Search

Cancer progression is driven in part by genomic alterations

Background Trastuzumab deruxtecan (T-DXd) is an antibody-drug conjugate (ADC) composed of an HER2-directed antibody and a topoisomerase I inhibitor covalently linked via a tetrapeptide-based cleavable linker. In DESTINY-Breast01, tolerability and efficacy of T-DXd including overall response rate, progression-free survival, duration of response and overall survival have been demonstrated for HER2-positive metastatic and/or unresectable BC in patients (pts) relapsing after 2 or more anti-HER2-based regimens. In France, rapid and fair access to innovative drugs outside clinical trials, prior to their marketing authorization in a given indication, is granted by the French Health Agency, ANSM (Agence Nationale de Sécurité du Médicament et des Produits de Santé), through cATU program. Here we report first real world evidence data from cATU program in HER2+ BC pts treated with T-DXd. Methods T-DXd 5.4mg/kg was given intravenously in monotherapy every 3 weeks in HER2-positive metastatic and/or unresectable BC pts who had previously received at least 2 lines of anti-HER2 regimens in the metastatic setting. Eligible pts needed to have normal neutrophil count and no left ventricular dysfunction. Pts with active or history of interstitial lung disease (ILD), pneumonitis, severe pulmonary disease were excluded. Clinical, biological and safety data were collected until the end of cATU, as well as treatment response according to RECIST 1.1., dose modification, treatment interruption and discontinuation. Analysis was performed on March 31th, 2021 on the basis of available collected data. Results From September 30th, 2020 to March 31th, 2021, 155 centers requested at least one ATU for a total of 539 adult pts; 468 requests were accepted and 71 were refused as they did not meet eligibility criteria. T-DXd was received by 459 pts with the following characteristics: 99.1% were women, median age was 58 years, 90.4% had a ECOG score of 0-1, 98.9% had initial HER2-positive BC (IHC 3+ or IHC 2+/ISH+), 67% were hormone receptor positive. The main sites of metastases were bones (57.3%), lymph nodes (51.6%), lungs (36.2%), liver (33.1%), brain (28.1%) and cutaneous/subcutaneous (13.9%). Median time between initial diagnosis of primary BC and inclusion was 6.6 years (range: 6.6 months - 33.9 years). 81.7% of pts had previously received radiotherapy and 76.5% underwent surgery. The median number of prior cancer regimens in the metastatic setting was 4 (range: 2-22). 21.1% received 2 prior lines of metastatic treatments, 19.6% received 3 lines and 59.3% received 4 lines or more. 94.8% pts received prior trastuzumab emtansine, and 79.3% had prior pertuzumab. During follow-up, data on tumor assessment were available for 160 pts. Of these, 56.7% had complete or partial response and 12.1% had progression. Of the 459 treated pts, 97 pts (21.1%) experienced ≥ 1 Adverse Drug Reaction (ADR) including 41 pts (8.9%) with ≥ 1 serious ADR. Most frequent ADRs were related to gastrointestinal toxicity (35.4%). During cATU, 17 cases (3.7%) with ILD or considered as ILD were reported but no cases had a fatal outcome (only grade 1 or 2 when reported by physicians). 13 fatal cases were reported (no drug-related deaths, attributed by physician). ADRs leading to T-DXd discontinuation were reported in 4 pts (0.9%). Dose reductions were reported in 17 pts (3.7%) and 21 pts (4.6%) had temporary interruptions. Conclusions We report here the first real world data from the French cATU in HER2-positive BC pts treated by T-DXd. The enrolment of 468 pts in 6 months illustrated the unmet medical need for this population. T-DXd had antitumor activity with a similar response rate to that reported in previous clinical studies. T-DXd was well tolerated and no new safety signals were observed. Citation Format: Thierry Petit, Nawale Hajjaji, Eric-Charles Antoine, Marc-Antoine Benderra, Michel Gozy, Cyril Foa, Jean-Loup Mouysset, Julien Grenier, Mireille Mousseau, Audrey Mailliez, Mahasti Saghatchian, Emma Lachaier, Isabelle Desmoulins, Audrey Hennequin, Patricia Maes, Delphine Loirat, Francesco Ricci, Véronique Diéras, Dominique Berton, Florence Lai Tiong, Luis Teixeira, Nadine Dohollou, Christelle Lévy, Thomas Bachelot, Jean-Yves Pierga. Trastuzumab deruxtecan in previously treated HER2-positive metastatic or unresectable breast cancer (BC): First real-life data from the cohort temporary authorization for use (cATU) program in France [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 P2-13-26.

Background Cancer heterogeneity is a main obstacle for the development of effective therapies, as its replication in in vitro preclinical models is challenging. Around 96% of developed drugs are estimated to fail from discovery to the clinical trial phase probably because of the unsuitability and unreliability of current preclinical models (Front Pharmacol 9:6, 2018; Nat Rev Cancer 8: 147–56, 2008) in replicating the overall biology of tumors, for instance the tumor microenvironment. Breast cancer is the most frequent cancer among women causing the greatest number of cancer-related deaths. Breast cancer can typically be modeled in vitro through the use of tumoroids; however, current approaches using mouse tumoroids fail to reproduce crucial aspect of human breast cancer, while access to human cells is limited and the focus of ethical concerns. New models of breast cancer, such as companion dogs, have emerged given the resemblance of developed spontaneous mammary tumors to human breast cancer in many clinical and molecular aspects; however, they have so far failed to replicate the tumor microenvironment. The present work aimed at developing a robust canine mammary tumor model in the form of tumoroids which recapitulate the tumor diversity and heterogeneity. Results We conducted a complete characterization of canine mammary tumoroids through histologic, molecular, and proteomic analysis, demonstrating their strong similarity to the primary tumor. We demonstrated that these tumoroids can be used as a drug screening model. In fact, we showed that paclitaxel, a human chemotherapeutic, could kill canine tumoroids with the same efficacy as human tumoroids with 0.1 to 1 μM of drug needed to kill 50% of the cells. Due to easy tissue availability, canine tumoroids can be produced at larger scale and cryopreserved to constitute a biobank. We have demonstrated that cryopreserved tumoroids keep the same histologic and molecular features (ER, PR, and HER2 expression) as fresh tumoroids. Furthermore, two cryopreservation techniques were compared from a proteomic point of view which showed that tumoroids made from frozen material allowed to maintain the same molecular diversity as from freshly dissociated tumor. Conclusions These findings revealed that canine mammary tumoroids can be easily generated and may provide an adequate and more reliable preclinical model to investigate tumorigenesis mechanisms and develop new treatments for both veterinary and human medicine.

PURPOSE Imaging reports in oncology provide critical information about the disease evolution that should be timely shared to tailor the clinical decision making and care coordination of patients with advanced cancer. However, tumor response stays unstructured in free-text and underexploited. Natural language processing (NLP) methods can help provide this critical information into the electronic health records (EHR) in real time to assist health care workers. METHODS A rule-based algorithm was developed using SAS tools to automatically extract and categorize tumor response within progression or no progression categories. 2,970 magnetic resonance imaging, computed tomography scan, and positron emission tomography French reports were extracted from the EHR of a large comprehensive cancer center to build a 2,637-document training set and a 603-document validation set. The model was also tested on 189 imaging reports from 46 different radiology centers. A tumor dashboard was created in the EHR using the Timeline tool of the vis.js javascript library. RESULTS An NLP methodology was applied to create an ontology of radiographic terms defining tumor response, mapping text to five main concepts, and application decision rules on the basis of clinical practice RECIST guidelines. The model achieved an overall accuracy of 0.88 (ranging from 0.87 to 0.94), with similar performance on both progression and no progression classification. The overall accuracy was 0.82 on reports from different radiology centers. Data were visualized and organized in a dynamic tumor response timeline. This tool was deployed successfully at our institution both retrospectively and prospectively as part of an automatic pipeline to screen reports and classify tumor response in real time for all metastatic patients. CONCLUSION Our approach provides an NLP-based framework to structure and classify tumor response from the EHR and integrate tumor response classification into the clinical oncology workflow.

Breast cancer is the most frequent cancer among women causing the greatest number of cancer-related deaths. Cancer heterogeneity is a main obstacle to therapies. Around 96% of the drugs fail from discovery to the clinical trial phase probably because of the current unreliable preclinical models. New models emerge such as companion dogs who develop spontaneous mammary tumors resembling human breast cancer in many clinical and molecular aspects. The present work aimed at developing a robust canine mammary tumor model in the form of tumoroids which recapitulate the tumor diversity and heterogeneity. We conducted a complete characterization of these canine mammary tumoroids through histologic, molecular and proteomic analysis, demonstrating their strong similarity to the primary tumor. We demonstrated that these tumoroids can be used as a drug screening model. In fact, we showed that Paclitaxel, a human chemotherapeutic, could killed canine tumoroids with the same efficacy as human tumoroids with 0.1 to 1 μM of drug needed to kill 50% of the cells. Due to easy tissue availability, canine tumoroids can be produced at larger scale and cryopreserved to constitute a biobank. We have demonstrated that cryopreserved tumoroids keep the same histologic and molecular features (ER, PR and HER2 expression) as fresh tumoroids. Two techniques of cryopreservation were compared demonstrating that tumoroids made from frozen tumor material allowed to maintain a higher molecular diversity. These findings revealed that canine mammary tumoroids can be easily generated at large scale and can represent a more reliable preclinical model to investigate tumorigenesis mechanisms and develop new treatments for both veterinary and human medicine.

Background Breast cancer (BC) remains a leading cause of cancer-related death among women worldwide. The complexity of this disease, especially its heterogeneity, have prevented its eradication and driven resistance to treatments. To reach precision oncology to eradicate BC, therapy needs to be specific to each tumor clone. Yielding enough molecular information from tumor clones to identify new drug targets represents a technical challenge due to sample size limitation or loss of spatial resolution. Matrix-assisted laser desorption/ionization (MALDI) imaging mass spectrometry combined with microproteomics enables a spatially-resolved unlabeled tumor imaging of its protein distribution, thus revealing proteomic clones. Our aims were to analyze the clonal proteome of luminal breast cancers, and explore its potential to expand new drug target discovery and drug repurposing. Methods A retrospective study at the Comprehensive Cancer Centre Oscar Lambret (Lille, France) was conducted to analyze 76 FFPE luminal HER2 negative tumors: 52 primary tumors from patients with early BC and 24 BC metastases. Patients gave their informed consent and the study was approved by the local institutional review board. MALDI mass spectrometry imaging and spatially-resolved on-tissue shotgun microproteomics were performed on FFPE slides of tumor tissue to determine the proteomic profile of selected clones using nanoLC-MS & MS/MS. Protein identification was performed using MaxQuant software against the Uniprot database. Functional annotation and characterization of the identified proteins were performed using Panther software. Candidate druggable targets were searched using DrugCentral druggable genome database, and their druggability level was assessed using the classification by the Illuminating the Druggable Genome Knowledge Management Center. The clonal proteome dataset was compared to publically available TCGA, BC360, and CDx datasets. Results The clonal proteome analysis identified a total of 2868 different proteins; 780 proteins were found in more than 50% of the patients. Panther analysis showed that 22% of the proteins were classified as enzymes, 15% were related to DNA processes, 6% were structural proteins, and less than 2% were related to immunity. Panther identified 139 pathways in the clonal proteome dataset. The clonal proteome analysis yielded the highest number of pathways compared to TCGA, BC360, and CDx datasets. 41 pathways (mainly metabolic pathways) were exclusive to the clonal proteome dataset. 1495 proteins of this dataset had an entry and were druggable in DrugCentral database, with 52% of them with known mechanisms of action and drug interaction. The main target classes were enzymes (60%), kinases (23%) and transporters (7%), whereas kinases were dominant in TCGA, BC360, and CDx datasets (46% to 77%). To explore the clonal proteome potential for repurposing anticancer drugs in luminal breast cancers, protein targets matching approved antineoplastic agents were searched using DrugCentral database. 97 approved anticancer drugs were identified, mostly chemotherapy (33%) or protein kinase inhibitors (28%), of whom only 17 were approved for breast cancer treatment. Compared to publically available TCGA, BC360, and CDx datasets, the clonal proteome analysis yielded the highest number of drug target candidate. Conclusion Mass spectrometry-based analysis of BC proteomic clones provides the technological means to access large functional molecular information at a clonal level to develop clone specific strategies for drug target discovery and drug repurposing in BC. Citation Format: Nawale Hajjaji, Soulaimane Aboulouard, Delphine Bertin, Yves Marie Robin, Isabelle Fournier, Michel Salzet. Proteomic profiling of specific tumor clones using spatially resolved mass spectrometry technologies for precision oncology [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 PS18-36.

HR+ breast cancers are defined by the prominence of signaling pathways dependent on the estrogen receptor. Endocrine therapy is the standard treatment for these advanced diseases. Resistance to these treatments, called hormone resistance, appears invariably with biological mechanisms that have led to the development of therapeutic opportunities. An exhaustive literature review was carried out concerning the biology of the hormone resistance pathways, the therapeutic options before the era of CDK4/6 inhibitors, the rise of CDK4/6 inhibitors and the therapeutic prospects in a situation of hormone resistance. Various biological abnormalities have been identified in the mechanisms of hormone resistance such as changes in the estrogen receptor, mutations in the ESR1 gene, aberrant activation of the PI3K pathway or cell cycle deregulations. Historical strategies for circumventing this hormone resistance have been based on hormonal manipulation, on the development of new endocrine therapy such as fulvestrant (selective estrogen receptor inhibitor, SERD), on combinations of treatments such as everolimus, a mTOR inhibitor. This strategy combining endocrine therapy and targeted therapy has led to the development of combinations with CDK4/6 inhibitors which have now become a standard treatment in the hormone resistance phase. The future of this therapeutic era remains to be written with new combinations of hormone therapy and targeted therapy such as PI3K inhibitors or even with the positioning of new SERDs in clinical development.

Background: Acute severe forms of COVID-19 infection are more likely in cancer patients and growing attention has been given to the persistent symptoms of the disease after severe COVID-19. However, mild illness is the dominant clinical presentation of COVID-19 infection. To investigate patients’ behavior and the short- and longer-term pattern of the disease in cancer patients with mild COVID infection, a longitudinal online survey was conducted for 16 months during the pandemic in a large cohort of cancer patients from a French COVID-19 hot spot. An online questionnaire was administered at three time points between the first wave of the pandemic in France and the fourth wave. The questionnaire was completed by 1415 to 2224 patients, which queried patients’ demographics, their behavior, and COVID infection patterns. Seventy percent of the patients were female, and 40% had a comorbid condition. More than one-third of the participants had breast cancer, and half were survivors. The rate of infection was 30% during wave 1 and 10% in wave 4; most patients had a mild COVID-19 infection. Twenty-five percent of infected patients during wave 4 did not seek medical advice. At wave 4, 87% of the patients received at least one dose of vaccine. Systematic compliance to shielding measures decreased over time. The short-term pattern of mild COVID changed between wave 1 and wave 4. Twenty-two percent of infected patients experienced persistent signs for more than 6 months with a negative impact on sleep, social behavior, and increased consumption of stress-relieving drugs. Our results showed a high prevalence of long-lasting symptoms in cancer patients with mild COVID-19 infection and inadequate behavior toward the disease and prevention measures among patients.

Integrating tumor heterogeneity in the drug discovery process is a key challenge to tackle breast cancer resistance. Identifying protein targets for functionally distinct tumor clones is particularly important to tailor therapy to the heterogeneous tumor subpopulations and achieve clonal theranostics. For this purpose, we performed an unsupervised, label-free, spatially resolved shotgun proteomics guided by MALDI mass spectrometry imaging (MSI) on 124 selected tumor clonal areas from early luminal breast cancers, tumor stroma, and breast cancer metastases. 2868 proteins were identified. The main protein classes found in the clonal proteome dataset were enzymes, cytoskeletal proteins, membrane-traffic, translational or scaffold proteins, or transporters. As a comparison, gene-specific transcriptional regulators, chromatin related proteins or transmembrane signal receptor were more abundant in the TCGA dataset. Moreover, 26 mutated proteins have been identified. Similarly, expanding the search to alternative proteins databases retrieved 126 alternative proteins in the clonal proteome dataset. Most of these alternative proteins were coded mainly from non-coding RNA. To fully understand the molecular information brought by our approach and its relevance to drug target discovery, the clonal proteomic dataset was further compared to the TCGA breast cancer database and two transcriptomic panels, BC360 (nanoString®) and CDx (Foundation One®). We retrieved 139 pathways in the clonal proteome dataset. Only 55% of these pathways were also present in the TCGA dataset, 68% in BC360 and 50% in CDx. Seven of these pathways have been suggested as candidate for drug targeting, 22 have been associated with breast cancer in experimental or clinical reports, the remaining 19 pathways have been understudied in breast cancer. Among the anticancer drugs, 35 drugs matched uniquely with the clonal proteome dataset, with only 7 of them already approved in breast cancer. The number of target and drug interactions with non-anticancer drugs (such as agents targeting the cardiovascular system, metabolism, the musculoskeletal or the nervous systems) was higher in the clonal proteome dataset (540 interactions) compared to TCGA (83 interactions), BC360 (419 interactions), or CDx (172 interactions). Many of the protein targets identified and drugs screened were clinically relevant to breast cancer and are in clinical trials. Thus, we described the non-redundant knowledge brought by this clone-tailored approach compared to TCGA or transcriptomic panels, the targetable proteins identified in the clonal proteome dataset, and the potential of this approach for drug discovery and repurposing through drug interactions with antineoplastic agents and non-anticancer drugs.

Glioblastoma (GBM) is the most common and devastating malignant brain tumor in adults. The mortality rate is very high despite different treatments. New therapeutic targets are therefore highly needed. Cell-surface proteins represent attractive targets due to their accessibility, their involvement in essential signaling pathways, and their dysregulated expression in cancer. Moreover, they are potential targets for CAR-based immunotherapy or mRNA vaccine strategies. In this context, we investigated the GBM-associated surfaceome by comparing it to astrocytes cell line surfaceome to identify new specific targets for GBM. For this purpose, biotinylation of cell surface proteins has been carried out in GBM and astrocytes cell lines. Biotinylated proteins were purified on streptavidin beads and analyzed by shotgun proteomics. Cell surface proteins were identified with Cell Surface Proteins Atlas (CSPA) and Gene Ontology enrichment. Among all the surface proteins identified in the different cell lines we have confirmed the expression of 66 of these in patient’s glioblastoma using spatial proteomic guided by MALDI-mass spectrometry. Moreover, 87 surface proteins overexpressed or exclusive in GBM cell lines have been identified. Among these, we found 11 specific potential targets for GBM including 5 mutated proteins such as RELL1, CYBA, EGFR, and MHC I proteins. Matching with drugs and clinical trials databases revealed that 7 proteins were druggable and under evaluation, 3 proteins have no known drug interaction yet and none of them are the mutated form of the identified proteins. Taken together, we discovered potential targets for immune therapy strategies in GBM.

Background Estimating the risk of relapse in early luminal breast cancers is paramount to determine the therapeutic strategy. Genomic signatures have come along clinical/pathological factors to refine this estimation. However, cases of tumors with discordant genomic and clinical risk estimates have highlighted the difficulty to delineate which risk parameter, clinical or genomic, should prevail. The aims of the study were to determine the frequency of discordant cases in a real life cohort, and provide their functional molecular characterization using mass spectrometry imaging associated to spatially-resolved tissue microproteomic. Methods Discordant cases were identified from a cohort of 222 early breast cancer patients with hormone receptor positive and HER2 negative tumors. All patients were treated at our institution and had a PAM50-based genomic assay (Prosigna®) to estimate their genomic risk of distant recurrence. The clinical estimation of the risk of relapse was based on tumor size and grade in node negative patients, and on the number of nodes involved (one versus two to three) in node positive patients. The tissue microproteomic profiling of both cancer cells and stroma was performed using a spatially resolved on-tissue shotgun mass spectrometry technology after MALDI mass spectrometry imaging. Results Nearly 30% of cases had discordant genomic and clinical risk estimates, both in node negative and node positive tumors. Proteins overexpressed in low clinical and high genomic risk tumors, compared to high clinical and high genomic risk tumors, were involved in angiogenesis, inflammatory processes, B and T cell activation, or tumor promoting signaling pathways including EGFR, PDGF, FGF, GnHR and Ras network. Proteins overexpressed in high clinical and low genomic risk tumors, compared to low clinical and low genomic risk tumors, were involved in DNA replication, inflammation or integrin signaling. Within each genomic risk group, comparison of low and high clinical risk tumors showed distinct proteomic profiles, both in tumor cells and stroma. The proteomic clusters differentially expressed between low and high clinical risk tumors varied according to the genomic risk group and nodal status. Functional analyses showed that proteins of these clusters were involved in several pathways or processes, such as nucleotides metabolism, apoptosis, tumorigenesis, integrin signaling pathways and inflammation. Conclusion Mass spectrometry microproteomics revealed distinctive tumor and microenvironment molecular profiles between low and high clinical risk tumors within homogeneous genomic risk groups. This suggests that clinical risk factors bring additional biological insight, and should be further integrated to genomic parameters. Citation Format: Nawale Hajjaji, Soulaimane Aboulouard, Yves Marie Robin, Delphine Bertin, Jacques Bonneterre, Isabelle Fournier, Michel Salzet. Functional proteomic characterization of early luminal breast cancers with discordant genomic and clinical risk estimates [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 P4-07-08.

Integrating tumor heterogeneity in the drug discovery process is a key challenge to tackle breast cancer resistance. Identifying protein targets for functionally distinct tumor clones is particularly important to tailor therapy to the heterogeneous tumor subpopulations. For this purpose, we performed an unsupervised, label-free, spatially resolved shotgun proteogenomic guided by MALDI mass spectrometry imaging (MSI) on 124 selected tumor clonal areas from early luminal breast cancers, tumor stroma, and breast cancer metastases. 2868 proteins were identified. The main protein classes found in the clonal proteome dataset were enzymes, cytoskeletal proteins, membrane-traffic, translational or scaffold proteins, or transporters. As a comparison, gene-specific transcriptional regulators, chromatin related proteins or transmembrane signal receptor were more abundant in the TCGA dataset. Moreover, 26 mutated proteins have been identified. Similarly, expanding the search to alternative proteins databases retrieved 126 alternative proteins in the clonal proteome dataset. The majority of these alternative proteins were coded mainly from non-coding RNA. To fully understand the molecular information brought by our approach and its relevance to drug target discovery, the clonal proteomic dataset was further compared to the TCGA breast cancer database and two transcriptomic panels, BC360 (nanoString®) and CDx (Foundation One®). We retrieved 139 pathways in the clonal proteome dataset. Only 55% of these pathways were also present in the TCGA dataset, 68% in BC360 and 50% in CDx. Seven of these pathways have been suggested as candidate for drug targeting, 22 have been associated with breast cancer in experimental or clinical reports, the remaining 19 pathways have been understudied in breast cancer. Among the anticancer drugs, 35 drugs matched uniquely with the clonal proteome dataset, with only 7 of them already approved in breast cancer. The number of target and drug interactions with non-anticancer drugs (such as agents targeting the cardiovascular system, metabolism, the musculoskeletal or the nervous systems) was higher in the clonal proteome dataset (540 interactions) compared to TCGA (83 interactions), BC360 (419 interactions), or CDx (172 interactions). Thus, we described the non-redundant knowledge brought by this approach compared to TCGA or transcriptomic panels, the targetable proteins identified in the clonal proteome dataset, and the potential of this approach for drug discovery and repurposing through drug interactions with antineoplastic agents and non-anticancer drugs.SignificanceSpatially resolved mass spectrometry guided by MALDI MS imaging is a precision oncology tool to map and profile breast cancer proteomic clones with the aim of integrating tumor heterogeneity in the target discovery process to develop clone-tailored therapeutic strategies in breast cancer.HighlightsSpatially resolved mass spectrometry guided by MALDI mass spectrometry imaging allows an in-depth proteomic profiling of breast cancer functional clones.This unsupervised and unlabeled technology performed on intact tumors provides a multidimensional analysis of the clonal proteome including conventional proteins, mutated proteins, and alternative proteins.The rich clonal proteomic information generated was not redundant with TCGA or transcriptomic panels, and showed pathways exclusively found in the proteomic analysis.A large proportion of the proteins in the clonal proteome dataset were druggable with both antineoplastic agents and non-anticancer drugs, showing the potential application to drug repurposing.A significant number of the proteins detected had partially or not yet known drug interactions, showing the potential for discovery.

Background Breast cancer is a heterogeneous disease with a wide range of outcomes that are not fully predicted by routine clinical and pathologic features. The risk of recurrence of hormone receptor positive early-stage breast cancer, the most frequent tumor subtype, increases continuously over time. Genomic-based signatures have been developed to categorize patients according to their risk of recurrence and guide therapeutic decisions in this tumor subtype. However, the main genomic-based assays used in clinical practice classify up to 40% of the patients in an intermediate risk group. Many interrogations remain about the optimal strategy in this group. The aim of this study was to refine the molecular characterization of the intermediate risk group and determine the proportion of shared features with the low or high-risk groups using a mass spectrometry based proteomic approach. Methods Tumors with available routine PAM50 assay (PROSIGNA) results were selected from a cohort of breast cancer patients treated at Oscar Lambret Cancer Center (France). Fifteen tumors evenly split between PAM50 low, intermediate and high-risk groups were analyzed to determine the proteomic profiles of both cancer cells and stroma using MALDI mass spectrometry imaging combined with microproteomics, a spatially resolved proteomic technology. Results PAM50-Intermediate risk tumors had a distinctive proteomic profile compared to low and high-risk tumors. Heterogeneous nuclear ribonucleoproteins, 4-aminobutyrate aminotransferase, and pleckstrin homology-like domain family A member 1 are discriminating proteins between intermediate risk and low risk tumors. Differences were observed in expression of integrin beta-1, DNA replication licensing factors, splicing factors and interleukin enhancer-binding factor 2 between intermediate and high-risk tumors. Proteomic profiles of stroma according to tumor risk groups also showed differential protein expressions mainly between intermediate and high-risk groups. Breast cancer markers such as nuclear mitotic apparatus protein 1(NUMA1), C-1-tetrahydrofolate synthase (MTHFD1), cystatin-C (CST3), and T-cell immune regulator 1 (TCIRG1) were identified in high-risk tumors. Specific protein profiles were identified in stroma versus tumor. Immunoglobulin kappa chain, IGHG1, IGHM, IGHM, MMP2, ORM1 & ORM2, podocan, asprorin, immunoglobulin superfamily containing leucine-rich repeat protein (ISLR) were detected in stroma. By contrast, squamous cell carcinoma antigen recognized by T-cells 3 (SART3), shootin-1, mitotic checkpoint protein BUB3, XRCC5 &XRCC6, membrane-associated progesterone receptor component 2 (PGRMC) and hepatocyte growth factor-regulated tyrosine kinase substrate (HGS) were specifically detected in tumor. Further analyses on an expanded cohort of patients will be presented. Conclusion MALDI mass spectrometry proteomics reveal distinctive tumor and microenvironment profiles in PAM50 intermediate risk early breast cancers. Citation Format: Hajjaji N, Aboulouard S, Robin YM, Bertin D, Fournier I, Bonneterre J, Salzet M. Proteomic profile of PAM50 intermediate risk early breast cancers [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P3-08-15.

Background Metastases from breast cancers have not yet reached curability mainly because of the continuous evolution of cancer cells leading to treatment resistance and tumor progression. Despite the identification of the main genomic drivers of breast cancer resistance, and the development of specific targeted therapies, resistance has not been eradicated. A deeper understanding of tumor biology underlying treatment escape and tumor progression is necessary to find new targets. The aim of this study was to determine whether mass spectrometry based imaging and microproteomics are able to capture changes in protein expression and pathways occurring during metastasis progression. Methods Ten tumor samples from five progressing metastatic breast cancer patients treated at Oscar Lambret Cancer Center (France) were selected based on the availability of paired metastasis biopsy performed at two time points during the evolution of the disease. The proteomic profiles of paired tumor samples were analyzed using MALDI mass spectrometry imaging combined with microproteomics, a spatially resolved proteomic technology. Results Comparison of paired samples showed gain and loss of proteins associated with tumor progression. However, few were shared between patients. The pathways and biological processes mainly represented during tumor progression, and which were found in at least two patients, were those involved in the negative regulation of leukocyte mediated cytotoxicity, the metabolism of carboxilic acid, peptides and nucleosides, in response to stress, oxidation-reduction process, endocytosis, catabolic processes, actin regulation, and extracellular matrix organization. During metastasis evolution, few shared proteins were identified in at least 3 patients such as SRPX2, CILP1, collagen alpha-2(V) chain, Ras-related C3 botulinum toxin substrate 1, filamin-C, PDZ and LIM domain protein 2 and chloride intracellular channel protein 4. The main proteins lost during progression and found in at least 2 patients were cell surface glycoprotein MUC18, collagen alpha-2(IV) chain, polyadenylate-binding protein 2, and latent-transforming growth factor beta-binding protein 4 (LTAGP4). Results from an expanded cohort will be presented at the meeting. Conclusion MALDI mass spectrometry proteomics identified private and shared changes in proteins and biological processes associated with breast cancer metastasis progression. Citation Format: Hajjaji N, Aboulouard S, Robin YM, Bertin D, Fournier I, Bonneterre J, Salzet M. Proteomic tracking of breast cancer metastasis progression [abstract]. In: Proceedings of the 2018 San Antonio Breast Cancer Symposium; 2018 Dec 4-8; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2019;79(4 Suppl):Abstract nr P3-08-19.

International audience; Background: Multigene signatures refine the risk of recurrence and guide adjuvant chemotherapy decision in luminal breast cancers. The decision to perform the assay is highly variable among oncologists. In order to guide the appropriate clinical group in whom to perform a genomic signature, our study analyzed in a homogeneous cohort which clinical risk groups triggered the use of the PAM50-based signature and their concordance with the genomic risk. Methods: A real life cohort of 222 early breast cancer patients with hormone receptor positive and HER2 negative disease had a commercial PAM50-based assay (Prosigna®) performed at our institution. The assay provided the risk group, the 10-year risk of distant recurrence and the intrinsic molecular subtype of breast cancer. Results: Based on nodal involvement, Ki67, tumor grade, mitotic index, and tumor size, no clinical pattern could identify a specific genomic risk group. The discordance with the genomic risk was high in patients with clinical low risk tumors, both in node negative and node positive patients. Up to 60% of them had a 10% or more risk of distant recurrence. Moreover, we identified a subgroup of luminal A tumors with a high genomic risk of recurrence. Genomic risk and intrinsic subtype were strong determinants of chemotherapy decision. Conclusions: Clinical profiles could not reliably identify genomic risk groups and guide the decision to use a multigene signature. Significant discordance with the genomic risk was observed within low clinical risk and luminal A tumors.

e13063 Background: Breast cancer mortality is expected to rise by almost 30% by 2030 worldwide, mainly due to the occurrence of distant metastases. The development of drugs specifically targeted at tumor drivers has not yet curbed resistance to treatment, which prevents metastases curability. There is a need for new molecular approaches to tackle metastases complex biology, particularly tumor heterogeneity, a main determinant of resistance. The aim of this study was to use a proteomic mass spectrometry-based approach to reveal functionally heterogeneous’ tumor subpopulations in breast cancer metastases, and identify clone specific drug targets. Methods: Metastasis biopsies (n = 21) were collected retrospectively from patients with advanced breast cancer treated at Oscar Lambret Cancer Center (Lille, France). Tumor heterogeneity was analyzed directly on FFPE tissue sections using MALDI mass spectrometry imaging (MSI) on a RapifleX Tissuetyper. Unsupervised spatial segmentation was performed to reveal tumor subpopulations with distinct proteomic profiles within each metastasis. The full proteomic characterization of these tumor clones was further performed with spatially resolved proteomic mass spectrometry. Results: MSI revealed that breast cancer metastases contained 2 to 5 functionally distinct tumor clones (proteomic clones). Although the clone profiles within a metastasis were correlated, unsupervised hierarchical clustering showed a clear distinction between them and specific proteomic signatures. Enrichment analysis showed that differentially expressed proteins were involved in a variety of biological processes or pathways including regulation of histone acetylation, extracellular matrix degradation, DNA repair, NOTCH pathway, estrogen-responsive target genes or exocytosis. The evolution of the proteomic clones profile during disease progression was also determined by comparison of paired biopsies. To identify the candidate treatments best fitted to metastasis heterogeneity, the specific proteomic signatures of the clones were matched against a druggable genome database. It was possible to unveil candidate drug targets personalized to each metastasis functional clone. Conclusions: MALDI mass spectrometry imaging combined with spatially resolved proteomics has the potential to tackle breast cancer metastases heterogeneity, and identify candidate drug targets specific to functional clones to personalize treatments.

3077 Background: Breast cancer is a heterogeneous disease with a wide range of outcomes. Among the intrinsic breast cancer subtypes, luminal A tumors are considered to have a favorable prognosis. However, molecular studies characterizing the genomic landscape of luminal A tumors revealed a molecular heterogeneity within this subtype, which also translated to variability in survival. A better understanding of the biology of this tumor subgroup is therefore needed to determine the appropriate therapeutic strategy. The aims of the study were to determine the frequency of high-risk luminal A tumors in a real life cohort of early breast cancers and provide a proteomic characterization of this subgroup using a mass spectrometry approach. Methods: 222 early breast cancer patients with hormone receptor positive and HER2 negative tumors treated at our institution had a PAM50-based genomic assay Prosigna to estimate their risk of recurrence. This assay assigned each tumor sample to an intrinsic molecular subtype of breast cancer. Luminal A and B tumors were analyzed with MALDI mass spectrometry imaging combined with microproteomics, a spatially-resolved on-tissue shotgun proteomic technology, to determine the proteomic profiles of both cancer cells and stroma. Results: Among the 129 luminal A breast cancers identified in our cohort, 67 (51%) had a risk of distant recurrence of 10% or more (32% had a 10% to 15% risk, and 19% a risk greater than 15%). High-risk luminal A tumors had a distinctive proteomic profile compared to low-risk luminal A or to luminal B tumors. Overexpression of the methionine biosynthesis pathway was the main differential protein expression observed in cancer cells and stroma of high-risk luminal A. Inflammation mediated by chemokine and cytokine signaling pathway and integrin signaling were also overexpressed in high risk luminal A compared to luminal B. In the stroma of luminal B tumors, EGR signaling, Ras and FGF pathways and angiogenesis were overexpressed compared to high-risk luminal A tumors. Conclusions: Real life data showed a significant proportion of high-risk luminal A breast cancers. MALDI mass spectrometry proteomics revealed distinctive tumor and microenvironment profiles in this breast cancer subgroup.

Despite great improvements, a significant proportion of cancer patients still die, mainly because of the development of metastases. At this stage, current treatments still rely heavily on conventional chemotherapy for most cancers. The efficacy of chemotherapy is dose-dependent, which is limited by toxicity to non-tumor tissues, as a result of its poor tumor selectivity. To improve survival length and preserve quality of life, the challenge is to develop approaches aimed at increasing chemotherapy toxicity to tumor tissue while not affecting non-tumor tissues. Marine-derived lipids, docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA), have the potential to differentially sensitize tissues to chemotherapy. These lipids enhance the cytotoxicity of 15 anticancer drugs (antimetabolites, alkylating or intercalating agents, microtubule stabilizers, Abl tyrosine kinase inhibitor and arsenic trioxide) to a variety of cancer cell lines or tumors in animals, used as models for breast, prostate, colonic, lung, cervical, ovarian cancers, neuroblastomas, leukemia or lymphomas. However, DHA and EPA do not sensitize non-tumor tissues to anticancer drugs, which suggests that the effect of these lipids is tumor selective. Two phase II clinical trials already support these results, and randomized, phase III trials are ongoing. In this review, we discuss the double-faceted properties of these lipids, and then focus on their potential for transfer to the patient in the light of current therapeutic strategies. Should their beneficial effects be confirmed, the consequences could be considerable by opening up the prospect of systematic supplementation during cancer treatment, a significant shift in current cancer therapeutic paradigms.

Body weight loss during the course of cancer disease has been associated with poor prognosis. Beside cancer-associated cachexia, weight loss can also result from chemotherapy. This work explored whether a model of mammary tumors in female Sprague Dawley rats could be appropriate to study the effect of doxorubicin on body weight, described weight change in this model, and assessed the effect of DHA on weight during chemotherapy. After tumor induction, rats were randomly assigned to a control or a DHA-enriched diet, and treated with doxorubicin or placebo twice a week for 2.5 wk (n = 6 in each group). Body weight, food intake, and tumor growth were monitored. Neither the induction of tumors nor their initial development impaired body weight gain. No reduction in food intake was observed. Tumor growth was similar between groups from day 1 to day 11. Although doxorubicin induced body weight loss from day 4 compared to placebo (P< 0.01) in rats fed the control diet, it did not induce body weight loss in rats fed the DHA-enriched diet (P = 0.02), indicating that DHA had a protective effect. These results indicate that doxorubicin can induce body weight loss in this model and that a DHA-enriched diet can prevent this effect.

Developing approaches that will increase the selectivity of anticancer drugs remains a challenge. Docosahexaenoic acid (DHA) has the potential to increase tumor sensitivity to chemotherapy with no sensitization of normal tissues. This study was aimed at exploring the mechanism involved in this differential sensitization with a focus on oxidative stress, one of the main determinants involved in DHA enhancement of anthracycline-based chemotherapy.In a previously validated model of chemically induced mammary tumors in rats where supplemental DHA sensitized tumors to epirubicin, DHA level, lipid hydroperoxides (LPO), total antioxidant activity, glutathione peroxidase (GPx), and superoxide dismutase (SOD) activities were measured in tumors, intestine, liver, and heart in rats treated with DHA+epirubicin and in control rats (palm oil, no chemotherapy).At baseline, the level of LPO was similar in tumors, liver, heart, and small intestine. The proportion of DHA was higher in non-tumor tissues compared to tumors (liver or heart, p0.0001; intestine, p = 0.01). DHA and epirubicin induced a significant increase in LPO in tumors (p = 0.03), but no increase was detected in liver, heart, or intestine. The difference in LPO production between these tissues was associated with differences in their antioxidant defenses. In tumors, the adjustment of GPx activity was possible but limited, and no adjustment of their total antioxidant and SOD activities was observed compared to other tissues.Supplementing DHA during an anthracycline-based chemotherapy induced a selective increase of LPO in tumors. This selectivity might result from a differential handling of oxidative stress between tumor and non-tumor tissues.

BACKGROUND: Breast cancer becomes lethal when visceral metastases develop. At this stage, anti-cancer treatments aim at relieving symptoms and delaying death without resulting in additional toxicity. On the basis of their differential anti-oxidant defence level, tumour cells can be made more sensitive to chemotherapy than non-tumour cells when membrane lipids are enriched with docosahexaenoic acid (DHA), a peroxidisable and oxidative-stress-inducing lipid of marine origin. METHODS: This open-label single-arm phase II study evaluated the safety and efficacy (response rate), as primary end points, of the addition of 1.8g DHA daily to an anthracycline-based chemotherapy (FEC) regimen in breast cancer patients (n ¼25) with rapidly progressing visceral metastases. The secondary end points were time to progression (TTP) and overall survival (OS). RESULTS: The objective response rate was 44%. With a mean follow-up time of 31 months (range 2–96 months), the median TTP was 6 months. Median OS was 22 months and reached 34 months in the sub-population of patients (n ¼12) with the highest plasma DHA incorporation. The most common grade 3 or 4 toxicity was neutropaenia (80%). CONCLUSION: DHA during chemotherapy was devoid of adverse side effects and can improve the outcome of chemotherapy when highly incorporated. DHA has a potential to specifically chemosensitise tumours.

Summary Objective To describe the main features of bone infection associated with Cat-scratch disease (CSD). Methods We searched for articles indexed in the international literature databases by using the following key words: " Bartonella ", "bone", "cat-scratch", "osteomyelitis" and "osteolytic". Results Cases of 47 patients were reviewed. The median age was 9years, with an equal sex distribution. Bone pain and fever were the main symptoms. The presence of fever and increased age were more common in patients with bone infection than classically reported in uncomplicated (i.e. nodal) CSD. The vertebral column and pelvic girdle were the most common sites of infection. Radiological examination typically confirmed bone osteolysis. All patients recovered without complications or chronic infection, although they received a various combination antibiotic regimen and duration therapy. The mechanism by which infection might spread to the bone is via the haematogenous route, accounting for most of the disseminated cases and via the lymphatic route, for those with regional limited extension. Conclusions Bone infection is rare but should be considered when bone pain and fever are present in a patient with nodal CSD. The prognosis is good, whatever treatment is given. Thus bone biopsy should be recommended only in a difficult diagnostic setting, when other bacteria or malignant disease are suspected.

Background Ever since the discovery of the PI3K/Akt/mTOR pathway key role in breast cancer resistance, several molecules have been developed to target this pathway. However, mechanisms of resistance restrain their efficacy. An in-depth understanding of the interactions of PI3K pathway is essential to improve its blockade and avoid resistance. Preclinical studies reported that lipid composition of cancer cells altered the activation of PI3K pathway, thus suggesting another possible molecular interaction with lipid metabolism. The aim of this study was to analyze the relationship between the activation of PI3K pathway and tumor lipid composition in breast cancer patients. Methods Frozen and formalin-fixed paraffin-embedded tumor specimens were retrieved for 50 French patients with stage I to III primary breast cancer treated uniformly with surgery, adjuvant chemotherapy, radiation therapy and endocrine therapy. The tumor specimens were analyzed for PIK3CA gene mutation, PTEN protein expression by immunohistochemistry, and pAktS473, pAktT308, mTOR, pS6K, and S6RP by protein array. Tumor total fatty acid (FA) and phospholipid FA compositions were analysed using mass spectroscopy. Pearson test was used for correlation analyses. Results Mean age was 53 years. All tumors were ER or PR positive and HER-2 negative, and 76% were node positive. PI3K mutation was observed in 14% of cases, and 34% were PTEN negative or had low PTEN expression. Mean tumor satured fatty acid (SFA) level was 33.78 mol%, monounsaturated FA (MUFA) level was 46.46 mol%, n-6 polyunsaturated FA (PUFA) level was 18.08 mol%, and n-3 PUFA level was 1.68 mol%. Large and high grade tumors had a higher SFA content. MUFA rich tumors were frequently low grade. There was a positive correlation between d18:1/C20:0 ceramide level and Akt, mTOR and p70S6K expression. mTOR expression also correlated with 18:0 lysophosphatidylserine (lysoPS), 36:1 PS, 38:3 PS, 38:5 PS, 40:5 PS, 40:6 PS, 34:0 phosphatidylcholine (PC), 34:1 PC, and 38:3 PC. Conclusions Our results show that SFA associated with a more aggressive tumor phenotype and that PI3K pathway activation significantly correlated with fatty acid composition of specific tumor phospholipids. These results suggest a potential role of tumor lipid metabolism in PI3K pathway activation. Citation Format: Nawale Hajjaji, Flavie Arbion, Alain Fautrel, Claire Villalva, Lucie Karayan-Tapon, Marie-Lise Jourdan. Activation of PI3K pathway in breast cancer associates with tumor phospholipid fatty acid composition. [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 1051.

International audience; Docosahexaenoic acid (DHA), upon incorporation into tumor tissue, has the potential to sensitize tumors to the effects of chemotherapy or radiation therapy. Although DHA has usually been supplied to tumor tissue in the diet, appropriate dietary conditions required to obtain optimal tumor levels have not been established. Hence, we studied mammary tumor tissue responses in rats fed various durations and doses of DHA. Rats fed a palm oil enriched diet (diet 0) were switched to diets providing either 0.8 g DHA/day (diet 1) or 1.5 g DHA/day (diet 2). Tumor tissue fatty acid composition was analysed at baseline (diet 0), at weeks 1, 4 and 9 during diet 1 and at week 4 during diet 2. Dietary DHA supplementation differentially increased DHA within phospholipids (PL) and triacylglycerol (TAG) fractions in tumors. DHA level equilibrated between 2 and 4 weeks in PL while DHA increase was more progressive in TAG and did not reach a steady state. A higher dose of DHA further increased DHA content in tumor PL and TAG (P = 0.018 and P < 0.001, respectively). DHA concentration in plasma PL was positively correlated with DHA in tumor PL (r = 0.72; P = 0.0003) and TAG (r = 0.64; P = 0.003). We conclude that dietary DHA supplementation enhances tumor content of DHA in a time- and dose-dependent manner, and that the DHA level in plasma PL could be used as a proxy for tumor DHA. These findings have implications for dietary DHA supplementations in cancer patients.

Lifestyle and nutritional factors have been recognized to influence breast cancer survival, irrespective of genomic alterations that are the hallmarks of the disease. The biological and molecular mechanisms involved in the effects of dietary polyunsaturated fatty acids and breast cancer response to treatments in clinical and preclinical studies have been reviewed. Among nutrients, rumenic acid, a naturally occurring CLA isomer and n-3 docosahexaenoic acid (DHA) a highly unsaturated fatty acid, have emerged due to their potential to increase cancer treatment efficacy without additional side effects. In this review, we analyze the literature evidence that breast cancer treatment and outcome could be improved through an adjuvant dietary supplementation. Such an original approach would involve two successive phases of breast cancer treatment: an initial sensitization of residual tumor cells to chemotherapy and to radiation therapy with dietary DHA; then a prevention of metastatic re-growth with a prolonged rumenic acid supplementation. Safety is not anticipated to be a critical issue, although it has to be assessed in the long term. Dietary supplements, used in combination to anti-cancer agents, should be provided under medical prescription. Such an original use of fatty acids in breast cancer treatment could provide the lipid field with a new avenue to impact public health.

The potential for dietary fat to prevent breast cancer makes identification of defined molecules a mandatory step. In order to circumvent the limitations and/or bias of dietary exposure assessment tools, we have used the fatty acid composition of white adipose tissue as biomarker of past lipid intake. When considered separately, candidate fatty acids identified as favourable on the basis of their association with breast cancer risk have usually led to inconsistent results in dietary intervention studies carried out in rats. This inconsistency indicates that any approach based on a single fatty acid should be abandoned for an integrated view over the complex lipid interactions, which finally determines the lipidome, the lipid profile that is found in individuals. We reappraised the role of the complete lipid profile through a comprehensive study of adipose tissue fatty acids obtained in patients with benign or malignant breast tumors. Rather than a single fatty acid, a composite indicator combining elevated monounsaturates and low n-6/n-3 fatty acid ratio was associated with decreased breast cancer risk. The lipidome may provide the opportunity to quantify the modifiable part of the risk of breast cancer. The lipidome may be used as a template for designing proper dietary modifications in order to delay the occurrence of breast cancer. Which dietary modifications should be undertaken in order to bring a pertinent change to the lipidome with respect to the risk of breast cancer is currently unknown. The lipidome may allow the individualization of a high risk population of women, who may be targeted for a dietary prevention of breast cancer. The setting and validation of a high-throughput lipidomic station with analytical capabilities fitted to the need of mass screening is required. These two locks must be resolved before a primary prevention of breast cancer by diet could be contemplated.

e15598 Background: The mammalian target of rapamycin (mTOR) signalling pathway regulates major cellular fonctions including metabolism. Everolimus, a mTor inhibitor, is an effective treatment for m...

Purpose Weight change (gain or loss) in early breast cancer patients is associated with worse survival. However, limited research investigates the impact of weight change in the neoadjuvant setting in women with HER2 overexpressing breast cancer. The goals of this study were to examine i) the relationship between weight change during trastuzumab-based neoadjuvant chemotherapy and pathologic response, and ii) determine whether weight change during post-operative trastuzumab was associated with outcome. Patients and methods From March 2004 to March 2013, 52 patients diagnosed with invasive non-metastatic HER2 overexpressing breast cancer were recruited from H.S. Kaplan Cancer center, France. They received trastuzumab-based neoadjuvant chemotherapy (NAC) before surgery, followed by adjuvant trastuzumab. Patients were distributed in the “weight loss” or “weight gain/stable” categories according to their relative weight change after NAC, and after adjuvant trastuzumab. Logistic regression was used to examine associations between weight change and pathologic response. The Kaplan-Meier method and Cox proportional hazards regression analysis were used to evaluate the relationship between weight change and progression free survival (PFS). Results Median age was 50 years (range 21 to 71 years); 25% of women experienced body weight loss during NAC, 40% had stable body weight, and 35% gained weight. In multivariate analysis, weight loss was significantly associated with worse pathologic complete response (pCR) in tumor compared with the weight gain/stable group (OR = 0.18; 95% CI, 0.04 to 0.84). There was no difference in pCR in nodes according to weight change (OR = 0.34; 95% CI, 0.06 to 1.76). Body mass index was not associated with pCR. At a median follow up time of 3 years, PFS was not associated with weight change during NAC but with weight change during adjuvant trastuzumab. PFS was worse in patients loosing weight during adjuvant trastuzumab (OR = 0.05; 95% CI, 0.01 to 0.34). Conclusion Weight loss was associated with worse pCR to trastuzumab-based neoadjuvant chemotherapy. In addition, weight loss during adjuvant trastuzumab was associated with worse PFS. These results suggest that weight control and promotion of weight maintenance in women with HER2 overexpressing breast cancer should be integrated within cancer treatment to optimize patient outcomes. Citation Format: Nawale Hajjaji, Dorothee Chocteau, Sylvie Gerard, Marie Brigitte Orgerie. Relationship between weight change and response to trastuzumab-based neoadjuvant chemotherapy among women with operable HER2 overexpressing 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 3842. doi:10.1158/1538-7445.AM2014-3842