pts with early BC.

Patients and Methods: The analyzed population consisted of 1683 pts who were premenopausal and ≤ 50 (out of 4524 enrolled in both trials). In PACS05 node-negative BC pts were randomized to 4 or 6 cycles of FE100C (standard arm); in PACS04 node-positive pts were randomized to 6 cycles of FE100C or 6 cycles of Epirubicin 75mg/m2 and Docetaxel 75 mg/m2 (ED75). Endocrine therapy (ET) (Tamoxifen) x 5 years was mandatory for ER+ BC. Variables significantly associated with MR in the univariate analysis (P&lt;0.20) were included in the multivariate analysis. Using this data set, a logistic regression-based nomogram was developed to predict MR at 6 and 18 mos.

Results: Pts' characteristics were: median age 43 (22-50), median body mass index (BMI) at baseline 22.6 (15.6-54.7), at the end of chemotherapy 22.8 (15.8-58.6). ED75 was administrated to 517 (30.7%), while 802 (47.7%) received 6FE100C, 364 (21.6) 4FE100C. Trastuzumab was given to 122 (7.2%), ET to 1229 (73%) pts. CT-induced amenorrhea was observed in 1407 (83.6%) pts. Factors associated to MR were assessed on 1210 pts (excluding pts who recovered menses during CT or of whom date of recovery was not specified). At a median follow-up of 90 mos, 28.2% (342/1210) of pts had recovered menstrual cycles: 11% (133/1210) at 6 mos and 24.3% (294/1210) at 18 mos. Multivariate analysis showed that younger age, higher BMI at the end of CT, non-alkylating agents and absence of ET were independently associated to MR.

Table 1 Multivariate Cox regression analysis of menses recoveryVariablesHR (95%CI)P valueAge1.49 (1.16-1.93)&lt; 0.002Age2*0.99 [0.98-0.99]&lt;0.0001BMI after CT1.02 (0.99-1.04)0.07Alkylating agents0.72 (0.57-0.90)0.004Endocrine Therapy0.50 (0.40-0.62)&lt;0.001* The quadratic term in the age variable accounts for the non-linearity of the relation between the age and the probability of recovering menses. Overall this probability tend to decrease when age increase with a greater decrease for the older patients.

Nomogram concordance-index was 0.749 and 0.750 for predicting MR at 6 and 18 mos respectively. A better calibration was observed at 18 mos, comparing nomogram predictions with the actual probability of MR in the 1210 women.

Conclusion:Our analysis confirmed the possibility of developing a user-friendly nomogram for predicting menses recovery after adjuvant chemotherapy. As next step, we will externally validate our nomogram on CANTO premenopausal population, one of the biggest national cohorts aiming to assess the long-term impact of cancer treatments toxicities (UNICANCER NCT01993498 - http://etudecanto.org/).

Citation Format: Pistilli B, Mazouni C, Zingarello A, Faron M, Saghatchian M, Grynberg M, Spielmann M, Kerbrat P, Roché H, Lorgis V, Bachelot T, Campone M, Levy C, Goncalves A, Lesur A, Veyrat C, Vanlemmens L, Lemonnier J, Delaloge S. MAAT: Menses after adjuvant treatment. Prediction of menses recovery after chemotherapy for early breast cancer (BC) by using a nomogram model in UNICANCER PACS04 and PACS05 trials [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 PD7-06.

Abstract P5-20-03: Impact of prior adjuvant trastuzumab (aT) on clinical characteristics, patterns of recurrence and outcome in 2863 patients with Her2 positive (HER2+) metastatic breast cancer (MBC)- Results from the French ESME UNICANCER program

Background: The management of HER2+ BC has changed dramatically with the introduction and widespread use of HER2-targeted therapies, especially in the adjuvant setting. However, there is relatively limited real-world information on the impact of adjuvant Trastuzumab (aT) on patterns of recurrence and outcome of HER2+ MBC.

Methods: In 2014, the 18 French Cancer Centers launched the Epidemiological Strategy and Medical Economics (ESME) program to provide real-world data on MBC patients (pts). All pts who started a 1st-line treatment for MBC between 01-Jan-2008 and 31-Dec-2014 were included. We examined clinical characteristics and outcomes (overall survival [OS] and time to next treatment [TNT]) in patients with HER2+ MBC pretreated with trastuzumab in the adjuvant setting (aT) compared with trastuzumab-naïve patients (nT) and patients with de novo HER2+ MBC (dn). Multivariate analyses adjusted for baseline demographic, prognostic factors and year of diagnosis (prior or after 2005, when aT was approved and widely administered in France for early HER2+ breast cancer).

Results: Among the 15170 pts of the ESME database, 2863 (19%) were HER2+: 1093 pts (38%) had de novo and 1765 pts (62%) recurrent MBC; 63% were Hormone Receptor (HR) +; 54%, 25% and 21% had respectively 1, 2, or &gt; 2 metastatic sites (68% visceral and 12% brain). Median time to 1st metastasis was 43.4 months (m) (95% CI: 24.6-84.4): 54 m in HR+ and 30 m in HR-. Among pts with recurrent MBC, 55% (995) had received aT. As 1st-line therapy for MBC, 90 % of pts received HER2-targeted agents (73% T-based). With a median follow-up of 46 m, median OS is 45 m (95% CI: 42.5-48). OS is significantly higher in de novo compared to recurrent MBC: 54 m (95% CI: 50.2-60.4) vs. 38.4 m (95% CI: 36.7-41.9), (p &lt; 0.0001). Among pts with recurrent cancers, median OS is inferior in pts who had received aT, as compared to those who had not: 33.4 m (95% CI: 29.6-36.7) vs. 49.5 m (95% CI: 44.3-56.8), (p &lt; 0.0001). Statistically significant differences persist after adjustment for age at MBC, disease-free interval, metastatic sites and RH status in the multivariate model (HR=1.45, 95% CI: 1.26-1.67) but not after adjustment for year of diagnosis (prior or after 2005) (HR=0.90, 95% CI: 0.70-1.15).

Conclusions: These large-scale real-world data in patients with HER2+ MBC provide evidence that the survival outcome remain similar in patients with failure of adjuvant trastuzumab compared with trastuzumab-naïve patients after adjustment for year of diagnosis. De novo HER2+ MBC pts have the best outcomes. Data on clinical characteristics of metastasis and time to next treatment for the three subgroups will be presented at the meeting.

Citation Format: Saghatchian M, Carton M, Piot I, Pérol D, Pistilli B, Brain E, Ghouadni A, Ricci F, Vanlemmens L, Loeb A, Levy C, Goncalves A, Dalenc F, Lefeuvre-Plesse C, Campone M, Jaffre A, Gourgou S, Cailliot C, Robain M, Dieras V. Impact of prior adjuvant trastuzumab (aT) on clinical characteristics, patterns of recurrence and outcome in 2863 patients with Her2 positive (HER2+) metastatic breast cancer (MBC)- Results from the French ESME UNICANCER program [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 P5-20-03.

Abstract PD8-12: Mutational processes, genome evolution and outcome in metastatic breast cancers

Background: to determine the distribution and evolution of mutational processes in metastatic breast cancers (mBC), together with their clinical relevance Methods: Whole exome sequencing (Hi-Seq, Illumina) and determination of copy number alterations (CNA) (CGH array / SNP6.0) were performed in 240 and 692 metastatic breast cancers respectively. Mutational processes were defined according to Alexandrov (Nature, 2013). Homologous Recombination Deficiency (HRD) was determined by genome wide assessment of loss-of-heteroygosity (LOH) on SNP6.0 (n = 210). Finally, genomic instability was assessed by the % of genome altered assessed by CGH / SNP6.0 Results: Whole exome sequencing showed that HR+/Her2- metastatic breast cancer presented an increased contribution of APOBEC-related signatures, as compared to early breast cancer (TCGA) (58% of the mutations vs 31%, p &lt; 0.0001). Twelve percent of the HR+/Her2- mBC acquired an hypermutator genotype ( &gt; 200 non-synonymous mutations). This acquisition of an hypermutator genotype was confirmed in five paired primary-metastatic samples. An operational APOBEC-related signature 13 was associated with a poor outcome in a multivariate analysis (HR: 1.75, 95%CI: 1.1-2.7, p = 0.017). High LOH score (HRD) was observed in 30% of HR+/Her2- mBC as compared to 13% of early HR+/Her2- early BC (p &lt; 0.0001). The opposite was observed in TNBC (43% in mTNBC versus 58% in early TNBC ,p = 0.032). High LOH score was associated with a trend for poor outcome in HR+/Her2- mBC (multivariate 1.67, 95%CI: 0.949-2.951, p = 0.075).The % of genome altered was associated with a poor outcome in multivariate analyses both in the overall and HR+/Her2- mBC (HR / 10 increase:1.144, 95%CI:1.038-1.261, p = 0.007 and HR:1.18, 95%CI:1.037-1.344, p = 0.012 respectively). Copy number analyses identified 143 genes that are more frequently amplified as compared to early breast cancers (FDR &lt; 0.01) Conclusions: metastatic HR+/Her2- metastatic breast cancer present an increased in APOBEC-related mutational burden and in LOH score as compared to early breast cancers. APOBEC-related signature 13 and genome instability are associated with a poor outcome and could be used in the future to better stratify metastatic breast cancer patients.

Citation Format: Patsouris A, Filleron T, Jacquet A, Goncalves A, Bonnefoi H, Letourneau C, Bachelot T, Jimenez M, Andre F. Mutational processes, genome evolution and outcome in metastatic breast cancers [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 PD8-12.

Abstract P5-06-02: Ex vivo CSC assays for personalized testing of drug susceptibility in advanced breast cancer

In the developing area of personalized medicine, targeted therapies are mainly based on genomic characterization of each tumor, and is currently proposed as promising strategies for advanced breast cancer (ABC). Despite the promises of advanced genome sequencing, many patients still fail therapy, resulting in disease progression, recurrence, and metastases. Cancer stem cells (CSCs) concept illustrates the non-genetic intrinsic resistance, recapitulates tumor heterogeneity that creates hierarchically organized tumor tissues where a subpopulation of self-renewing cancer stem cells (CSCs) sustains the long- term clonal maintenance of the neoplasm. Evidences indicate that CSCs survive many commonly employed cancer therapeutics. Patient-derived tumor xenograft (PDXs) models recapitulate tumor complexity and heterogeneity at cellular and molecular level.

We aimed to specifically address the therapeutic sensitivity in ABC, by using an ex vivo assay based on PDX prospective collection, fully characterized for genomic alterations.

In this work, we aim at defining for each tumor the best therapy to target breast cancer intratumor heterogeneity, the CSC component. For that, we defined a panel of 44 FDA-approved compounds used for cancer treatment, including breast and other types of cancer, cancer stem cell drugs, chemo or targeted therapies. For each drug, we screened the differential sensitivity of the bulk tumor cells and the CSC components for 12 PDX models using an ex vivo screening approach on short term culture. To assess intra tumor heterogeneity, we set up an original dual strategy: for the bulk cells, an ex vivo assay based on IC50, and for breast CSC component a miniaturized Aldefluor assay. First, we demonstrate that bulk cells and CSCs sensitivity may be dissociated for the same drug in the same PDX models. Then, we observed that whereas bulk cell sensitivity may be correlated to tumor genomic abnormalities, CSC drug sensitivity seems not to follow the rule.CSC are selectively sensitive to specific compounds. We are exploring the pathways that sustain this selective sensitivity in the CSCs components. We are currently identifying targets using mass spectrometry in CSCs and bulk cells.Then, we validated the hits predicted from ex vivo screening assays by in vivo treatment of using PDX models for the selected drugs, and in a patient with ABC.

In that work, we demonstrated that CSCs display different sensitivity profiles than bulk cells to the same agents, irrespective to their genomic background and are identifying the CSC specific targets. Here, we propose a new model of precision medicine based on ex vivo CSC assays for personalized testing of drug susceptibility in advanced breast cancer.

Citation Format: Charafe-Jauffret E, Wicinski J, Cabaud O, Lopez M, Audebert S, Adelaide J, Chaffanet M, Guille A, Goncalves A, Bertucci F, Birnbaum D, Ginestier C. Ex vivo CSC assays for personalized testing of drug susceptibility in advanced 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 P5-06-02.

Pancreatic Adenocarcinoma Therapeutic Targets Revealed by Tumor-Stroma Cross-Talk Analyses in Patient-Derived Xenografts

Preclinical models based on patient-derived xenografts have remarkable specificity in distinguishing transformed human tumor cells from non-transformed murine stromal cells computationally. We obtained 29 pancreatic ductal adenocarcinoma (PDAC) xenografts from either resectable or non-resectable patients (surgery and endoscopic ultrasound-guided fine-needle aspirate, respectively). Extensive multiomic profiling revealed two subtypes with distinct clinical outcomes. These subtypes uncovered specific alterations in DNA methylation and transcription as well as in signaling pathways involved in tumor-stromal cross-talk. The analysis of these pathways indicates therapeutic opportunities for targeting both compartments and their interactions. In particular, we show that inhibiting NPC1L1 with Ezetimibe, a clinically available drug, might be an efficient approach for treating pancreatic cancers. These findings uncover the complex and diverse interplay between PDAC tumors and the stroma and demonstrate the pivotal role of xenografts for drug discovery and relevance to PDAC.