Search

Introduction: Germline pathogenic variants in genes in the homologous recombination (HR) pathway such as BRCA1, BRCA2, ATM, PALB2, RAD51C and RAD51D confer an increased risk of breast and ovarian cancers. Screening for germline variants in these cancer susceptibility genes is critical as prophylactic measures and monitoring can be performed in these individuals to minimize their risk of developing breast and/or ovarian cancers. More recently, it has been recognized that cancers arising in carriers of germline pathogenic (P)/ likely pathogenic (LP) variants in HR genes often show a homologous recombination (HR) deficient (HRD) phenotype and can be targeted with platinum agents or PARP inhibitors. RAD51B is a RAD51 paralog that binds to RAD51C and functions as a heterodimer in the HR pathway. Whilst RAD51B germline variants have been reported in isolated cases of breast and ovarian cancers, it is unclear as to whether RAD51B P/LP germline variants would confer predisposition to these cancer types. Materials and Methods: We screened 9,287 consecutive unselected cancer patients who consented for both tumor and germline testing using the MSK-IMPACT platform for the presence of RAD51B germline truncating variants. Selected RAD51B germline mutant breast cancers identified by MSK-IMPACT were subjected to whole-exome sequencing (WES) analysis to determine the dominant mutational signatures using DeconstructSigs. Functional assays were performed to ascertain the impact of RAD51B loss on HR DNA repair and PARP inhibitor sensitivity using genome editing methods in non-malignant breast epithelial cell lines. Results: Out of the 9,287 cancer patients, we detected likely pathogenic loss of function RAD51B germline variants in 11 patients (0.12%), which was similar to the frequency detected in the gnomAD database (0.09%). All female carriers of RAD51B loss of function variants (n=8) had breast or ovarian cancers (8/1,619 breast or ovarian cancers, 0.5%). The observed carrier frequencies of germline truncating variants in RAD51B was statistically enriched in breast and ovarian cancer patients compared to individuals in the gnomAD database (0.5% vs 0.09% P=0.0003; odds ratio = 5.06 (95% CI: 2.1-10.3)). Although segregation studies were not available, 9/11 of the RAD51B germline loss of function variant carriers had a personal or family history of breast or ovarian cancers. All five breast and ovarian cancers from RAD51B mutation carriers investigated by WES were found to harbor RAD51B bi-allelic inactivation through loss of heterozygosity of the RAD51B wild-type allele. In addition, these five cases were found to display genomic features of HRD including high large-scale state transition scores and a dominant mutational signature 3. CRISPR/Cas9 genome editing of RAD51B in a non-malignant breast epithelial cell model revealed that RAD51B deficient cells display PARP inhibitor sensitivity similar to that reported in BRCA1 and BRCA2 deficient cell lines. Conclusion: RAD51B loss-of-function germline variants confer susceptibility to breast and ovarian cancer development. Breast and ovarian cancers occurring in the context of RAD51B germline mutations harbor bi-allelic inactivation of RAD51B through loss-of-heterozygosity of the wild-type allele, genomic features of HRD and are likely sensitive to PARP inhibition. Albeit rarely germline mutated, RAD51B should be considered as an addition to clinical germline testing panels for hereditary breast and ovarian cancer syndrome patients. Citation Format: Diana SReis-FilhoReis-Filho Mandelker, Semanti Mukherjee, Jeremy Setton, Pier Selenica, Yelena Kemel, Ozge Ceyhan-Birsoy, Margaret Sheehan, Kaitlyn Tkachuk, David N Brown, Simon Powell, Britta Weigelt, Mark E Robson, Nadeem Riaz, Kenneth Offit, Jorge S Reis-Filho. RAD51B loss-of-function variants confer susceptibility to hereditary breast and ovarian cancers and result in tumors with genomic features of homologous recombination repair defects [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 P6-09-01.

Background: Male breast cancer is rare, with approximately 2,600 cases diagnosed annually in the United States. Data are scant regarding the genomics and pathophysiology of male breast cancer, especially in the metastatic setting, requiring most treatment recommendations in male breast cancer to be made by inference from breast cancer in women. Methods: We performed prospective genomic profiling of primary and metastatic tumor samples from men with breast cancer treated at Memorial Sloan Kettering Cancer Center using the MSK-IMPACT targeted-DNA-sequencing panel for somatic mutations. Comprehensive demographic, clinical, and pathologic data were collected on all included patients. Statistics are descriptive. Results: Genomic sequencing was performed on 45 samples from 41 men (31 primary samples and 14 from metastatic sites). Median age at time of sample collection was 61 years, with a range of 27-92 years. Thirty-seven (90.2%) men had ER+/HER2- breast cancer, 3 (7.3%) had ER+/HER2+ breast cancer and 1 (2.4%) had triple negative disease. Thirty-nine (95.1%) had ductal carcinoma, and no cases of lobular carcinoma were identified. Forty patients underwent germline testing, and 12 (30%) were found to have pathogenic germline mutations (6 BRCA2 mutations, 2 BRCA1 mutations [one of whom had a concurrent CHEK2 mutation], and one mutation each in PALB2, MUTYH, and MSH6). Overall, the pattern of genomic alterations in male breast cancer was similar that in women. Twelve (29.3%) patients had PIK3CA mutations, 9 (22%) had GATA3 mutations, 3 (7.3%) had TP53 mutations, 3 (7.3%) had ARID1A mutations, 3 (7.3%) had KMT2C mutations, 2 (4.9%) had FOX1A mutations, 2 (4.9%) had RB1 mutations, and 2 (4.9%) had TERT promoter hotspot mutations. Eleven (26.8%) patients had CCND1 amplification, 8 (19.5%) had MYC amplification, 6 (14.6%) had FGFR1 amplification, and 5 (12.2%) had MDM2 amplification. All other findings were present in ≤ 1 patient. All included patients had normal mutational burden, and all samples were microsatellite stable. PIK3CA mutations occurred in 33% of primary samples vs. 15% of metastatic samples, CCND1 amplification occurred in 23% of primary samples vs. 38% of metastatic samples, and TERT hotspot promoter mutations were found only in metastatic samples. Of note, we observed a single ESR1 D538G mutation in the metastatic sample of a patient with significant prior exposure to aromatase inhibitors in the adjuvant and metastatic settings. We further found concurrent ERBB2 mutation and amplification in the post-treatment metastatic samples of an ER+/HER2- patient, who was treated with neratinib for 14 weeks with clinical response. Lastly, we report a heavily pretreated patient with metastatic secretory breast carcinoma who was found to have an ETV6-NTRK3 fusion gene. This patient was treated with a first-generation TRK inhibitor and continues to exhibit an ongoing clinical response at 8.6 months. Conclusions: Based on our data, the overall genomic landscape of male breast cancer appears comparable to that of breast cancer in women, as has been previously reported. However, despite the small number of metastatic cases examined, several previously unreported and treatment-informing signatures were discovered, especially in those patients with less common male breast cancer variants. Further study is warranted to confirm these findings in a larger cohort. Citation Format: Joshua Z Drago, Cristian Serna-Tamayo, Carlos H Dos Anjos, David N Brown, Shanu Modi, Komal Jhaveri, David B Solit, Tiffany A Traina, Sarat Chandarlapaty, Jorge S Reis-Filho, Mark E Robson, Ayca Gucalp, Pedram Razavi. Genomic profiling of primary and metastatic breast cancer in men [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 P4-17-01.

Background: Despite therapeutic advancements, a substantial number of patients with treated early-stage invasive lobular carcinoma (ILC) of the breast will still relapse as late as 20 years after the initial tumor diagnosis, thus making follow-up a challenging task. Here, using patient-matched primary and metastatic tumors from patients with ILC, we aimed to decipher the evolution of lobular breast cancer through time and explore metastatic dissemination events and underlying heterogeneity. Materials and Methods: We retrospectively identified 38 patients with metastatic ILC, of which 6 with de novo metastatic disease. In total, 99 formalin-fixed paraffin embedded tissue samples of primary tumor and metastasis, together with well-annotated clinicopathological data, were available. Low coverage whole-genome sequencing was performed to infer somatic copy number aberrations (CNA). Phylogenetic reconstruction was performed with Integer Linear Program for the Copy-Number Tree Problem algorithm. Results: Reconstruction of the phylogenetic trees showed different patterns of dissemination. For patients with multiple metastatic samples sequenced (4/4 patients), the metastases were found to be seeded from a common metastatic precursor. For a subset of patients with multiple primaries (5/10 patients), distant metastases were seeded from specific clones inside the primary tumor. Detailed analysis of known driver genes showed that alterations in common driver genes such as TP53, MYC and CCND1 were shared between primary tumor and metastases in a large proportion of the patients (47%, 32% and 21% of patients respectively) whilst CNA in other driver genes including FGFR2, PTEN or AKT2 among others, were frequently private in the metastases. Interestingly, we observed that in approximately 30% of the patients, the metastases harbored less CNA than the matched primary tumor, resulting in shorter metastatic branches and possibly indicating a decelerated tumor progression. Further analysis of the genomic and clinicopathological characteristics showed that these patients were mostly de novo metastatic (p-value=0.002), metastasizing primarily to the liver or bone and had significantly more aberrations in NF1, AKT1 and FGFR (p-value=0.04, 0.01 and 0.03 respectively), when compared to the rest of the cohort. Finally, most of the patients (63%) had a lower number of CNA accumulated after the metastatic lineage separation compared to the common trunk, implying a late dissemination event. Conclusions: In this study, we described different metastatic dissemination events taking place in ILC. Our analyses allowed us to identify a subset of tumors characterized by slow tumor progression, indicating potential tumor dormancy, as well as evidence of late dissemination evidence for most of the patients, thus highlighting the importance of early detection through cancer screening. These findings further shed light on how ILCs evolves through time and could potentially influence the clinical management of invasive lobular breast cancer. Citation Format: Danai Fimereli, David Venet, Mattia Rediti, David N Brown, Bram Boeckx, Marion Maetens, Samira Majjaj, Ghizlane Rouas, Maria Capra, Giuseppina Bonizzi, Federica Contaldo, Christine Galant, Martine Piccart, Giancarlo Pruneri, Denis Larsimont, Diether Lambrechts, Christine Desmedt, Françoise Rothé, Christos Sotiriou. Portraying tumor evolution of lobular breast cancer through phylogenetic analysis [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 PD14-05.

Introduction: Invasive lobular carcinoma (ILC) is the most frequent special histologic subtype of breast cancer (BC). ILC is identifiable by pathologic assessment given its distinctive discohesive growth pattern, largely caused by CDH1 inactivation. Compared to common forms of BC, ILCs display lower responses to chemotherapy and selective estrogen receptor modulators. The low interobserver agreement for the diagnosis of ILC, however, renders the inclusion of histologic subtyping in therapeutic decision-making challenging. Artificial intelligence (AI)-based algorithms hold promise for improving pathologic diagnosis; their performance, however, depends on the ground truth labeling used. Here, we seek to develop an AI-based methodology for detection of ILC using ‘CDH1 biallelic mutations’ (i.e., mutation + loss-of-heterozygosity of the wild-type allele or two pathogenic somatic mutations) as ground truth, reasoning that in BC, >95% of CDH1 bi-allelic inactivation is found in ILCs.Materials and methods: We developed a convolutional neural network system to detect CDH1 biallelic genetic inactivation (AI-CDH1) using whole slide images (WSI) of 1,100 primary BCs with available targeted sequencing data. The model was trained using a 10-fold cross-validation method to detect biallelic mutations. The mean number of positive and negative samples in the training set was 85.2 (SD=2.57) and 562.8 (SD=10.51) per fold, respectively. The evaluation set consisted of a mean of 14.2 (SD=2.04) positive and 93.8 (SD=9.13) negative samples. We evaluated the performance of the AI-CDH1 classifier to predict the lobular phenotype and CDH1 status using original and revised labels, following a histopathologic re-review of the histologic type and CDH1 status curation. The latter was conducted by incorporating information on biallelic CDH1 inactivation beyond CDH1 mutations (homozygous deletions, deleterious structural rearrangements, and loss-of-heterozygosity and gene promoter methylation).Results: The AI-CDH1 classifier predicted biallelic CDH1 mutations with an area under the curve (AUC)=0.944 (95 CI: 0.925-0.963), sensitivity=91.6% and specificity=85.9%, PPV=49.8%, NPV=98.5% and accuracy=86.7%, and the original ‘lobular phenotype’ with an AUC=0.941 (95 CI: 0.922-0.960), sensitivity=89%, specificity=86.7%, PPV=55.6%, NPV=97.7% and accuracy=87.1%. Review of the CDH1 gene status revealed that 7/957 BCs lacking CDH1 biallelic mutations harbored biallelic CDH1 inactivation by promoter methylation, homozygous deletions or structural rearrangements. The AI-CDH1 classifier detected all seven reclassified BCs and predicted the revised CDH1 biallelic inactivation with an AUC=0.948 (95 CI: 0.930-0.966), sensitivity=92%, specificity=86.5%, PPV=52.3%, NPV=98.5% and accuracy=87.2%. Upon histologic re-review, which resulted in reclassification of 36/927 non-lobular BCs as ‘lobular’ and 5/173 ‘lobular’ BCs as ‘non-lobular’, the AI-CDH1 classifier detected the ‘lobular phenotype’ with an AUC=0.953 (95 CI: 0.935-0.971), sensitivity=90.7%, specificity=89.7%, PPV=66.8%, NPV=97.7% and accuracy=89.9%. Using the revised histologic re-classification and CDH1 biallelic inactivation status labels, the AI-CDH1 classifier predicted the lobular phenotype irrespective of CDH1 status (P>0.05).Conclusions: By training a machine learning system to detect ‘CDH1 biallelic mutations’, as ground truth rather than histologic diagnosis of lobular carcinoma, which might be confounded by human subjectivity, we developed an AI-based system that can detect ILCs accurately, providing a new paradigm for the development of AI-based cancer classification systems. Citation Format: Jorge S Reis-Filho, Fresia Pareja, Fatemeh Derakhshan, David N Brown, Jillian Sue, Pier Selenica, Yi Kan Wang, Arnaud Da Cruz Paula, Monami Banerjee, Zahra Ebrahimzadeh, Manuel Isava, Matthew Lee, Ran Godrich, Adam Casson, Ruben Padron, George Shaikovski, Alexander van Eck, Antonio Marra, Higinio Dopeso, Hannah Y Wen, Edi Brogi, Matthew G Hanna, Chris Kanan, Jeremy D Kunz, Felipe C Geyer, Carla Leibowitz, David Klimstra, Leo Grady, Thomas J Fuchs. An artificial intelligence-based predictor of CDH1 biallelic mutations and invasive lobular carcinoma [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 PD11-01.

Background In the NeoALTTO trial, the combination of trastuzumab and lapatinib for the neoadjuvant treatment of HER2 positive early breast cancer patients (pts) nearly doubled the rate of pathological complete response (pCR) compared with either anti-HER2 treatment alone. In the same trial, pCR was shown to be a surrogate for long-term outcome. Expression of ERBB2, ESR1 and immune signatures were the main drivers of pCR (Fumagalli et al. JAMA Oncol 2016). The main aim of the current study was to investigate the relevance of copy number aberrations (CNAs) for pCR and event-free survival (EFS). Methods For 271 out of the 455 pts enrolled in the study with sufficient tumour content, DNA was hybridized on CytoScan HD arrays. The log2 ratio intensities and B allele frequencies were segmented jointly using the multitrack segmentation algorithm. Integer level estimates of total copy number and major allele were obtained using GAP. Recurrent CNAs were found with GISTIC2. The genome instability index (GII) was defined as the median absolute deviation of the normalized copy number. Tumor infiltrating lymphocytes (TILs) and gene expression were obtained as described previously. The correlations between differents parameters were assessed using Spearman correlations (ρ). The Mann-Whitney test was used to relate binary and numerical features. EFS analysis was performed using the Cox proportional hazard model. Results CNAs estimates were obtained for 185 out of 271 pts. CNA distribution was similar to the ones observed in HER2+ pts from the METABRIC and the TCGA datasets. There were 64% of diploid pts, 26% of triploid and 9% of tetraploid or more. Aneuploidy level was not associated with pCR or EFS. Of interest, there were many significant differences in CNA profiles between ER+ and ER- pts. Those differences mirror those observed between ER+ and ER- among HER2- pts in TCGA (ρ=63%) and METABRIC (ρ=56%). ERBB2 amplification was predictive of high pCR (p=0.0007), albeit less so than ERBB2 expression, and ceased to be significant correcting for ERBB2 expression. The pCR rate increased with the GII (p=0.03), independently of ERBB2 amplification. The effect was stronger in ER+ patients (p=0.01). GISTIC analysis revealed 159 recurrent CNA regions. Amplification of 2 regions on 6q23-24 was significantly associated with higher pCR (p=0.00005 and p=0.00087, FDR=0.006 and 0.05). The most significant segment of 6q23-24 contained 39 genes, some whose expression level also predicted pCR (e.g. MAP3K5, p=10-4). Gene ontology analysis of the genes correlated with this segment highlighted the category 'response to interferon-alpha' (p=4.3x10-7). No amplified region or gene was found to be predictive of EFS, after multiple testing correction. Conclusions The amplification of ERBB2 was shown to be predictive of pCR, however its expression was more predictive. High genomic instability was associated with higher rate of pCR in ER+ subgroup. Of interest, a novel amplified region, involving chromosome 6 was shown to be predictive of pCR, which may warrant further investigation. Citation Format: Sotiriou C, Rothé F, Maetens M, Fumagalli D, Brown DN, Salgado R, Bradbury I, Pusztai L, Harbeck N, Ignatiadis M, Sarp S, de la Pena L, de Azambuja E, Huober J, Nuciforo P, Baselga J, Piccart M, Loi S, Venet D. Copy number aberration analysis to predict response to neoadjuvant anti-HER2 therapy: Results from the NeoALTTO phase III trial [abstract]. In: Proceedings of the 2017 San Antonio Breast Cancer Symposium; 2017 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2018;78(4 Suppl):Abstract nr GS1-04.

Background Epidemiological evidence indicates a clear relationship between pregnancy and breast cancer (BC) risk. However, little is known regarding the impact of pregnancy on BC biology. DNA copy number aberrations (CNAs) play an important role in breast carcinogenesis. BC during pregnancy is a rare disease but yet could serve as a good model to study the impact of pregnancy on BC biology. Methods We retrospectively included 54 pregnant and 113 non-pregnant BC patients matched for age and stage with complete clinico-pathological, gene expression and 5-year follow-up data. CNAs were assessed using Affymetrix OncoScan FFPE arrays. We identified the CNAs associated with pregnancy using a multivariate logistic regression adjusted for classical clinico-pathological features. We further evaluated their impact on gene expression. Results After quality control, CNA profiles were obtained for 38 pregnant and 87 non-pregnant BC patients. We identified 13 regions with copy number gains, 11 of which were more frequently gained in pregnant compared to non-pregnant controls and 5 regions with copy number loss, 3 of which were more frequently lost in pregnant patients (p≤0.05). Of interest, we identified 4 genes previously identified as driver event associated with CNAs in breast cancer (S. Nik-Zainal et al, Nature 2016). AKT1 and CDKN2A/B were more frequently gained in the pregnant compared to the non-pregnant (23.7% vs. 8.0%, p=0.068 and 18.4% vs. 4.6% p=0.036) and ARID1B was less frequently gained in the pregnant cohort (2.6% vs. 13.8%, p=0.02). Interestingly, PAPPA which had been previously identified as a pregnancy-dependent oncogene (Takabatake Y. et al, EMBO Mol Med. 2016) was also more frequently gained in the pregnant compared to the non-pregnant patients (21.1% vs 5.8%, p=0.03). We next evaluated the effect of these CNAs on their own gene expression levels and found that AKT1 and CDKN2A/B CNAs were affected by gene-dosage effect. Conclusions In this study, we were able to identify several genomic alterations associated with pregnancy that could further elucidate the impact of pregnancy on BC risk. Moreover, by combining CNAs with gene expression, we were able to identify genes whose expression were associated with CNAs and therefore could be considered potential drivers of this rare disease. Citation Format: Nguyen B, Brown DN, Rothé F, Desmedt C, Majjaj S, Pruneri G, Peccatori F, Azim Jr HA, Sotiriou C. Interrogating the impact of pregnancy on breast cancer biology using DNA copy number profiling [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P1-05-17.

This abstract was withdrawn by the authors.

Introduction: There is evidence to suggest that the biology of breast cancer (BC) might vary according to age, as BCs in elderly patients might be associated with accumulation of DNA damage and BCs in younger individuals might be enriched for genetic alterations affecting cancer predisposition genes. Whether the genetic landscape of BC in elderly individuals differs is yet to be determined. Here, we sought to describe the spectrum of somatic genetic alterations and mutational signatures in BC according to age. Materials and Methods: We conducted a re-analysis of two cohorts of BC: i) MSK-IMPACT targeted sequencing cohort (primary, n=918; metastatic, n=1,000) and ii) whole-exome sequencing data of primary BCs from The Cancer Genome Atlas (TCGA) BC study (n=1087). BCs of younger ( Results: Primary BCs in elderly and younger patients, matched by ER/HER2 status, from the MSK-IMPACT (n=220 and n=440) and TCGA (n=290, each) cohorts, and metastatic BCs from elderly (n=98) and younger (n=196) patients from the MSK-IMPACT cohort, matched by ER/HER2 status, were compared. In both cohorts of primary BCs, PIK3CA, TP53 and CDH1 were the most frequently mutated genes. Compared to younger patients, primary BCs in elderly patients from the MSK-IMPACT cohort displayed a lower frequency of mutations in TP53 (23% vs 34%, P Conclusion: Our findings show that BCs in elderly patients harbor a lower frequency of somatic mutations in TP53, AKT1 and MAP2K4 than younger individuals. In contrast, BCs in elderly patients more frequently harbor somatic mutations in the transcription factors RUNX1 and NCOR1. Bi-allelic alterations affecting HRD-related genes was numerically more frequent in younger than in older patients. As expected, BCs in elderly patients show a numerically higher incidence of the aging signature, whereas primary BCs affecting younger patients displayed a higher frequency of the APOBEC signature than primary BCs from elderly individuals. Citation Format: Pier Selenica, Fresia Pareja, Audree Tadros, Lorenzo Ferrando, David N Brown, Hong Zhang, Pedram Razavi, Diana Mandelker, Mark E Robson, Sarat Chandarlapaty, Britta Weigelt, Jorge S Reis-Filho. Genomic landscape of breast cancer occurring in elderly individuals [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 P4-05-08.

Introduction: Mucinous carcinoma of the breast (MCB) is a rare histologic form of estrogen receptor (ER)-positive/HER2-negative breast cancer characterized by tumor cells floating in lakes of mucin. As compared to invasive ductal carcinoma of no special type (IDC-NST), we have recently shown that MCBs harbor a lower frequency of PIK3CA mutations and lack concurrent 1q gains and 16q losses, hallmark genetic alterations of ER-positive breast cancer. Despite their distinctive phenotype and previous efforts to characterize their genomic landscape by whole-genome, whole-exome and RNA-sequencing, no pathognomonic somatic mutation or fusion gene underpinning MCBs or the mucinous phenotype have been identified. In this study we sought to determine whether MCBs would be defined by specific epigenetic changes. Materials and methods: Thirty-six MCBs were subjected to DNA methylation profiling using Infinium MethylationEPIC arrays. Following quantile normalization, low quality probe filtering, and data background and dye bias correction using the ‘noob’ algorithm, methylation profiling data of MCBs were compared to those of ER-positive/HER2-negative IDC-NSTs from The Cancer Genome Atlas (TCGA) lacking concurrent 1q gains/16q losses or PIK3CA hotspot mutations and matched according to age and menopausal status at a 1:2 ratio (n=72). An enrichment analysis of the differentially methylated targets in gene promoters and enhancers was conducted using Minfi and MethylGSEA R packages. A subset of 12 MCBs of this cohort was subjected to RNA-sequencing and compared to age, menopausal status and molecular features-matched ER-positive/HER2-negative IDC-NSTs from TCGA (1:2 ratio; n=24) using gene set enrichment analysis (GSEA). Results: Enrichment analysis of differentially methylated probes revealed a significant enrichment of targets in promoters and enhancers of mucin-encoding genes and in genes of the mTOR signaling pathway between MCBs and matched IDC-NSTs from TCGA (P Conclusions: Taken together, our data suggest that MCBs display a high expression of mucin-encoding genes, due to hypomethylation of promoters and enhancers and could provide the basis for their distinctive phenotype. Moreover, our findings suggest that despite the lack of genetic alterations affecting genes of the PI3K/mTOR signaling pathway in MCBs, mTOR signaling might be constitutively active in these tumors via epigenetic mechanisms, supporting the notion that, in the absence of pathognomonic genetic alterations, a disruption of the epigenetic landscape is a critical driver in the development of this rare breast cancer type. Citation Format: Fresia Pareja, Lorenzo Ferrando, Edaise M da Silva, Arnaud Da Cruz Paula, Anthe Stylianou, David N Brown, Pier Selenica, Jonathan Serrano, Hannah Y Wen, Hong Zhang, Edi Brogi, Larry Norton, Matjia Snuderl, Jorge S Reis-Filho, Britta Weigelt. Methylation profiling of mucinous breast cancer [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 P4-05-11.

Introduction: Breast cancers (BCs) with microsatellite instability (MSI)/DNA mismatch repair (MMR) deficiency are rare. Their identification is key, given that these tumors might be sensitive to immune checkpoint blockade. Here we sought to define the characteristics of primary and metastatic MSI-high (MSI-H) BCs using massively parallel sequencing. Materials and methods: To define MSI, MSISensor was applied to MSK-IMPACT targeted sequencing data of 918 primary BCs (pBCs) and 1,000 metastatic BCs (mBCs), and MSISensor and MANTIS to whole-exome sequencing (WES) data of 1,084 pBCs from The Cancer Genome Atlas (TCGA) BC project. BCs were classified as MSI-H if they displayed an MSISensor score ≥3.5 for WES or ≥10 for targeted sequencing data, or a MANTIS score ≥0.4 for WES data. The dominant mutational signature was inferred using SigMA in cases harboring at least 5 single nucleotide variants (SNVs). Mann-Whitney U and Fisher’s exact t-tests were employed for statistical analyses. Results: MSISensor classified 0.3% (3/903) of pBCs and 0.3% (3/965) of the mBCs subjected to MSK-IMPACT sequencing as MSI-H. The MSI-H pBCs were invasive ductal carcinomas (IDC), preferentially ER-positive/HER2-negative (66.6%). Compared to MSS pBCs, MSI-H pBCs displayed a higher tumor mutational burden (TMB; 11.2 vs 2.6 somatic mutations) and were enriched for dominant MMR mutational signatures 20 or 26 (33% vs 1.1%). MSI-H mBCs were all ER-positive/HER2-negative IDCs, displayed similar TMBs, percentage of small insertions and deletions (indels) and fraction of the genome altered (FGA) as compared to MSS mBCs. Only one of the three MSI-H mBCs had enough SNVs for inference of a dominant mutational signature, which was aging. MSISensor and MANTIS classified 1.5% (11/735) and 1.8% (15/836) of pBCs from TCGA as MSI-H. We observed a moderate agreement between MSISensor and MANTIS (Cohen’s Kappa = 0.552) with 1% (7/735) of pBCs being classified as MSI-H by both algorithms. pBCs from TCGA classified as MSI by either MSISensor or MANTIS were IDCs (90.9% and 85.7%, respectively) and invasive lobular carcinoma (ILC; 9.1% and 14.3%, respectively), and were either ER-positive/HER2-negative (45.5% and 42.9%, respectively) or ER-negative/HER2-negative (45.5% and 42.9%, respectively). These MSI-H pBCs (TCGA) displayed a higher TMB (MSISensor: 12.7 vs 0.8; MANTIS: 3.6 vs 0.8; both P Conclusions: We observed a moderate agreement between different MSI detection algorithms in the classification of pBCs as MSI-H based on WES analysis. pBCs classified as MSI-H were of common histologic types, either ER-positive/HER2-negative or ER-negative/HER2-negative, and displayed features characteristic of MSI, including a high TMB, high percentage of indels and an enrichment for MMR deficiency mutational signatures. mBCs classified as MSI-H by targeted capture sequencing were found to be of ER-positive/HER2-negative phenotype. Citation Format: Pier Selenica, Fresia Pareja, Lorenzo Ferrando, David N Brown, Nadeem Riaz, Britta Weigelt, Joshua Z Drago, Mark E Robson, Pedram Razavi, Sarat Chandarlapaty, Jorge S Reis-Filho. Massively parallel sequencing analysis of microsatellite instability in breast cancer [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 P4-05-01.

Background: Local and regional lymph node metastases are associated with worse prognosis in a variety of cancers, including breast cancer. However, to date it still remains debated whether axillary lymph node metastases can seed distant metastases or whether they merely represent a surrogate marker for tumor aggressiveness. Here, using multiple matched samples from primary tumor, axillary lymph nodes and distant metastases from breast cancer patients, we explored to which extent lymph nodes are transit points for further dissemination events to distant organs and assessed their role in disease progression using copy number phylogenetic reconstruction. Samples and Methods: We retrospectively identified 30 estrogen receptor-positive lymph node-positive metastatic breast cancer patients with 238 formalin-fixed paraffin-embedded tissue samples available together with well-annotated clinico-pathological and outcome data. Eligible patients had at least one primary tumor, one axillary lymph node metastasis, one distant metastatic sample as well as histologically normal tissue sample as germline reference. Low coverage whole genome sequencing was performed to infer somatic copy number aberrations. After rigorous filtering of the raw data, copy number phylogenetic reconstruction was performed with Integer Linear Program for the Copy-Number Tree Problem algorithm. Bootstrapping of the samples and of the segmented copy number data was performed to assess the robustness of the phylogenetic tree classification. Results: Phylogenetic tree analyses revealed that two different routes of metastatic dissemination occur in breast cancer. Unexpectedly, in only 24% of the patients, lymph nodes and distant metastases shared a common origin implying that dissemination occurred via the axillary lymphatic system whereas in the majority of the cases (76%) lymph node and distant metastases showed a distinct origin suggesting a direct seeding from the primary tumor to a distant organ. Of note, histological grade was the only clinic-pathological feature significantly associated with the route of dissemination, low grade tumors being associated with the common origin dissemination route (p=0.04). Interestingly, the patients sharing a common origin dissemination route were found to have significantly worse prognosis than the remaining patients (p=5.3x10−4), despite being low grade. Bootstrapping analyses showed identical topology as the original phylogeny in about 80% of the cases demonstrating the robustness of our method. Conclusions: In this study, we showed for the first time that the predominant route of metastatic dissemination operating in estrogen receptor-positive breast cancer is the direct seeding of tumor cells from the primary tumor to distant organ, independently of lymph node metastasis. These findings shed light on the routes on which tumor cells metastasize and could potentially improve the clinical management of lymph node-positive breast cancer patients. Citation Format: David Venet, David Brown, Danai Fimereli, Bram Boeckx, Marion Maetens, Samira Majjaj, Ghizlane Rouas, Maria Capra, Giuseppina Bonizzi, Federica Contaldo, Christine Galant, Martine Piccart, Giancarlo Pruneri, Denis Larsimont, Diether Lambrechts, Françoise Rothé, Christine Desmedt, Christos Sotiriou. Phylogenetic reconstruction of advanced breast cancer reveals two different routes of metastatic dissemination associated with distinct clinical outcome [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 PD8-02.

Introduction: Breast cancer (BC) is heterogeneous and comprises diverse entities with distinctive phenotypes, biology and clinical presentations. There are 21 special histologic types, accounting for approximately 20% of invasive BCs. Large-scale sequencing studies have focused preferentially on common histologic forms, and data on the genomic landscape of special histologic types, particularly in the metastatic setting, are scarce. Here, we sought to define the repertoire of somatic genetic alterations of metastatic special histologic types of BC. Materials and methods: We conducted a reanalysis of MSK-IMPACT targeted sequencing data of 346 special histologic types of BC from Razavi et al (Cancer Cell 2018), including invasive lobular (ILC; primary, n=129; metastasis, n=150), mucinous (primary n=11, metastasis n=5), micropapillary (primary n=8, metastasis n=13) and metaplastic carcinomas (primary, n=6; metastasis, n=11). Somatic single nucleotide variants (SNVs) and copy number alterations (CNAs) were inferred using a validated bioinformatics pipeline, and mutational signatures were assessed with SigMA. Genetic alterations of metastatic special histologic type BCs were compared to those of primary tumors and metastatic invasive ductal carcinomas of no special type (IDC-NSTs) matched by age, menopausal status and estrogen receptor (ER)/HER2 status at a 1:2 ratio for ILCs and at a 1:3 ratio for the remaining BCs. Results: The most frequently mutated genes in metastatic ILCs were CDH1 (71%), PIK3CA (49%), ESR1 (20%) and KMT2C (19%). ESR1 mutations were more frequent in metastatic than in primary ILCs (20% vs 3%; P0.05). Metastatic mucinous BCs had a numerically higher frequency of mutations in ESR1 (40% vs 6%; P>0.05) and GATA3 (40% vs 0%; P>0.05), and lower frequency of TP53 mutations (20% vs 53%; P>0.05) than matched metastatic IDC-NSTs. PIK3CA and TP53 were the most recurrently mutated genes in metastatic micropapillary BCs and were numerically more frequent in metastatic than in primary tumors (PIK3CA, 61% vs 25%; TP53, 46% vs 25%; P>0.05). PIK3CA was more frequently mutated in metastatic micropapillary BCs than IDC-NSTs (61% vs 23%; P Conclusions: The repertoire of somatic genetic alterations of primary and metastatic special histologic type BCs markedly overlaps. Nonetheless, there is an enrichment of mutations in ESR1 in metastatic BCs of typically ER-positive special histologic type. Metastatic BCs of special histologic type show an enrichment of similar genetic alterations as primary tumors, such as a higher frequency of mutations in CDH1, FOXA1 and TBX3 in ILCs, a lower frequency of PIK3CA and TP53 mutations in mucinous BCs and a higher frequency of PIK3CA and TP53 mutations in micropapillary and metaplastic BCs. Citation Format: Fresia Pareja, Simon SK Lee, Francisco Beca, Lorenzo Ferrando, Pier Selenica, David N Brown, Hannah Y Wen, Hong Zhang, Edi Brogi, Pedram Razavi, Sarat Chandarlapaty, Britta Weigelt, Jorge S Reis-Filho. The genomic landscape of metastatic special histologic types of breast cancer [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 PD8-06.

Background: The CDK4/6 inhibitors (i) palbociclib (Palbo), ribociclib, and abemaciclib remarkably improved the outcome of patients with metastatic ER+/HER2- breast cancer (BC) and are now under clinical investigation in early BC. Despite high efficacy, de novo and acquired resistance to CDK4/6i is common. Elucidating the molecular basis for sensitivity and resistance to CDK4/6i is crucial to identify predictive biomarkers and novel therapeutic targets to improve patient outcome. Materials and Methods: MCF7, T47D and ZR75-1 parental (P) BC cells and their derivatives made resistant to tamoxifen, estrogen deprivation (EDR), or fulvestrant were used. The P and EDR models of MCF7 and T47D cells were chronically exposed to increasing concentrations of Palbo to generate derivatives with acquired resistance to Palbo (PalboR). The transcriptomic profiles of P, endocrine-resistant (EndoR) and PalboR models were determined by RNA-seq. IC50s were determined by exposing MCF7, T47D, and ZR75-1 P and EndoR lines (n=12) to increasing concentrations of Palbo. Cell growth was assessed by methylene blue staining, and changes in the mRNA and protein levels of key cell cycle molecules were assessed by RT-PCR and Western blot, respectively. Gene expression data from the Cancer Dependency Map (DepMap), baseline tumors from the NeoPalAna (NCT01723774) and neoMONARCH (NCT02441946) neoadjuvant trials, as well as the TCGA and METABRIC datasets were interrogated for correlations of gene signatures and patient outcome (by KMPlot). Results: Palbo treatment resulted in a dose-dependent inhibition of the growth of P and EndoR BC cell lines, with varying degree of sensitivity among the models. GSEA analysis comparing the least sensitive (IC50>350nM) vs. the most sensitive (IC50 Citation Format: Carmine De Angelis, Xiaoyong Fu, Maria Letizia Cataldo, Agostina Nardone, Valerie M. Jansen, Jamunarani Veeraraghavan, Sarmistha Nanda, Lanfang Qin, Vidyalakshmi Sethunath, Resel Pereira, Matteo Benelli, Ilenia Migliaccio, Luca Malorni, Joshua Donaldson, Pier Selenica, David N. Brown, Britta Weigelt, Jorge S. Reis-Filho, Ben H. Park, Sara A. Hurvitz, Dennis J. Slamon, Mothaffar F. Rimawi, Rinath Jeselsohn, Kent Osborne, Rachel Schiff. High levels of interferon-response gene signatures are associated with de novo and acquired resistance to CDK4/6 inhibitors in ER+ breast cancer [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 GS2-01.

CDK4/6 inhibitors (CDK4/6i) have proven benefits in estrogen receptor-positive (ER+) metastatic breast cancer (MBC), however the activity is variable among patients and biomarkers of sensitivity are lacking. We sought to determine whether any genomic alterations might characterize extreme responders and non-responders from over 500 patients with ER+ MBC who received CDK4/6i plus aromatase inhibitors or fulvestrant. We first analyzed a cohort of 348 patients who has MSK-IMPACT sequencing performed on a pre-treatment tumor sample and asked whether any somatic mutation or copy number alteration was associated with altered progression-free survival (PFS) on subsequent CDK4/6i-based therapy. As expected, common alterations in CCND1, PIK3CA or ESR1 were not associated with altered PFS on CDK4/6i. Also as anticipated, RB1 loss was associated with resistance to CDK4/6i (median PFS: 3.6 months, 95% confidence interval [CI]: 2.2, not reached, p=0.00041). Unexpectedly, alterations in components of the Hippo pathway (FAT1, YAP, LATS1/2, and NF2) were mutually exclusive with each other and with RB1 and were collectively associated with resistance to CDK4/6i in ER+ MBC (Hazard Radio [HR]: 3.6; 95% CI: 1.7, 7.8, p=0.0044). FAT1 was the most frequently mutated gene in the Hippo pathway with alterations resulting in FAT1 loss associated with a markedly shortened PFS (median: 2.4 months, HR: 11.1; 95% CI: 4.65, 26.62; p=2.2x10-11). No alterations were clearly linked with extremely prolonged PFS on CDK4/6i. Additional analyses from a larger cohort of patients with genomic sequencing prior to starting CDK4/6i are now underway and will be presented. To further understand the mechanisms of Hippo pathway regulation of cell cycle, cell line and patient derived xenograft models were utilized. These studies revealed that different mechanisms of suppression of Hippo signaling such as FAT1 knockdown (CRISPR or shRNA) and NF2 knockdown (shRNA) lead to YAP activation and YAP-dependent CDK4/6i resistance. Moreover, in all cases, resistance was mediated by upregulation of CDK6 as knockdown of CDK6 prevented drug resistance. Finally, we have sought to understand the implications of these distinct mechanisms of resistance to CDK4/6i. Clinical follow up on subsequent lines of therapy and duration of treatment after CDK4/6i (e.g. chemotherapy or other hormone therapy) are now being analyzed based on class of resistance alteration and will be presented. In conclusion, RB1 loss and Hippo pathway alterations mediate resistance to CDK4/6i in patients with ER+ MBC, nominating these as potential biomarkers for this class of drugs. Citation Format: Pedram Razavi, Qing Li, Zhiqiang Li, David N. Brown, Christina Ping, Ronglai Shen, Francisco Sanchez-Vega, Nikolaus D. Schultz, Jose Baselga, Maurizio Scaltriti, Jorge S. Reis-Filho, Sarat Chandarlapaty. Genomic profiling of metastatic breast cancer identifies the Hippo pathway to be a key determinant of CDK4/6 inhibitor sensitivity [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 927.

BACKGROUND: We have previously shown that TILs are predictive of benefit to trastuzumab and chemotherapy in HER2+ BC in the FinHER study, an adjuvant, phase III study in early-stage BC, where HER2+ patients were randomized to 9 weeks of trastuzumab or no trastuzumab in addition to chemotherapy (Loi et al, ASCO 2012). We sought to further confirm this positive association as well as to understand the composition of TILs. METHODS: The association between TILs and response to trastuzumab with chemotherapy (epirubicin/cyclophosphamide with docetaxel with or without capecitabine) was evaluated in 156 HER2+ patients from the neoadjuvant GeparQuattro trial. The primary correlative endpoint was the association between TILs (quantified using the same method as previously published) with pathological complete response rates (pCR), adjusted for clinicopathological characteristics. To understand the composition of TILs, correlations between TILs and gene expression levels of 13 pre-defined immune markers were evaluated from 202 HER2+ samples from the FinHER study. These represented T and B cell infiltration (CD3D, IGKC), Th1 (IFNG, CD8A), chemoattractants (CXCL9, CXCL13), immunosuppression (VEGFA, FOXP3, IDO1) and T-cell checkpoint receptors and ligands (PD-1, PD-L1, CTLA-4, CD80). Prognostic associations and interactions with trastuzumab were studied in Cox regression models for distant-disease free survival (DDFS). Preclinical mice models of HER2 mammary cancer were used to investigate combination therapies. RESULTS: In the GeparQuattro trial data, each 10% increment in TILs was associated with higher rates of pCR (adjusted OR:1.14 95%CI:1.01-1.29;P = 0.037) after neoadjuvant trastuzumab and chemotherapy supporting the findings from the FinHER study. Gene expression analyses using the FinHER samples revealed IDO1 and CXCL13 were most highly correlated with TILs (R = 0.58 and 0.51;p We next postulated that enhancing T-cell responses would be synergistic with trastuzumab. Trastuzumab in combination with several inhibitors of T-cell negative regulation (anti-CTLA4, anti-PD-1, anti-PD-L1) resulted in higher tumor regressions compared with monotherapy in a mouse transplant model of HER2+ mammary cancer. In particular, anti-PD1 (P = 0.02) and anti-PD-L1 (P = 0.008) were most effective. In a BALB/c-MMTV-neu transgenic mice mouse model (i.e. immune tolerant to HER2 antigen), the combination of trastuzumab and anti-PD1 antibody could significantly delay mammary gland tumor formation (p CONCLUSIONS: We confirm that TILs are associated with higher responses to trastuzumab and chemotherapy. Furthermore, tumor-mediated immunosuppression is evident in the lymphocytic infiltrate with PD-1 and IDO1 significantly predictive of trastuzumab benefit. The addition of a T-cell checkpoint inhibitor in a preclinical model significantly enhanced trastuzumab responses. These combinations warrant evaluation in the clinical setting. Citation Information: Cancer Res 2013;73(24 Suppl): Abstract nr S1-05.

Standard treatment for clinical stage II/III rectal adenocarcinoma (READ) includes a combination of chemotherapy with pyrimidine analogues, such as capecitabine, and radiation therapy, followed by surgery after 6 to 12 weeks. Currently no predictors of benefit from neoadjuvant chemoradiotherapy (NACT) exist, and transcriptomic studies, prevalently focused on coding transcripts or miRNAs have so far failed to find clinically useful gene signatures. In the present analysis we assessed the expression levels of 8,127 long non-coding RNAs (lncRNAs), poorly studied in the context of READ, to infer whether they could stratify pre-treatment READ samples according to patients' pathological complete response (pCR). Methods: We collected and analysed with Agilent SurePrint G3 Human v2 8x60K microarrays, a consecutive series of 49 specimens from READ patients undergoing NACT with a combination of capecitabine (825 mg/m2) and radiation (50.4 Gy) cycles for 6 weeks. We performed a regularized variable selection regression (“glmnet” R package) to identify potential lncRNAs predicting pCR. We then analysed, with a bivariate correlation, the top lncRNAs predictors with the expression of 20,560 coding transcripts represented in the microarrays we used, and performed pathway enrichment analysis using Ingenuity® Pathway Analysis (Qiagen, Inc.). Results: Out of 49 patients for whom we could obtain snap-frozen biopsies with complete clinical and pathological data, we analysed 30 patients, 12 with minor or absent pathological response and 18 with major or complete response to NACT. We identified a signature of 11 lncRNAs able to successfully stratify responder vs. non-responder patients. By investigating the correlation of these 11 lncRNAs with coding transcripts, we identified one of them, lnc-ZNF-366-6, presenting an unusually high correlation with more then 1136 coding genes (FDR < 0.01). By gene enrichment analysis we found an enrichment in cancer-related pathways, especially alterations in solid and colon adenocarcinoma pathways (p < 0.01), as well as mitochondrial dysfunction, oxidative phosphorylation and sirtuin signalling pathways (p < 0.00001), defined by the positively and negatively correlated variables with lnc-ZNF-366-6. Moreover, we identified potential drivers of disease or therapeutic targets, either correlated or anti-correlated with lnc-ZNF-366-6. Conclusions: LncRNAs have the potential to classify responder versus non-responder READ patients undergoing NACT. Lnc-ZNF-366-6 deserves further experimental validation and functional analyses to clarify its exact contribution to resistance to DNA damaging agents, such as radiation, and to inhibitors of DNA/RNA synthesis, such as capecitabine. Citation Format: Lorenzo Ferrando, Gabriella Cirmena, A Garuti, Stefano Scabini, Federica Grillo, Luca Mastracci, Edoardo Isnaldi, Ciro Marrone, Roberto Murialdo, David Norman Brown, Roberto Fiocca, Emanuele Romairone, Alberto Ballestrero, Gabriele Zoppoli. Long non-coding RNA ZNF-366-6 predicts lack of response to neoadjuvant chemoradiotherapy in locally advanced rectal adenocarcinoma and presents meaningful biological associations with potential coding cancer driver transcripts related to gastrointestinal cancer pathways [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 2465.

Background Over the past years, tumor derived alterations in plasma has emerged as a promising cancer biomarker. The clinical applications envisioned with liquid biopsies range from monitoring of treatment response to disease progression and acquired resistance. However, the analysis of circulating tumor DNA (ctDNA) is challenging owing to the small fraction of tumor specific DNA relative to background levels of non-tumor derived DNA and therefore requires highly specific and sensitive techniques. Experimental procedures The Ion AmpliSeqTM Cancer Hotspot Panel v2, covering 2,800 COSMIC mutations from 50 cancer genes was used to analyze 31 primary tumor and metastatic lesions along with 19 matched plasma samples from 13 breast cancer (BC) patients receiving neoadjuvant treatment and 6 advanced stage metastatic BC patients. Digital PCR (dPCR) experiments were set up using mutation-specific assays (Biorad) for the detection of 13 hotspot mutations in PIK3CA, TP53, PTEN, AKT1, CDKN2A and STK11 genes previously identified in the solid tumor samples using next generation sequencing (NGS). Results The specificity and sensitivity of the mutation-specific dPCR assays were tested in dilution curves using synthetic oligonucleotides and tumor samples bearing the mutation of interest. Mutations could be robustly detected in plasma samples using dPCR with a sensitivity as low as 0.1% and a sensitivity of 0.5% using NGS. All mutations previously identified using NGS were validated using dPCR in tumor (n = 27) as well as in plasma samples (n = 5). Moreover, dPCR identified 4 additional mutations in 4 plasma samples in the neoadjuvant setting and 4 in the metastatic setting, 2 in tumor and 2 in plasma samples. The mutant allelic frequencies (MAF) of these newly identified mutations were very low (range 0.2%-1.8%) close to the detection limit of the technology. Through the use of dPCR, the detection rate of plasma ctDNA increased from 8% using NGS to 38% in the neoadjuvant setting and from 67% using NGS to 100% in the metastatic setting thereby highlighting the greater sensitivity of the dPCR technology. The MAF assessed by NGS and dPCR were highly correlated both in the neoadjuvant and metastatic (Spearman ρ = 0.84 and ρ = 0.97) settings, a higher concordance between NGS and dPCR being observed when the MAF are high. Conclusions The use of dPCR allows the detection of low-frequency tumor-specific mutations in plasma with a greater sensitivity compared to NGS providing the high accuracy and sensitivity essential for liquid biopsies. Citation Format: Françoise Rothé, Ghizlane Rouas, Michelle Coureau, Marion Maetens, David Brown, Christos Sotiriou, Michail Ignatiadis. Comparison of NGS and digital PCR for the detection of point mutations in breast tumor and plasma samples. [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 3968.

Background In the era of precision medicine, data obtained from sequencing the cancer genome of a patient can be integrated into clinical decision-making. Liquid biopsy using ctDNA is a non-invasive tool, which can circumvent unfeasible or unrepresentative tumor biopsies and allow continuous monitoring of the disease thereby contributing to rational drug delivery. Experimental procedures Precision-f is a feasibility molecular screening study that enrolled 34 patients with metastatic melanoma, non-small cell lung cancer (NSCLC) and colorectal cancer (CRC). Results from targeted gene sequencing (TGS) of metastatic biopsies were presented at AACR 2015. The patients had a blood draw concomitantly with the metastatic biopsy for plasma isolation. Similar to biopsies, somatic mutations (substitutions, small insertions and deletions) were assessed using the Ion AmpliSeq™ Cancer Hotspot Panel v2, designed to amplify 207 amplicons covering 2800 COSMIC mutations from the 50 most commonly reported oncogenes and tumor suppressor genes with a sensitivity of 1%. All mutations identified in the biopsy and/or plasma samples using TGS are currently being validated using Droplet Digital PCR (ddPCR) technology. Results Evaluable ctDNA-derived sequencing data were available for 32 patients. The sequencing results between biopsies and plasma were fully concordant in 63% (20/32) of the patients. TGS of ctDNA failed to detect clinically relevant mutations (EGFR, PIK3CA, KRAS and BRAF) identified in the metastatic biopsies in 19% (6/32) of the patients. In contrast, biologically relevant mutations (STK11, KRAS, PIK3CA and SMO) not detected in tumor biopsies were found in the ctDNA in 9% (3/32) of the patients. Interestingly, in one CRC patient a KRAS G13D mutation still potentially associated with benefit to anti-EGFR therapy was detected only in the tumor biopsy whereas a KRAS A146T mutation mediating resistance to anti-EGFR therapy was detected solely in ctDNA. Tumor biopsies and plasma are currently being tested using ddPCR to validate the findings. The results of these analyses will be presented at the meeting. Conclusion Although technical challenges remain, ctDNA profiling appears to be a promising tool for precision oncology by allowing the identification of genomic alterations associated with drug resistance and tumor progression. Citation Format: Philippe G. Aftimos, Marion Maetens, Françoise Rothé, David Brown, Yacine Barèche, Alexandre Irrthum, Thierry Berghmans, Joseph Kerger, Alain Hendlisz, Michail Ignatiadis, Martine Piccart, Christos Sotiriou, Ahmad Awada. Circulating tumor DNA (ctDNA) as an alternative to tumor biopsies for precision oncology: The Jules Bordet Institute experience. [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 3962.

Background: Multifocal breast cancer (MFBC), defined as multiple synchronous unilateral lesions of invasive breast cancer, is relatively frequent and has been associated with more aggressive features than unifocal cancer. Here, we aimed to investigate the genomic heterogeneity between MFBC lesions sharing similar histo-pathological parameters.Material and methods: The characterization of different lesions from 36 patients with ductal MFBC involved the identification of non-silent coding mutations in 360 protein-coding genes (171 tumour and 36 matched normal samples). We selected only patients with lesions presenting the same grade, ER and HER2 status. Mutations were classified as ‘oncogenic’ in case of recurrent substitutions reported in COSMIC or truncating mutations affecting tumour suppressor genes. All mutations identified in a given patient were further interrogated in all samples from that patient through deep re-sequencing using an orthogonal platform. Whole genome rearrangement screen was further conducted in 8/36 patients.Results: Twenty-four patients (67%) had substitutions/indels shared by all their lesions, of which 11 carried the same mutations in all lesions, and 13 had lesions with both common and private mutations. Three-quarters of those 24 patients shared oncogenic variants. The remaining 12 patients (33%) did not share any substitution/indels, with inter-lesion heterogeneity observed for oncogenic mutation(s) in genes such as PIK3CA, TP53, GATA3 and PTEN. Genomically heterogeneous lesions tended to be further apart in the mammary gland than the homogeneous ones. Genome-wide analysis of a limited number of MFBC nevertheless identified a common somatic background in all studied MFBC, including those with no mutation in common between the lesions. Conclusion and perspectives: As the number of molecular targeted therapies increases and trials driven by genomic screening are ongoing, our findings, based on the targeted sequencing of cancer genes in 36 MFBC tumors, highlight the presence of genomic inter-lesion heterogeneity in one-third of the cases despite similar pathological features. This implies that deeper molecular characterization of all MFBC lesions is warranted for the adequate management of those cancers. [CD, DF, and EP contributed equally to this work. PJC and CS contributed equally to this work.] Citation Format: Christine Desmedt, Debora Fumagalli, Elisabetta Pietri, Gabriele Zoppoli, Serena Nik-Zainal, Gunes Gundem, David Brown, Francois Rothe, Samira Majjaj, Anna Garuti, Enrico Carminati, Sherene Loi, Thomas Van Brussel, Marion Maetens, Laura Mudie, Delphine Vincent, Naima Kheddoumi, Luigi Serra, Ilaria Massa, Alberto Ballestrero, Dino Amadori, Roberto Salgado, Alexandre de Wind, Diether Lambrechts, Martine Piccart, Denis Larsimont, Peter J. Campbell, Christos Sotiriou. Uncovering the genomic heterogeneity of multifocal breast cancer. [abstract]. In: Proceedings of the 106th Annual Meeting of the American Association for Cancer Research; 2015 Apr 18-22; Philadelphia, PA. Philadelphia (PA): AACR; Cancer Res 2015;75(15 Suppl):Abstract nr CT135. doi:10.1158/1538-7445.AM2015-CT135

Background: Microarrays have revolutionized breast cancer research by enabling gene expression comparisons on a transcriptome-wide scale. They have provided clinically useful information such as the development of the intrinsic classification as well as several gene prognostic signatures. Recently, with the advent of next generation sequencing technology, RNA-Seq has emerged as a promising technology for digital read-out of the transcriptome. Here, we will present a comparison of microarray and RNA-Seq data on a series of 58 breast cancer samples. Methods: 58 tumors representing the different molecular subtypes [17 triple negative (ER−, PgR−, HER2−); 14 HER2 positive (HER2+); 17 luminal A (ER+, HER2−, histological grade 1), 10 luminal B (ER+, HER2−, histological grade 3)] were obtained from patients recruited between 2007 and 2011 at Jules Bordet Institute. Their RNA was profiled using the Affymetrix HG-U133 Plus 2.0 chips and sequenced with the ILLUMINA platform, producing ∼30 million 2*50 bp paired-end reads per sample. The reads alignment and expression quantification were performed using Tophat 2 and Cufflinks 2, respectively, on the basis of the Ensembl genome annotations hg19. Affymetrix microarray data were normalized using fRMA and probesets were selected using JetSet. A selection of 7 clinically relevant gene signatures was evaluated. Results: About 16000 genes were retained for analysis. Comparison of Affymetrix and RNA-Seq data revealed that 50%, 25% and 5% of these genes had a Spearman correlation superior to 0.7, 0.82 and 0.92, respectively. The genes with the highest correlation coefficients between Affymetrix and RNA-Seq tended to have medium/high expression levels (median Affymetrix log2 expression of 7.5) consistent with the fact that microarrays have a limited dynamic range compared to RNA-Seq. We observed an excellent correlation between Affymetrix and RNA-Seq for the clinically relevant genes ER (0.97), PgR (0.95) and HER2 (0.90) respectively. We observed, however, a few discordances between ER and PgR quantification by immunohistochemistry and RNA-Seq/Affymetrix. We next measured the consistency of two molecular classifiers, SCMGENE (Haibe-Kains JNCI2012) and PAM50 (Parker JCO2009), between Affymetrix arrays and RNA-seq. A kappa coefficient of 0.93 (CI 0.85–1) and 0.79 (CI 0.66–0.91) was observed for SCMGENE and PAM50, respectively, with the majority of the discrepancies for PAM50 concerning the HER2 subtype. The first generation proliferation-based gene expression signatures showed an excellent correlation between Affymetrix and RNA-Seq: Genomic Grade Index (0.98), Mammaprint (0.98) and Oncotype (0.97). The immune and stroma gene signatures (Desmedt et al. 2008) showed a correlation of 0.97 and 0.79 respectively. The lower correlation for the stroma signature could partly be explained by the fact that RNA-Seq and Affymetrix are measuring different splice variants. Conclusions: This is to our knowledge the first study to report a systematic comparison between RNA-Seq and Affymetrix HG-U133 Plus2.0 data in breast cancer for individual genes and several prognostic gene signatures. According to these results, RNA-Seq can reliably be used for the quantification of breast cancer transcriptome. Citation Information: Cancer Res 2012;72(24 Suppl):Abstract nr P3-04-10.

Background: The aim of this study was to interrogate primary/metastatic breast cancer (BC) genetic heterogeneity using next generation whole exome sequencing and copy number analysis from an autopsy series of BC patients (pts). Methods:To reconstruct the trajectories of BC progression, we performed exome sequencing (Illumina HiSeq2000, Truseq DNA sample preparation kit v2 and Exome Enrichment Kit v2; alignement done with BWA, substitutions and indels called with GATK and DINDEL respectively) coupled to validation by Sequenom and deep re-sequencing (ongoing), as well as copy number (CN) analyses (Oncoscan, Affymetrix) on DNA from matched primary (n=1-6/pt), axillary lymph node (ALN, n=1-6 for 4 pts), local recurrence (for 1 pt) and distant metastatic (n=1-5/pt) FFPE samples obtained from an autopsy series of 10 BC pts. Results:1/ The degree of primary/metastatic genetic heterogeneity is proportional to the time elapsed between the diagnosis of the primary tumor and the emergence of the metastases (mutations: corr=0.76, CN: corr=0.63); metastases from pts with a longer cancer history being genetically more different from their corresponding primary tumor than those from pts with a shorter cancer history. 2/ The phylogenetic analyses revealed that in all but 1 pt, the various distant metastases are present on the same branch of the phylogenetic tree. This suggests that the majority of the metastases arise from a single metastasizing event, with one or more distant sites further re-seeding additional organs. ALN metastases are genetically less evolved compared to distant metastases. 3/ Distant metastases from a same patient only share between 11 and 50% metastasis-specific mutations. This implies that metastasis-specific potentially targetable genetic alterations might either be present in all metastases from a given patient (such as EGFR, PDGFRA amplification for pt 3, AR amplification for pt 8) or only in some metastases (such as NOTCH2 mutation in the liver metastasis of pt 6). 4/ Primary tumor samples and metastases from one ER+/HER2- pt were hypermutated (696 unique mutations in total). The substitutions were enriched for the APOBEC-related mutational signature, possibly linked to the somatic APOBEC3B D316N mutation observed in all samples from this pt. Conclusions: Here we report for the first time that primary/metastatic BC genetic heterogeneity is proportional to the time elapsed between the diagnosis of the primary tumor and the emergence of the metastases. These results therefore suggest that it is extremely relevant to interrogate distant metastatic lesions, multiple if possible to account for inter-metastatic heterogeneity, to guide treatment, especially but not only, in those pts relapsing a few to many years after initial diagnosis. Citation Format: Christine Desmedt, David Brown, Borbala Szekely, Dominiek Smeets, Marcell A. Szasz, Pierre-Yves Adnet, Françoise Rothé, Zsofia I. Nagy, Zsofia Farago, Anna-Maria Tokes, Dimitrios Zardavas, Gabriele Zoppoli, Michail Ignatiadis, Lajos Pusztai, Martine Piccart, Denis Larsimont, Diether Lambrechts, Janina Kulka, Christos Sotiriou. Unraveling breast cancer progression through geographical and temporal sequencing. [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 986. doi:10.1158/1538-7445.AM2014-986