Perez, Edith A.

529 Background: Multiple prognostic models exist to predict late relapse risk in early stage hormone receptor-positive (HR+) breast cancer (BC). The CTS5 is one such model that has been validated in HR+ HER2-negative BC. The value of this model in HR+ HER2+ has not been established. Here, we assessed CTS5 in patients (pts) with early stage HER2+ BC treated in the NCCTG N9831 (Alliance) trial. Methods: Pts with stage I-III HER2+ HR+ BC who survived ≥ 5 years were included. The online CTS5 calculator was used to determine CTS5 score and risk group (low, intermediate, and high) based on age, tumor size, grade, and number of involved nodes. Kaplan-Meier (KM) estimates, Cox regression models, and C index were used for analysis. Results: From 3,130 pts, 1,204 pts met the criteria and were included. Median age was 49 (22-79) years and median tumor size was 2.4 (0.1-12) cm. 63.6% had grade 3 tumors, 33.6% grade 2, and 2.8% grade 1. Median follow up was 10.89 (5.01-15.32) years. Based on CTS5, 821 (68.2%) pts were classified as high risk, 289 (24%) as intermediate risk, and 94 (7.8%) as low risk. Overall, using univariate Cox regression analysis, there was no statistically significant difference in recurrence free survival (RFS) among pts with intermediate vs. low (HR 0.47 95%CI 0.18-1.22, p = 0.12) and high vs. low (HR1.23 95%CI0.57-2.67, p = 0.6) with the C index of 0.58. Among pts who received concurrent trastuzumab (H) with HR+ BC, there was also no statistical difference in RFS between high vs. low (HR 0.68 95%CI0.24-1.97, p = 0.48) with the C index of 0.55. Paradoxically, pts with intermediate risk had better RFS than low risk (HR 0.18 95%CI0.03-0.97, p = 0.05). As a continuous variable, there is also no significant improvement in RFS per 1 unit increase in CTS5 score (HR 1.19 95%CI 0.73-1.96, p = 0.49) with the C index of 0.54. After 5 years, 7.06% (n = 30/425) of HR+ pts treated with concurrent H recurred. Conclusions: The CTS5 model is not prognostic in pts with early stage HR+ HER2+ BC receiving adjuvant H. While most HR+ HER2+ pts are classified as high risk by CTS5, the recurrence between years 5-10 was low in pts who received adjuvant H. This study highlights the need to develop a new predictive model for risk of late relapse in this specific group of pts to enable clinicians to determine which pts would benefit from extended adjuvant endocrine therapy. Support: BCRF-19-161, U10CA180821, Genentech. https://acknowledgments.alliancefound.org Clinical trial information: NCT00005970 .

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 .

Abstract Introduction: Although HER2+ breast cancer (BC) is considered a moderately immunogenic tumor, several studies have shown a role of pre-existing immunity associated with favorable long-term prognosis and better response to treatment. In this study, we performed exploratory analyses to assess whether the efficacy of HER2 targeted treatment in the MARIANNE trial correlated with immune gene expression. Methods: MARIANNE (NCT01120184) is a phase 3 study in patients (pts) with centrally confirmed HER2+ local advanced/metastatic BC naïve to prior treatments in the advanced disease. Pts were randomized (1:1:1) to trastuzumab+taxane (HT), T-DM1, or T-DM1+Petuzumab (P) and the trial showed noninferior PFS of T-DM1 and T-DM1+P vs HT. Gene expression (RNA) analysis was performed on tumor samples by a custom 800-gene codeset on the nCounter platform. PD-L1, CD8 expressions and immune gene signatures (sign) analyses were assessed by multivariate Cox regression models using median (cut-off) as categorical variable and adjusted by prior HT, presence of visceral disease, world region, baseline ECOG, measureable disease at baseline, therapy setting, HER2 mRNA expression, PIK3CA mutation status. Results: MARIANNE randomized 1095 pts (HT, n=365; T-DM1, n=367; T-DM1+P, n=363). Gene expression results were available for 671 pts (61.3% of the intent-to-treat [ITT] population) which was representative of ITT. In ITT, HR below 1 was observed when comparing pts with high (>median) vs low (≤median) immune gene expression by clinical outcome suggesting a potential association of high immune marker expression with improved PFS (Table 1) and to some extent with OS (data not shown). This association was primarily observed in the T-DM1 arm where the HR suggested a risk reduction of disease progression(PD)/death especially in the high Teff, high PD-L1 and high CD8 subgroups, and to some extent in the HT arm (Table 1). When assessing the predictive impact on PFS by comparing T-DM1 vs HT, HR below 1 was observed especially in pts with high Teff signature, high PD-L1 and high CD8 expressions (HR 0.67 (95% CI (0.47-0.95)), HR 0.68 (95% CI (0.48-0.97), and HR 0.64 (95%CI 0.44-0.93), respectively). When comparing T-DM1+P vs. HT, HR below 1 was observed especially in pts with low Teff signature and low PD-L1 expression (HR 0.70 (95% CI (0.50-0.99), and HR 0.68 (95% CI (0.48-0.96) respectively). No clear differences between immune gene expression subgroups was observed when comparing treatment arms in regards to OS (data not shown). Conclusions: In the exploratory analysis from the MARIANNE study, high immune gene expression, especially in the high PD-L1, CD8 and Teff subgroups, showed an association with improved clinical benefit with HRs reflecting for a risk reduction of PD/death for PFS and partially for OS. This association was less obvious in the T-DM1+P arm. When comparing the treatments effect, the data showed a potential impact of high Teff signature, and high CD8 and PD-L1 expressions on T-DM1 and less on HT. The potential opposite association of low Teff signature and low PD-L1 expression with improved benefit in the T-DM1+P arm was unexpected and needs further investigation. Table 1: Prognostic biomarker effect on PFSBiomarker by categories (>Median vs ≤Median)HR (95% CI) ITT n=671HR (95% CI) HT n=220HR (95% CI) T-DM1 n=227HR (95% CI) T-DM1+P n=224Teff sign0.89 (0.73-1.09)0.97 (0.68-1.38)0.64 (0.45-0.91)1.09 (0.75-1.58)Th1 cytokine sign0.91 (0.74-1.11)0.92 (0.64-1.31)0.78 (0.55-1.11)0.96 (0.67-1.36)Checkpoint inhibitor sign0.95 (0.78-1.15)0.91 (0.64-1.29)0.90 (0.64-1.26)1.02 (0.71-1.47)PD-L10.80 (0.66-0.98)0.79 (0.55-1.13)0.62 (0.44-0.87)1.07 (0.74-1.55)CD80.91 (0.75-1.11)1.10 (0.77-1.57)0.66 (0.46-0.93)0.98 (0.68-1.41) Citation Format: Edith A Perez, Sanne Lysbet de Haas, Carlos H Barrios, Wolfgang Eiermann, Masakazu Toi, Young-Hyuck Im, Pier Franco Conte, Miguel Martin, Tadeusz Pienkowski, Xavier B Pivot, Howard A Burris III, Chiara Lambertini, Silke Hoersch, Monika Patre, Paul Anthony Ellis. Association of immune gene expression with outcome in the MARIANNE phase 3 clinical trial in HER2-positive metastatic breast cancer (MBC) [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr PD5-11.

Abstract Background: Growing evidence supports the critical role of preexisting immune response in triple negative breast cancer (TNBC). However, there are limitations with current evaluation approaches: inability to functionally assess the type of immune infiltration with traditional pathologic evaluation and loss of spatial distribution in conventional high-plex gene or protein expression analyses from the whole tumor section. Here, we report the initial analysis of immune protein expression as a function of spatial distribution and clinical outcomes in TNBC samples. Methods: NanoString GeoMxTM Digital Spatial Profiling (DSP) was used to quantify 39 immune-related proteins in stromal and tumor segments from 44 TNBC samples from the FinXX trial. Samples were matched for patient characteristics, treatment arm (capecitabine vs. 5-fluorouracil), and outcome based on recurrence-free survival (RFS) with 22 samples from patients who recurred and 22 samples from patients with durable RFS. Regions of interest (ROIs) were selected based on expression of cytokeratin (tumor), CD45 (leukocytes), or CD68 (macrophages). Each ROI was segmented into tumor (pancytokeratin-positive area) and stroma (pancytokeratin-negative/nuclear SYTO13-positive area). The general linear model was used for statistical analysis of differential expression with RFS as a categorical variable (recur yes or no). Results: A total of 950 tumor and stroma segments were included in this initial analysis. In both tumor and stroma segments, over-expression of T cell activation markers (CD137, GITR) was associated with better outcome, whereas T cell markers (CD3, CD4, CD8) were not significantly associated with outcome. In tumor segments alone, improved outcome was significantly associated with increased protein expression [> 2-fold change (FC) at p<0.001] of CD56, PD-L2, HLA-DR, CD137, GITR, and CD40. In CD45-enriched stroma, improved outcome was associated with elevated expression (FC>2.0, p<0.001) of PD-L2; whereas durable RFS was associated with elevated PD-L2 and IDO1 expression in CD68-enriched stroma. In contrast, macrophage/dendritic cell markers CD68 and CD11c were not associated with outcome. In tumor cells adjacent to CD45-enriched stroma, durable RFS was associated with increased abundance (FC>2.0, p<0.001) of PD-L2, CD56, CD27, GITR, CD20, HLA-DR, and IDO1. Similarly, in tumor segment associated with CD68-enriched stroma, proteins associated with RFS included PD-L2, HLA-DR, CD56, GITR, and CD137. Among 39 immune function proteins, only elevated CTLA4 expression in CD68-enriched stroma was associated with recurrence (FC=0.55, p<0.001), whereas tumor-segment CTLA4 was associated with RFS. Conclusions: Using an in-depth analysis to precisely quantify the abundance of multiple immune function proteins in a spatially defined manner, we observed that PD-L2, IDO1, and T cell activation markers were robustly associated with durable RFS in both tumor and stromal segments. In contrast, MHC components (HLA-DR, beta-2-microglobulin), B cell markers (CD20), and NK cell markers (CD56) were strongly associated with favorable outcome in tumor but not in stromal segments. Our study highlights the conclusion that the immune landscape of TNBC is far too complex to be encompassed by any single molecular marker, and more detailed analyses of the DSP data reported here are ongoing with a view towards using quantitative multiplex analyses to refine our understanding of how therapeutic outcome is influenced by interactions among immune cells and between immune and tumor cells. Acknowledgements: Supported by the Breast Cancer Research Foundation (BCRF18-161), Bankhead Coley (6BC05 Florida Department of Health), 26.2 with Donna Foundation, US National Cancer Institute (CA15083), and the Canadian Cancer Center Citation Format: Saranya Chumsri, Douglas Hinerfeld, Jennifer M. Kachergus, Yaohua Ma, Heather A Brauer, Sarah Warren, Xue Wang, Torsten O. Nielsen, Karama Asleh, Heikki Joensuu, Edith A. Perez, E. A. Thompson. Digital spatial mapping of the immune landscape of triple negative breast cancer reveals novel features of immune-tumor cell interaction [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr PD5-06.

TPS1136 Background: Hsp90 is a molecular chaperone protein required for the stabilization and activation of many proteins, referred to as Hsp90 ‘clients’, including those commonly implicated in breast tumorigenesis such as HER2, EGFR, ER, PI3K, AKT, P53 and VEGFR. Ganetespib is a novel triazolone inhibitor of Hsp90 being studied in over 20 clinical trials, with over 700 patients treated to date. Ganetespib is ~50x more potent than 1st-generation Hsp90 inhibitors, and has been well tolerated in clinical trials with a favorable safety profile. Mild to moderate, transient diarrhea is the most common adverse event associated with ganetespib infusion and is manageable with appropriate supportive care. Ganetespib has shown activity in preclinical models of HER2+, ER+/PR+ and triple negative breast cancer (TNBC). In a phase I trial, ganetespib demonstrated single-agent clinical activity in HER2+ disease with an objective response rate (ORR) of 15% and a disease stabilization rate of 46% in heavily pretreated patients. This efficacy-screening study is designed to provide further evidence of ganetespib activity and identify potentially predictive biomarkers in metastatic breast cancer. Methods: The ENCHANT-1 trial is an international, first-line Phase II study in two cohorts of breast cancer patients: Cohort A, HER2 amplified (n=35), and Cohort B, TNBC (n=35). HER2+ patients must have received prior anti-HER2 therapy in the adjuvant setting. Patients are treated with ganetespib at 150 mg/m2 twice weekly on a three out of four-week regimen. Primary endpoint: ORR assessed using RECIST1.1 criteria. Key secondary endpoints include metabolic effects as assessed by PET/CT at week 3. Tumor genetic signature and proteomic profiling are performed on patient’s tumors in an effort to develop biomarkers of response. At the time of submission 6 patients have been enrolled into this study. Clinical trial information: NCT01677455.

1055^ Background: Key properties of QoL instruments in a clinical trial setting are reliability, the ability to detect a change, and content validity. Using data from the Study 301 phase III breast cancer trial, we report here content validity and ability to detect a change for the EORTC QLQ-C30 and breast cancer-specific QLQ-BR23 questionnaires. Methods: Patients with locally advanced or metastatic breast cancer were randomized to 21-day cycles of either eribulin mesylate 1.4 mg/m2 given on Days 1 and 8, or capecitabine 1.25 g/m2BID orally on Days 1-14. QoL questionnaires were completed at baseline and at 6 weeks, 3, 6, 12, 18, and 24 months. Objective tumor response was evaluated (complete response [CR]; partial response [PR]; stable disease [SD]; progressive disease [PD]). Univariate and multivariate longitudinal analyses using weighted generalized estimating equations were employed to assess the responsiveness of the QoL scales to objective tumor response. Results: 1,102 patients were randomized (554 eribulin, 548 capecitabine). Global health status (GHS)/QoL scores were low at baseline (55). GHS/QoL scores were highest for patients with CR (70.8), followed by those patients with PR (63.5), SD (60.5), and PD (58.1). Physical functioning followed the same pattern: CR (98.3); PR (79.1); SD (72.8); PD (71.0). Role and social functioning scores were also responsive. Pain increased, while fatigue and body image worsened, with poorer tumor responses. Using the weighted generalized estimating equations, there were improvements in physical (34.78; p<0.01), cognitive (27.29; p<0.01), and social (22.04; p<0.01) functioning, and future perspective 11.47 (p<0.01), in patients who responded (CR and PR) to treatment compared with non-responders. Pain decreased significantly by 28.62 (p<0.01) on a 0-100 scale. Patients who responded also gained appetite and had fewer breast symptoms. Conclusions: These results suggest content validity of the EORTC QLQ-30 and QLQ-BR23 questionnaires as they correlate with changes in objective tumor assessments. Clinical trial information: NCT00337103.

1007 Background: Determining optimal adjuvant chemotherapy for early stage breast cancer depends on efficacy and toxicity. We sought to determine if T is equivalent to AC but with reduced toxicity. Methods: Pts with operable breast cancer with 0-3 positive nodes were enrolled on a 2x2 factorial design study which addressed (1) superiority of 6 vs. 4 cycles of therapy (previously reported, Shulman, JCO 2012) and (2) equivalence of single-agent T to standard AC, defined as upper bound of 95% confidence interval (CI) of hazard ratio (HR) of T vs. AC < 1.30 for the primary endpoint of relapse-free survival (RFS). A planned target of 567 RFS events required 4,646 pts with 4 yrs FU. At activation in 2002, T (80mg/m2) was q1wk for 12 or 18 wks and AC (60/600 mg/m2) was q3wk for 4 or 6 cycles. In 2003 (570 pts enrolled) schedules were revised to 4 or 6 cycles q2wk for both T (175 mg/m2) and AC. The 6-cycle arms were dropped in 2008 (3,171 pts enrolled) due to slow accrual. Relative effectiveness of T to AC is shown by hazard ratio (HR). Logrank p-values are measures of discordance but are not relevant for the equivalence question and are not adjusted for multiple comparisons. Results: After enrolling 3,871 pts, the study closed in 2010 due to slowing accrual. With a median follow-up of 6.1 yrs there are 437 RFS events. The HR of 1.26 (95% CI: 1.05-1.53; p = 0.02) does not allow a conclusion of equivalence of T with AC. With 266 deaths the HR for overall survival (OS) is 1.27 (95% CI=1.00-1.62; p = 0.05), favoring AC. The estimated absolute advantage of AC at 5 yrs is 3% (91 vs. 88%) for RFS and 1% (95 vs. 94%) for OS. All 9 treatment-related deaths were in pts receiving AC and are included in the survival analysis. The incidence of Grade 3+ toxicity for AC vs T was 33% vs. 4% for hematologic toxicity and 36% vs 22% for non-hematologic toxicity. Conclusions: This trial did not show equivalence of T to AC, a conclusion that is very unlikely to change with additional follow-up. T was less toxic than AC. Clinical trial information: CDR0000069444/NCT00041119.

520 Background: There is considerable interest in developing HER2 testing criteria for adjuvant H. We used the 21-gene assay to examine the relationship of HER2 mRNA to benefit from H. Methods: N9831 compared adjuvant chemotherapy AC-T to concurrent chemotherapy-trastuzumab AC-TH in stage I-III HER2+ breast cancer. Recurrence Score (RS) and HER2 mRNA expression were determined by Oncotype DX (neg<10.7, equiv 10.7 to <11.5, and pos ≥11.5 log2 expression units). Cox regression was used to assess the association of HER2 expression with H benefit for distant recurrence. Results: Median follow-up: 7.4 yrs. Of 1,940 total pts, 901 had consent and sufficient tissue. HER2 by RT-PCR was neg in 130 (14%), equiv in 85 (9%), and pos in 686 (76%) pts. Concordance between HER2 assessments was 95% for RT-PCR vs central IHC (>10% + cells = +), 91% for RT-PCR vs central FISH (≥2.0 = pos) and 94% for central IHC vs central FISH. In the primary analysis, the association of HER2 expression with H benefit was marginally non-significant (P=0.057). In hormone receptor pos pts (local IHC) the association was significant (P=0.002). The association was nonlinear with the greatest estimated benefit at lower and higher HER2 mRNA expression levels. The observed treatment benefit in low HER2 pts was not due to imbalance between arms in RS and individual gene expression values. Conclusions: Concordance among HER2 assessments by central IHC, FISH, and RT-PCR was high. Association of HER2 mRNA expression with H benefit was marginally non-significant. A consistent benefit of trastuzumab irrespective of mHER2 levels was observed in the pts with either IHC+ or FISH+ tumors. Benefit was observed in pts with high HER2 by RT-PCR but also observed for the small groups of pts with negative results by quantitative RT-PCR or FISH (Table). Plausible mechanisms for this observation will be discussed. [Table: see text]

e13501 Background: Loss of estrogen receptor alpha gene expression has been associated with insensitivity to endocrine therapy in human breast cancer patients. Histone deacetylase (HDAC) inhibitors have recently been found to restore sensitivity to the estrogen receptor by modulation of the estrogen and progesterone receptors. This had been shown with both aromatase and tamoxifen refractory and in triple negative cell lines . We performed a Phase I study of the combination of LBH589 (panobinostat) and letrozole to evaluate safety and tolerability in patients with metastatic breast cancer prior to the performance of a phase II trial. Methods: We enrolled postmenopausal women with metastatic breast cancer, ECOG PS 0 or 1, ANC>1500/mm3, platelets>100,000/mm3, normal total bilirubin, and ALT/AST adequate laboratory tests were eligible. Letrozole dose was 2.5 mg/day orally. Dose of LBH589: Level 1, 20 mg orally three times weekly; Level 2, 30 mg orally three times a week. Results: 12 patients (dose level 1:6 patients, dose level 2: 6 patients) have been enrolled. 43 cycles of treatment have been given to these12 patients. Initial cohort of 3 patients at the 20 mg dose level had no dose limiting toxicity (DLT). At the 30 mg dose level 3/6 patients had DLT (thrombocytopenia grade 1: 2 pts; grade 3:1 pt; grade 4: 1 pt; and diarrhea grade 3: 1 pt. One pt at the 30 mg dose level has a confirmed partial response and remains on study after 6 cycles of treatment. Subsequent cohort of 3 patients had 1 dose DLT with doubling of the creatinine. The main DLT was thrombocytopenia in 3/6 pts at the 30 mg dose level. Conclusions: The recommended dose for phase II testing of LBH589 is 20 mg orally 3 times per week in combination with standard dose letrozole.

1021 Background: At 5 years, AT did not improve disease free survival or overall survival and RS was a more accurate predictor of relapse than standard clinicopathologic characteristics for patients with hormone receptor (HR) positive tumors. Methods: A Phase III Intergroup trial tested adjuvant AT vs. AC. Women with 1-3 N + or N - and T-size > 1cm were randomized to 4 cycles of AT (60 mg/m2/60 mg/ m2) or AC (60 mg/m2/600 mg/m2) q 3 wk x 4. Patients(pts) with ER + and/ or PR + tumors received tam for 5 yrs. Pts were stratified by nodal, HR (ER+ PR+, ER+PR-, ER-PR+, ER-PR-, ER/PR unk) and menopausal status. The primary endpoint was DFS. A sample of 465 pts with HR + breast cancer with 0 to 3 positive axillary nodes who did (N =116) or did not have a recurrence had tumor tissue evaluated using the 21- gene assay. Grade and HR expression were evaluated locally and centrally. Results: 2952 pts were randomized between 7/30/98 and 1/21/00. 2883 were eligible and analyzable. Arms were balanced for age, HR, menopause, nodes, surgery, grade and T-size: median age 51; 64% ER +; 65% LN-; grade: 10% low, 38% int., 46% high; and median T-size - 2.0 cm. At a median follow-up of 11.5 years the DFS/OS results are shown in the table below. RS was a highly significant predictor of recurrence including node negative and node positive disease (P < .0001) and predicted recurrence more accurately than clinical variables. Conclusions: At 11.5 yrs. median follow-up, there remains no difference in DFS or OS, although there continue to be fewer events in the AT arm in the prespecified ER/PR negative subgroup. At 10 years, the RS continues to be a more accurate predictor of relapse than standard clinical features. [Table: see text]