Search

Background: Invasive lobular carcinoma (ILC) represents the second most common subtype of breast cancer after invasive breast cancer of no special type (NST). In this retrospective analysis of the MINDACT trial, we aimed at identifying/refining the transcriptomic differences between: 1) estrogen receptor positive/HER2-negative (ER+/HER2-) ILC versus ER+/HER2- NST, 2) classic and non-classic ER+/HER2- ILC, and, 3) recurring and non-recurring ER+/HER2- ILC in the subgroup of patients with a low clinical and low genomic (cL/gL) risk (as defined by a modified version of Adjuvant Online! and the 70-gene signature). Patients and methods: Central pathology review was performed for histological subtype, grade and Ki67 (G.V.) for 5929/6693 (88.6%) of the patients included in the MINDACT trial (NCT00433589). Analysis of transcriptomic data adjusted for age and grade was performed using the R/Bioconductor package ‘limma’ to identify differentially expressed genes (DEGs). DEGs having absolute log-fold change (logFC)≥ 0.2 and FDR-adjusted p-value (q-value) < 0.05 were considered. Gene set enrichment analyses (GSEA) of MSigDB hallmark gene sets were performed. Adjusted Cox regression models were used to evaluate the association of these hallmarks with disease free survival (DFS) and distant recurrence free survival (DRFS). Results: After central pathological review, 464 patients with ER+/HER2- ILC and 3798 patients with ER+/HER2- NST were identified. Patients with ILC were significantly older at diagnosis, had larger tumors, less axillary nodal involvement, more grade 2 tumors than patients with NST. At the transcriptomic level, we observed a high number of DEGs between these 2 subgroups, confirming their distinct phenotype. CDH1, the gene coding for E-cadherin, was as expected the most highly overexpressed gene in NST versus ILC. We further observed an increased expression of leptin (LEP), leptin receptor (LEPR), lipoprotein lipase (LPL), and the fatty acid transporter CD36 in ILC. This could suggest that ILC relied on increased lipid uptake thanks to the increased contact of ILC tumor cells with the adipocytes. IGF1 was also overexpressed in ILC versus NST, as a potential consequence of high LEP and high LEPR expression. Differences were also evident with regard to the extracellular matrix (ECM), with many collagens, matrix metalloproteinases (MMPs) and other key enzymes (e.g. LOXL1) being differentially expressed. We confirmed a decreased ER-signaling and increased PI3K/Akt signaling in ILC versus NST. Out of the 464 ER+/HER2- ILC tumors, 253 (55%) were classic ILC and 211 (45%) non-classic ILC. There were more grade 3 tumors, more highly proliferative tumors and more nodal involvement in patients with non-classic versus classic ILC. At the transcriptomic level, differences were subtler than the differences seen above. Still, a significant enrichment of the hallmarks related to cell cycle in the non-classic ILC, and of the hallmarks related to epithelial-to-mesenchymal transition, hypoxia, adipogenesis and IL6/JAK/STAT3 signaling in classic ILC was observed. Finally, 216/464 patients with ER+/HER2- ILC (47%) were assigned to the cL/gL risk group and did not receive chemotherapy. 28/216 of these patients (13%) relapsed (DFS, median FU: 8.7 years). Enrichment of hallmarks related to apoptosis, inflammatory response, hypoxia and oncogenic signaling (PI3K/Akt, Ras, c-Myc) was associated with worse survival. Conclusion: This represents, to the best of our knowledge, the largest set of gene expression data for patients with ILC, issued from a clinical trial where histology was reviewed centrally. These results could be used to personalize treatment for patients with ILC. This project is funded by the Breast Cancer Research Foundation. Citation Format: Christine Desmedt, Ha-Linh Nguyen, François Richard, Sabine Linn, Otto Metzger, Coralie Poncet, Jelle Wesseling, Florentine Hilbers, Kim Aalders, Mauro Delorenzi, Suzette Delaloge, Jean-Yves Pierga, Etienne Brain, Suzan Vrijaldenhoven, Peter A Neijenhuis, Karen Van Baelen, Marion Maetens, Emiel Rutgers, Martine Piccart, Laura Van ’t Veer, Giuseppe Viale, Fatima Cardoso. Transcriptomic insights into lobular breast cancer biology: a retrospective analysis of the MINDACT clinical trial [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr P5-14-01.

Background: Breast cancer with low HER2 expression (HER2-low) is of high clinical relevance because of new therapeutic options with antibody-drug conjugates. We have recently shown in a large cohort from neoadjuvant clinical trials that HER2-low breast cancer has different molecular characteristics as well as different clinical outcomes compared to HER2-zero. Considering the positive correlation between HER2-low expression and hormone receptor positivity observed consistently in many investigations, we have extended our analysis to HR+ tumors from the post-neoadjuvant PenelopeB trial. In PenelopeB, patients with HR+ breast cancer and residual disease after neoadjuvant chemotherapy (NACT) were randomized to post-neoadjuvant palbociclib versus placebo in addition to endocrine therapy. We evaluated the molecular phenotype and clinical outcomes of HER2-low compared to HER2-zero patients. Methods: A total of 1250 patients were randomized, HER2 status was available for 1151 tumors from pretherapeutic core biopsy, determined mainly by local pathology, and from 1213 tumors from the post-NACT sample, determined as part of central pathology. For 1119 patients a paired HER2-status was both available. HER2-zero was defined as IHC0 and HER2-low-positive was defined as IHC1+ or IHC2+/ISH-. Gene expression analysis of 2549 genes using the HTG oncology biomarker panel was performed in 620 pretherapeutic biopsies and 780 post-NACT residual tumor samples, with 539 paired gene expression samples. Breast cancer subtypes were determined using the AIMS approach. Results: In pretherapeutic biopsies, 695 tumors (60%) were HER2-low and 457 (40%) were HER2-zero. A HER2-low status in the biopsy was significantly linked to improved iDFS (HR 0.76 (0.60-0.96; p=0.02). In residual tumors, 632 tumors (60%) were HER2-low and 581 (40%) were HER2-zero, without any prognostic impact of HER2 low status. In addition, a shift of HER2-low-status comparing core biopsy and residual tumor was observed in 415 (37%) of 1119 tumors. 161 (14%) had a shift from HER2-zero to HER2-low and 254 (23%) shifted from HER2-low to HER2-zero. A shift from HER2-zero to HER2-low in the post-NACT samples was significantly linked to reduced iDFS (HR 1.43 [95%CI 1.01-2.01]), p=0.04), compared to HER2-low group, while a shift from HER2-low to HER2-zero was associated with better iDFS compared to HER2-zero group, although not statistically significant (p=0.17). We did not observe a significant correlation of HER2-low status and AIMS molecular subtypes. In particular, the HER2-enriched (HER2E) subtype was assigned to only 4.3% of HER2-zero and 3.1% of HER2-low tumors. Significant iDFS differences were observed for HER2-low-status in combination with AIMS subtypes (lumB/basal/HER2E vs. lumA/normL; overall p-value < 0.0001) for both pretherapeutic biopsies and residual tumor. Patients with post-NACT HER2-low tumors had an improved survival in the subgroups of aggressive AIMS subtypes (lumB/basal/HER2E), but not in the less aggressive AIMs subtypes (lumA/normL), with a positive test for interaction (p=0.02). For the pre-NACT samples a similar, but non-significant trend was observed. We evaluated a total of 620 core biopsies for differences in gene expression comparing HER2-low and HER2-zero tumors. A total of 417 genes were statistically significantly different, but in a hierarchical clustering there was no clear separation of HER2-low and HER2-zero tumors. Conclusions: In the PenelopeB cohort of HR+ tumors, a HER2-low status in pretherapeutic core biopsies is related to improved disease-free survival, especially for those tumors that have a more aggressive intrinsic subtype. A shift of HER2-low status was observed before and after chemotherapy, indicating an adaptation of the pathway activity to therapy-induced stress, which might become relevant for future diagnostic and therapeutic approaches. Citation Format: Carsten Denkert, Miguel Martín, Michael Untch, Hervé R. Bonnefoi, Erik S. Knudsen, Seock-Ah Im, Angela DeMichele, Agnieszka Witkiewicz, Laura Van ’t Veer, Sung-Bae Kim, Harry D. Bear, Nicole McCarthy, Karen Gelmon, Frederik Marmé, José Ángel García-Sáenz, Nicholas Turner, Federico Rojo, Martin Filipits, Lesley-Ann Martin, Peter A. Fasching, Christian Schem, Catherine M. Kelly, Toralf Reimer, Masakazu Toi, Hope Rugo, Michael Gnant, Andreas Makris, Yuan Liu, Karsten Weber, Sivaramakrishna Rachakonda, Sibylle Loibl. HER2-06 Outcome analysis of HER2-zero or HER2-low hormone receptor-positive (HR+) breast cancer patients - characterization of the molecular phenotype in combination with molecular subtyping [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr HER2-06.

Background: Breast cancer (BC) is one of the cancer types recognized as an obesity-associated disease. Current understandings of molecular mechanisms underlying the BC-obesity connection however largely came from experimental models while systematic investigation of the impact of obesity on BC biology in large patient series is still lacking. The purpose of this study is to discover changes in the transcriptomic profile of primary BC according to patients’ body mass index (BMI). Data and Methods: Bulk and single-cell gene expression data from treatment-naïve primary breast tumors from non-underweight patients were retrieved from the MINDACT trial (NCT00433589; N = 1481) and the pre-treatment cohort of the BioKey trial (NCT03197389, N = 36), respectively. Three categories were considered for BMI: lean, overweight and obese. The main analyses focused on the invasive carcinoma of no special type (NST) estrogen receptor-positive/HER2-negative (ER+/HER2-, N_bulk = 735, N_single-cell(sc) = 10) and NST ER-/HER2- (N_bulk = 118, N_sc = 15) subgroups. The bulk expression data was subjected to differential gene expression analyses according to BMI which was adjusted for menopausal status and tumor grade, then followed by gene set enrichment analyses. Clustering and cluster annotation were performed on the single-cell profiling data before differentially expressed genes according to BMI were identified for each of the present cell types. Results: Obesity-associated differences in the transcriptomic profile of breast tumors, which were subtle but potentially indicative of a biological relationship, were revealed by the bulk data. In both investigated subgroups, tumors from obese patients were shown to be enriched in cell cycle hallmarks. In ER-/HER2- tumors, adiposity further increased MYC signaling. We also observed different obesity-associated changes according to the ER status. Among ER+/HER2- tumors, those from obese patients were enriched in hallmarks related to inflammatory response compared to those from lean patients. In contrast, these hallmarks appeared to be enriched in the ER-/HER2- tumors from lean patients. Our investigation of the single-cell data further revealed shifts in the cell composition of tumor tissue and cell type-specific transcriptomic differences according to BMI which were more pronounced than those detected from the bulk data. ER+/HER2- tumors from obese patients have a higher frequency of immunosuppressive and pro-tumoral cell subpopulations such as dendritic cells (DC) enriched in immunoregulatory molecules (p = .03), LYVE1+ macrophages (p = .02) and myofibroblasts (p = .03) than those from lean patients. Overexpression of Cyclin D1 and CD24 was found in cancer cells in ER+/HER2- tumors from obese patients. A reduction in anti-tumor immune responses was evident with downregulation of multiple interferons in CD8+ and CD4+ T cells as well as B cells. We observed in the ER-/HER2- subgroup increased infiltration of plasmacytoid DC (p = .01), CCL2+ macrophages (p = .01) in tumors from obese versus lean patients, while fibroblasts showed an opposite tendency. Additionally, significant obesity-associated downregulation of major histocompatibility complex (MHC) molecules class I in cancer cells and MHC class II molecules in B cells could be suggestive of deficient antigen presentation and activation of cytotoxic and helper T cells. Conclusion: We highlighted the impact of obesity on the remodeling of tumor and tumor microenvironment which might generally lead to a suppression of anti-tumor immune responses, albeit potentially via diverse axes according to the ER status. Although investigation on a larger cohort is warranted, our current results suggest that obesity-associated transcriptomic changes in BC could be highly cell type-specific, hence we recommend single-cell approaches in addition to spatial multi-omics analysis to further elucidate the interplay between obesity and BC. Citation Format: Ha-Linh Nguyen, Tatjana Geukens, Marion Maetens, Karen Van Baelen, Maxim De Schepper, Coralie Poncet, Mauro Delorenzi, Marjanka K. Schmid, Emiel Rutgers, Laura Van ’t Veer, Martine Piccart, Fatima Cardoso, Giuseppe Viale, Ayse Bassez, Hanne Vos, Patrick Neven, Ines Nevelsteen, Kevin Punie, Hans Wildiers, Giuseppe Floris, Diether Lambrechts, Ann Smeets, Elia Biganzoli, François Richard, Christine Desmedt. PD4-06 Obesity-associated changes in transcriptomic profile and immune landscape of primary breast cancer revealed by bulk and single-cell gene expression data [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr PD4-06.

Introduction: One of the principal limiting factors to achieving cures in patients with cancer is drug resistance. Drug repositioning is the application of FDA-approved drug compounds for novel indications beyond the scope of the drug’s original intended use. This approach offers advantages over traditional drug development by reducing development costs and providing shorter paths to approval, as drug safety has already been established during the drug’s original regulatory process. One approach for computational drug repositioning involves generating a disease gene expression signature and then identifying a drug that can reverse this disease signature. In this study, we extracted drug resistance signatures from the I-SPY 2 TRIAL by comparing gene expression profiles of responder and non-responder patients stratified by treatment and molecular subtype. We then applied our drug repositioning pipeline to predict compounds that can reverse the gene expression profiles of these drug resistance signatures. We hypothesize that reversing these drug resistance signatures will resensitize tumors to treatment and improve patient outcome. Methods: We first generated the drug resistance signatures by performing differential expression between responders (RCB 0/I) and non-responders (RCB III) within treatment arms and molecular subtypes. An optimal log fold-change cutoff was selected for each signature by identifying the cutoff that best separates the responder and non-responder samples using k-means clustering. We then applied our drug repositioning pipeline to identify compounds that significantly reverse these signatures using the drug perturbation profiles generated in a breast cancer cell line in the Connectivity Map v2 dataset. Briefly, the pipeline uses a non- parametric, rank-based pattern-matching strategy based on the Kolmogorov-Smirnov (KS) statistic to assess the enrichment of resistance genes in a ranked drug gene expression list. Significance of each prediction is estimated from a null distribution of scores generated from random gene signatures. Results: We found that few individual genes are shared among the resistance signatures across the treatment arms and molecular subtypes, with the most common genes present in only 5/17 of the treatment arm and molecular subtype groups. At the pathway-level, however, we found that immune-related pathways are generally enriched among the responders and estrogen-response pathways are generally enriched among the non-responders. Although most of our drug predictions are unique to treatment arms and molecular subtypes, our drug repositioning pipeline identified the selective estrogen receptor degrader (SERD) fulvestrant as a compound that can potentially reverse resistance across a majority of the treatment arms and molecular subtypes. Conclusion: We applied our drug repositioning pipeline to identify novel agents to sensitize drug-resistant tumors in the I-SPY 2+ clinical trial and identified a SERD, fulvestrant, as a potential candidate for multiple molecular subtypes and treatment arms. Citation Format: Katharine Yu, Amrita Basu, Christina Yau, Denise Wolf, Gillian Hirst, Laura Sit, Nicholas O’Grady, Thelma Brown, I-SPY 2 TRIAL Investigators, Angela DeMichele, Don Berry, Nola Hylton, Doug Yee, Laura Esserman, Laura van 't Veer, Marina Sirota. Computational drug repositioning for the identification of new agents to sensitize drug-resistant breast tumors across treatment arms and molecular subtypes [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr PS11-04.

Background: I-SPY 2 is a neoadjuvant platform trial open to patients with locally advanced, molecular high-risk breast cancer. In a concerted pursuit of mid-therapy response biomarkers, we evaluated inter-regimen biopsies, to identify patients who may be candidates for treatment de-escalation. In a pilot study, we observed that absence of carcinoma in an inter-regimen biopsy may predict pathologic complete response (pCR). In this expanded study of 100 participants, we sought to confirm that finding and assess pathologic features of the inter-regimen biopsy as predictors of tumor response to neoadjuvant therapy. Methods: Digital H&E images of 100 inter-regimen (12 week) image-guided breast biopsies +/- ancillary immunohistochemistry (p63 and/or cytokeratin) were reviewed by 9 I-SPY affiliated pathologists to record 1) tumor bed and 2) presence/absence of residual invasive carcinoma (IC) (with tumor cellularity scored as 0-100%). The data set included 393 cores (mean 3.9 (2-4) cores/biopsy). Fisher’s exact t-test was used for association of presence/absence of IC with pCR, and tumoral hormone receptor (HR) and HER2 status. Association between biopsy tumor cellularity and residual cancer burden (RCB) indices used Pearson’s correlation. Results: In the biopsy set, 84 (84%) had ≥80% inter-observer diagnostic agreement on both 1) presence of tumor bed and 2) presence/absence of IC (53 IC+ /31 IC-). IC+/IC- biopsies had equal numbers of evaluable tissue cores. The primary tumors were 63% HR+/37% HR-. The presence of IC in the biopsy correlated with tumoral HR/HER2 status (p=0.0014: 74%: HR+HER2-; 62%: TN; 60%: HR+HER2+; 10%: HR-HER2+). Of 31 patients with IC- biopsies, 25 (80%) went on to pCR, whereas only 7/53 (13%) of patients with IC+ biopsies had pCR, conferring an odds ratio for pCR of 26, Fisher p=7.5E-10. Overall, IC- biopsies had a positive predictive value (PPV) for pCR of 81%, with a PPV for HR- tumors of 94% vs. 67% for HR+ tumors (Table 1). In the 6 IC- biopsies from patients with non-pCR (“false-negatives”), most were HR+ (5/6, Table 1), and tumor bed size in the resection specimen was smaller than for IC+ biopsies with non-pCR: 276 mm2 (0.4-1000 mm2) vs. 1166 mm2 (1-11960 mm2). In contrast, the 46/53 IC+ biopsies in patients with non-pCR had a PPV for predicting non-pCR of 86%, (PPV for HR+ tumors: 94% vs. PPV for HR- tumors: 66%. Tumor cellularity in the biopsy (mean 37%, [2.5-93%]) did not correlate with RCB index (p=0.57) or RCB breast-only index (p = 0.17) at resection. Conclusion: In this 100 biopsy set from the I-SPY2 trial, the absence of residual carcinoma in inter-regimen biopsies was highly predictive of pCR, particularly for HR- tumors. The “false-negative” biopsies (IC-/non-pCR) were predominantly HR+ tumors with small residual tumor beds at resection. Conversely, the presence of carcinoma in inter-regimen biopsies was highly predictive of non-pCR, particularly for HR+ tumors. These data demonstrate the utility, and the limitations, of the inter-regimen biopsy as one tool to identify patients who may benefit from therapeutic de-escalation. Table 1: PPV for pCR/non-PCR by Inter-regimen Biopsy StatusInter-regimen biopsy with or without Invasive carcinoma (IC+/-)pCRnon-pCRPPV (Sensitivity) for pCR(IC- Biopsies)PPV (Sensitivity) for non-pCR(IC+ biopsies)IC- biopsiesAll25681% (78%)-HR+10567% (83%)-HR-15194% (75%)-IC+ biopsiesAll746-86% (88%)HR+236-94% (88%)HR-510-66% (91%) Citation Format: Jodi M Carter, Molly E Klein, Sara J Venters, Kimmie Rabe, I Tolgay Ocal, Kamaljeet Singh, Denise M Wolf, Sunati Sahoo, Shuko Harada, Laila Khazai, Malini Harigopal, Alexander D Borowsky, Gregor Krings, Ronald Balassanian, Yunn-Yi Chen, Kimberley Cole, Sonal Shad, Amy Delson, Lamorna Brown-Swigart, I-SPY 2 TRIAL Consortium, Laura Esserman, Laura van ‘t Veer, W Fraser Symmans. Pathologic features of the inter-regimen biopsy predict response to neoadjuvant therapy in the I-SPY 2 TRIAL [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr PS4-09.

Background: The I-SPY 2 TRIAL, open to patients with locally advanced, molecular high-risk breast cancer, aims to bring each patient to pathologic complete response (pCR) with a minimum of toxicity. Here we test the hypothesis that imaging (MR volume predictors) combined with core biopsy may be used to accurately select candidates who show early response and provide an option of treatment de-escalation at mid-therapy (12 weeks). Methods: Of 100 I-SPY 2 patients with pathologist-assessed core biopsies at the inter-regimen time point (~12 weeks through treatment) and pCR data, 87 also had serial MR images and were considered in this study. Eleven I-SPY 2 TRIAL pathologists independently provided a digital assessment of the presence or absence of residual invasive cancer from H&E stained, and any requested ancillary IHC, images from imaging-guided core biopsies. Pathology predicts pCR if there is a consensus of no invasive residual disease. We generated predictions for all (55) unique pairs over the 11 pathologists, where pCR is predicted if both pathologists find no invasive cells. MRI pCR prediction models were previously developed on an independent dataset of ~990 I-SPY 2 patients, and applied to this cohort. Volume-based prediction models were previously optimized within each subtype and predicted probability thresholds were selected over a range of positive predictive value (PPV). In this study, MR predicts pCR (positive test) if the predicted probability is above a threshold that yields a given PPV value. For each pathologist pair, we combined pathology-based and MR-based predictors into a predictive-RCB (pre-RCB); and pre-RCB predicts a patient as pCR (RCB0) if both MR and pathology predicts pCR. Predictive performance is assessed by calculating the mean and range of PPV and sensitivity.Results: 39% (34/87) of the patients in this study achieved pCR. Over all pairs of pathologists, on average 80% of pathology-only predicted pCRs were true pCRs (mean PPV = 80% [range: 69-92%]), and 74% of patients who achieved pCR were predicted pCR by pathology alone (mean sensitivity = 74% [65-82%]). We assessed combinations with MR probability thresholds at PPV levels 50%-70%; and observed the best balance of PPV and sensitivity for the pre-RCB when MR thresholds were set at 50% PPV level. At this threshold setting, the pre-RCB achieved a PPV = 92% [83-100%], meaning on average 92% of predicted pCRs were true pCRs, and this improvement in positive predictive performance over pathology alone is achieved with a lower but still-reasonable 53% sensitivity [33-62%]. Conclusion: Pre-RCB, which predicts a patient as pCR if both MR and inter-regimen pathology predicts pCR, provides clinically actionable accuracy for treatment de-escalation for early responders (PPV>90%). Adding a final MR review at the time of early surgery may further improve performance. Resulting from data presented in this abstract, the pre-RCB algorithm, including the final MR review, has been operationalized and will be used prospectively to identify patients who are highly likely to have already achieved pCR by the inter-regimen timepoint. Citation Format: Sara J Venters, Wen Li, Denise M Wolf, Jodi M Carter, Molly E Klein, Kamaljeet Singh, Kimmie Rabe, I Tolgay Ocal, David Newitt, Christina Yau, Natsuko Onishi, Jessica Gibbs, Sunati Sahoo, Shuko Harada, Laila Khazai, Malini Harigopal, Alexander D Borowsky, Gregor Krings, Ronald Balassanian, Yunn-Yi Chen, Kimberley Cole, Sonal Shad, Barbara LeStage, Amy Delson, Sandra Finestone, Lamorna Brown-Swigart, I-SPY 2 Imaging Working Group, I-SPY 2 TRIAL Consortium, Laura Esserman, Laura van ‘t Veer, W Fraser Symmans, Nola M Hylton. Serial MRI and pathology combined to select candidates for therapy de-escalation in the I-SPY 2 TRIAL [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr PS4-10.

Background: Disseminated tumor cells (DTCs) in bone marrow detected after treatment may represent occult residual disease. We enumerated DTCs after neoadjuvant chemotherapy (NACT) in patients (pts) diagnosed with high-risk early stage breast cancer and examined the relationship of these cells with response and survival. Methods: I-SPY SURMOUNT is a sub-study of the I-SPY 2 TRIAL (NCT01042379). Pts enrolled on I-SPY 2, who signed consent for this sub-study, had bone marrow aspirates (BMA) collected after NACT at the time of surgery. DTCs were isolated and enumerated from BMA using immunomagnetic enrichment/flow cytometry (IE/FC). DTCs were defined as EPCAM-positive and CD45-negative nucleated cells. Samples were considered positive using a predetermined threshold of >4 DTCs per mL (Magbanua et al, unpublished data). Pathologic response was assessed using the residual cancer burden (RCB) method at local sites, and pts underwent standard adjuvant therapy if indicated and follow up for recurrence events and death. Relationship of DTCs with clinicopathologic variables was examined using Chi-squared test. Group means were compared using t tests. The log-rank test was used to compare survival curves. Results: A total of 73 patients were enrolled, 51 of whom had successful DTC assessment. The median DTC per mL was 4 (interquartile range 1.2-11.6). 24/51 (47%) were DTC-positive. Clinical characteristics by DTC status are shown in the table. DTC-positive pts were significantly younger (p=0.02) and had larger pretreatment tumors (longest diameter by magnetic resonance imaging) compared to DTC-negative pts (p=0.032). DTCs were not associated with receptor subtype. Thirty pts (41%) achieved a pathologic complete response (pCR). DTCs were not associated with pCR (p= 0.166); however, DTC-positive patients were significantly more likely to have residual cancer (RCB-II/III) after NACT compared to DTC-negative patients (OR 3.3, p=0.037). Median follow up of this cohort was 2.8 years (range: 0.9-4.8). Interim survival analysis showed that DTCs were not significantly correlated with EFS (p=0.6) or DRFS (p=0.41). Conclusions: Detection of DTCs at surgery after NACT is significantly more common in young patients, those with larger tumors, and those with residual disease at surgery. While these associations suggest higher risk for later recurrence, larger studies and longer follow up are necessary to determine if DTCs add prognostic value over pathologic evaluation alone for pts receiving NACT. Citation Format: Mark Jesus M Magbanua, Laura van 't Veer, Amy Clark, A. Jo Chien, Judy Boughey, Heather Han, Anne Wallace, Heather Beckwith, Minetta Liu, Christina Yau, E. Paul Wileyto, Lamorna Brown Swigart, Jane Perlmutter, Lauren Bayne, Shannon Deluca, Stephanie Yee, Erica Carpenter, Laura Esserman, John Park, Lewis Chodosh, Angela DeMichele. Outcomes associated with disseminated tumor cells at surgery after neoadjuvant chemotherapy in high-risk early stage breast cancer: The I-SPY SURMOUNT study [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr PS2-07.

Background: The Athena Breast Health Network is a multi-site effort aimed at integrating cutting-edge research and clinical care for breast cancer screening and treatment. At the University of California, San Francisco (UCSF) Breast Care Center (BCC), an Athena-based electronic Health Questionnaire System (eHQS) is distributed to all new patients. This questionnaire assesses self-reported personal and familial health history, comorbidities, and Patient-Reported Outcome Measurement and Information System (PROMIS) quality of life domains. Based on these responses, patients are proactively triaged to genetic counseling, psycho-oncology, onco-fertility, smoking cessation, peer support, nutritional counseling, behavioral sleep, and social work. These referrals are based on pre-defined thresholds, such as Ashkenazi Jewish Ancestry (referral to Genetic Counseling), or PROMIS Depression and Anxiety scores greater than or equal to 9 and 11, respectively. This analysis aims to assess the eHQS as a means to refer patients to supportive care services and assess if high-need patients (those with the highest symptom burden as assessed by quality of life PROMIS scores) are referred to services more frequently. Patients and Methods: Since Athena’s launch in 2013, over 6,000 patients have completed the online intake questionnaire at the UCSF BCC, 4,984 of whom have consented to research. 3,627 patients completed all relevant data to this study and were used in our analysis. A review of patient referrals was conducted to assess the number of referrals as a result of eHQS completion. PROMIS T-Scores in 8 domains (depression, anxiety, fatigue, sleep-related impairment and disturbance, cognitive function, applied cognition and physical function) were calculated with the Health Measures system and compared across patients stratified by age and cancer type and against the US general population. Results: When compared to the US population as a whole, PROMIS scores indicate that UCSF BCC patients have impaired quality of life in every assessed PROMIS domain except for depression. As age increases, PROMIS scores across multiple domains (depression, anxiety, sleep disturbance, cognitive function) increase, signifying lower symptom burden. Importantly, high symptom burden in the anxiety PROMIS domain was suggestive of lower quality of life scores in 6 out of the other 7 assessed PROMIS domains. 39% (1408 patients) of all analyzed BCC patients received at least one referral to a supportive care service. 721 of 1450 (50%) patients with an Invasive Breast Cancer (IBC) diagnosis and 246 of 571 (43%) patients with a DCIS diagnosis triggered at least one referral. 31% of all patients received a referral to genetic counseling, 16.3% to psycho-oncology, 11% to behavioral sleep, 8% to nutritional counseling, 7.6% to social work, 2.4% to smoking cessation, and 1.6% to onco-fertility. When compared to patients with DCIS or without a cancer diagnosis, patients with IBC are most likely to receive a referral to genetic counseling and psychosocial services. Referrals to supportive care services are less frequent as age increases. Conclusion: The Athena online intake form at the UCSF BCC effectively triages patients with high levels of anxiety (PROMIS > 8) and depression (PROMIS >10) to supportive care services. Preliminary results regarding the correlation between anxiety and overall poor quality of life suggest that referrals to psycho-oncology may be most efficacious in decreasing symptom burden. Using these analyses, we aim to refine our use of the eHQS to tailor the most relevant and effective referrals to the highest-need patients. Citation Format: Madeline Matthys, André Dempsey, Amrita Basu, Emily Wong, Nickolas Dreher, Irene Acerbi, Suzanne Ackerson, Laura Esserman, Athena Breast Health Network Investigators and Advocate Partners, Michelle Melisko, Laura van 't Veer. Triaging breast care center patients to supportive care services based on the Athena Breast Health Network online intake form and the patient-reported outcomes measurement information system (PROMIS) [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 P2-12-08.

Background: A variety of investigational HER2-inhibitor agents/combinations have been tested in I-SPY 2, including neratinib (N), TDM1 combined with pertuzumab (P) (TDM1/P), and trastuzumab (H) combined with pertuzumab (H/P; prior to this combination becoming standard of care), all with trastuzumab as control (Ctr). All three experimental arms graduated, showing improved efficacy over control in one or more receptor subsets (HR+HER2+, HR-HER2+, or/and HER2+). Here we assess 10 biomarkers in the HER2, ER/PR, and proliferation pathways on multiple levels of resolution (expression, protein, phospho-protein) as predictors of response in these four arms, hypothesizing that highly HER2-activated, proliferative tumors may be more sensitive to HER2-inhibition than those that are more luminal and quiescent. Methods: 192 HER2+ patients were considered in this analysis: (31 Ctr, 65 N, 52 TDM1/P, and 44 H/P). 10 biomarkers relating to HER2, ER, or proliferation were evaluated from pre-treatment biopsies: HER2 IHC (n=145), 3 expression signatures (n=192), BluePrint subtype (n=192), and 5 protein/phospho-protein analytes by RPPA (n=175). Each biomarker was tested for association with pCR in the whole population and within each arm using a logistic model. This analysis was adjusted for HR status and treatment arm as covariates, and performed within receptor subtypes. This analysis does not adjust for multiplicities of other biomarkers. Results: In the population as a whole, HER2 and HER2-signaling biomarkers, evaluated at multiple levels of resolution: IHC, total-/phospho-protein by RPPA, and mRNA (HER2 amplicon module) - are highly correlated (rho=0.8 [0.65-0.92]). Higher HER2 levels and activity are associated with response: HER2 IHC 3+ status (LR p=0.00032), total quantitative ERBB2 protein by RPPA (LR p=5.4E-09), ERBB2 activation levels (LR p=6.58E-06 (pERBB2 (Y1248)) and 9.95E-06 (pEGFR Y1173)), and the ERBB2 amplicon expression signature (LR p=2.38E-08). In contrast, higher average ER/PR expression associates with non-response to HER2-targeted therapy (LR p=4.28E-08). Both HER2 and ER/PR signaling phenotypes are captured by BluePrint subtyping; and consistent with the individual pathway markers, tumors classified Luminal-type had a lower pCR rate relative to those classified as Her2-type (or Basal-type) (LR p=4.84E-11). These associations all retain significance in a model adjusting for HR status and treatment arm, and in the HR+HER2+ subset. In addition, we quantitatively assessed proliferation markers at the total protein (RPPA: Ki67), phospho-protein (pAURKA) and mRNA (proliferation signature Module11_Prolif) levels. All three proliferation biomarkers predict response overall; but this association is strongest within the HR+HER2+ subset (LR p=0.0012 (Module11_prolif), 0.0036 (pAURK), and 0.045 (Ki67)). None of the biomarkers tested were associated with response in the HR-HER2+ subset. Numbers are small within individual arms. Within the HR+HER2+ subtype, higher HER2 and lower ER/PR is observed in responders in all experimental arms; but the proliferation markers Module11_Prolif (LR p=0.0031) and Ki67 total protein (LR p=0.0029) are associated with response to TDM1/P but not H/P or N. Conclusion: High HER2 signaling at the expression, protein, and phospho-protein levels, and low ER signaling, all predict response to HER2-inhibition across treatment arms. Proliferation markers may be useful for prioritizing therapies in the HR+HER2+ subset. Citation Format: Denise M Wolf, Christina Yau, Julia Wulfkuhle, Lamorna Brown-Swigart, Smita M. Asare, Gillian L. Hirst, Laura Sit, Jane Perlmutter, I-SPY 2 TRIAL Consortium, Minetta Liu, John Park, Angela DeMichele, Douglas Yee, Don Berry, Laura Esserman, Emanuel Petricoin, Laura van t' Veer. HER2 signaling, ER, and proliferation biomarkers predict response to multiple HER2-targeted agents/combinations plus standard neoadjuvant therapy in the I-SPY 2 trial [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 P4-10-02.

Background: While new treatments and improved subtyping schemas are anticipated to improve treatment response in triple-negative breast cancer (TNBC) patients, therapeutic resistance remains a significant challenge. Moreover, there is an urgent need for additional research model systems to study resistance and residual disease in breast cancer, including aggressive subtypes of breast cancer. Organoid culture is a promising technology used for growing breast cancer cells with high efficiency; however, the extent to which treatment resistance can be modeled using this system is unknown. This research used patient-derived organoid cultures in the context of computational analyses of large molecular and clinical datasets to study resistance mechanisms, biomarkers, and alternative treatment strategies to overcome drug resistance in early-stage TNBC. Methods: Organoid cultures were derived from breast tumor samples (taken from lumpectomy, mastectomy, or core biopsy samples), digested to the organoid level using collagenase, and grown in three dimensional cultures using a basement membrane extract and a fully-defined organoid medium (Dekkers et al. Nat Protoc 2021). An evaluation of all available I-SPY2 biomarker data (Wolf et al. Cancer Cell 2022) was performed focusing on genes, proteins, and pathways associated with resistance. These were then used to study whether resistance biomarkers identified in patient tumors are also present in organoids propagated from breast cancer post-treatment residual disease. To this end, bulk RNA sequencing data of organoids were normalized and merged with the TCGA dataset (Hoadley et al. Cell 2018) to enable analysis in a larger context, and immunofluorescence staining of organoids was performed. A high-throughput 386 anti-cancer drug compound screen and subsequent synergy testing with the most promising compounds were performed to identify and predict alternative treatment strategies. Additional assays to explore kinome activity in this organoid model are in progress. Results: A TNBC organoid biobank was established (n=31), which was enriched for inflammatory breast cancer (IBC; n=15), an aggressive form of breast cancer. Most organoids were derived from residual disease after neoadjuvant therapy. Bulk RNA sequencing analysis performed on 10 TNBC organoids revealed 3 subsets that were characterized predominantly by either normal-like/luminal androgen receptor or basal-like features or expressed distinct gene expression profiles, with IBC cases present in all 3 subsets. Intriguingly, the IBC organoids were characterized by higher expression of a number of immune-related signatures, despite an absence of clear immune cells in culture. A residual disease IBC/TNBC organoid resistant to chemotherapy was used to perform the 386-drug compound screen. The organoid model showed resistance to veliparib-cisplatin, consistent with the expression of gene/protein biomarkers predictive of drug resistance found in I-SPY2 (low PARPi7 levels and high pFOXO1 and pMEK1/2 expression). In addition, the screen identified multiple classes of inhibitors as promising synergistic/additive candidates for overcoming resistance to cisplatin. Conclusion: In this proof-of-principle study, we demonstrated the utility of matching I-SPY2 resistance biomarkers and signatures to residual disease tumor organoid cultures. We show that tumor organoid cultures modeling drug resistance states are a useful complement to existing research models of breast cancer and can be used for compound testing. We have developed a pipeline to propagate residual tumors from patients enrolled in I-SPY2 into organoid cultures to create a broader platform for preclinical drug testing informed by tumor biology with the ultimate goal of enabling faster, more successful translational studies and increased treatment options for resistant disease. Citation Format: Tam Binh V. Bui, Denise M. Wolf, Kaitlin Moore, Isaac S. Harris, Pravin Phadatare, Christina Yau, Lamorna A. Brown Swigart, Laura J. Esserman, Jean-Philippe Coppe, Julia Wulfkuhle, Emanuel F. Petricoin, Michael Campbell, Laura M. Selfors, Deborah A. Dillon, Beth Overmoyer, Filipa Lynce, Laura Van ’t Veer, Jennifer Rosenbluth. PD5-02 An Organoid Model System to Study Resistance Mechanisms, Predictive Biomarkers, and New Strategies to Overcome Therapeutic Resistance in Early-Stage Triple-Negative Breast Cancer [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr PD5-02.

BACKGROUND WISDOM (Women Informed to Screen Depending on Measures of risk) is a pragmatic trial comparing the safety, efficacy, cost, and patient acceptability of personalized versus annual breast cancer screening in women ages 40-74 years with no personal history of breast cancer. Cancer registry reporting is considered the gold standard for ascertaining cancer diagnosis data; however, the time lag between diagnosis and reporting (1-2 years or more) by the California Cancer Registry (CCR) presents challenges. The WISDOM study must quickly ascertain and report accurate cancer diagnoses to the Data Safety Monitoring Board; delays could lead to preventable harm, unnecessary costs, and unwarranted early trial termination. To ascertain cancer diagnoses in a timely manner, we used multiple data collection methods and compared self-reported with Electronic Health Records (EHR) information, and in some cases verified with chart review. Specifically, we conducted a procedural study with WISDOM participants who sought care at a University of California (UC) health system by extracting their data from the University of California Data Warehouse (UCDW) to 1) verify self-reported breast cancer diagnoses in WISDOM; 2) ensure that all on-study breast cancer diagnoses have been captured within the WISDOM study cohort; and 3) assess the accuracy and reliability of self-reported data. The UCDW provides harmonized EHR data from the six University of California (UC) health systems and constitutes ~120 billion data points from 7.8 million patient records. METHODS We provided the UCDW with a list of WISDOM participants who indicated they had sought care from a UC health system. Participants with recorded breast cancer diagnostic code (ICD-10CM C50.) were matched with the WISDOM self-reported cancer dataset to uncover participants who had no recorded breast cancer diagnosis in the study system. For those individuals with no recorded self-report diagnosis, coordinators conducted manual chart review to determine whether the participants had been diagnosed with breast cancer. RESULTS This cohort included 11,314 enrolled WISDOM participants who self-reported care at a UC and had at least one diagnostic or procedural breast care code. Among these participants, 160 had a ICD-10-CM C50 breast cancer diagnostic code of which 132 breast cancer cases were confirmed: 111 were already self-reported through WISDOM and 21 additional confirmed breast cancer cases were identified through the UCDW. As a standard process, only WISDOM self-reported cancer cases that were confirmed by chart review are entered into the study database. The percentage of confirmed UCDW-identified cases that were not also self-reported was greater for recent diagnoses (16% overall and 6% for period prior to June 2020). Among the 11,314 participants in our cohort, the CCR identified 61 pre-June 2020 cancer diagnoses. Of these 61 diagnoses, 92% (56) were identified by the UCDW procedural process. If the UCDW had been used as the single source for cancer diagnosis, 21 participants without cancer and 7 participants with unclear status (unlocated or inaccessible records) would have been classified as cancer cases. DISCUSSION/CONCLUSION Self-reported data provides quick ascertainment with relative accuracy compared to cancer registry. Cancer ascertainment can be further improved by combining self-reported data with EHR data from a health system data warehouse registry, particularly for self-reported questionnaire issues such as timing and lack of response. Accuracy of self-reported cancer diagnosis from annually distributed questionnaires improves over time. Identifying cancer diagnosis discordance between data sources can inform processes to improve self-reported study accuracy. Citation Format: Katherine Leggat-Barr, Rita H. Ryu, Allison Stover Fiscalini, Tomiyuri Lewis, Rohini S. Bulusu, Samrrah A. Raouf, Hannah Lui Park, Alyssa N. Rocha, Liliana Johansen, Laura Van ’t Veer, Laura J. Esserman, Michael A. Hogarth. The WISDOM Study: Early Ascertainment of Breast Cancer Diagnoses by Utilizing Self-Reported Questionnaires and Data Warehouse Electronic Health Records [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr P3-03-24.

Background: I-SPY2 is a multicenter, phase 2 trial using response-adaptive randomization within biomarker subtypes defined by hormone-receptor (HR), HER2, and MammaPrint (MP) status to evaluate novel agents as neoadjuvant therapy for high-risk breast cancer. The primary endpoint is pathologic complete response (pCR). Cemiplimab (Cemi) is a PD-1 inhibitor approved for the treatment of NSCLC, cutaneous basal, and squamous cell cancer. Here, we report current efficacy rates of Cemi in combination with paclitaxel followed by AC. Methods: Women with tumors ≥ 2.5cm were eligible for screening. Only HER2 negative (HER2-) patients were eligible for this treatment; HR positive (HR+) patients had to be MP high risk. Treatment included paclitaxel 80 mg/m2 IV weekly x 12 and Cemi 350 mg IV given q3weeks x 4, followed by doxorubicin/cyclophosphamide (AC) every 2 weeks x 4. The control arm was weekly paclitaxel x 12 followed by AC every 2-3 weeks x 4. All patients undergo serial MRI imaging; and imaging response (at 3 weeks, 12 weeks and prior to surgery) were used along with accumulating pCR data to continuously update and estimate pCR rates for trial arms. Analysis was modified intent to treat. Patients who switched to non-protocol therapy count as non-pCR. The goal is to identify (graduate) regimens with ≥85% Bayesian predictive probability of success (i.e. demonstrating superiority to control) in a future 300-patient phase 3 neoadjuvant trial with a pCR endpoint within responsive signatures. Cemi was eligible to graduate in 3 pre-defined signatures: HER2-, HR-HER2-, and HR+HER2-. To adapt to changing standard of care, we constructed “dynamic controls” comprising ‘best’ alternative therapies using I-SPY 2 and external data and estimated the probability of Cemi being superior to the dynamic control. Results: 60 HER2- patients (28 HR+ and 32 HR-) received Cemi arm treatment. The control group included 357 patients with HER2- tumors (201 HR+ and 156 HR-) enrolled since March 2010. Cemi graduated in HR-/HER2- signature. Estimated pCR rates (as of June 2022) are summarized in the table. Immune-related endocrine disorders include: hypothyroid (14.5%), adrenal insufficiency (10%), hyperthyroid (4.8%),) and thyroiditis (3.2%). Only one grade 3 adrenal insufficiency was observed. All immune related AE’s were manageable. Additional biomarker analyses are ongoing and will be presented at the meeting. Response predictive subtypes (Immune+ vs Immune-) and additional predictive biomarkers were assessed. Associations with pCR will be presented at SABCS. Conclusion: The I-SPY 2 study aims to assess the probability that investigational regimens will be successful in a phase 3 neoadjuvant trial. Anti-PD-1 therapy with Cemi resulted in a higher predicted pCR rate in HR-/HER2- 55 rate% disease compared to control at 29%. Immune-mediated AE’s were observed. This data is consistent with previously published data using check point inhibitors in early-stage HR-/HER2- breast cancer. Estimated pCR rates Citation Format: Erica Stringer-Reasor, Rebecca A. Shatsky, Jo Chien, Anne Wallace, Judy C. Boughey, Kathy S. Albain, Hyo S. Han, Rita Nanda, Claudine Isaacs, Kevin Kalinsky, Zahi Mitri, Amy S. Clark, Christos Vaklavas, Alexandra Thomas, Meghna S. Trivedi, Janice Lu, Smita Asare, Ruixiao Lu, Maria Pitsouni, Amy Wilson, Jane Perlmutter, Hope Rugo, Richard Schwab, W. Fraser Symmans, Nola M. Hylton, Laura Van ’t Veer, Douglas Yee, Angela DeMichele, Donald Berry, Laura J. Esserman, I-SPY Investigators. PD11-01 Evaluation of the PD-1 Inhibitor Cemiplimab in early-stage, high-risk HER2-negative breast cancer: Results from the neoadjuvant I-SPY 2 TRIAL [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr PD11-01.

Background: In previous work we leveraged the I-SPY2 trial to create treatment response predictive subtypes (RPS) incorporating tumor biology beyond clinical HR/HER2, to better predict drug responses in an expanded treatment landscape that includes platinum agents, dual HER2-targeting regimens and immunotherapy [1]. We showed that best performing schemas incorporate Immune, DRD and HER2/Luminal phenotypes, and that treatment allocation based on these would increase the overall pCR rate to 63% from 51% using HR/HER2-based treatment selection. The RPS schema has been selected for prospective evaluation in I-SPY2. Using the RPS, one would prioritize platinum-based therapy for HER2-/Immune-/DRD+, immunotherapy for HER2-/Immune+, and dual-anti-HER2 for HER2+ that are not luminal. HER2+/Luminal patients have low response rates to dual-anti-HER2 therapy but may respond better to anti-AKT. However, there is still a considerable ‘biomarker-negative’ group of resistant cancers (HER2-/Immune-/DRD-) with very low pCR rates to all tested agents, that require a new therapeutic approach. Here we characterize the protein signaling architecture of these tumors to identify new target candidates. Methods: 987 I-SPY 2 patients from 10 arms of the trial were considered for this analysis. All have gene expression, pCR and RPS; 944 have distant recurrence free survival (DRFS) data; and 736 have reverse phase protein array (RPPA) data from laser capture microdissected tumor epithelium. These data – known collectively as the I-SPY2-990 mRNA/RPPA Data Resource - were recently made public on NCBI’s Gene Expression Omnibus [GEO: GSE196096]. We focus on HER2-/Immune-/DRD- tumors, applying Wilcoxon and t-tests to identify phosphoproteins that differ between HR+HER2-/Immune-/DRD- and other HR+HER2- tumors; and between TN/Immune-/DRD- and other TNs. The Benjamini-Hochberg (BH) method is used to adjust p-values for multiple hypothesis testing. In addition, the Kaplan-Meier method is used to estimate DRFS. Results: 201/736 I-SPY 2 patients with RPPA data are classified HER2-/Immune-/DRD- (HR+HER2-: n=138; TN: n=63). Of these, 8.5% (17/201) achieved pCR. Non-responding HER2-/Immune-DRD- had worse outcomes than responders (~75% vs. ~95% DRFS at 5 years). 60/139 phospho-proteins differ significantly between HR+HER2-/Immune-/DRD- and other HR+HER2- tumors (n=122). These tumors are relatively ‘cold’, in that 90% (54/60) of the phosphoprotein activities characterizing this group are at lower levels than in the overall HR+HER2- population; including immune (e.g. pPDL1, pJAK/STAT) and proliferation (e.g., Ki67, CyclinB1, pAURK) endpoints. Phosphoproteins showing higher levels in this subset include ERBB2 (BH p=1.7E-06), Cyclin D1 (BH p=1.4E-05), pAR (BH p=1.4E-05), and ER (BH p=3E-04). Within the TN subset, only 3/139 phospho-proteins differed significantly between TN/Immune-/DRD- and other TN tumors (n=189). These were all immune-related (pPDL1, pSTAT1, and HLA DR), with lower expression in the TN/Immune-/DRD- group. Conclusion: HR+HER2- and TN patients who are Immune-Low and DRD-Low have very low pCR rates to all tested therapeutics in I-SPY2 including standard chemotherapy, platinum, and immunotherapy. Senolytics (possibly targeting Cyclin D1), HER2low agents, and AR modulators may overcome resistance in HR+HER2-/Immune-/DRD-, whereas an immune activator beyond checkpoint inhibition is suggested for TN/Immune-/DRD- patients. [1] Wolf et. al., Redefining Breast Cancer Subtypes to Guide Treatment Prioritization and Maximize Response: Predictive Biomarkers across 10 Cancer Therapies. Cancer Cell 2022 Citation Format: Denise M. Wolf, Christina Yau, Julia Wulfkuhle, Rosa I. Gallagher, Lamorna A. Brown Swigart, Gillian L. Hirst, Jean-Philippe Coppe, Mark Jesus M. Magbanua, Rosalyn Sayaman, I-SPY2 Investigators, Laura Sit, Nola M. Hylton, Angela DeMichele, Donald A. Berry, Lajos Pusztai, Douglas Yee, Laura J. Esserman, Emanuel F. Petricoin, Laura Van ’t Veer. PD5-04 Characterizing the HER2-/Immune-/DNA repair (DRD-) response predictive breast cancer subtype: the hunt for new protein targets in a high-needs population with low response to all I-SPY2 agents [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr PD5-04.

Background: The detection of circulating tumor DNA (ctDNA) may serve as an early predictor of response and recurrence. In this study, we used a tumor-informed ctDNA test to monitor clinical outcomes in patients with high-risk hormone receptor-positive HER2-negative (HR+HER2-) tumors who received neoadjuvant chemotherapy (NAC) on the I-SPY 2 trial (NCT01042379). Methods: We collected blood samples at pretreatment, during (at 3 and 12 weeks after initiation of paclitaxel-based treatment with or without an investigational drug), after NAC prior to surgery, 4 weeks after surgery, and annually until clinical diagnosis of recurrence. Cell-free DNA was isolated from plasma (N=329 samples) and ctDNA was detected using a personalized, tumor-informed multiplex polymerase chain reaction next generation sequencing-based test (SignateraTM). All patients were at high risk for recurrence by MammaPrint. The response endpoints were pathologic complete response (pCR) and residual cancer burden (RCB), and the survival endpoint was event-free survival (EFS). Results: This analysis included 66 patients with HR+HER2- breast cancer who had blood samples collected before, during, after NAC and had at least one blood sample after surgery with sufficient plasma for analysis. 57.1% (32/56) had grade III disease; 72.4% (42/58) were node-positive; 36.2% (21/58) had T3/T4 disease; and 33.3% (22/66) were MammaPrint High 2. The percent ctDNA positivity rates at pretreatment, after NAC prior to surgery, and 4 weeks after surgery were 79.7% (47/59), 6.5% (4/62), and 2% (1/50), respectively. Significantly higher ctDNA positivity rates at pretreatment were observed in patients with larger tumors (95% in T3/T4 vs. 69% in T1/T2, Fisher’s exact p=0.0387), higher grade tumors (94% in Grade III vs. 67% in Grade I/II, p=0.0147) and by MammaPrint score (100% in High 2 vs. 71% in High 1, p=0.0052). In this high-risk HR+/HER2- cohort, 10/66 (15.2%) achieved pCR/RCB 0, who were all ctDNA-negative at surgery. 56/66 (84.8%) had no-PCR, with RCB I (limited residual cancer), II (moderate) and III (extensive) in 7 (10.6%), 31 (47.0%) and 18 (27.3%), respectively. ctDNA-positivity after paclitaxel-based treatment was significantly associated with RCB II/III status (Fisher’s exact p=0.01). All patients in this cohort with persistent ctDNA subsequently had RCB II or III at surgery. 47 patients had paired samples collected after NAC prior to surgery and at 4 weeks after surgery. Of the 47, 91.5% (43/47) were ctDNA-negative at both time points and 8.5% (4/47) were discordant; 1 was ctDNA-negative and later tested ctDNA-positive, while 3 were ctDNA-positive and later tested ctDNA-negative. 61/66 patients had EFS data with a median of 1.6 years of follow up (range: 0.6 to 5.6). 5 tested ctDNA-positive in at least one time point after surgery. Of these, 2 experienced a recurrence (one local relapse and one distant metastasis) and both tested positive at the time of recurrence. For the patient who developed a distant recurrence it was the only blood sample available at a follow-up time point; for the patient who developed a local recurrence, blood from two earlier follow-up time points had tested negative. To date, no recurrences have been observed in those whose test(s) after surgery were negative for ctDNA. Conclusions: The persistence of ctDNA during neoadjuvant therapy is associated with the extent of residual disease in a cohort of patients with HR+HER2- breast cancer in the I-SPY 2 trial and thus may be useful in identifying patients who are not having an optimal response to therapy. I-SPY 2.2 will test whether ctDNA has utility in redirecting therapy to improve surgical outcome and subsequent prognosis. Citation Format: Mark Jesus M. Magbanua, Hope Rugo, Lamorna A. Brown Swigart, Ziad Ahmed, Gillian L. Hirst, Denise M. Wolf, Ruixiao Lu, Ekaterina Kalashnikova, Derrick Renner, Angel Rodriguez, Minetta C. Liu, Christina Yau, Laura J. Esserman, Laura Van ’t Veer, Angela DeMichele. Monitoring for response and recurrence in neoadjuvant-treated hormone receptor-positive HER2-negative breast cancer by personalized circulating tumor DNA testing [abstract]. In: Proceedings of the 2022 San Antonio Breast Cancer Symposium; 2022 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2023;83(5 Suppl):Abstract nr P5-05-05.

Background: Identification of women at elevated risk for highly proliferative, poor prognosis breast cancers could have important implications for screening and prevention. Genome-wide association studies (GWAS) have found >200 single nucleotide polymorphisms (SNPs) associated with breast cancer risk, with many SNPs differentially associated with ER status or intrinsic subtype. We hypothesized that some of these SNPs are preferentially associated with more proliferative tumors, while others are preferentially associated with less proliferative tumors. In this study, we aimed to build a polygenic risk score (PRS) predictive of proliferative tumors, using the GWAS-identified SNPs. Methods: We used data from 3 studies that included array-based SNP genotyping and tumor transcriptomic data: The Cancer Genome Atlas, METABRIC, and the I-SPY 2 Trial (total n=2,467). Our outcome was the risk of recurrence score weighted on proliferation (ROR-P), a validated tumor prognostic signature. Using the breast cancer risk SNPs, we built respective linear regression models to predict ROR-P, with genetic ancestry, study, and ER status as covariates. We performed 5-fold cross-validation and used the model r2 to identify the optimal p-value threshold for including SNPs in the PRS. To decrease uncertainty of our estimates, we performed 100 repeats across the pooled datasets. To test whether this model predicted poor prognosis breast cancers, we first used it to impute ROR-P among genotyped breast cancer cases in UK Biobank (UKB). We then examined the association between “genetically predicted” ROR-P and breast cancer-specific survival using Cox proportional hazards models adjusted for genetic ancestry and age at diagnosis. Results: Associations between 224 breast cancer SNPs and ROR-P were tested. The best-performing model in cross-validation contained 96 SNPs, each associated with ROR-P at p < 0.45, with model r2 0.054. The SNPs with the strongest positive correlations with ROR-P included those discovered in GWAS for HER2-positive and ER-negative cancers, both of which tend to be highly proliferative. Among 7,247 incident cancers in UKB, higher genetically predicted ROR-P was associated with shorter survival, with a per-standard deviation hazard ratio of 1.14 (95% CI 1.05-1.24, p = 0.002). Conclusions: We used breast cancer susceptibility SNPs to construct a PRS fitted to ROR-P, a prognostic signature recapitulating tumor proliferation. This PRS was associated with worse clinical outcomes in breast cancer cases from UKB. Our results suggest that correlations between SNPs and tumor gene expression can be used to “tune” PRS to tumor phenotype, e.g. proliferation. Highly proliferative tumors are more likely to present as advanced cancers even among women getting routine screening. If replicated in other datasets, our findings could be used to identify women who may especially benefit from tailored screening and prevention. Citation Format: Yiwey Shieh, Jacquelyn Roger, Scott Huntsman, Donglei Hu, Jovia L. Nierenberg, Pooja Middha Kapoor, Christina Yau, Gillian Hirst, Laura van 't Veer, Laura Esserman. Development and testing of a polygenic risk score for proliferative breast cancers [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 5883.

Background: We compared the predictive and prognostic value of ctDNA dynamics in high-risk hormone receptor-positive/HER2-negative (HR+/HER2-) and triple negative breast cancer (TNBC) receiving neoadjuvant chemotherapy (NAC) enrolled in the I-SPY 2 trial (NCT01042379). To our knowledge, this is the largest ctDNA study in breast cancer in the neoadjuvant setting. Methods: Blood samples were collected at pre-treatment (T0), during treatment (T1 at 3 weeks, and T2 at 12 weeks) and after NAC (T3 at 24 weeks) from 106 HR+/HER2- and 97 TNBC patients. Plasma samples (n=734) were analyzed using a personalized and tumor-informed mPCR NGS-based ctDNA test (SignateraTM). Patients, all high risk for recurrence by MammaPrint, received paclitaxel-based treatment +/- experimental therapy followed by anthracycline. The median follow-up was 3.0 years (0.5 to 6.5). The predictive and prognostic value of ctDNA dynamics and status at different timepoints were examined. Our analysis is exploratory and does not adjust for other biomarkers. Results: Pretreatment ctDNA positivity (Fisher p Conclusions: The predictive value of ctDNA for prediction of pCR and RCB differed between subtypes (HR+/HER2- vs. TNBC), while similar prognostic value was observed. In TNBC, early clearance of ctDNA at 3 weeks was a significant predictor of favorable response to NAC. Compared to patients who were ctDNA-positive after NAC, ctDNA-negative status in both subtypes was associated with improved DRFS even in patients with residual cancer (no pCR or RCB-II/III). These findings could inform on the design of future studies that seek to demonstrate the utility of ctDNA in the curative setting. Predictive and prognostic significance of ctDNA in early breast cancer in the neoadjuvant setting HR+HER2- (n=106) TNBC (n=97) Predictive value for prediction of pCR and RCB Fisher p-value Fisher p-value Early ctDNA clearance (between T0 and T1) and pCR 0.4521 Citation Format: Mark Jesus Mendoza Magbanua, Lamorna Brown Swigart, Derrick Renner, Svetlana Shchegrova, Gillian L. Hirst, Christina Yau, Denise M. Wolf, Hsin-Ta Wu, Ekaterina Kalashnikova, Amy L. Delson, A. Jo Chien, Debu Tripathy, Smita Asare, Raheleh Salari, Angel Rodriguez, Bernhard Zimmermann, Himanshu Sethi, Alexey Aleshin, Paul Billings, Rita Nanda, Hope S. Rugo, Laura J. Esserman, Minetta C. Liu, Angela DeMichele, Laura van 't Veer. Comparison of the predictive and prognostic significance of circulating tumor DNA in patients with high risk HER2-negative breast cancer receiving neoadjuvant chemotherapy [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr LB111.

Background Transcriptomic immune-related gene signatures have been associated with achievement of pathologic complete response (pCR) and prognosis in the neoadjuvant setting. I-SPY 2 is a multicenter, phase 2 platform trial using response-adaptive randomization within subtypes defined by receptor status (HR/HER2) and MammaPrint (MP) risk to evaluate novel agents as neoadjuvant therapy for women with high-risk breast cancer. Given racial disparities in mortality from breast cancer and the paucity of racial demographic data from clinical trials, we aimed to evaluate the association between racial groups and baseline characteristics, including expression-based subtypes and immune signatures, treatment response, and prognosis of patients enrolled in the I-SPY 2 TRIAL. Methods Our study population included 990 I-SPY 2 patients. 15 patients identified as part of a racial group with Citation Format: Beverly Kyalwazi, Christina Yau, Olufunmilayo Olopade, A. Jo Chien, Anne Wallace, Andres Forero-Torres, Lajos Pusztai, Erin Ellis, Kathy Albain, Anne Blaes, Barbara Haley, Judy Boughey, Anthony Elias, Amy Clark, Claudine Isaacs, Rita Nanda, Hyo Han, Rachel Yung, Debu Tripathy, Kristen Edmiston, Rebecca Viscusi, Donald Northfelt, Qamar Khan, Ashish Sanil, Scott Berry, Smita Asare, Amy Wilson, Gillian Hirst, Nola Hylton, Michelle Melisko, Jane Perlmutter, Hope Rugo, Fraser Symmans, Laura van ‘t Veer, Donald Berry, Laura Esserman. Analysis of clinical outcomes and expression-based immune signatures by race in the I-SPY 2 trial [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr GS4-02.

Background: Liquid biopsies have emerged as effective diagnostic tools in disease monitoring and minimal residual disease detection. Circulating tumor DNA (ctDNA) was recently shown to be a predictor of poor response and recurrence in breast cancer. However, ctDNA shedding from breast tumors can rapidly decrease during treatment, resulting in reduced sensitivity in measuring early changes in tumor response or residual cancer burden (RCB) after neoadjuvant chemotherapy (NAC). We recently reported the discovery of orphan non-coding RNAs (oncRNAs), a class of small RNAs that are not present in healthy cells, but emerge from cancer cells. Similar to ctDNA, tumor-released oncRNAs can be used to detect the presence of an underlying tumor; however, since they are actively released by cancer cells, their abundance in the cell-free compartment is substantially higher than ctDNA. Therefore, we hypothesized that monitoring circulating oncRNAs in blood permits a more sensitive approach to measuring treatment response (i.e., pathologic complete response, or pCR) and estimating RCB. Patients and Methods: Cell-free RNA (cfRNA) was extracted from ~1 ml sera of 72 breast cancer patients treated in the neoadjuvant I-SPY 2 TRIAL with NAC alone or combined with MK-2206 (AKT inhibitor) treatment. For each patient, treatment-naïve samples (T0) were compared with samples from post-treatment and prior to surgery (T3) time-point. RNA samples were subjected to small RNA sequencing (SMARTer), and the presence and abundance of cell-free oncRNA species were then determined by identifying and counting the reads that map to oncRNA loci across samples. Notably, oncRNAs species were pre-annotated from the Cancer Genome Atlas (TCGA), and our approach does not require bespoke personalized assays. We used a machine-learning model to compare abundance of cfRNA species before and after treatment (i.e., T3-T0) to predict pCR and RCB. For this, we split our cohort into a training and a testing set (48 and 24) and trained a model to simultaneously learn the presence of residual disease (pCR vs. no pCR) and its extent (RCB). We then measured the performance of our model on the held-out test data and the entire dataset. To confirm the robustness of our model, we also employed a leave-one-out strategy, whereby pCR and RCBIndex of each patient was predicted using a model that was trained on the other patients in the cohort. Finally, to assess the ability of our oncRNA-based model to risk-stratify patients who fail to achieve pCR (without having been explicitly trained on relapse data), we used the model’s oncRNA score to predict patients at the highest risk of distant recurrence (n=8 out of 36) and performed a multivariate Cox analysis, controlling for HR/Her2 status (median follow-up time was 4.8 years). Results: The model’s accuracy for predicting pCR—based on changes in circulating oncRNA species between T3 and T0—was 85% for the training data and 79% for the held-out test data (positive predictive value of 75% and negative predictive value of 83%) with combined accuracy of 83%; precision 86% and recall 83%; Pearson R=0.5 for RCB. A leave-one-out strategy showed similar performance (area under ROC of 0.77 versus 0.81 in train-test split). Finally, among the patients who failed to achieve pCR, we observed a significantly higher risk of distant recurrence in those with the highest scores (DRFS: hazard-ratio = 8.4, ANOVA P Citation Format: Hani Goodarzi, Albertas Navickas, Jefferey Wang, Kristle Garcia, Mark J Magbanua, Lisa Fish, Lamorna Brown Swigart, Gillian Hirst, Denise Wolf, Christina Yau, Jo Chien, Carol Simmons, Amy Delson, Laura Esserman, Laura van 't Veer. Tumor-released circulating orphan non-coding RNAs reflect treatment response and survival in breast cancer [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr PD9-04.

Background: Invasive lobular carcinoma (ILC) is the second most common type of breast cancer after invasive ductal carcinoma (IDC). ILC has unique features, such as a diffuse growth pattern due to characteristic loss of E cadherin, and a different pattern of disease metastasis compared to IDC. Prior investigators have shown increased numbers of circulating tumor cells (CTCs) in patients with metastatic ILC versus IDC. We explored the distribution of CTCs and disseminated tumor cells (DTCs) in early stage ILC versus IDC. Methods: We performed a secondary data analysis of the TIPPING study, an institutional review board approved study that pre-operatively collected blood and bone marrow samples from 655 treatment-naïve early-stage breast cancer patients. We analyzed data from 284 patients who had CTCs and DTC enumerated by an EPCAM-based method involving immunomagnetic enrichment and flow cytometry (IE/FC) (61 patients with ILC and 223 patients with IDC). We compared CTC and DTC counts by histology using the Welch Two Sample t-test, linear regression models, as well as ANOVA tests. Multivariate Cox regression analyses were performed to assess association between levels of CTCs/DTCs and clinical outcomes (distant recurrence-free survival [DRFS] and breast cancer-specific survival [BCSS]). Results: ILC tumors were lower grade than IDCs and had a higher proportion of HR+HER2- subtypes (92.00% vs. 75.30%; p Citation Format: Silver Alkhafaji, Denise Wolf, Mark Magbanua, Laura Van 't Veer, John Park, Laura J. Esserman, Rita A. Mukhtar. Differences in levels of circulating tumor cells (CTC) and disseminated tumor cells (DTC) in early-stage lobular versus ductal breast cancer [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr P2-02-02.

Cancers have been associated with a diverse array of genomic alterations, many of which are rare with unknown significance. To understand the cellular mechanisms impacted by such alterations in breast invasive carcinoma, we have applied affinity-purification mass spectrometry to delineate comprehensive biophysical interaction networks for 40 frequently altered breast cancer proteins across three human breast cell lines, providing a novel resource of context-specific and shared protein-protein interaction networks in breast cancer cells. These networks interconnect and enrich for common and rare cancer mutations, and are substantially rewired by mutations. Our analysis identifies novel PIK3CA-interacting proteins which repress AKT signaling, and UBE2N emerges as a BRCA1 interactor predictive of clinical response to PARP inhibition. We also show that Spinophilin interacts with and dephosphorylates BRCA1 to promote DNA double-strand break repair. Thus, cancer protein interaction landscapes provide a framework for recognizing oncogenic drivers and drug vulnerabilities. Citation Format: Minkyu Kim, Jisoo Park, Mehdi Bouhaddou, Kyumin Kim, Ajda Rojc, Maya Modak, Margaret Soucheray, Patrick O'Leary, Denise Wolf, Dominique C. Mitchell, Fan Zheng, John D. Gordan, Jean-Philippe Coppé, Danielle L. Swaney, Laura van' t Veer, Alan Ashworth, Trey Ideker, Nevan J. Krogan. The protein interaction landscape of breast cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 2308.

Background: In the I-SPY 2 TRIAL, the addition of P to standard NAC resulted in more than doubling of the pathologic complete response (pCR) rates for both hormone receptor-positive (HR+)/HER2- and triple-negative (TN) early breast cancer (EBC) patients (pts) compared to NAC only (Nanda et al, JAMA Oncol, 2020). At 3 years, distant recurrence-free survival (DRFS) rates in pts with pCR following NAC+P was >95%. We hypothesized that ctDNA can serve as a predictive biomarker of response and survival in pts treated with NAC. Methods: A personalized ctDNA test (Signatera) was performed on 511 serial plasma samples from 138 pts with high-risk HR+/HER2- (n=77) or TN (n=61) stage II/III EBC. Pts received P with paclitaxel (Tx) followed by AC (P arm, n=42) or standard NAC only (n=96), an exploratory subset of pts evaluated for P efficacy. Plasma was collected; pretreatment (T0), 3 weeks after treatment initiation (T1), between Tx+/-P and AC regimens (T2), and prior to surgery (T3). ctDNA was deemed positive with a minimum of 2 of the pt specific tumor mutation fragments detected in cfDNA. Association of ctDNA with response and survival was analyzed using logistic and Cox regressions with pCR and DRFS as endpoints. Median follow-up was 2.8 years. Results: Detection of ctDNA decreased over time (P arm: T0-81%, T1-50%, T2-19%, T3-3%) and NAC only: T0-82%, T1-65%, T2-26%, T3-10%). ctDNA data at T0 and T1 was available for 96% (132/138) of pts in P arm or NAC only (Table). Among ctDNA+ patients at baseline, clearance at T1 was significantly associated with pCR (OR=1.92, ctDNA+/-; OR=0.27, ctDNA+/+; LR p ctDNA data at T0, T1, and T2 was available for 86% (118/138) pts. (Table). Among all ctDNA+ pts at baseline, dynamics through T2 was associated with pCR (OR=1.44, ctDNA+/-/-; OR=0.33, ctDNA+/+/-, OR=0.12, ctDNA+/+/+; LR p=0.0011). This association remained significant when adjusted for HR status and treatment (LR p All pts who achieved pCR were ctDNA- at T3 (n=34). Among those who failed to achieve pCR (n=81), DRFS was significantly better in ctDNA- (n=72/81; 20 in P and 52 in NAC) versus ctDNA+ pts (n=9/81; 1 in P and 8 in NAC) (adjusted HR 0.13; 95% CI 0.05-0.37). Conclusions: These exploratory results align with our previous findings that early clearance of ctDNA during NAC treatment was significantly associated with increased likelihood of achieving pCR. Additionally, we show that ctDNA clearance can be an early surrogate marker for therapy response assessment. Residual ctDNA after neoadjuvant treatment was a significant predictor of metastatic recurrence and death. Personalized monitoring of ctDNA during the course of NAC is feasible and provides information that can be combined with imaging and pathology, and may help to optimize decision making for de-escalation or escalation of therapy. Larger studies are ongoing. ctDNA dynamics and pCRctDNA status at T0 and T1 (n=132)ctDNA status at T0, T1, and T2 (n=118)ctDNA-/-ctDNA+/-ctDNA+/+ctDNA-/-/-ctDNA+/-/-ctDNA+/+/-ctDNA+/+/+Total, n (%)24 (18)28 (21)80 (61)22 (19)24 (20)43 (36)27 (23)pCR, n (%)9 (38)15 (54)11 (14)9 (41)12 (50)8 (19)2 (7)No pCR, n (%)15 (63)13 (46)69 (86)13 (59)12 (50)35 (81)25 (93) Citation Format: Mark Jesus M Magbanua, Denise Wolf, Derrick Renner, Svetlana Shchegrova, Lamorna Brown Swigart, Christina Yau, Gillian Hirst, Hsin-Ta Wu, Ekaterina Kalashnikova, Antony Tin, Amy Delson, Douglas Yee, Angela DeMichele, Raheleh Salari, Angel Rodriguez, Bernhard Zimmermann, Himanshu Sethi, Alexey Aleshin, Paul Billings, Laura Esserman, Minetta Liu, Rita Nanda, Laura van ‘t Veer, I-SPY 2 Investigators. Personalized ctDNA as a predictive biomarker in high-risk early stage breast cancer (EBC) treated with neoadjuvant chemotherapy (NAC) with or without pembrolizumab (P) [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr PD9-02.

Background: The goal of I-SPY 2 is to rapidly test novel therapies in addition to standard of care in high-risk breast cancer patients. It has resulted in increasing response rates, where pCR rates in TNBC and HR-HER2+ subsets have reached ~60% and ~75%, respectively. Yet, there remains a sizeable subset of non-responders, especially among HR+ patients. Identification of ‘universal’ resistance mechanisms may guide rational selection of agents to improve these patient's outcomes. Thus, we analyzed reverse phase protein array (RPPA) based quantitative protein/phosphoprotein data across arms to assess whether there are common mechanisms rendering these cancers resistant to all agent classes tested to date. Methods: 736 patients (260 HR+HER2-, 252 TN, 142 HR+HER2-, and 82 HR-HER2+; over 8 arms: 194 Ctr, 105 neratinib (N), 63 veliparib/carboplatin (VC), 128 AMG386 (anti-ANG1/2), 87 MK2206 (anti-AKT), 43 TH/pertuzumab (P), 49 TDM1/P, and 67 pembrolizumab (Pembro)) with pCR and RPPA data at the pre-treatment time point were considered for this analysis. 141 RPPA endpoints representing key cancer pathways were assessed for association with pCR using logistic regression modeling, with HR, HER2 and treatment arm as covariates (likelihood ratio test; p Citation Format: Julia Wulfkuhle, Denise Wolf, Christina Yau, Lamorna Brown-Swigart, Rosa I Gallagher, Gillian Hirst, Laura Sit, Smita Asare, I-SPY 2 TRIAL Investigators, Nola Hylton, Angela DeMichele, Douglas Yee, Jo Chien, Hope Rugo, John Park, Kathy Albain, Rita Nanda, Debu Tripathy, Richard Schwab, Don Berry, Laura Esserman, Laura van t' Veer, Emanual Petricoin, III. Identification of biomarkers associated with therapeutic resistance: Quantitative protein/phosphoprotein analysis of ~750 patients across 8 arms of the neoadjuvant I-SPY 2 TRIAL for high-risk early stage breast cancer [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr PD9-04.

Background: Recent studies have demonstrated independent validation of the prognostic relevance of residual cancer burden (RCB) after neoadjuvant chemotherapy. However, a pooled subject-level analysis of multiple cohorts is needed to determine estimates of long-term prognosis for each class of RCB in each phenotypic subtype of breast cancer (BC) to better inform on patient outcomes. Also, a pooled subject-level analysis allows more detailed analyses of generalizability of the prognostic meaning of RCB assessments in a broader experience of practice settings. Method: Subject-level RCB results, with relevant clinical and pathologic stage, tumor subtype and grade, demographic, treatment and follow-up data from 11 institutes/trials are being collected for combined analysis. The association between the continuous RCB index and event-free survival (EFS), and distant recurrence free survival (DRFS) were assessed using mixed effect Cox models with the incorporation of random RCB coefficients to account for between-study heterogeneity. We will also allow for differences in baseline hazard across biological BC subtypes and, if needed, across studies as well. In addition to this stratified mixed effect model, a multivariate analysis adjusting for age, T-category, nodal status and grade was performed within each subtype. In addition, mixed effect Cox models will be employed to evaluate association between RCB index with EFS and DRFS within each HR/HER2 subtype. Kaplan Meier estimates of EFS and DRFS at 5 and 10 years were computed for each RCB class within subtype. Results: We analyzed subject-level data from 9 institutes/trials representing 4077 patients currently available from an anticipated final total of 4,800 patients (to be presented at the meeting). There were 950 EFS and 876 DRFS events during follow up (median 65 months, IQR: 70 months). RCB index (continuous) was independently prognostic within each subtype: HR+/HER2- (EFS HR (per unit increase in RCB index) =1.64, 95%CI 1.48-1.82; DRFS HR=1.68, 1.51-1.87), HR+/HER2+ (EFS HR=1.80, 1.57-2.05; DRFS HR=1.93, 1.67-2.24), HR-/HER2+ (EFS HR=2.15, 1.76-2.62; DRFS HR=2.10, 1.77-2.50), and HR-/HER2- (EFS HR=2.05, 1.89-2.22; DRFS HR=2.16, 1.90-2.46); and remained prognostic in multivariate models adjusting for age, grade, and clinical T and N stage at diagnosis. Table 1 contains the response rate and estimated EFS at 5 years and 10 years for each RCB class within each HR/HER2 phenotype (DRFS results were similar). Conclusions: Long-term prognosis after pCR was similarly excellent in all phenotypic subtypes. RCB index and classification was independently and strongly prognostic in all subtypes, and generalizable to multiple practice settings. Prognostic differences by RCB class occurred within 5 years in HR- BC, but extended to 10 years in HR+ BC. RCB-I had slightly worse EFS than pCR in HR- BC and HR+/HER2+ BC (after 5 years), but the same EFS as pCR in HR+/HER2- BC. Complete analysis of all subjects, including neoadjuvant treatments, will be presented at the meeting. PhenotypeOutcomepCRRCB-IRCB-IIRCB-IIIHR+/HER2-Frequency (%)11%10%52%27%(N=1467)5 yr EFS (95% CI)91% (86-96)93% (89-98)82% (79-85)70% (65-75)10 yr EFS (95% CI)84% (75-93)88% (82-95)71% (67-75)52% (46-58)HR+/HER2+Frequency (%)38%18%35%9%(N=762)5 yr EFS (95% CI)94% (91-97)93% (88-98)78% (73-84)49% (37-65)10 yr EFS (95% CI)91% (86-96)79% (70-90)65% (59-73)42% (29-60)HR-/HER2+Frequency (%)66%11%18%5%(N=550)5 yr EFS (95% CI)93% (90-96)88% (79-97)60% (50-71)45% (30-69)10 yr EFS (95% CI)90% (86-94)84% (74-95)56% (46-68)45% (30-69)HR-/HER2-Frequency (%)41%13%33%13%(N=1293)5 yr EFS (95% CI)92% (90-94)85% (79-91)68% (63-72)28% (21-36)10 yr EFS (95% CI)87% (82-91)80% (72-88)63% (58-68)24% (18-33) Citation Format: Christina Yau, Marieke van der Noordaa, Jane Wei, Marie Osdoit, Fabien Reyal, Anne-Sophie Hamy, Marick Lae, Miguel Martin, Maria del Monte, I-SPY 2 TRIAL Consortium, Judy C Boughey, Rebekah Gould, Jelle Wesseling, Tessa Steenbruggen, Maartje van Seijen, Gabe Sonke, Stephen Edge, Stephen-John Sammut, Elena Provenzano, Jean Abraham, Peter Hall, Ashley Graham, Lorna Mackintosh, David Cameron, Alice Wang, Priyanka Sharma, Kimberly Cole, Lajos Pusztai, Mi-Ok Kim, Laura van ‘t Veer, Laura Esserman, W. Fraser Symmans. Residual cancer burden after neoadjuvant therapy and long-term survival outcomes in breast cancer: A multi-center pooled analysis [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 GS5-01.

Background: We developed a 7-gene DNA-repair deficiency signature (PARPi-7) that predicts breast cancer cell line sensitivity to the PARP inhibitor olaparib [PMID: 22875744]. We hypothesized that this signature would also predict response to other PARP inhibitors including veliparib. In the I-SPY 2 TRIAL, HER2- patients were randomized to receive standard chemotherapy or the oral PARP inhibitor veliparib in combination with carboplatin (V/C) and chemotherapy. V/C graduated in the triple-negative (TN) signature. Here we assess the PARPi-7 as a specific biomarker of V/C response. Methods: 115 HER2- patients (V/C: 71 and concurrent controls: 44) were considered in this analysis. The PARPi-7 signature score is computed from Agilent 44K array data as published using expression levels of BRCA1, CHEK2, MAPKAPK2, MRE11A, NBN, TDG, and XPA. We assess association between PARPi-7 and response in the V/C and control arms alone (Wald p < 0.05), and relative performance between arms (biomarker x treatment interaction, likelihood ratio p < 0.05) using a logistic model. In an exploratory analysis, we dichotomized patients by the PARPi-7 score using the published in vitro derived cutpoint (0.037). To assess PARPi-7 in the context of the graduating signature, we added the PARPi-7 High patients to the graduating TN subset and evaluated the treatment effect in this ‘biomarker-positive’ group. Our study is exploratory with no claims for generalizability of the data. Statistical calculations are descriptive (e.g. p-values are measures of distance with no inferential content). Our analyses do not adjust for multiplicities of other biomarkers in the trial but outside this study. Results: The PARPi-7 signature associates with patient response in the V/C arm (OR = 3.9, p=0.00056) but not in the control arm (OR = 0.87, p=0.68). There is a significant biomarker x treatment interaction (OR in V/C arm relative to control arm = 4.48, p=0.0028), which remains significant upon adjusting for HR status (p=0.0018). In an exploratory analysis, PARPi-7 dichotomized using the published in vitro derived cutpoint yields 62 PARPi-7 Low and 53 PARPi-7 High patients. 26% of PARPi-7 High patients are not TN. The distribution of pCR rates among PARPi-7 dichotomized groups are in Table 1. V/C (n=71)Control (n=44)PARPi-7 Low (n=38)PARPi-7 High (n=33)PARPi-7 Low (n=24)PARPi-7 High (n=20)TN (n=59)5 / 1317 / 251 / 74 / 14HR+HER2- (n=56)2 / 253 / 84 / 170 / 6 When the PARPi-7 High patients are added to the graduating TN subset, the OR associated with V/C is 5.12, which is comparable to that of the TN signature (OR: 4.29), while increasing the prevalence of biomarker-positive patients by ∼12%. Evaluation of PARPi-7 in the context of the graduating signature under the I-SPY 2 Bayesian model is pending. Conclusion: Our sample size is small. Our pre-specified analysis suggests the PARPi-7 signature shows promise for predicting response to veliparib/carboplatin combination therapy relative to control. If verified in a larger trial, this cell-line derived signature may contribute to the selection criteria of PARP inhibitor trials in the future. Citation Format: Denise M Wolf, Christina Yau, Ashish Sanil, Anneleen Daemen, Laura Heiser, Joe Gray, Lamorna Brown-Swigart, Susan Flynn, Gillian Hirst, I-SPY 2 TRIAL Investigators, Meredith Buxton, Angela DeMichele, Nola Hylton, Fraser Symmans, Doug Yee, Melissa Paoloni, Laura Esserman, Don Berry, Hope Rugo, Olufunmilayo Olopade, Laura van 't Veer. Evaluation of an in vitro derived signature of olaparib response (PARPi-7) as a predictive biomarker of response to veliparib/carboplatin plus standard neoadjuvant therapy in high-risk breast cancer: Results from the I-SPY 2 TRIAL [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P3-06-05.

Background: Triple-negative breast cancer (TNBC) accounts for approximately 10 to 15% of all breast cancer cases. The management of TNBC cases will significantly improve once molecular mechanisms specific to TNBC cells will be identified and treated accordingly. Evidence suggests that TNBC cells display deregulated kinase-dependent signaling cascades that differ from non-triple-negative breast cancer cells. We hypothesized that uniquely divergent phospho-circuits could be distinguished between TNBC and non-TNBC cell lines. By revealing such unique, dysfunctional phospho-signaling network, our long-term objective is to identify kinases that underpin triple-negative breast cancer development, and can be inhibited using targeted therapy. Methods: The kinome activity of TNBC and non-TNBC cell lines was identified (HCC70, MDA-MB-231, MDA-MB-436 compared to AU565, MCF-7, T47D). The functional phospho-signature of each breast cancer cell was analyzed using a high throughput experimental platform that monitors the level of activity of myriad kinases at once. This technique uses 242 phospho-sensing probes and 78 controls in an aqueous-based assay to simultaneously and directly measure the phospho-catalytic activity of phosphorylating enzymes in cell lysates. We mapped the most significantly deranged phospho-signaling cascades and the related kinases. Results: Using 6 cell lines tested under various conditions, we generated 72 phospho-signatures, out of a total of 23,040 data points. After validating the repeatability and robustness of the assay, the kinase activity signature of each breast cancer cell line was analyzed and compared to each other using unsupervised hierarchical clustering. The phospho-sensing assay revealed the heterogeneity of kinase activity networks among breast cancer cells. These data also established that phospho-signaling cascades related to AKT, ERK, and SRC kinases were differentially altered in TNBC and non-TNBC cell lines. Conclusions: We successfully identified unique phospho-circuits of TNBC and non-TNBC cell lines. Our goal is now to test whether specific kinase inhibitors can efficiently kill or prevent the growth of TNBC cell lines in culture and animal models. We will expand our approach into a high-content, functional kinome-screening platform to characterize the phospho-fingerprint of breast cancer cells and tissues, and explore the druggable, kinase-dependent mechanisms critical to triple-negative breast tumors. Citation Format: Bo Pan, Miki Mori, Pei Rong Evelyn Lee, Marij Hartog, Qiang Sun, Laura van 't Veer, Jean-Philippe Coppe. Functional mapping of the oncogenic kinome activity of triple-negative breast cancer cells [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P2-05-09.

Background: Veliparib, an inhibitor of Poly(ADP-ribose) polymerase (PARPi), in combination with DNA-damaging agents showed significant efficacy, especially in triple negative breast cancer (TNBC) patients. However, in I-SPY 2, approximately 42% of TNBC did not optimally respond to the veliparib based treatment. To exert therapeutic effect, drugs such as veliparib or the DNA damaging agent carboplatin must reach cancer cells at an adequate concentration. Differences within the microenvironment of the tumor (e.g. reduced vascular density and leaky endothelium) are responsible for variability in the uptake and distribution of the drug in the tumor. We hypothesize that heterogeneous or insufficient drug concentrations in the solid tumor lead to inadequate response to PARPi in some TNBC patients. First we performed a mouse xenograft study to determine the penetration and distribution of veliparib and carboplatin in TNBC tumors in mice. Methods: 1*106 MDA.MB.231, HCC70 or MDA.MB.436 cells were injected bilaterally into mamamary fat pad of a total of 36 Beige SCID mice. When the tumors exceeded 200 mm3, veliparib (20mg/kg or 60mg/kg) or placebo was orally administered 3 times daily for 3 days + Carboplatin 60mg/kg on day 1+2. These doses and dosing schedules were estimated to result equivalent plasma and tumor concentrations in mice compared to the doses currently used in clinical trials. Mice were euthanized 3 hours after the last dose of veliparib and tumors, and muscle tissues were obtained. In addition, control tissues were spiked with a broad range of concentrations of veliparib. The spatial distribution of the drugs in the tumor tissue was measured using Matrix-assisted laser desorption/ ionization mass spectrometric imaging (MALDI). The protonated molecular ion at m/z 245.1 was used to construct 2D ion maps of the tissue showing relative quantitation of drug levels. H&E staining was performed to characterize the tumors. Results: Veliparib in control tissues was detected at very low concentrations with a range of detection between 100 fmol-1nmol. After dosing, veliparib penetrates into the tumors and was detectable at both dose levels. However, drug distribution within the tumors was observed to be inhomogeneous. In spots where the drug accumulated, necrosis was observed. Veliparib accumulated in spots in the tumor and near the rim of the tumor. Differences between cell lines were observed, with the largest accumulation at the rim in BRCA mutated cells. Conclusions: Veliparib can be measured using MALDI with good specificity and sensitivity and using concentrations equivalent to patients. TNBC tumors show largely heterogeneous distribution of PARPi, with accumulation in the pushing margings of the tumor. Future analyses will determine whether this heterogeneity in drug distribution explains variability in response to PARPi therapies. Citation Format: Imke H Bartelink, Brendan Prideaux, Byron Hann, Gregor Krings, Jean-Phillippe Coppe, Lamorna Swigart-Brown, Pei Rong Evelyn Lee, Laura Esserman, Douglas Yee, Denise Wolf, Rada M Savic, Laura van t Veer. Spatial distribution of veliparib measured by matrix assisted laser desorption ionization mass spectrometry in triple negative breast cancer tumors [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P6-11-02.

Background: Further stratification of the 70-gene MammaPrintTM signature into ‘high’ and ‘ultra-high’ risk groups may help predict chemo-sensitivity. In I-SPY 2, patients were classified as MammaPrint High1 (MP1) or MammaPrint (ultra) High2 (MP2), with MP2 defined as MP_score Methods:115 HER2- patients (V/C: 71 and concurrent controls: 44) were considered in this analysis. We assess association between MP1/MP2 and response in the V/C and control arms alone using Fisher’s exact test, and relative performance between arms (biomarker x treatment interaction, likelihood ratio p < 0.05) using a logistic model. This analysis is also performed adjusting for HR status as a covariate. To assess MP1/MP2 in the context of the graduating signature, we added the MP2 patients to the graduating TN subset and evaluated the treatment effect in this ‘biomarker-positive’ group. Our study is exploratory with no claims for generalizability of the data. Statistical calculations are descriptive (e.g. p-values are measures of distance with no inferential content). This analysis does not adjust for multiplicities of other biomarkers in the trial but outside this study. Results: In the V/C arm vs. concurrent controls, there were 66 MP1 (V/C: 32, Control: 34) and 49 MP2 patients (V/C: 39, Control: 10), 78% of which are TN. The distribution of pCR rates among MP1/MP2 dichotomized groups are summarized in Table 1. V/C (n=71)Control (n=44)MP1 (n=32)MP2 (n=39)MP1 (n=34)MP2 (n=10)TN (n=59)3 / 819 / 303 / 132 / 8HR+HER2- (n=56)1 / 244 / 94 / 210 / 2 The OR between MP1/MP2 risk groups for predicting pCR is 9.71 in the V/C arm (p=6.63E-05), in comparison to an OR of 0.97 in the control arm (p=1). There is a significant biomarker x treatment interaction (p=0.023), which remains upon adjusting for HR status (p= 0.028). Based on the I-SPY 2 Bayesian model, a Phase III trial with 300 MP2 patients has a 95% predictive probability of success. When the MP2 patients are added to the graduating TN subset, the OR associated with V/C is 4.36, which is comparable to that of the TN signature (OR: 4.29), while increasing the prevalence of biomarker-positive patients by ∼10%. Conclusion: In our exploratory analysis, MP2 suggests higher sensitivity to V/C combination therapy relative to controls. This observation has prompted an investigation into the biological mechanisms distinguishing the MP1/MP2 subtype that may account for this specificity. Citation Format: Denise M Wolf, Christina Yau, Ashish Sanil, Jo Chien, Anne Wallace, Angela DeMichele, Hank Kaplan, Doug Yee, Claudine Isaacs, Kathy Albain, Rebecca Viscuzi, Judy Boughey, Stacey Moulder, Steven Chui, Qamar Khan, Toncred Styblo, Kirsten Edmiston, Donald Northfelt, Anthony Elias, Barbara Haley, Debu Tripathy, Lamorna Brown-Swigart, Susan Flynn, Gillian Hirst, Meredith Buxton, Nola Hylton, Melissa Paoloni, Fraser Symmans, Laura Esserman, Don Berry, Hope Rugo, Olufunmilayo I. Olopade, Laura van 't Veer. MammaPrint High1/High2 risk class as a biomarker of response to veliparib/carboplatin plus standard neoadjuvant therapy for breast cancer in the I-SPY 2 TRIAL [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P3-06-25.

Background: Further stratification of the 70-gene MammaPrintTM signature into ‘high’ and ‘ultra-high’ risk groups may help predict chemo-sensitivity. In I-SPY 2, patients were classified as MammaPrint High1 (MP1) or MammaPrint (ultra) High2 (MP2), with MP2 defined as MP_score AC). HER2- patients were randomized to receive N+T- >AC vs. T->AC. For HER2+ patients, neratinib was administered in place of trastuzumab (N+T->AC vs. H+T->AC). Here, we assess the performance of MP1/MP2 class as a specific biomarker of neratinib response. Methods: 115 patients in the neratinib arm and 76 concurrently randomized controls had Agilent 44K microarrays and pCR data available for analysis. We assess association between MP1/MP2 and response in the neratinib and control arms alone using Fisher’s exact test, and relative performance between arms (biomarker x treatment interaction, likelihood ratio p < 0.05) using a logistic model. This analysis is also performed adjusting for HR status as a covariate, and in receptor subsets. Our study is exploratory with no claims for generalizability of the data. Statistical calculations are descriptive (e.g. p-values are measures of distance with no inferential content). Our analyses do not adjust for multiplicities of other biomarkers in the trial but outside this study. Results: There are 133 MP1 patients (neratinib: 74, Control: 59) and 58 MP2 patients (neratinib: 41, Control: 17), 84% (49) of which are Her2-. The distribution of pCR rates among MP1/MP2 dichotomized groups are summarized in Table 1. Neratinib (n=115)Control (n=76)MP1 (n=74)MP2 (n=41)MP1 (n=59)MP2 (n=17)HER2- (n=105)0 / 1715 / 337 / 395 / 16HER2+ (n=86)22 / 574 / 85 / 200 / 1 MP2, one of the 10 eligible signatures, did not meet the graduation threshold; and MP1/MP2 did not show a significant biomarker x treatment interaction (OR in neratinib relative to control arm = 1.25). The MP1/MP2 x treatment interaction remains non-significant after adjustment for HR and HER2 status (p=0.54). In HER2- patients receiving neratinib, 45% (15/33) of MP2 patients achieved a pCR, compared to 0% (0/17) of MP1 patients. In the HER2- controls, there is a 31% pCR rate in MP2 (5/16) vs. 18% in MP1 (7/39) patients (OR=2.14). This difference in performance between treatment arms appears significant (p=0.041). 90% of HER2+ patients are MP1, thus MP1/MP2 status x treatment interaction within the HER2+ subtype cannot be evaluated. Conclusion: Within the I-SPY 2 population as a whole, MP1/MP2 stratification does not appear to be a specific biomarker of response to neratinib relative to the control arm. The number of HER2- patients is small and precludes any definitive conclusion, but these data motivate further investigation of the biological mechanisms distinguishing MP1 from MP2 to better understand chemotherapy and/or neratanib responsiveness. Citation Format: Christina Yau, Denise M Wolf, Ashish Sanil, Jo Chien, Anne Wallace, Judy Boughey, Doug Yee, Debu Tripathy, Angela DeMichele, Rita Nanda, Steven Chiu, Claudine Isaacs, Kathy Albain, Hank Kaplan, Stacey Moulder, Rebecca Viscusi, Donald Northfelt, Kirsten Edmiston, Anthony Elias, Toncred Styblo, Barbara Haley, Lamorna Brown-Swigart, Susan Flynn, Gillian L Hirst, Meredith Buxton, Nola Hylton, Melissa Paoloni, W Fraser Symmans, Laura Esserman, Don Berry, Minetta C Liu, John W Park, Laura van 't Veer. MammaPrint High1/High2 risk class as a biomarker of response to neratinib plus standard neoadjuvant therapy for breast cancer in the I-SPY 2 TRIAL [abstract]. In: Proceedings of the Thirty-Seventh Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2014 Dec 9-13; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2015;75(9 Suppl):Abstract nr P3-06-29.

Background: Triple negative breast cancer (TNBC), generally considered aggressive with few options beyond standard chemotherapy, is currently experiencing a period of optimism, with multiple trials showing increased efficacy of platinum containing regimens (with and without PARP-inhibition), and more recently, immunotherapy. In I-SPY 2, veliparib/carboplatin (VC) and the immune checkpoint inhibitor pembrolizumab (Pembro) graduated in the TN subtype. As part of the I-SPY 2 biomarker program, we previously identified the signature PARPi7, a measure of DNA repair deficiency (DRD), as predicting response to VC. As well, immune signatures, including one representing dendritic cells, predicted response to Pembro. Here we investigate the overlap between these biomarkers to identify patient subgroups more likely to respond to immunotherapy or platinum-based therapy and estimate response rates to VC and Pembro in patients positive for one, both, or neither biomarker. Methods: 153 TNBC patients (Control: 85; VC: 39; Pembro: 29) were considered in this analysis. Continuous DNA repair deficiency and immune signatures assayed at the pre-treatment time point were evaluated as published (Wolf, 2017; Danaher, 2017). To identify optimal dichotomizing thresholds, 2-fold cross-validation was repeated 500 times. Bayesian logistic regression was used to estimate pCR rates by arm for patient subsets defined by biomarker combinations. Our study is exploratory with no claims for generalizability; and this analysis does not adjust for multiplicities. Results: Using the optimal dichotomizing cutpoints, 54% of TNBC patients are classified as DRD+, and 67% Immune+. As expected, DRD+ patients have a high estimated pCR rate to VC (75%); and Immune+ patients have a high estimated pCR rate to Pembro (82%). Combining these biomarkers, 40% of TNBC are positive for both biomarkers, 40% for only one biomarker (26% Immune+/DRD-, 14% Immune-/DRD+), and 20% for neither. The Immune+/DRD+ TN subset predicted sensitive to both VC and Pembro has a higher estimated pCR rate to both VC and Pembro relative to control (estimated pCR rates: VC: 75%, Pembro: 73% vs. control: 18%). In contrast, the Immune+/DRD- group, predicted to be sensitive to Pembro, has the highest pCR rate to Pembro (VC: 48%, Pembro: 83% vs. control 21%), whereas the Immune-/DRD+ group is most responsive to VC (VC: 63%, Pembro: 41% vs. control 29%). For the 20% of Immune-/DRD- TNBC patients, their estimated response rates are lower than the biomarker-positive patients (VC: 27%, Pembro: 43% vs. control: 19%). Conclusion: Molecular phenotypes capturing immune activation and DNA repair deficiency may predict response of TNBC and help prioritize selection of platinum or immunotherapy containing regimens. Citation Format: Denise M. Wolf, Christina Yau, Julia Wulfkuhle, Emanuel Petricoin, Lamorna Brown-Swigart, Gillian Hirst, Smita Asare, I-SPY 2 Consortium, Douglas Yee, Angela DeMichele, Hope Rugo, Olufunmilayo Olopade, Rita Nanda, Minetta Liu, Laura Esserman, Laura van t' Veer. Integration of DNA repair deficiency and immune biomarkers to predict which early-stage triple-negative breast cancer patients are likely to respond to platinum-containing regimens vs. immunotherapy: The neoadjuvant I-SPY 2 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 2679.

Background: I-SPY2 is a multicenter, Phase II trial using response-adaptive randomization within biomarker subtypes to evaluate novel agents as neoadjuvant therapy for high-risk at least T2N0 breast cancer. The primary endpoint is pathologic complete response (pCR) at surgery. The goal is to identify regimens that have ≥ 85% Bayesian predictive probability of success in a 300-patient phase 3 neoadjuvant trial defined by hormone-receptor (HR) and HER2 status, and MammaPrint (MP). Regimens may leave the trial for futility (< 10% probability of success), maximum sample size accrual (with probability of success ≥ 10% and < 85%), or as recommended by the independent DSMB. For HER2- subjects the control arm is weekly paclitaxel x12 then doxorubicin & cyclophosphamide (AC) q2-3 weeks x4. For this arm, paclitaxel was omitted and replaced with maximum tolerated dose PARPi talazoparib with synergy dosed irinotecan (25mg/m2). Paclitaxel could be given adjuvantly for these subjects and non-responding subjects could be taken off of experimental therapy. Methods: Women with tumors ≥ 2.5cm were eligible for screening. MP low/HR+ tumors were ineligible. MRI scans (baseline, 3 cycles after start of therapy, prior to AC, and prior to surgery) were used in a longitudinal statistical model to predict pCR for individual patients. Talazoparib was given at 1mg daily with 25mg/m2 irinotecan q2wks. Analysis was intention to treat. Subjects who switched to non-protocol therapy count as non-pCR. Subjects on experimental therapy at time of arm closure are non-evaluable. Talazoparib/irinotecan (TI) was open only to HER2- tumors and eligible for graduation in 3 of 10 pre-defined signatures: HER2-, HR+HER2- and HR-/HER2-. Results: TI did not meet criteria for graduation and was stopped at the recommendation of the DSMB based on expectations of limited activity beyond that seen with standard treatment. Maximum sample size had been reached at the time of this recommendation and subjects currently receiving TI were allowed to continue or change to standard therapy. Exploratory “as treated” analysis for response in gBRCA mutation carriers showed 6/10 gBRCA carriers attained a pCR in the TI arm. Except for 1 patient these gBRCA pCR subjects had >90% tumor reduction by MRI after TI and prior to AC (range: 68-96%). In the TI arm pCR rates were also higher in subjects with a PARPi7-High/MP2 gene expression signature (0.344 vs 0.146). Expected differences in toxicity were seen between arms including g3/4 peripheral neuropathy on control therapy which included paclitaxel (2.6% vs none) and g3/4 neutropenia with TI (30.2% vs 8.2%). Notably gBRCA mutation carriers receiving TI had higher rates of g3/4 neutropenia (60% vs 25.9%). Conclusion: The I-SPY2 study finds the probability that investigational regimens will be successful in a Phase III neoadjuvant trial; TI did not reach the efficacy threshold of 85% probability of success in Phase III in any of the 3 signatures. However by adding talazoparib with synergy dosed irinotecan we were able to omit paclitaxel and observe similar estimated pCR rates. This informs current work to evolve the I-SPY2 trial design to reduce toxicity without compromising outcomes and develop successful combinations targeted to biology, including DNA repair deficiency. Citation Format: Richard Schwab, Amy S. Clark, Christina Yau, Nola Hylton, Wen Li, Denise Wolfe, A Jo Chien, Anne M. Wallace, Andres Forero-Torres, Erica Stringer-Reasor, Rita Nanda, Nora Jaskowiak, Judy Boughey, Tufia Haddad, Heather S. Han, Catherine Lee, Kathy Albain, Claudine Isaacs, Anthony D. Elias, Erin D. Ellis, Payal Shah, Julie E. Lang, Janice Lu, Debasish Tripathy, Kathleen Kemmer, Douglas Yee, Barbara Haley, Melanie Majure, Erin Roesch, Christos Vaklavas, Cheryl Ewing, Teresa Helsten, W Fraser Symmans, Jane Perlmutter, Hope S. Rugo, Michelle Melisko, Amy Wilson, Ruby Singhrao, Laura van 't Veer, Angela DeMichele, Smita Asare, Don Berry, Laura J. Esserman. Evaluation of talazoparib in combination with irinotecan in early stage, high-risk HER2 negative breast cancer: Results from the I-SPY 2 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 CT136.

BACKGROUND: The presence of disseminated tumor cells (DTCs) in bone marrow of early breast cancer (EBC) patients is a strong predictor of poor prognosis. We assessed heterogeneity in copy number, PIK3CA mutation, and gene expression in DTCs to shed light on the molecular biology of these cells. METHODS: We isolated EPCAM-positive DTCs in 71 EBC patients using immunomagnetic enrichment and fluorescence-activated cell sorting (IE/FACS). Isolated DTCs, along with corresponding primary tumors (n=16), were subjected to genome-wide copy number profiling (n=47) by array comparative genomic hybridization, and mutation screening of the PIK3CA gene (n=53) by Sanger sequencing. The expression of 64 cancer-related genes in DTCs was analyzed by microfluidic-based multiplexed RT-QPCR (n=30). DTC expression profiles were compared with available gene expression data from circulating tumor cells (CTCs) of metastatic breast cancer patients. RESULTS: Copy number profiles of DTCs were less aberrant and distinct from their corresponding primary tumors. PIK3CA mutations detected in 26% of DTCs were mutually exclusive to those found in matched primary tumors. Expression profiles of DTCs were distinguishable from marrow leukocytes, and displayed up-regulation of oncogenes MYC and CCNE1. Unsupervised hierarchical clustering analysis revealed two subtypes of DTCs: (1) luminal with dual epithelial-mesenchymal properties (high ESR1 and VIM/CAV1 expression), and (2) basal-like with proliferative/stem cell-like phenotype (low ESR1 and high MKI67/ALDH1A1 expression). DTCs possessed gene expression signatures that were unique from those of CTCs. ALDH1A1, CAV1 and VIM were up-regulated in DTCs relative to CTCs. ESR1/ER and ERBB2/HER2 status in DTCs vs. corresponding primary tumors showed high discordance (40% and 43%, respectively), suggesting shift in biomarker status in micrometastatic cells in the bone marrow. CONCLUSIONS: We demonstrate the feasibility of isolation and comprehensive molecular characterization of DTCs from bone marrow of EBC patients. Comparative genomic analysis suggests that DTCs disseminate early and acquire genomic aberrations independently of the primary tumor. Expression profiling revealed two distinct subpopulations of DTCs. Validation in larger cohorts is needed to confirm the presence of these molecular subtypes and to evaluate their biological and clinical significance. Citation Format: Mark Jesus M. Magbanua, Hope Rugo, Louai Hauranieh, Ritu Roy, Janet Scott, Jen Chieh Lee, Feng Hsiao, Eduardo Sosa, Denise Wolf, Laura van't Veer, Laura Esserman, John Park. Genomic and expression profiling reveals molecular heterogeneity of disseminated tumor cells in early breast cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 5578.

Background: Previous studies suggest that within the triple positive HR+HER2+ subtype, patients classified as BluePrint (BP) Luminal subtype are more responsive to pertuzumab and trastuzumab (P/H) as opposed to trastuzumab (H) alone. In the I-SPY2 TRIAL, HER2-targeted treatment arms include H, P/H, neratinib (N), and T-DM1/Pertuzumab (T-DM1/P); and patients were classified by BP molecular subtyping in addition to conventional receptors. We evaluated BP subtype as a predictor of response in HR+HER2+ patients and assessed pathway differences between BP molecular subtypes. Methods: 125 HR+HER2+ patients (N: 42; P/H: 29, T-DM1/P: 35; H: 19) with pre-treatment Agilent microarrays and BP subtype assignments were considered. We assess association between BP subtypes and pCR using Fisher's exact test. To identify genes associated with BP Luminal vs. BP HER2 subtype, we (1) apply a Wilcoxon rank sum test and (2) fit a logistic model, with the Benjamini-Hochberg (BH) multiple testing correction (BH p Results: Of the 125 HR+HER2+ patients, 71 were BP HER2-type and 50 were BP Luminal-type. The distribution of pCR rates in BP Luminal/ HER2 subtypes are as follows: Distribution of pCR rates in BP Luminal/ HER2 subtypes by treatment armN (n = 40)P/H (n = 29)T-DM1/P (n = 34)H (n = 18)All arms (n = 121)BP Luminal (n = 50)0 / 9 (0%)1 / 12 (8.3%)2 / 16 (12.5%)1 / 13 (7.7%)4 / 50 (8%)BP HER2 (n = 71)12 / 31 (38.7%)13 / 17 (76.5%)15 / 18 (83.3%)1 / 5 (20%)41 / 71 (57.7%)Odds Ratio6.1; p = 0.12730.2; p = 0.00048429.6; p = 8.60E-052.8; p = 0.49015.3; p = 1.04E-08 In a whole transcriptome analysis, 1725 genes were differentially expressed. By DAVID enrichment analysis, the most significantly enriched pathways were related to immune function, with the BP HER2 subgroup showing higher expression. Conclusion: Our analysis suggests that HR+HER2+ BP Luminal subtype is associated with lower response rates to HER2-targeted agents, including P/H, and may need an alternative strategy. BP HER2 subtype appears associated with higher expression of immune-related genes, relative to BP Luminal; and suggests that immune signaling may contribute to HER2-targeted therapy sensitivity. Citation Format: Pei Rong Evelyn Lee, Zelos Zhu, Denise Wolf, Christina Yau, William Audeh, Annuska Glas, Lamorna Brown-Swigart, Gillian Hirst, Angela DeMichele, I-SPY 2 TRIAL Investigators, Laura Esserman, Laura van 't Veer. BluePrint Luminal subtype predicts non-response to HER2-targeted therapies in HR+/HER2+ I-SPY 2 breast cancer patients [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 2612.

Triple-negative breast cancer (TNBC) accounts for approximately 15% of all breast cancer cases, with over 35,000 newly diagnosed women per year in the USA. TNBC patients are at high risk of recurrence, and neoadjuvant standard chemotherapy leads to pathologic complete response in only about 30% of patients. No targeted therapy has yet been conclusively established to improve outcome. The management of TNBC will significantly improve once mechanisms responsible for TNBC resistance to chemotherapies are identified. Here, we applied a new functional proteomic strategy to reveal which (dys)regulated phospho-signaling circuits are the effective dependencies of chemotherapy-resistant TNBC cells. The high-throughput kinase activity-mapping (HT-KAM) assay is our new screening technology to assess the catalytic activity of many kinases in parallel. HT-KAM relies on collections of biologic peptide probes that are computationally derived from PhosphoAtlas (Olow and Chen et al., Cancer Research 2016) and are physically used as combinatorial sensors to measure the activity of kinase enzymes in biologic extracts. The HT-KAM system provides access to a vast, untapped resource of meaningful measurements, whether readouts are interpreted irrespective of which enzymes phosphorylate which probes, or analyzed to convert global phospho-signatures into functional profiles of kinase activities. Kinome maps reveal how signaling networks are rewired by drug interventions in the context of different cellular backgrounds or exogenously mutated proteins/pathways, and provide insight into potentially targetable pathways. We previously successfully showed that the HT-KAM platform identifies new, actionable kinases responsible for intrinsic or acquired targeted-therapy resistance in BRAFV600E colorectal cancer cells and melanoma tumors from patients. To study TNBC, we used a 615-peptide sensor library that supports the mapping of >100 kinases and represents >900 functional kinase-substrate interactions relevant to tumor biology, and matches hundreds of druggable components of signaling circuits. We characterized the phospho-catalytic signatures of 10 TNBC cell lines, either untreated or treated with chemotherapeutic drugs (5-FU, carboplatin, doxorubicin) at IC50 concentrations. We also included a panel of 8 luminal BC cells. Based on our advanced preliminary data, we anticipate showing a comprehensive map of the oncogenic kinome of TNBC, and how it rewires its signaling network in response to chemotherapies. The most hyperactive and conserved pathways establish a priority map of strategic kinase hot spots to explore as therapeutic candidates. We find that chemotherapy-induced kinases such as SRC, AKT, CHEK, or regulating NFkB or PARP pathways can be inhibited to improve therapeutic sensitivity in TNBC. Citation Format: Jean-Philippe F. Coppe, Pan Bo, Carolien van der Borden, Nina Koemans, Changjun Wang, Denise Wolf, Christina Yau, Sietske Bakker, Marij Hartog, Miki Mori, Ana Ruiz-Saenz, Zhongzhong Chen, Aleksandra Olow, Pei Rong Evelyn Lee, Laura van 't Veer. Systematic identification of the actionable kinase dependencies of chemotherapy-resistant triple-negative breast cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 956.

Background: The angiogenesis (ANG1/2) inhibitor trebananib (TR) was one of the experimental agents evaluated in I-SPY 2. In I-SPY 2, all patients received at least standard chemotherapy (paclitaxel followed by doxorubicin/cyclophosphamide: T->AC). HER2- patients were randomized to receive TR+T->AC vs. T->AC. For HER2+ patients, TR was administered with trastuzumab (TR+H+T->AC vs. H+T->AC). We hypothesized that genes/signatures in the ANG/TIE signaling axis may specifically predict response to TR, and tested expression levels of 11 genes: TIE1/2, ANGPT1/2/4, AGPTNL1/3, VEGFA, ICAM1, PECAM1 and MMP2. We also evaluated angiogenesis and hypoxia expression signatures, based on the hypothesis that hypoxic tumors with a fragile blood supply may be vulnerable to drugs in this class. Methods: Data from 266 patients (TR: 134 and concurrent controls: 132) were available. Pre-treatment biopsies were assayed using Agilent 44K (32627) or 32K (15746) expression arrays; and these data were combined using ComBat. All I-SPY 2 qualifying biomarker analyses follow a pre-specified analysis plan. We use logistic modeling to assess biomarker performance. A biomarker is considered a specific predictor of TR response if it associates with response in the TR arm, and if the biomarker x treatment interaction is significant (likelihood ratio test, p Results: ANGPT1, a direct target of trebananib, associates with pCR in the TR arm but not the control arm, and shows a significant interaction with treatment that retains significance in a model adjusting for HR and HER2. In addition ICAM1, expressed on endothelial and immune cells, strongly associates with response in the TR arm, but also in the control arm in the population as a whole. In the HR+HER2- subset, both ICAM1 and PECAM1 associate with pCR in the TR arm and not the control arm, with a trend toward treatment interaction. Interestingly, in the TN subset, where pCR rates were highest in the TR arm relative to control, these mechanism-of-action biomarkers did not appear to predict response. Rather, in exploratory whole genome analysis, response of TN's strongly associates with immune related genes (e.g. HLA's, IL21R, CCL13). Conclusion: Following our pre-specified analysis, ANGPT1 succeeds as a specific predictor of response to trebananib in I-SPY 2. In addition, ICAM1 and PECAM1 associate with response in the HR+HER2- subset; and in exploratory analysis immune signaling predicts response in the TN subset. These biomarkers may merit further evaluation in future trials. Citation Format: Denise M. Wolf, Christina Yau, Lamorna Brown-Swigart, Gillian Hirst, I-SPY 2 TRIAL Investigators, Smita Asare, Richard Schwab, Don Berry, Laura Esserman, Kathy S. Albain, Brian Leland-Jones, Laura van 't Veer. Evaluation of ANG/TIE/hypoxia pathway genes and signatures as predictors of response to trebananib (AMG 386) in the neoadjuvant I-SPY 2 TRIAL for Stage II-III high-risk breast cancer [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2018; 2018 Apr 14-18; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2018;78(13 Suppl):Abstract nr 2611.

BACKGROUND: In the I-SPY 2 TRIAL, HER2-negative patients received standard chemotherapy alone or with the PARP inhibitor veliparib and carboplatin (VC). VC graduated in the triple-negative (TN) subtype, and we’ve previously shown that MammaPrint High1/High2 (MP1/2) risk class and the PARPi-7 signature may predict VC response. Here we evaluate whether combining these signatures identifies a subset of TN patients especially likely to respond to VC. METHODS: This analysis includes 60 TN patients (VC: 39 and controls: 21). PARPi-7 and MP1/2 signature scores are computed from Agilent 44K arrays. We further stratify TN patients by VC-sensitivity biomarkers (MP2, PARPi7-high). We use Bayesian modeling to estimate pCR rates in each arm and the predictive probability of VC demonstrating superiority to control in a 1:1 randomized phase 3 trial of 300 ‘biomarker-positive’ patients. Our study is exploratory and does not adjust for multiplicities of biomarkers outside this study. RESULTS: Though 90% of TNs are PARPi7-high or MP2, concordance between these biomarkers is 50%. The estimated pCR rates to VC are 69% in PARPi7-high and 64% in MP2 TN patients, compared to 53% in the entire TN subgroup. TN patients positive for both sensitivity markers (assessed as PARPi7-high and MP2) achieved an estimated pCR rate of 79% in the VC arm vs. 23% in the control arm, with a predictive probability of success in phase 3 of 99.6%. In contrast, TN patients negative for at least one VC sensitivity marker (PARPi7-low and/or MP1) only had an estimated response rate to VC of 35%. Estimated pCR rates in biomarker subsetsBiomarker subsets within TNEstimated pCR rate in V/C [95% CI]Estimated pCR rate in controls [95% CI]Predictive probability of phase 3 success (300 pt)Unselected TN (n = 60)53% [39 - 67]27% [13 - 43]0.904PARPi7-high and MP2 (n = 24; 40%)79% [60 - 93]23% [5.8 - 49]0.996PARPi7-low and/or MP1 (n = 36; 60%)35% [17 - 55]29% [13 - 49]0.386 CONCLUSION: Our analysis suggests TN patients who are also MP2 and PARPi7-high may be more sensitive to V/C than patients with fewer markers in the ‘sensitive’ state. Citation Format: Denise M. Wolf, Christina Yau, Ashish Sanil, Lamorna Brown-Swigart, Gillian Hirst, Meredith Buxton, Melissa Paoloni, I-SPY 2 TRIAL Investigators, Olufunmilayo Olopade, Hope Rugo, Angela De Michele, Fraser Symmans, Don Berry, Laura Esserman, Laura van ‘t Veer. Combining sensitivity markers to identify triple-negative breast cancer patients most responsive to veliparib/carboplatin: results from the I-SPY 2 TRIAL. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 858.

BACKGROUND: Further stratification of the 70-gene MammaPrint(TM) prognostic signature into ‘high’ and ‘ultra-high’ risk groups may help predict chemo-sensitivity. In I-SPY 2, patients were classified as MammaPrint High1 (MP1) or MammaPrint (ultra) High2 (MP2), using a threshold predefined by the median cut-point of I-SPY 1 participants who fit the eligibility criteria for I-SPY 2. MP1/2 classification was added to HR and HER2 to define the subtypes used in the I-SPY 2 adaptive randomization engine. The first two experimental agents/combinations to graduate from I-SPY 2 were veliparib/carboplatin (V/C) in the TN subset, and neratinib (N) in the HR-HER2+ subset. MP2 was found to be a sensitivity marker for V/C but not N, whereas MP1 class appears associated with resistance to N within the HER2- subset. Despite these associations with response, it remains unclear whether MP1/MP2 classification represents differences in tumor biology. Here, we present exploratory analysis to elucidate the pathway differences between the MP classes. METHODS: 263 patients (V/C: 71, N: 115, and controls: 77) with pre-treatment Agilent 44K microarrays and MP1/2 class assessments were considered in this analysis. To identify signature genes associated with MP1 vs. MP2 class, we (1) apply a Wilcoxon rank sum test and (2) fit a logistic model. P-values are corrected for multiple comparisons using the Benjamini-Hochberg (BH) method, with a significance threshold of BH p RESULTS: 63% (165/263) of patients are MP1 class and 37% (98/263) MP2. MP1/2 class is associated with receptor subtype (Fisher's exact test: p CONCLUSION: MP2 class appears associated with higher expression of cell cycle & DNA repair genes. Association between MP2 class and response to V/C suggests that higher cell cycle activity may contribute to V/C sensitivity. Citation Format: Denise M. Wolf, Christina Yau, Lamorna Brown-Swigart, Gillian Hirst, Meredith Buxton, Melissa Paoloni, I-SPY 2 TRIAL Investigators, Olufunmilayo Olopade, Angela De Michele, Fraser Symmans, Hope Rugo, Don Berry, Laura Esserman, Laura van ‘t Veer. Gene and pathway differences between MammaPrint High1/High2 risk classes: results from the I-SPY 2 TRIAL in breast cancer. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 859.

Background: The U.S. Preventive Services Task Force recommends that women with a >3% five-year risk of developing breast cancer consider taking selective estrogen receptor modifiers (SERMs) or aromatase inhibitors (AIs) to reduce their risk. Polygenic risk score (PRS), calculated by adding the individual breast cancer risk association for each common genetic variant (SNP), has been found to predict women at low- to high-risk of breast cancer. We analyze associations between SNP risk alleles and known breast cancer risk factors (ethnicity, family history of breast cancer and number of biopsies); furthermore, we quantify the likely impact on chemoprevention recommendations by adding the PRS to known risk models in a subset of women participating in the University of California 100,000 women Athena Breast Health Network. Methods: Our research cohort included 838 women with no previous diagnosis of breast cancer from the University of California, San Francisco, and was enriched for women determined to be at elevated risk for developing breast cancer by the Gail model. A panel of 75 breast cancer risk SNPs were evaluated on saliva and blood samples (Akesogen Inc; COGS oncochip array). The PRS for each patient was calculated by converting the odds ratio for each SNP into a likelihood ratio (LR) and combining LR's across SNPs. Breast Cancer Surveillance Consortium (BCSC), Gail, BCSC-PRS and Gail-PRS scores (risk models incorporating PRS within a Bayesian framework), were evaluated for each patient. Associations between variables were assessed using t-test or ANOVA. A threshold of p Results: Women in this study carry an average of 65 risk allele SNPs (of 150, 2 per locus). By ANOVA, there is a statistically significant association between the SNPs risk allele count and ethnicity (p = 0.014), with a trend towards association with a family history of a first-degree relative with breast cancer (p = 0.053). PRS is significantly associated with a family history breast cancer (p = 0.031); neither SNP allele count nor PRS associates with previous biopsy status. We found by adding PRS that 12% (86/707) and 13% (104/776) of patients with a prior BCSC or Gail score 3% five-year risk, respectively, changed classifications by adding PRS and would no longer be eligible for chemoprevention. Conclusion: The addition of SNP based PRS to BCSC and Gail models significantly changes how women are classified and as a result changes whether risk reducing agents are recommended. PRS will be combined with BCSC and genetic test results for 9 breast cancer genes to calculate a women's breast cancer risk in the PCORI-funded Athena WISDOM study of 100,000 women, comparing risk-based vs. annual mammography screening. Citation Format: Sarah Theiner, Sarah D. Sawyer, Paige Kendall, Alexandra S. Perry, Denise Wolf, Scott Huntsman, Bo Pan, Jeffery A. Tice, David A. Pearce, Thomas Cink, Laura Esserman, Elad Ziv, Laura van ‘t Veer. Common genetic variants associated with breast cancer risk used in the Athena study to enhance models identifying women for breast cancer chemoprevention. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 2623.

Background: The MINDACT TRIAL investigated the clinical utility of the 70-gene profile (MammaPrint®), when confronted to the standard clinical pathological (CP) criteria, for the selection of patients unlikely to benefit from adjuvant chemotherapy (CT).Methods: From 2007 to 2011, 11288 patients were screened in 112 centers from 9 countries, of whom 6693 were enrolled. Enrolled patients had undergone successful determination of their genomic risk G (with MammaPrint®) and clinical risk C (modified version Adjuvant! Online). Patients were evaluated as “low clinical risk”, if their 10-year disease specific survival (without CT or endocrine therapy (ET), as estimated by Adjuvant! Online, was greater than 88% for ER-positive patients, and greater than 92% for ER-negative patients. Patients characterized in both assessments as “low- risk” were spared CT while for those “high- risk”, CT was advised. Those with discordant results were randomized to use either C or G risk evaluation to decide on CT. The main hypothesis was that G-assay would outperform the C-criteria by reducing the administration of adjuvant CT without impairing outcome. The CT and ET regimen randomization will not be part of the current presentation. The primary endpoint for the CT (yes/no) decision was distant metastasis free survival (DMFS). Secondary endpoints were the proportion of women treated per C over G, the overall survival (OS) and disease free survival (DFS). In the G-low/C-high risk group, a null hypothesis of a 5-year DMFS of 92% was tested (one-sided a = 2.5%) in the patients who were randomized to use the 70-gene signature prognosis (no CT), and who complied with this “no CT” assignment. Results: Median age of enrolled patients was 55 years. Eighty percent were LN- negative. 58% had tumor size between 1 and 2cm, 88% were HR-positive and 10% were HER2 positive. The enrolled patients were categorized in 4 risk groups; G-low/C-low [2745 patients (41%)], G-high/C-low [592 (9%)], G-low/C-high [1550 (23%)] and G-high/C-high [1806 (27%)]. In the G-low/C-high group, 51% of patients were randomized to CT (49% to no CT) and in the G-high/C-low group 50% to CT. At 5- years, DMFS was 94.7% (95%CI, 92.5-96.2%) for patients who were C- high/ G- low and randomized to use G and received no CT. This CI exceeds 92%, thereby rejecting the primary null hypothesis of MINDACT. There was an absolute 14% reduction in adjuvant CT administration when using the G versus C treatment strategy. Conclusion: C-high/G-low patients had a 5-year DMFS rate in excess of 94%, whether they were randomized to adjuvant CT or no CT. Using the G assay for CT treatment decision led to a 14% reduction in adjuvant CT administration in MINDACT. When using the G assay MammaPrint® among the C-high patients, there was a 46% (1550/3356) reduction in CT prescription, providing level 1A evidence for the clinical utility of MammaPrint® for assessing the lack of a clinically relevant chemotherapy benefit in the clinically high risk (C-high) population. Citation Format: Martine Piccart, Emiel Rutgers, Laura van’ t Veer, Leen Slaets, Suzette Delaloge, Giuseppe Viale, Jean Yves Pierga, Peter Vuylsteke, Etienne Brain, Suzan Vrijaldenhoven, P. Neijenhuis, Bruno Coudert, Tineke Smilde, Miguel Gil, Alastair Thompson, Isabel Rubio, Rodolfo Passalaqua, Erika Matos, Urlike Nitz, Mauro Delorenzi, Geraldine Thomas, Theodora Goulioti, Carolyn Straehle, Konstantinos Tryfonidis, Jan Bogaerts, Fatima Cardoso, on behalf of TRANSBIG consortium and MINDACT investigators. Primary analysis of the EORTC 10041/ BIG 3-04 MINDACT study: a prospective, randomized study evaluating the clinical utility of the 70-gene signature (MammaPrint) combined with common clinical-pathological criteria for selection of patients for adjuvant chemotherapy in breast cancer with 0 to 3 positive nodes. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr CT039.

Background: I-SPY 2 is a multicenter, phase II neoadjuvant trial in women with high-risk stage II/III breast cancer using adaptive randomization within biomarker subtypes to evaluate novel agents added to standard chemotherapy. Primary endpoint is pathologic complete response (pCR). Goal is to identify regimens that meet a high Bayesian predictive probability of statistical significance in a neoadjuvant 300-patient phase III trial defined by hormone-receptor (HR), HER2 status, and MammaPrint (MP). Experimental regimens may “graduate” in 1 of 10 signatures, with a maximum of 120 patients. We report efficacy results for neratinib (N). Methods: Tumors ≥2.5cm by clinical exam & ≥2cm by imaging are eligible for screening. MP low risk/HR+/HER2- tumors are ineligible for randomization. Patients receive chemotherapy (paclitaxel qwk x 12, doxorubicin and cyclophosphamide q2-3 wk x 4, T->AC). HER2- pts were randomized to N+T->AC vs. T->AC and HER2+ pts to N+T->AC vs. trastuzumab+T->AC. Analysis is intent-to-treat with pts who switch to non-protocol therapy regarded as non-pCRs. We provide estimated pCR rates (95% Bayesian probability intervals), probabilities of superiority of neratinib over control, and Bayesian predictive probabilities of success in an equally randomized phase III trial. Results: Neratinib met the predictive probability criterion in HR-/HER2+, “graduated”, and accrual ceased [115 N patients (65 HER2+), 78 concurrently randomized controls (22 HER2+)]. The table shows results for all 10 signatures. Two patients (1 N and 1 control) withdrew consent and are not included. Conclusion: I-SPY 2's standing trial mechanism efficiently evaluates agents in biomarker-defined patient subsets. In a modest number of patients, adaptive randomization successfully identified a biomarker signature (HR-/HER2+) for neratinib's further development. All HER2+ and MP+ tumors may also benefit from this regimen, consistent with preclinical data. Evaluation in I-SPY 3, a phase III registration trial, is planned. SignatureEstimated pCR Rate (95% probability interval)Probability NeratinibIs Superior to ControlPredictive Probability of Success in Phase IIINeratinibControlALL32% (28-36%)23% (19%-28%)92%44%HR+23% (18%-28%)17% (11%-22%)81%40%HR-43% (37%-49%)31% (24%-38%)89%53%HER2+39% (33%-45%)23% (15%-30%)95%73%HER2-26% (21%-32%)24% (18%-29%)63%20%MP+*45% (38%-53%)29% (20%-39%)91%66%HR-/HER2-36% (29%-43%)30% (23%-38%)72%34%HR-/HER2+55% (46%-64%)32% (22%-43%)94%78%HR+/HER2+31% (24%-37%)17% (10%-24%)91%65%HR+/HER2-14% (8%-19%)16% (10%-21%)39%12%* MammaPrint 44K Array Low/High score was adjusted to MammaPrint High1 (MP-) or High2 (MP+), using a pre-defined median cut-point of I-SPY 1 participants, who fit the eligibility criteria of I-SPY 2. Citation Format: John W. Park, Minetta C. Liu, Douglas Yee, Angela DeMichele, Laura van 't Veer, Nola Hylton, Fraser Symmans, Meredith B. Buxton, A. Jo Chien, Amy Wallace, Michelle Melisko, Richard Schwab, Judy Boughey, Debashish Tripathy, Hank Kaplan, Rita Nanda, Stephen Chui, Kathy S. Albain, Stacy Moulder, Anthony Elias, Julie E. Lang, Kirsten Edminston, Donald Northfelt, David Euhus, Qamar Khan, Julia Lyandres, Sarah E. Davis, Christina Yau, Ashish Sanil, Laura J. Esserman, Donald A. Berry. Neratinib plus standard neoadjuvant therapy for high-risk breast cancer: Efficacy results from the I-SPY 2 TRIAL. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr CT227. doi:10.1158/1538-7445.AM2014-CT227

Background: The OMIC revolution has produced tens of thousands of published gene expression signatures, pathways, and modules associated with cancer biology or patient outcome. Thanks to the TCGA effort to systematically collect molecular data from multiple cancer types, it is now possible to compare these signatures within and across cancers, and to collapse them into a smaller set of non-redundant features. Methods: We scored 3602 TCGA samples representing 12 cancer types for expression of a compendium of 6,898 published gene expression prognostic signatures, co-expression modules, amplicons and pathways, using RNAseq expression profiles. 6477 of these features were from the database MsigDB, and the rest from the literature. To define non-redundant gene-programs, we applied a bimodality filter and used Weighted Gene Correlation Network Analysis (WGCNA) to aggregate the bimodal signatures and pathways into coherent clusters. Associations were assessed using standard statistical methods. Results: From the thousands of cancer signatures in our analysis we identified 22 non-redundant gene-programs, defined as groups of gene signatures with high correlation across the Pan-Cancer dataset. These gene programs represent many of the hallmarks of cancer, including sustained proliferative signaling and DNA repair; immune system signaling; altered glycolytic and fatty acid metabolism yielding resistance to stress; apoptosis evasion; epithelial to mesenchymal transition; co-opted squamous and ‘stemness’ developmental programs; adhesion and cell-cell communication programs, some using plasma membrane and lipid vesicles; and self-sufficiency in growth signals employing estrogen, EGF, and MYC. More than a third of the original signatures clustered together in the proliferation gene-program. In addition to cell cycle and proliferation pathways, this program includes prognostic signatures for lymphoma, bladder, lung, breast, ovarian and other cancers, suggesting that most prognostic signatures are sensing a common proliferative signal associated with cancer aggression. The presence of two dominant immune signals showed a clear divergence in immune signaling between T cell/B cell immune activation and interferon responsive signaling, with the former correlated with immuno-modulatory treatment targets like CTLA4 and PD1. Interestingly, basal, hypoxic, and squamous differentiation programs were associated with poor outcome in many cancer types; in addition, expression of a more novel program dominated by tumor suppressing miRNA associated with good prognosis. Conclusion: The compendium of gene expression signatures is correlated, and can be reduced to a small set of non-redundant gene-programs that mirror the cancer hallmarks and may provide opportunities for investigating cancer heterogeneity and treatment approaches beyond tissue of origin. Citation Format: Denise M. Wolf, Cheng Fan, Katherine A. Hoadley, Christina Yau, Artem Sokolov, TCGA Network, Josh Stuart, Charles Perou, Laura van ’t Veer. Thousands of published cancer signatures and pathways can be collapsed into a handful of non-redundant gene programs: a TCGA pan-cancer analysis. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 360. doi:10.1158/1538-7445.AM2014-360

Background: Improper control of phosphorylation networks represents a powerful regulator and biomarker of cell dys-function leading to malignancy. Discovering which active mechanisms of disease progression can be efficiently targeted, and knowing whether kinase networks circumvent therapeutic interventions, are challenges researchers and clinicians face. Surprisingly however, measuring the phosphorylating activity of kinases, and potentially monitoring the functionality of the entire human phospho-reactome at once, remains largely unexplored. We developed a semi-high throughput assay to monitor the phospho-catalytic activity of kinase enzymes, using their biological targets as phospho-sensors. We successfully used this assay to identify the active, oncogenic phospho-signature of breast cancer cells. Methods: We first built a library of peptide sensors established from confirmed kinase substrates' phosphorylation sites. Precisely, we used computational methods to create a unique phospho-repertoire cataloguing 3,408 peptide sequences established from validated human proteins' phosphorylation sites, curated from 38 public databases. Second, we experimentally used these biologically relevant probes in multiplex assays to measure the catalytic state of kinases. Specifically, 228 peptides were developed into an ATP-consumption screen to identify the activity signatures of AKT, EGFR, MAPK, ABL and SRC family kinases. We developed analysis methods to derive phospho-signatures from semi-high throughput measurements. Finally, we validated the bio-applicability of the assay using isogenic culture model of basal-like breast cancer (HMT-3522 S1 and T4-2), and cell lines harboring EGFR/HER2-oncogenic alterations (MDA-MB-231, AU565, MCF7, T47D). Results: The differential phosphorylation activity of 25 recombinant, active kinase enzymes was successfully captured on 228 phospho-sensing peptides. In cancer cell extracts, hyper-activated AKT, SRC, EGFR, or MAPK kinases originally identified by immuno-detection were reliably and specifically detectable using the peptide-based kinase-activity assay. The phospho-sensing assay revealed the heterogeneity of kinase activity networks in breast cancer cells, and possible new drug-targeting opportunities. Conclusion: This unique strategy and resources allows to comprehensively measure the catalytic activity of a multitude of kinases at once, ultimately representing a new molecular dimension to characterize biological samples. Such new phospho-reactome profiling system suggests how chemotherapeutic interventions affect the activity of kinases and lead to reprogramming of phospho-circuits. Our efforts now concentrate on expanding this approach into a high-content, functional kinome-screening platform to map phosphorylation signatures that cause breast malignancies, therapeutic resistance, or recurrence. Citation Format: Jean-Philippe F. Coppé, Miki Mori, Evelyn Lee, Bo Pan, Aaron Boudreau, Zhongzhong Chen, Laura van 't Veer. Phospho-reactome profiling reveals the heterogenic, targetable kinase signature of breast cancer. [abstract]. In: Proceedings of the 105th Annual Meeting of the American Association for Cancer Research; 2014 Apr 5-9; San Diego, CA. Philadelphia (PA): AACR; Cancer Res 2014;74(19 Suppl):Abstract nr 5309. doi:10.1158/1538-7445.AM2014-5309

Background: Co-expression modules in cancer are cohesive groups of genes with highly correlated expression patterns, presumably regulating phenotypes important in carcinogenesis, metastasis, or response to treatment. With the growing database of gene expression data, it is now possible to deduce co-expression modules active in breast cancer subpopulations. Methods: We assembled 74 breast-cancer related gene expression datasets containing ∼5,500 samples altogether. Per dataset, we identified genes with bimodal expression analyzed using mixture-model clustering to ultimately find gene groups that are consistently co-regulated across multiple datasets. Functional and pathway enrichment of modules was assessed using DAVID and g:profiler software tools. Associations between module expression and patient outcome, chemotherapy response, clinical variables, epithelial vs. stromal expression, intrinsic subtype and other published signatures were assessed using standard statistical methods. Results: Our meta-analysis identified 11 modules ranging in size from ∼5-200 genes. As expected, there were modules representing estrogen signaling, cell proliferation, and ERBB2 signaling that correlate with intrinsic subtype and receptor status. In addition, we found modules associated with immune signaling, development, histone modification, and the ECM. The immune modules were highly correlated to previously published Tcell/Bcell, STAT1, and IFN immune signatures. In multivariate analysis, the combination of a downregulated Tcell/Bcell immune module and an upregulated proliferation module associated with recurrence, suggesting that cancers with a high proliferation rate in the absence of an activated immune system are prone to recur. This association was especially strong in TN and basal cancers. Comparing modules associated with response to chemotherapy to those associated with the prognosis of untreated patients, the most common pattern was that of a module associating with good prognosis or a good response to chemotherapy (but not both). For instance, high expression of the estrogen module associates with a good prognosis but a poor response to chemotherapy, whereas upregulation of the proliferation module associates with a poor prognosis but a good response to chemotherapy. Another pattern we observed, of a biomarker that associates with good prognosis without chemotherapy and a good response to chemotherapy, was found in the T/B cell immune modules. Conclusion: Co-expression modules provide a high-level functional view of breast cancer that complements the ‘cancer hallmarks’. These results suggest that in some high-immune patients, the same host processes contributing to an excellent response to chemotherapy might preclude its necessity, and support a treatment strategy boosting anti-tumoral immunity in low-immunity patients with highly proliferating tumors. Citation Format: Denise M. Wolf, Christina Yau, Aaron Boudreau, Laura Esserman, Laura Van ‘t Veer, Marc Lenburg. Co-expression modules as hallmarks of breast cancer survival and response. [abstract]. In: Proceedings of the 104th Annual Meeting of the American Association for Cancer Research; 2013 Apr 6-10; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2013;73(8 Suppl):Abstract nr 5221. doi:10.1158/1538-7445.AM2013-5221

The oncogenic activity of kinases harnesses myriad signalling cascades driving tumorigenesis. Improper control of phosphorylation networks represents one of the most relevant regulator and biomarker of cell dys-functions leading to malignancies. Surprisingly however, measuring the phosphorylation activity of kinases, and potentially monitoring the functionality of the entire kinome at once to identify molecular signatures causing cancers and guide therapeutic interventions, remains largely unexplored. Specifically, a highly convenient tool for clinical applications or research purposes would be the advent of a microarray-like chip to map phosphorylation networks. We undertook to develop such a kinase-sensing platform, and present here unique strategies, technical advances and data -including the detection of oncogenic Src and HER kinases activity in biological samples- supporting such potential breakthrough. The devised kinase activity-sensing system relies on peptide probes that act as beacons to evaluate enzymes functionality, and emit distinctive optical signals before and after (de-)phosphorylation. The following steps have been achieved. First, in order to use baits relevant to human cancer etiology, we created a protein and peptide repertoire cataloguing 6173 distinctive kinase-substrate active nodes and compiling 2702 unique biological kinase peptide targets that all represent latent nano-sensors usable to comprehensively track kinase signaling networks. Second, we established that our label-free detection approach (namely Raman scattering of probes monitored via surface-enhanced Raman spectroscopy (SERS)) is a functional system that reliably and robustly identifies the phosphorylation state of peptide probes. Third, real-time kinetics and kinase inhibitor experiments using biochemical samples and biological extracts from mammalian cells demonstrated that our approach could directly assess the functionality of kinases. Fourth, in an attempt to identify hyperactive Src kinase in tumor cells, our assay discerned normal from oncogenic Src-mutated cells. Finally, using breast cancer cell models of tumor progression and tissue samples from patients, we were able to measure EGFR kinases (hyper-)activity in vivo. Together, our strategy allows for highly sensitive, specific, rapid, one-step, label-free, multiplex capture and measurement of phosphorylation events. Our efforts will now focus on expending this approach into a fully functional kinomic-screening platform to explore oncogenic kinase networks in human cancers. We will translate such device into the clinic, and map phosphorylation signatures that cause breast malignancies to eventually guide therapeutic interventions that best match breast cancer patients’ disease characteristics and predicted responses to (chemo-)therapies. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 103rd Annual Meeting of the American Association for Cancer Research; 2012 Mar 31-Apr 4; Chicago, IL. Philadelphia (PA): AACR; Cancer Res 2012;72(8 Suppl):Abstract nr 4790. doi:1538-7445.AM2012-4790

Background: Breast cancer is a heterogeneous disease. Clinically useful predictors for breast cancer management may therefore need to consider the variety of molecular processes that can impact chemotherapy response and survival. With the growing database of gene expression data, it is now possible to deduce co-expression modules active in breast cancer subpopulations. Viewed from a combinatorial perspective supporting ‘OR’ and ‘AND’ reasoning, the activities of these modules might provide insights into breast cancer biology and form the basis for improved predictors. Methods: We assembled 74 breast-cancer related gene expression datasets containing ∼5,500 samples altogether. Per dataset, we identified genes with bimodal expression analyzed using mixture-model clustering to ultimately find gene groups that are consistently co-regulated across multiple datasets. We scored the AFFY U133 ISPY1 dataset (117 pts) for module expression, and explored the relationships between module state, chemotherapy response, patient outcome, receptor status, intrinsic subtype, and other signatures. To derive logical functions consisting of nested ‘AND’ and ‘OR’ gates relating module expression to clinical variables, we discretized the module scores and applied a methodology in the spirit of Karnaugh maps for digital circuit design. Results: Our meta-analysis identified 11 modules ranging in size from ∼5–200 genes. The expression level of some modules were associated with known molecular markers such as intrinsic subtype or wound signature scores, whereas others – three immune related modules, two ECM related modules, a mixed immune/ECM module, and a small module encoding histones – do not. Disease free survival of patients with triple negative disease is predicted (p=1.93e-08) by the logical function fDFS_TN=[ER_module>LOW] OR [TN_Immune_ECM=HIGH] OR [(AURK/PROLIF_module=HIGH) AND ((Histone_module>LOW) OR (TN_Immune_ECM=HIGH)], where TN_Immune_ECM is a composite function over the three immune modules and the immune/ECM hybrid module. The best predictor of chemo-sensitivity for triple negatives is a hybrid over subtype and module state, fRCB01_TN=Subtype_basal AND TN_immune_ECM, correctly classifying 95% of the RCB01 sensitives and 90% of the RCBIII resistants. Conclusion: In triple negatives, high proliferation can be rescued by immune/ECM upregulation, though neither immune rescue nor low RCB is necessary for survival if the estrogen module is even slightly upregulated or if the histone module is high. Similar analyses of ER/Pr+ and Her2+ patients demonstrate receptor-subtype specificity in the logic predicting response. Citation Format: {Authors}. {Abstract title} [abstract]. In: Proceedings of the 102nd Annual Meeting of the American Association for Cancer Research; 2011 Apr 2-6; Orlando, FL. Philadelphia (PA): AACR; Cancer Res 2011;71(8 Suppl):Abstract nr LB-311. doi:10.1158/1538-7445.AM2011-LB-311