Liu, Minetta C.

Abstract Purpose: Blood-based candidate biomarkers of disease can be monitored by analyzing CTCs and/or circulating cfDNA isolated from the peripheral blood. Our primary objective is to understand the relative contributions of these circulating factors (i.e., CTCs and cfDNA) to the overall disease profile in MBC. Methods: Clinically archived FFPE tumor tissue and prospective blood samples are collected through a minimal risk protocol approved by the Mayo Clinic IRB (#16-001540) from patients with MBC and objective evidence of disease progression. Blood samples include 20 mL whole blood in Streck blood collection tubes (BCTs) for platelet poor plasma (PPP); 20 mL whole blood in AccuCyte BCTs for CTCs, WBCs, and PPP; and 10 mL whole blood in EDTA BCTs for PPP. Nucleated, EpCAM+/cytokeratin+/CD45- CTCs are identified, assessed for ER/HER2 status, and isolated using a centrifugation and direct imaging platform that allows for single cell retrieval (RareCyte). DNA is extracted from PPP, CTCs, WBCs, and FFPE tumor tissue using established methods. Targeted sequencing for SNVs/indels is performed on paired WBCs and CTC-DNA, AccuCyte-cfDNA, Streck-cfDNA, EDTA-cfDNA, and tumor tissue derived DNA using the same NGS panel and informatics pipeline (65 genes; CleanPlex OncoZoom; Paragon Genomics). Results: Tissue and blood samples were collected from 40 patients with metastatic breast cancer. 10 cases were selected for initial analyses on the basis of CTC yield (range 3-113 per 3.75 mL blood); up to 5 CTCs per subject were isolated and pooled for DNA extraction. Plasma cfDNA yields and variant allele frequencies were highly comparable between AccuCyte and Streck collected blood samples. Mutations (range 1-3) were identified in CTC-DNA and/or cfDNA in 9 of 10 cases for a total of 18 detected mutations: 10 in CTC-DNA and cfDNA (BRCA2 N372H, PIK3CA E542K, PIK3CA E545K (x3), PIK3CA H1047R, PTEN R130P, RET G691S, TP53 C135W, TP53 Q192*); 5 in CTC-DNA only (EGFR R521K, EGFR T790M, PIK3CA H1047R, SMAD4 C363Y, TP53 N263D); and 3 in cfDNA only (DNMT3A W893S, DNMT3A S714C, TP53 Q136E). Parallel analyses of samples from 10 more subjects are in progress. Analysis of tumor tissue for all 20 subjects and of EDTA-cfDNA and single CTCs for a subset of cases is ongoing. Updated results will be presented at the meeting. Conclusions: It is feasible to isolate high quality CTCs and cfDNA from the same blood collection tube to perform targeted sequencing; this streamlines specimen processing, decreases overall costs, and minimizes required blood volumes. Importantly, there is overlap in the majority of mutations identified in CTC-DNA and cfDNA, but actionable mutations (e.g., PIK3CA, EGFR) were detected in CTC-DNA only. The clinical and theranostic relevance of these findings is unclear and warrants further investigation. Citation Format: Minetta C. Liu, Karthik V. Giridhar, Roberto A. Leon Ferre, Jamie L. Carroll, Matthew P. Goetz, Tufia C. Haddad, Deanne R. Smith, Siddhartha Yadav, Vidushi Kapoor, Guoying Liu, Tad George, Nolan Ericson, Arturo B. Ramirez, Eric Kaldjian, Keegan E. Haselkorn. Comparison of circulating tumor cell (CTC) derived DNA and circulating cell-free DNA (cfDNA) from simultaneous blood sampling of patients with metastatic breast cancer (MBC) [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 3119.

Abstract Background: Disease relapse is observed in a significant proportion of HER2-positive breast cancer (BC) patients with residual disease (RD) after neoadjuvant treatment, as well as in a subgroup of those achieving pathological complete response (pCR). As the host immune response plays a key role in modulating the activity of anti-HER2 agents, we investigated the association of T- and B-cell receptor (TCR and BCR) repertoires with pCR and event-free survival (EFS) in the NeoALTTO phase 3 trial. Methods: RNA sequencing (RNAseq) data from baseline tumor biopsies were available for 254 patients out of the 455 enrolled in the study. Among those, 166 did not achieve a pCR defined as ypT0/is. Matched RNAseq data from RD samples were available for 43 cases. TCR/BCR repertoires were extracted from RNAseq data using the MiXCR software. TCR and BCR read counts, number of clones, evenness, Shannon entropy, Gini index, length of the complementarity determining region 3, top and second top clone proportion were evaluated. Survival analysis was performed using univariate and multivariate (adjusted for tumor size, nodal status, grade, estrogen receptor [ER] status, age and treatment arm) Cox proportional hazard models, while logistic regressions were used for pCR. False discovery rate (FDR) was obtained using Benjamini & Hochberg method. Results: Baseline TCR top (odds ratio [OR]=0.63 [95% CI 0.46-0.87], FDR=0.021) and second top (OR=0.57 [0.41-0.79], FDR=0.004) clone proportion were significantly associated with a lower probability of achieving pCR in the multivariate analysis. BCR evenness (hazard ratio [HR]=1.5 [1.2-2], FDR=0.015) and Gini index (HR=0.66 [0.52-0.85], FDR=0.015) were significantly associated with EFS in the multivariate analysis. In residual disease, BCR evenness and Gini index showed a similar trend in the EFS univariate analysis, while TCR read counts, number of clones, and entropy were associated with lower HR (P<0.05), although FDR were borderline significant (0.05<FDR<0.1). A model to predict EFS including baseline BCR evenness, ER status, stromal tumor-infiltrating lymphocyte level, and pCR was able to identify 3 groups with good, intermediate and poor prognosis (Kaplan-Meier 5-year EFS rates of 95%, 80% and 58%, respectively). Of note, about 63%, 45% and 12% of the patients achieved a pCR in the three prognostic groups, respectively. Conclusions: In the NeoALTTO trial, the presence of an evenly distributed BCR repertoire was associated with worse EFS. A model integrating baseline immune-related and clinical features was able to identify patients with excellent prognosis despite RD or, conversely, with poor prognosis after pCR. We envision that our model has the potential to allow the personalization of post-operative treatment strategies after both pCR and RD in HER2-positive BC. Further validation of our findings is warranted. Citation Format: Mattia Rediti, David Venet, Françoise Rothé, Tao Qing, Marion Maetens, Ian Bradbury, Miguel A. Izquierdo, Serena Di Cosimo, Florentine Hilbers, Mohammed Bajji, Nadia Harbeck, Michael Untch, David L. Rimm, Stephen Chia, Minetta C. Liu, Cristina Saura, Jens Huober, Paolo Nuciforo, Roberto Salgado, Sherene Loi, Lajos Pusztai, Christos Sotiriou. Predictive and prognostic role of T- and B-cell receptor repertoire in HER2-positive breast cancer: An analysis of the NeoALTTO clinical trial [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 1998.

Abstract Background: Earlier cancer detection could identify tumors when they are more treatable and thus may improve clinical outcomes. A blood-based multicancer early detection test using circulating tumor cell-free DNA (cfDNA) offers the potential to address this unmet need. Our previous discovery work identified whole-genome bisulfite sequencing as outperforming whole-genome and targeted sequencing approaches for multicancer detection across stages. We thus developed a targeted methylation (TM) assay for multicancer detection and tissue-of-origin (TOO) localization. Methods: Participants were from the Circulating Cell-free Genome Atlas (CCGA; NCT02889978) and STRIVE (NCT03085888) studies, both prospective, multicenter, observational studies with longitudinal follow-up. cfDNA samples spanning >20 cancer types of all stages were divided into a cross-validated training set and an independent validation set; there were 3,583 evaluable samples in training (1,530 cancer; 2,053 noncancer) and 1,803 in validation (678 cancer; 1,125 noncancer). Performance in a prespecified set of 12 high-signal cancers was also reported (anus, bladder, colon/rectum, esophagus, head and neck, liver/bile-duct, lung, lymphoma, ovary, pancreas, plasma cell neoplasm, stomach). This multicancer early detection test evaluated cfDNA for the presence of cancer and localization of TOO; the TM classifier was trained to target >99% specificity. Precision was defined as the fraction of correct calls. Results: Participants with and without cancer were similar in age. Specificity across all cancer types in the training and validation sets was 99.8% (95% CI, 99.2-99.9%) and 99.3% (98.3-99.8%), respectively, which reflects a consistent (P=0.292) false positive rate of <1%. Across all cancer types, aggregate sensitivity was also consistent between the training and validation sets (55% [53-58%] and 55% [51-59], respectively; P=0.897). In the validation set, aggregate sensitivity in the prespecified group was 76% (73-78%); detection was 39% (27–52%) in stage I (n=62), 69% (56–80%) in stage II (n=62), 83% (75–90%) in stage III (n=102), and 92% (86–96%) in stage IV (n=130). Overall, the classifier assigned a TOO across >20 cancer types in 96% of samples (344/359); of these, the TOO was correct in 93% (321/344) of cases, which was consistent with training set analyses. Conclusions: This multicancer early detection test detected cancer signal across >20 cancer types with a single, fixed, low false positive rate and highly accurate TOO localization. Importantly, results in the independent validation set were consistent with the training set, demonstrating the robustness of machine learning classifier training, and confirms that data were not overfitted. These data support the feasibility of a single blood-based test that can detect multiple cancers, supporting further clinical development in preparation for the return of results. Citation Format: Anne-Renee Hartman, Geoffrey Oxnard, Eric Klein, Michael Seiden, Earl Hubbell, Oliver Venn, Arash Jamshidi, Nan Zhang, John Beausang, Samuel Gross, Kathryn Kurtzman, Eric Fung, Brian Allen, Alexander Fields, Hai Liu, Mikkael Sekeres, Donald Richards, Peter Yu, Alexander Aravanis, Minetta Liu. Multicancer detection of early-stage cancers with simultaneous tissue localization using a plasma cfDNA-based targeted methylation assay [abstract]. In: Proceedings of the AACR Special Conference on Advances in Liquid Biopsies; Jan 13-16, 2020; Miami, FL. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(11_Suppl):Abstract nr IA02.

e14018 Background: Cancer survival rates are improving. Therefore, management of cardiovascular complications has now become a crucial clinical concern. Cardio-oncology is the sub-specialty that assists in the overall management of cancer patients in a multi-disciplinary manner. Mayo Clinic cardio-oncology practice was initiated to work closely with our oncology colleagues for early detection of cardiovascular complications in response to cancer-therapy. Majority of the patients visit our cardio-oncology clinic once, we thought it is important to study the group of patients that visited frequently due to cardiovascular complications. Aim: To evaluate the most common cardiovascular complication in patients with 2 or more visits to our cardio-oncology clinic. Methods: From 2012-2017, there were > 2500 patients visits to our clinic, with 24 patients having 2 or more visits. Data including patients’ demographics, ethnicity, chemotherapeutic medications, primary cancer type, cardiovascular risk factors, echocardiography and clinical outcomes were collected. Cardiotoxicity was defined as the decrease in left ventricular ejection fraction (LVEF) of > 10% to a value of < 53%. Heart failure was diagnosed based on Framingham’s criteria or by a cardiologist. Results: There were 19 women (80%) and 5 men (20%). Median age at the time of diagnosis was 56 years [19-76]. The most common malignancy was breast cancer (70%), followed by B-cell lymphoma (12%) and acute myeloid leukemia (8%). Thirty percent had > 2 risk factors for cardiovascular disease. 75% of the patients had an LVEF of < 53, of these 67% developed heart failure with 58% preserved and 42% reduced ejection fraction. Those with heart failure had received a mean anthracycline dose of 305 ± 91.8mg/m2. With initiation of ACEI, B-Blockers, and diuretics (GDMT) 79% showed recovery of LVEF to ≥53 during the follow up. Conclusions: In our experience, most patients who were seen at least twice in the cardio-oncology clinic for heart failure had received a dose of > 300mg/m2 anthracycline. With GDMT over 75% of the patients recovered. Care in the cardio-oncology clinic plays a key role in optimizing these clinical outcomes.

TPS1114 Background: Immunotherapy combined with chemotherapy is being studied in metastatic breast cancer, and may have durable outcomes. Chest wall recurrence represents a difficult to treat subtype of breast cancer with a poor prognosis, with lymphovascular invasion in the primary tumor a significant risk factor. Given the inflammatory nature of this disease and the association of programmed cell death 1 (PD-1) expression with lymphovascular invasion, we hypothesized that the combination of pembrolizumab, an anti-PD-1 antibody, with carboplatin may be effective as treatment for breast cancer chest wall recurrences. Methods: This randomized phase II study is enrolling 84 patients with breast cancer involving the chest wall, who may also have distant metastases. Patients receive treatment with pembrolizumab 200 mg and carboplatin AUC 5 every 3 weeks for 6 cycles (Arm A, n = 56) followed by maintenance pembrolizumab +/- carboplatin (Arm Ax), or carboplatin AUC 5 every 3 weeks for 6 cycles (Arm B, n = 28) with an option to cross-over to pembrolizumab +/- carboplatin on progression (Arm Bx). Patients with all disease subtypes, triple-negative, hormone receptor positive/HER2- after 2 prior lines of hormone therapy, and HER2+ disease (with the option to continue trastuzumab) are eligible, with no limit on the number of prior therapies. Prior platinum chemotherapy is allowed in the absence of overt disease progression. Patients undergo clinical assessment with every cycle of treatment including chest wall photography, scans (CT chest, abdomen, and pelvis) every 2 cycles, and have peripheral blood and chest wall biopsies collected at baseline and the start of cycle 3 for correlative studies. The primary endpoint is the disease control rate in the chest wall and distant sites at 18 weeks of treatment based on RECIST 1.1. The study is powered to determine a 20% difference in disease control between arms (hazard ratio 0.52, α = 0.10, β = 0.20). Additional endpoints include response by tumor programmed death ligand 1 (PD-L1) status and irRECIST, progression-free survival, and toxicity. Chest wall tumor samples will be analyzed for changes in tumor immune composition, and PD-L1 and MYC oncogene expression, based on preclinical data to suggest that PD-L1 may be upregulated by MYC. Peripheral blood samples will be evaluated for changes in PD-L1 expression, cell-free DNA, and circulating tumor cells with treatment. The study is enrolling patients at 7 sites within the Translational Breast Cancer Research Consortium (TBCRC), with current enrollment of 38/84 patients. Clinical trial information: NCT03095352 .

1016 Background: The ph III EMBRACE trial of E vs physician’s choice of tx led to FDA approval of E as ≥3rd-line tx for MBC pts with prior exposure to anthracyclines/taxanes. The ph III BOLD 301 trial of E vs capecitabine in advanced BC treated with anthracyclines/taxanes showed a nonsignificant trend to improved median OS for all pts; pre-planned analysis by HER2 status revealed a nominally significant benefit for HER2- pts. Methods: RU011201I is an investigator initiated ph III trial with 1:1 randomization to E (1.4 mg/m2 D1,8 q21days) vs P (90 mg/m2 D1,8,15 q28days) within strata defined by (neo)adjuvant taxanes (yes/no), HR status (+/-), and line of tx (1st/2nd). Pts had measurable or non-measurable disease by RECIST v1.1; new or progressive mets; peripheral neuropathy (PN) gr<2; ≤1 prior chemotx regimens for advanced or MBC. Asymptomatic brain mets with stable MRIs for >3 mos were allowed. (Neo)adjuvant taxanes were allowed if >12 mos between tx completion and disease recurrence. Radiographic studies occur q12wks. Survival data are collected q12wks after the Off-Tx Visit. Pts reported side effects weekly; post-baseline symptomatic AE rates worse than baseline were compared between arms using Fisher’s exact tests. We report clinical outcomes and the primary objective related to PRO-CTCAE use. Results: 201 pts enrolled 3/2014 - 5/2018 with 33.8 mos median f/u; 3 are on tx as of 2/20/20 (1E, 2P). Pt characteristics were the same between E vs P: median age 62 yrs (range 27-85); 42% prior taxane; 78% ER+; 70% starting 1st line tx. Baseline lesion distribution was similar except for lung mets (37E, 53P). No difference was seen between E vs P in PFS (5.7 vs 5.9 mos; P=0.72), OS (18.1 vs 16.4 mos; P=0.75), TTF (5.3 vs 4.9 mos; P=0.82), DOR (10.8 vs 12.3 mos; P=0.84), or number of metastatic events (55 vs 54). 37E and 36P pts required ≥1 dose reduction. Hematologic toxicities ≥ gr 3 were higher with E (40 vs 22%). PN events were similar: 56 vs 58 total, 4 vs 7 motor, 52 vs 51 sensory. Median duration of PN and median time to 1st PN event were 74 vs 140 and 56 vs 41 days. Worsened numbness/tingling severity and other pt-reported AE rates were similar, but worsened numbness/tingling interference (63% vs 78%, P=0.04) and vomiting frequency (35% vs 57%, P=0.005) were lower with E. Conclusions: Clinical outcomes were equivalent with E vs P as 1st/2nd line tx for HER2- advanced BC. The nature and severity of PN were similar between arms, but time of onset, duration, and interference with daily living favor E. E may be a suitable treatment option in this setting. Clinical trial information: NCT02037529 .

511 Background: Clinicopathological and molecular features, including estrogen receptor (ER) status and PAM50 subtypes, have shown an association with immunogenicity and tumor-infiltrating lymphocyte (TIL) levels in breast cancer (BC). To investigate the complexity of the immune response in HER2+ BC, we explored the association of T- and B-cell receptor (TCR and BCR) repertoires with clinicopathological characteristics, PAM50 subtypes and outcome in the NeoALTTO phase III trial. Methods: RNA sequencing (RNAseq) data from baseline tumor biopsies were available for 254 out of the 455 patients enrolled. TCR and BCR repertoires were extracted from RNAseq data using the MiXCR software. Repertoire and diversity measures (read counts, number of clones, evenness, Gini index, Shannon entropy, length of the complementarity-determining region 3 [CDR3], top and second top clone proportions) were estimated. PAM50 subtypes were computed from RNAseq data. Univariate and multivariate (adjusted for clinicopathological characteristics, TIL levels dichotomized using the median value of 12.5% and treatment arm) Cox proportional hazard models were used for survival analysis, while logistic regressions were used for pathological complete response (pCR), defined as ypT0/is. All results reported had a false discovery rate (FDR) <0.05. Results: Higher BCR read counts, number of clones and Gini index were significantly associated with ER-negative as well as grade 3 tumors. Among the PAM50 subtypes, HER2-enriched (HER2-E) showed significantly higher BCR read counts, number of clones and Gini index along with lower evenness compared to luminal A and B, as well as higher length of CDR3 than luminal A. Of note, basal-like showed similar BCR diversity measures to HER2-E. No significant differences were noted for TCR diversity measures. In multivariate analyses, neither TCR nor BCR features were associated with pCR, while BCR evenness (HR 1.5; 95%CI 1.1-2.1) and Gini index (HR 0.66; 95%CI 0.5-0.88) were associated with event-free survival. Conclusions: BCR repertoire measures suggest a clonal expansion in HER2-E and basal-like PAM50 subtypes. Furthermore, the implementation of BCR-derived biomarkers can help to identify patients with an improved clinical outcome after neoadjuvant anti-HER2 treatment. Our findings highlight the heterogeneity of the immune response within HER2+ BC and provide support for biomarker-driven treatment strategies including immunotherapy in this BC subtype. Further validation is required. Clinical trial information: NCT00553358 .

510 Background: Several studies have established the critical role of preexisting immune response in triple negative breast cancer (TNBC). Most studies evaluated the tumor infiltrating lymphocytes in stroma. However, limited data are available with regards to the importance of specific subtypes and spatial distribution of these immune infiltrates. Methods: NanoString IO360 gene expression analysis and Digital Spatial Profiling (DSP) were used. DSP was used to quantify 39 immune-related proteins in stromal and tumor-enriched segments from 44 TNBC samples from the FinXX trial (NCT00114816) and 335 samples from the Mayo Clinic (MC) cohort of centrally reviewed TNBC (Leon-Ferre BCRT 2018). In FinXX trial, 22 patients with recurrence and 22 patients without recurrence were included. In MC cohort, 217/335 patients received adjuvant chemotherapy while 118 patients had surgery only without adjuvant chemotherapy. Regions were segmented based on pancytokeratin staining. The general linear model was used for statistical analysis of differential expression with recurrence free survival (RFS) as a categorical variable (recur yes or no). Kaplan-Meier estimates and Cox regression models were also used for analysis. Results: In the FinXX trial, using global gene expression analysis with IO360, there was no signature significantly associated with RFS. However, using DSP, high protein expression of CD56 in the tumor-enriched segments was associated with significant improvement in RFS (HR 0.26, 95%CI 0.09-0.78, p 0.01). Nevertheless, CD56 expression in the stroma (HR 0.66, 95%CI 0.29-1.53, p 0.33) and all segments (HR 0.53, 95%CI 0.23-1.25, p 0.14) was not significantly associated with improved outcome. We further validated these findings in the MC TNBC cohort where intratumoral CD56 expression was associated with a significant improvement in RFS (HR 0.23, p 0.002) but not stromal CD56 (p 0.79). Interestingly, when evaluating the MC TNBC cohort according to receipt of chemotherapy, intratumoral CD56 was associated with improved outcome only in patients who received chemotherapy (p 0.02 vs. 0.07). In both cohorts, higher expressions of intratumoral PD-L1, HLA-DR, and CD8 were associated with improved outcome. Conclusions: Using an in-depth analysis with spatially defined context, we identify that intratumoral CD56-positive NK cells are associated with improved outcome in TNBC. Our study highlights the potential role of NK cells in TNBC and future implications for biomarkers and therapeutic targets.Support: W81XWH-15-1-0292, P50CA116201-9, P50CA015083. Clinical trial information: NCT00114816 .