Search

Background Immunotherapies such as PD-L1 inhibitors have shown promising efficiency in breast cancer (BC) patients. However, treatment responses for immunotherapies are diverse since the immunogenicity of breast cancer is heterogeneous. Specific subtypes such as hormone receptor (HR)-positive, human EGF receptor 2 (HER2)-positive, and triple-negative breast cancer (TNBC) have shown heterogeneity in immunogenicity. Therefore, precisely identification of patients potentially benefit from immunotherapy is important. Previous studies have proved PD-L1 expression and tumor mutational burden (TMB) as predictive biomarkers for immunotherapy. Here, we report the PD-L1 expression and TMB status in Chinese BC patients with different subtypes. Using comprehensive molecular analysis, we also characterized the genomic features related to these biomarkers. Methods Tumor samples from 112 Chinese patients with BC were collected and subjected to next-generation sequencing (NGS) and immunohistochemical (IHC) analysis. NGS were performed in a laboratory accredited by College of American Pathologists (CAP) and certified by Clinical Laboratory Improvement Amendments (CLIA) using validated panel targeting 450 cancer genes. Genomic alterations, included including single base substitution, short and long insertions/deletion (Indel), copy number variation, gene fusion, and rearrangement, were assessed. Tumor mutational burden (TMB) was measured by an algorithm developed in-house. Tumor tissues were analyzed for PD-L1 expression by IHC with 22C3 or 28-8 antibodies, respectively. Results The 112 Chinese BC patients consisted of 87 HR+ (77.7%), 19 TNBC (17.0%) and 6 HR-/ HER2 + (5.3%), with a median age of 47.5 years old (range 24-81). Of all patients, 42.0% were positive for PD-L1 expression (CPS≥1), including 30.4% PD-L1 (1≥CPS>10) and 11.6% PD-L1 (CPS≥10). PD-L1 expression were observed in HR+ BCs (41.4%) as well as in TNBC (47.4%) and HR-/ HER2+ (33.3%) subtypes, no significant difference were observed among the three subtypes. 8.9% patients were TMB-High (≥10 muts/Mb) with median TMB of 3.4 muts/Mb (range 0-80.8). TMB was slightly correlated with PD-L1 expression (Kendall tau = 0.241; P =0.01). 17.0% of PD-L1 positive patients and 3.1% of PD-L1 negative patients were TMB-High (p=0.016). Patients with PD-L1 CPS≥10 had significantly higher median TMB (7.1 vs. 3.1 muts/Mb, p Conclusion Our data shows that the PD-L1 expression has no significant difference among HR+, TNBC and HR-/ HER2+ subtypes. These data provide the hypothesis that PD-L1 expression positive in different subtype breast cancers may benefit from immune treatment, including HER2 and Luminal subtypes. Otherwise, there was a correlation between PD-L1 expression and TMB, High TMB was also observed in PD-L1 negative patients, which may enrich the subgroup of patients who could benefit from immunotherapy. The verification of integrated predictive biomarkers for immunotherapy in BC is further needed. Citation Format: Ning Liao, Cao Li, Jundong Wu, Bo Chen, Guochun Zhang, Xueri Li, Liping Guo, Guangnan Wei, Jiali Lin, Yingzi Li, Yuchen Zhang, Hsiaopei Mok, Chongyang Ren, Yulei Wang, Kai Li, Cheukfai Li, Lingzhu Wen, Minghan Jia, Tengjiao Lin, Chunyan Cheng. Pd-l1 expression among different subtypes of chinese breast cancer patients [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 PS17-44.

Over the past decade, genomic advancements have afforded tremendous opportunities in cancer research; including characterization of gene expression networks that distinguish tumor types, and the resolution of genomic structure that drives tumor growth and treatment outcomes. Despite these advances, disparities in cancer outcomes persist as does our paucity in data to understand the breadth of diversity in tumor traits across populations. Most of the data that defines our current understanding of cancer genetics was generated within the context of mainly European ancestry. This bias is partially due to geographic convenience of recruitment within the proximity of the institutions that generate these data elements. This type of systemic bias has perpetuated disparities of already marginalized racial minority populations, compared to their European race-group counterparts. However, international consortia have led to new insights and access, with strategic recruitment and study designs that harness population genetic diversity. Germline genomic investigations have uncovered ancestral susceptibilities that are either due to prevalence of ancestry-specific risk alleles or ancestrally/evolutionarily conserved modifiers of molecular pathways that prime development of certain tumor subtypes. Tumor/somatic genomics have repeatedly uncovered gene signatures that reflect thematic differences in tumor immune microenvironment as well as copy number variations, which are more recently being investigated with high-resolution technologies in imaging and genomic architecture, linked to genetic ancestry. As we turn our lens to the genomic context of population diversity, we are realizing novel opportunities to develop transformative precision medicine applications. Specifically, for this plenary presentation, we will highlight the transition of GWAS studies to functionalize genetic ancestry differences, enhanced with inclusion of ancestral populations, as opposed to the traditional filtration of these features. We will overview the newest concepts of complex germline and somatic architecture in the context of disparities and how these are determining the mysterious function of genomic associations identified as population-private risk across multiple ethnicities; including Asian, Latin American and African ancestry clades. We will demonstrate the utility of Whole Genome Sequencing in the clinical genetics space to mitigate disparities by characterizing the pathological features (Genomic Phenotypes) of the insidious Variables of Unknown Significance (VUS) and how these findings can fast-track minority access to alternative treatments. These data will ultimately improve our interpretations of overall cancer risk and risk of recurrence test results and recalibrate risk models. Lastly, we give a call to action that resounds the importance of including not only the genetic history (ancestry) but also the social history (self-reported race) of cancer patients and how this multidisciplinary approach will unlock our understanding of the drivers of aggressive tumor biology and be a catalyst for implementation of customized prevention and survivorship interventions. Only with the “Power of Inclusion” can we hope to traverse the breadth of disparities and make leaps of progress in cancer research and treatment. Citation Format: Melissa B. Davis. Ancestry and cancer disparities: Genomic sequencing in diverse populations [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr PL01-03.

Background: Next generation sequencing (NGS) and the according target-based precision therapy has shown promising efficacy in many tumors. Here we report the genomic characteristics of Chinese breast cancer to support the development of more precise treatment strategy. Methods: Formalin fixed, paraffin embedded (FFPE) tumor samples and matched peripheral blood samples of 220 Chinese cancer patients including 218 females and 2 males were collected for NGS-based 450-genes panel assay. Pathological subtypes included 104 HR+/HER2-, 30 HR-/HER2+, 44 HR+/HER2+ and 42 triple negative breast cancers (TNBC). Assessed genomic alterations included single base substitution, short and long insertions/deletion, copy number variation, gene fusion and rearrangement. MSK breast cancer data was obtained from cBioPortal for comparing the difference between Chinese and Western patients. Results: The top mutated genes were TP53 (88.1%), PIK3CA (26.2%) and PTEN (21.4%) in TNBC; ERBB2 (86.7%), TP53 (86.7%), CDK12 (80.0%) and PIK3CA (23.3%) in HR-/HER2+ patients; ERBB2 (75%), TP53 (61.4%), CDK12 (54.5%) and PIK3CA (45.5%) in HR+/HER2+ patients. In HR+/HER2- patients, the most frequently mutated genes was PIK3CA (49.0%), followed by TP53 (30.8%) and GATA3 (27.9%), whereas the frequency of PIK3CA and GATA3 mutation were lower (41.6% and 19%, respectively) in Western patients. Although Alpelisib was approved by FDA in PIK3CA mutated HR+/HER2- patients, we observed PIK3CA mutation frequency had no difference among four subtypes. In this cohort, patients with PIK3CA mutation were significantly elder than patient without PIK3CA mutation (50 vs 47 years old, p=0.006). The hotspot mutations of PIK3CA (E542X, E545X and H1047X) accounted for 79.8% of PIK3CA mutations (71/89). Gene fusion/rearrangement was observed in 26% patients, in which 4 patients had gene fusions. Gene variations in homologous recombination pathway were found in 27.7% of patients. Among the 11.4% of patients with BRCA1/2 mutations, 11 patients harbored germline mutations and 19 patients had somatic mutations. Rearrangement accounted for 35% in somatic BRCA1/2 mutations. In the patients with germline BRCA1/2 mutation, 6 were HR+/HER2- patients who have been approved to use Olaparib. Based on the usage of CDK4/6 inhibitor in HR+/HER2- patients, this 450-genes panel enabled us to find that 41.3% HR+/HER2- patients had genes variations are related to CDK4/6 inhibitor resistance (CCND1, 18.3%; FGFR1 17.3%; NF1, 7.7%; MDM2, 6.7%; ESR1, 6.7% and RB1, 1.9%). The median TMB was 4.3 Muts/Mb in the whole cohort and patients with KMT2C mutations had significantly higher TMB (5.5 vs 3.8 Muts/Mb, p=0.004). In addition, Chinese breast cancer patients had a significantly higher frequency of KMT2C mutations compared to western patients (11.4% vs 1.4%, p Citation Format: Ning Liao, Guochun Zhang, Bo Chen, Yulei Wang, Kai Li, Chongyang Ren, Hsiaopei Mok, Li Cao, Lingzhu Wen, Minghan Jia, Cheukfai Li, Liping Guo, Guangnan Wei, Jiali Lin, Jiangguo Lai, Honglin Guo, Wenjing Wang, Shiyue Zhang, Zhijian Song, Jian Wang, Hui Chen, Jinwei Hu, Weifeng Wang, Weiwei Shi, Kai Wang. Comprehensive genomic analysis of Chinese breast cancer and clinical application [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 P4-07-03.

Background: Understanding the molecular influence of epidemiological risk factors can provide insights into breast cancer etiology and progression, and lead to better prevention efforts and treatment guidelines. The link between modifiable breast cancer risk factors and tumor genomic alterations remains unexplored. This study evaluated the association of tobacco use and alcohol consumption with somatic copy number variation (SCNV), total somatic mutation burden (TSMB), seven single base substitution (SBS) signatures (SBS1, SBS2, SBS3, SBS5, SBS13, SBS29, and SBS30) (1-3), and nine breast cancer driver mutations (CDH1, GATA3, KMT2C, MAP2K4, MAP3K1, NCOR1, PIK3CA, RUNX1, and TP53) (4,5), in a subset of The Cancer Genome Atlas (TCGA). Method: Clinical and genomic data were retrieved from the TCGA database, and breast cancer risk factor information was collected from four TCGA sites; in all, 219 female breast cancer cases were included. Pre-diagnostic tobacco use was defined as never or ever; alcohol consumption was defined as drinkers or non-drinkers. Total SCNV was obtained by summing all segments with ≥|0.2|. TSMB was calculated by summing all significant mutations for each case. SBS signatures were computed using whole exome sequencing data. Analyses were conducted in all tumors (i.e., pooled) and by stratifying according to estrogen receptor (ER) status using multivariate regression, adjusting for co-variables. Results: Tobacco users had higher TSMB compared to non-users (p Conclusion: This study provides preliminary evidence that tobacco use and alcohol consumption can influence breast tumor biology, in particular, DNA alterations. Future larger epidemiological studies are required to confirm these findings. References 1. Vineis P, et al. Mutational signatures associated with tobacco smoking in human cancer. Science. 2016;354(6312):618-22. 2. Alexandrov LB, et al. Clock-like mutational processes in human somatic cells. Nat Genet. 2015;47(12):1402-7. 3. Petljak M, Alexandrov LB. Understanding mutagenesis through delineation of mutational signatures in human cancer. Carcinogenesis. 2016;37(6):531-40. 4. Heng YJ, et al. The molecular basis of breast cancer pathologic phenotypes. J Pathol. 2017 Nov;241(3):375-91. 5. Ciriello G, et al. Comprehensive Molecular Portraits of Invasive Lobular Breast Cancer. Cell. 2015 Oct 10;163(2):506-19. Citation Format: Yujing J Heng, Susan E Hankinson, Jun Wang, Ludmil B Alexandrov, Christine B Ambrosone, Victor Piana de Andrade, Adam M Brufsky, Fergus J Couch, Tari A King, Francesmary Modugno, Celine M Vachon, A. Heather Eliassen, Rulla M Tamimi, Peter Kraft. Tobacco use, alcohol consumption, and breast cancer somatic genomic alterations [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 P5-08-07.

Background: Triple negative breast cancer (TNBC) is a highly metastatic and aggressive subtype of breast cancer which lacks receptors for-estrogen, progesterone and human epidermal growth factor 2. It accounts for approximately 10-20% of all breast cancer cases. It is commonly diagnosed in younger African-American women less than 40 years. There are no targeted therapies for this breast cancer subtype, therefore, chemotherapy and surgery remain the only options for treatment. However, there is a high chance of recurrence within three years of diagnosis and the majority of deaths occurs within the five years of diagnosis. High genomic aberration is a common event in this cancer where the differentially expressed genes in TNBC could be due to the variations in copy number of those genes. The dissemination of primary cancer cells to the lymphatic system represents one of the first signs of meta-static spread. In TNBC, any LN involvement is associated with worse disease-free and overall survival. Therefore, identification of copy number variations (CNVs) within LN metastases in this highly aggressive breast cancer subtype may serve as an indicator of prognosis. Objectives: The aim of the current study was to define regions of copy number gain or loss that are associated with metastasis to the lymph node in TNBC and their association with gene expression. Methods: Partek Genomic Suite was used to analyse the copy number variations in the study cohort containing 23 invasive ductal carcinoma (IDC), 12 lymph node metastasis (LNmets) and 3 normal adjacent tissues (NAT); as well as in validation cohort containing 70 IDC samples (Avery-Kiejda KA, Genom Data.2017;14:1-4).All the breast cancer samples were TNBC. The genomic segmentation algorithm was used for copy number detection. Moreover, their association with gene expression was also analysed using previous gene expression datasets (Accession No.GSE78758). Additionally, GO-enrichment and pathway analysis was performed to identify functional groups and pathways affected by copy number changes. Results: More CNV regions were amplified compared to loss in both the study and the validation cohort. The whole genome copy number profile was similar between IDC and lymph node metastases with 81% (263 in total) of amplified CNVs present in LNmets also present in matched IDCs, implying that many of the alterations in the primary tumour were carried over in metastasis. Very few genes were deleted and of those that were, only 2 were common to IDCs and LNmets. No significant CNVs were observed in NAT. Frequent amplification was mainly observed in chromosome 1q, 8q and 10p whereas the deletion in 4q, 5q and 14q was common in both IDC and metastasis samples. In total, 95 genes (66 gains and 29 losses) were only present in LNmet and not in IDCs from either the study or validation cohorts. According to GO-enrichment analysis, these LNmet associated genes were mainly enriched for DNA-binding transcription activity, cell fate specification and negative regulation of EGFR signalling pathway. Conclusion: This study has identified several regions of CNV in a well-defined TNBC cohort that could play major role in metastasis to lymph node. Citation Format: Mamta Pariyar, Andrea Mathe, Rodney Scott, Kelly Avery-Kiejda. Identification of copy number variation associated with lymph node metastasis in triple negative 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 P6-10-26.

Background: Phyllodes tumor (PT), a rare fibroepithelial neoplasm of the breast, poses a significant diagnostic challenge in breast pathology, as there are no generally accepted histological criteria or cutoffs to unambiguously differentiate PT variants (benign, borderline, malignant) and fibroadenomas (FA) of the breast. In recent years, genome-wide DNA methylation analysis has gained wide interest as an objective tool to distinguish all sorts of neoplasms based on their individual methylation signatures. For brain tumors, methylation-based classification already outperforms classification by conventional histology and has led to changes in the WHO tumor classification. Materials and Methods: We have so far analyzed fresh frozen as well as formalin-fixed, paraffin-embedded tissue samples of PT (n=31) and FA (n=2) on Illumina Infinium Methylation EPIC bead chips. The resulting data were preprocessed, normalized, mapped to the genome and converted into beta values. Top differentially methylated probes were determined by the calculation of standard deviations across the dataset. Such filtered sets of methylation beta values were then compared by a dimension reduction algorithm alongside TCGA methylome datasets as well as collections from GEO and the entire diagnostic database of our institute (n=14342). In parallel, the genome-wide copy number alterations were calculated from array data for each case. Results: In analogy to other tumor entities included in our reference cohort (brain tumors, sarcomas, carcinomas, melanomas), PT and FA of the breast show distinct DNA methylation patterns that contain a breast-specific signature as well as tumor-derived patterns. Their methylomes are distinct from breast adenocarcinoma as well as from healthy breast tissue. Two distinct groups within the PT/FA spectrum were identified so far: One group contains a high fraction of histologically malignant PT and is enriched for complex copy number alterations while the other group, also containing the currently two included fibroadenomas and, features less prominent to no copy number variants and a higher similarity to the overall normal breast tissue methylation signature. Conclusion: The distinct methylation signatures of the histological PT/FA spectrum will not only allow the diagnostic discrimination of PTs from histological mimics such as FA or carcinomas, but will also enable the distinction of benign, borderline and malignant PTs, thus paving the way to an optimized prognostic patient stratification and clinical management. Citation Format: Simone Muenst, Tatjana Vlajnic, Savas Deniz Soysal, Stefan Frank, Jürgen Hench. Distinct methylation and copy number alteration patterns in phyllodes tumors of the breast and its mimics [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 P1-21-02.

Introduction: Male breast cancer (BCa) is a rare disease accounting for less than 1% of all breast cancers (BC) and 1% of all cancers in males. The clinical management is largely extrapolated from female BCa. Few studies have examined the genomic landscape of male BCas, with six male BCas included in The Cancer Genome Atlas (TCGA). Familial studies of male BCas have shown genomic changes similar to female BCa, while a larger targeted sequencing study of 59 male breast cancers identified recurrent mutations affecting PIK3CA and GATA3. To date, there is still limited information regarding the genomic landscape of male BCas; particularly in the context of identifying targeted treatments. To reveal genomic changes that characterize male BCas in the context of known cancer driver genes linked to prognosis and targeted agents, we performed a targeted sequencing study on 248 male BCas from the International Male Breast Cancer Program. Methods: 248 primary M0, ER+ve, HER2-ve male BCas enrolled in the Part 1 (retrospective joint analysis) International Male Breast Cancer Program of 1483 patients diagnosed between 1990-2010 (Cardoso et al. Annals of Oncology, 2018) were processed for nucleic acid extraction from formalin-fixed paraffin embedded (FFPE) tissues. Using the Thermo Fisher Scientific Oncomine Comprehensive Assay v3 (OCAv3), a validated targeted sequencing panel currently used in the NCI-MATCH trial (NCT02465060), we evaluated mutational and copy number variations (CNVs) of genes that are prognostic or predictive to targeted therapies currently in use in the clinic or late-stage clinical trials. The OCAv3 DNA pan-cancer panel assays 115 genes for determining mutational status (48 full coding and 67 hotspot) as well as copy-number assessment in 43 genes. The OCAv3 uses Ampliseq-based technology linked to the Oncomine NGS workflow to identify actionable mutations and CNVs Results: Of the 248 samples assayed, 216 passed strict quality control parameters (87.1%). Using the Oncomine NGS workflow, actionable mutations at ≥5% variant allele frequency (VAF) were most frequently identified in PIK3CA (29.2%), BRCA2 (11.1%), NF1 (11.6%), indels in TP53 (10.6%), ATR (5.6%), ATRX (5.1%), indel BRCA2 (5.1%), TP53 point mutations (4.6%), MET (4.6%), ATM (4.6%), NOTCH2 (4.6%), CHEK1 (4.2%), FANCI (4.2%), PTEN (3.2%) with a number of additional genes identified at lower frequencies. Gene amplifications were most frequently detected in MYC (24.5%), FGFR1 (14.8%), CCND1 (12%), FGF3 (9.7%), FGF19 (9.7%), MDM2 (6.5%), CDK4 (1.4%), FGFR3, MDM4, ERBB2 (0.9% each), and FLT3, AR, MYCL, CDK6, IGF1R, FGFR4, KRAS, AKT3 and ESR1 (0.5% each). Although the results here describe the mutations and copy-number changes deemed to be actionable, further analysis of all non-actionable somatic mutations and CNVs will be presented and compared to female BCas previously assayed using the same panel. Conclusion: In this targeted sequencing study of the largest series of male BCas to our knowledge, we have revealed that PIK3CA continues to be a frequently altered gene in both male and female BCas. However, there is an enrichment of mutations in genes related to DNA repair in male BCs. Interestingly, while MYC is commonly amplified in female BCa, a higher frequency of amplified cases were seen in male BCas, in contrast to female BCas. Together with our previously generated transcriptional profiling data in this data set, we believe that both common and unique biological processes comprising male and female BCas will ultimately improve their clinical management and move towards the goal of precision medicine. This work has been funded by the Breast Cancer Research Foundation (BCRF). Citation Format: Jane Bayani, Coralie Poncet, Cheryl Crozier, Quang M Trinh, Megan Hopkins, Aime Lambert Uwimana, Tammy Piper, Carrie Cunningham, Monika Sobol, Stefan Aebi, Kim Benstead, Oliver Bogler, Lissandra Dal Lago, Florentine Hilbers, Ingrid Hedenfalk, Larissa Korde, Barbro Linderholm, John Martens, Lavinia Middleton, Melissa Murray, Catherine Kelly, Cecilia Nilsson, Monika Nowaczyk, Stephanie Peeters, Aleksandra Peric, Peggy Porter, Carolien Schröder, Isabel T Rubio, Kathryn J Ruddy, Christi van Asperen, Danielle Van Den Weyngaert, Carolien HM van Deurzen, Elise van Leeuwen-Stok, Joanna Vermeij, Eric Winer, Lincoln D Stein, Sharon H Giordano, Fatima Cardoso, John MS Bartlett. The genomic landscape of male breast cancers using the oncomine comprehensive assay for actionable mutations [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr P4-10-03.

Routine breast cancer screening has increased detection of pre-invasive lesions such as ductal carcinoma in situ (DCIS). Currently, DCIS cases are mostly treated with wide local excision combined with adjuvant radiotherapy and/or hormonal therapy, even though only 25% of DCIS treated with surgery alone recur with 50% of recurrences being invasive. At present, prediction of recurrence risk is based on histopathological features and patient characteristics but reliable identification of individual recurrence risk is still challenging. Certain genetic aberrations (i.e. copy number variation, somatic mutations) are associated with breast cancer progression. While most recurrences are assumed to be genetically related to the primary tumour (“clonal recurrence”), the prevalence of ipsilateral independent primary events (“non-clonal recurrence”) after treatment for DCIS is uncertain. If this is a substantial proportion of cases, their presence would compromise attempts to find tumour intrinsic biomarkers of recurrence. To enable the discovery of better recurrence biomarkers, this project initially aims to identify cases with primary DCIS and later clonal ipsilateral recurrence. Sections from formalin-fixed paraffin-embedded blocks from primary DCIS cases and their subsequent recurrences were obtained from Nottingham City Hospital (UK). Two pathologists reviewed representative sections to identify areas for microdissection. Tumour epithelial cells were needle-microdissected and DNA was isolated. Sequencing libraries were prepared using a) a custom capture panel of 110 breast cancer-related genes or b) NEBNext Ultra II for low-coverage whole-genome sequencing (LC-WGS) depending on DNA quality and quantity. Sequencing was then performed on Illumina NextSeq 500 system with an average coverage of a) 400x and b) 1.3x, respectively. Data was processed by an in-house bioinformatics pipeline to detect and filter for high confidence somatic variants. Copy number profiles were obtained using CopywriteR (panel off-target reads), PureCN (panel on-target reads) and Control-FREEC (LC-WGS). Of 33 patients initially ascertained on the basis of a recurrent breast event, 30 had an ipsilateral recurrent tumour after 4.7 years on average, and 24 primary/recurrence pairs had sufficient data available to determine clonality. Eighteen pairs (75%) were classified as clonal by the following criteria: occurring in similar locations, showing a similar or higher tumour grade, and sharing ≥ 3 copy number breakpoints or likely somatic variants in the genes included in the panel. Six pairs (25%) did not meet this criteria and were classified as non-clonal. Preliminary data analysis shows that invasive recurrences (n = 12) were detected on average 3 years later than DCIS recurrences (n = 6), despite constant surveillance. The most commonly observed somatic mutations occurred in the TP53 and PIK3CA genes. Genetic variants in TP53 and PIK3CA were equally common in primary tumours with clonal and non-clonal recurrence. Frequent copy number alterations involved losses on 16q as well as gains on 1q, 8q and 17q, whereat the 8q gain was less profound. The presence of 25% non-clonal recurrences shows that differentiating between clonal and non-clonal recurrences might be crucial for further recurrence risk biomarker discovery. Thus, this project will proceed with comparing genetic variation in primary DCIS with clonal recurrence to DCIS cases without later recurrence. These samples are currently being obtained and analysed. Citation Format: Dorothea Lesche, Hugo Saunders, Sakshi Mahale, Magnus Zenthoven, Siobhan Huges, Tanjina Kader, Niko Thio, Islam M Miligy, Jia-Min Pang, Andrew R Green, Emad A Rakha, Ian G Campbell, Stephen B Fox, Kylie Gorringe. Evaluation of genomic changes in ductal carcinoma in situ as potential biomarkers of recurrence risk [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 P6-10-21.

Background: Identification of genomic alterations present in cancer patients can aid in improving cancer diagnosis and prognosis. In this study, we aimed to identify clinically actionable variations present in stage IV breast cancer (BC) samples. Methods: DNA was extracted from formalin fixed paraffin embedded samples of breast cancer (n=41). DNA was sequenced using MammaSeqTM, a breast cancer specific NGS panel targeting 79 genes and 1369 mutations. Ion Torrent Suite 4.0 was used to make variant calls on the raw data. Resulting single nucleotide variants (SNVs) were annotated using CRAVAT toolkit. CNVkit was employed to identify copy number variations (CNVs). The PMKB and OncoKB Precision Oncology Databases were used to associate clinical significance with the identified variants. Results: A total of 41 BC patient samples were sequenced (read depth of 94 - 13340, median of 1529). In total, 96 different alterations (84 SNVs and 12 CNVs) were identified. Frequent SNVs were seen in PIK3CA, TP53, IKBKE, ATM and MAP3K1 genes. Moreover, amplifications were found in ERBB2, FGFR and AR, and deletions in BRCA2, RUNX1 and RB1. 3 CNVs (ERRB2, FGFR and AR amplifications) and 7 SNVs (IDH1.R132H, TP53.E204*, AKT1.E17K, PI3KCA.E545K, PI3KCA.H1047R, PI3KCA.R88Q and ERBB2.V777L) were reported as clinically significant by PMKB and OncoKB. Conclusion: We identified 96 alterations in 50 genes in 41 stage IV BC tissue samples using MammaSeqTM. Ten of these alterations were reported as clinically significant. This study highlights the potential use of cancer specific NGS panels in clinic to get better insight into the cancer patient-specific genomic alterations. Citation Format: Atilla Soran, Mustafa Sahin, Serdar Ugras, Esin Celik, Osama Shiraz Shah, Peter C Lucas, Adrian V Lee. Identifying genomic alterations in stage IV breast cancer patients using MammaSeq: An international collaborative study [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 P3-06-15.

Background: Germline copy number variants (CNV) of hereditary breast and ovarian cancer genes are a known class of clinically significant mutations. However, failure to detect CNV is a known limitation of Sanger sequencing and conventional amplicon next-generation sequencing (NGS). Multiplex ligation-dependent probe amplification (MLPA), while robust for CNV detection, is labor-intensive, costly to scale up and has limited breakpoint resolution at exon level only. We evaluated the clinical utility of molecular-barcode NGS and long-read nanopore sequencing in our genetic testing algorithm, specifically on CNV detection and breakpoint characterization of probands and family members. Methods: Both molecular-barcode NGS (BRCA1, BRCA2, TP53, PTEN, PALB2 and CDH1) and MLPA (BRCA1 and BRCA2) were performed in parallel on 200 high-risk breast and/or ovarian cancer probands and 11 CNV positive controls. CNV calling from MLPA and NGS data was performed by Coffalyser.Net and a novel in-house bioinformatics algorithm BAMClipper-SE, respectively. Nanopore amplicon or whole-genome sequencing was attempted to map the exact breakpoint of CNV controls and any newly identified CNV positive probands (up to May 2019). Personalized breakpoint PCR test for specific CNVs was designed and validated on probands and subsequently applied to corresponding family members for carrier testing. Results: Molecular-barcode NGS was 100% sensitive in detecting all BRCA1/2 CNVs from 11 positive controls. A total of 4 BRCA1/2 CNVs were detected by NGS from the 200 probands and matched the corresponding MLPA results. In addition, a PALB2 CNV (exons 4-6 deletion) was also detected by NGS during the parallel comparison. Up to May 2019, a total of 22 probands with detectable CNV in BRCA1, BRCA2, PALB2 or CHEK2 were accrued. Exact CNV breakpoint at nucleotide resolution was identified from 20 probands by nanopore and Sanger sequencing (BRCA1 n=12, BRCA2 n=5, PALB2 n=2 and CHEK2 n=1). High-resolution breakpoint mapping was applicable to inform variant pathogenicity and to enable recurrent CNV identification. Personalized breakpoint PCR designed for the 20 probands were applied to 42 family members, of whom 16 were CNV carriers and 26 were non-carriers. Breakpoint PCR results and any available MLPA results of these family members were 100% concordant. Conclusion: Molecular-barcode NGS allows a streamlined workflow to detect both sequence mutations and CNVs. It proves to be a reliable and cost-effective replacement of BRCA1/2 MLPA in genetic testing algorithm of new probands. Simultaneously the NGS workflow expands CNV detection coverage to additional genes (e.g. PALB2) without costly scale-up of MLPA gene panels. Nanopore sequencing is a practical alternative to MLPA in orthogonal validation of newly detected CNV. The high-resolution breakpoint mapping transforms CNV carrier testing to a personalized, simple and cost-effective breakpoint PCR. Citation Format: Ava Kwong, Chun H Au, Dona N Ho, Elaine YL Wong, Yvonne Chung, Fian BF Law, Cecilia YS Ho, Jiawei Chen, Isabella W Cheuk, Vivian Y Shin, Tsun L Chan, Hextan YS Ngan, Edmond SK Ma. Personalized and cost-effective detection of copy number variants by molecular-barcode next-generation sequencing and long-read nanopore sequencing [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 P6-08-13.

Background and Objectives HRpos, HER2neg early breast cancer is molecularly heterogeneous. Classifications that integrate genetic, genomic and clinical-pathologic data (nodal status, tumour size, tumour differentiation, Ki-67 status, PR expression) are lacking. In a previous study we have defined three different molecular groups of invasive ductal luminal breast cancer (IDLBC) using clinical prognostic variables, gene expression, copy number variation and mutation data from the METABRIC series [1, 2] and showed that IDLBC groups (IDLBC1, -2, -3) had distinct prognostic outcomes [3, 4]. The 3 groups differ by DNA damage load, predominant lesion type, characteristic lesion patterns, oncogenic driver mechanisms and time to first recurrence. Here we evaluated the prognostic and predictive value of these groups on patients treated on the BIG1-98 study. Methods We used our predefined subgroups (IDLBC1, -2, -3) to classify patients in the BIG1-98 set [5, 6]. A weighted Cox proportional hazards regression model was used to assess the association between IDLBC subtype, treatment on OS and DRFI endpoints in the BIG1-98 study. Weighted chi-squared tests were used for categorical variables and weighted t-tests for mutational burden. All statistical computations were carried out in R v.3.2.3. Results Five hundred thirty eight breast cancers (ductal 75%, lobular 17%, 8% other) were included in this study. The DNA damage burden of IDLBC1 or -2 type tumours was significantly lower compared to IDLBC3 (mean lesion numbers: 9, 11, 18.5). IDLBC1 type tumours (48%) harboured mainly oncogenic point mutations (predominantly PIK3CA 51%, MAP3K1 27%, NCOR1 22%, FANCD2 20%). IDLBC2 (34%) and IDLBC3 type tumours (18%) were both mutated (PIK3CA 53% vs 38%, GATA3 12% vs 30%, TP53 10% vs 37%, CDH1 20% vs 8%), 8p11-12 (IDLBC3 38% vs IDLBC2 21% or IDLBC1 5%, p value < 0.001) and/or 11q13-14 amplified (IDLBC3 46% vs IDLBC2 28% or IDLBC1 2%, p value < 0.001). Consistent with the higher DNA damage load patients with IDLBC3 type tumours had significantly shorter distant recurrence free intervals (DRFI) (HR (95% CI) 2.35 [1.99; 2.78] p-value= Conclusions We have validated our integrated clinical-pathologic and genomic classification method on an independent dataset of HRpos, HER2neg early stage breast cancer. Our classification also suggests differing magnitude of benefit for adjuvant Letrozole benefit. Prospective clinical studies will be required to validate and corroborate this novel classification approach. References 1. Pereira, Nat Commun (2016) 2. Curtis, Nat (2012) 3. Demanse, ENBDC Workshop (2018) 4. Demanse, manuscript in preparation (2018) 5. Loi, SABCS (2016) 6. Luen, JAMA Oncology (2018) Citation Format: Demanse D, Luen SJ, Pinet D, Regan M, Kammler R, Thuerlimann B, Viale G, Colleoni MA, Loi S, Hackl W. Analysis of DRFI and OS of HR-positive, HER2-negative breast cancer patients treated on the BIG1-98 study, classified according to a novel, integrated, clinical-pathologic and genomic classification method [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 P4-14-02.

Background: Chemotherapy can alter tumor cell populations by exerting selection pressure. This study was performed to evaluate the change in Shannon diversity index for c-MYC copy number variation after primary systemic therapy (PST) and its clinical implications in breast cancer patients treated with PST. Materials and Methods: One hundred and forty-four breast cancer patients who had residual disease after anthracycline- or anthracycline and taxane-based PST were included in the study. Associations between Shannon index for c-MYC copy number variation in pre- and post-PST breast cancer samples and clinicopathologic features of tumors as well as patient survival were analyzed. Results: Among the 119 patients with both pre- and post-PST samples available for comparison, 17 (14.3%) underwent change in c-MYC amplification status and 40 (33.6%) in c-MYC copy number gain status with most cases showing positive to negative conversion. In the whole group, Shannon index for c-MYC copy number variation was decreased in post-PST specimens compared to pre-PST specimens. Especially, the chemo-responsive group showed a more significant decrease in Shannon index after PST. However, there was no difference in diversity indices between pre- and post-PST specimens in the chemo-resistant group. c-MYC copy number gain and high Shannon indices for c-MYC copy number variation in both pre- and post-PST specimens were associated with adverse clinicopathologic features of breast cancer. In survival analyses, high Shannon indices for c-MYC copy number variation in post-PST samples as well as pre-PST samples were revealed as an independent prognostic factor for decreased disease-free survival. Upon subgroup analysis according to hormone receptor status, high Shannon indices before and after PST were associated with adverse clinical outcome in hormone receptor-positive group but not in hormone receptor-negative subgroup. Conclusions: These results suggest that a change in Shannon index for c-MYC copy number variation after PST reflects chemo-responsiveness and that Shannon index after PST can be used as a prognostic factor in breast cancer patients who receive PST. Citation Format: Kim G, Chung YR, Kim HJ, Kim M, Ahn S, Park SY. Clinicopathological significance of diversity index for c-MYC copy number variation after primary systemic therapy in breast cancer [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 P1-15-22.

Background: In depth sequencing of primary breast cancer (BC) has identified a heterogeneous repertoire of disease drivers, evidence of clonal evolution and underlying mutational processes. As cancer evolves over time and under treatment pressure, whole genome sequencing (WGS) of metastatic BC (MBC) tissue is crucial to gain further insight into its genetic make-up and driving forces, thereby allowing improved patient management. Methods: Metastatic tissue and matched germline DNA of patients with MBC (N=501) were prospectively recruited under the biopsy protocol of the Center of Personalized Cancer Treatment (CPCT-02; NCT01855477) and analyzed by WGS. Sequence reads were mapped to the reference genome to call somatic single nucleotide variants (SNV), small insertions and deletions (InDels) and copy number variations from which mutational signatures and tumor mutational burden (TMB; the number of SNV and InDels relative to the genome) were derived. The incidence of called aberrations in our cohort was compared to previously reported WGS data of 560 primary BC (BASIS cohort, Nik-Zainal et al. Nature 2016) (Table 1). Results: According to routine diagnostics 303 patients (60.5%) were ER+/HER2-, 70 (14%) triple negative (TNBC), 95 (19%) HER2+ and the remaining 33 (6.6%) had as of yet an unknown subtype. Top 5 recurrently affected genes were TP53 (47%), ATM (33%), MAP2K4 (32%), NCOR1 (31%), ERBB2 (30%). In the metastatic lesions, median TMB was 2.9 per million base pairs (IQR: 1.7-5.3). Interestingly, 53 (11%) patients had a high TMB (≥10). Compared to primary BC (BASIS cohort), we found (subtype-specific) enrichment of alterations in multiple genes such as ATM (0.4% to 33%), GPS2 (1.3% to 29%), MAP2K4 (6.4% to 32%), CBFB (2.7% to 25%), and, as previously reported, ESR1 (1.3% to 20%). APOBEC signature mutations appeared to be enriched in MBC while HRD signature mutations seemed less prevalent. Analyses to reveal additional genomic features is ongoing as well as the association of genomic alterations with uniquely collected information on prior treatments and response to treatment received directly after biopsy (i.e. endocrine therapy alone or combined with CDK4/6 inhibitors and chemotherapy). Also, to exclude methodological bias the raw data of the BASIS cohort will be processed through our pipeline. Table 1:Comparison of ER+/HER2- and TNBC subtypes: BASIS versus CPCT-02 MBC BASISCPCT-02 MBC ER+/HER2-(%)TNBC(%)ER+/HER2-(%)TNBC(%)Number of samples32615530370Median TMB0.962.632.732.91SNV burden/Mbp0.892.52.412.64InDel burden/Mbp0.060.120.210.3Top 5 affected genesPIK3CA (37)TP53 (83)TP53 (43)TP53 (61) TP53 (20)PTEN (37)ATM (43)MYC (37) CCND1 (20)MYC (26)MAP2K4 (37)CDKN2A (33) GATA3 (15)RB1 (24)NCOR1 (35)CDKN2B (33) MYC (14)PIK3CA (17)CDH1 (34)RB1 (33)Mutational signatures (median relative contribution) Age23.27.213.711.8APOBEC6.97.814.68.3Homologous-recombinant deficiency4.439.94.919.8DNA mismatch Repair Deficiency0.600.30 Conclusion: WGS of this unique cohort of patients with MBC shows a genetic make-up roughly similar to primary BC, but does show subtype-specific enrichment of selected driver mutations in metastatic disease. This study provides better insight into the tumor biology of MBC potentially improving management of these patients. Citation Format: Angus L, Wilting SM, van Riet J, Smid M, Steenbruggen TG, Tjan-Heijnen VC, Labots M, van Riel JM, Bloemendal HJ, Steeghs N, van de Werken HJ, Lolkema MP, Voest EE, Jager A, Cuppen E, Sleijfer S, Martens JW. The genomic landscape of 501 metastatic breast cancer patients [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 GS1-07.

Somatic copy number variations (CNV) may drive cancer progression through both coding and noncoding transcripts. However, noncoding transcripts resulting from CNV are largely unknown, especially for circular RNAs. By integrating bioinformatics analyses of alerted circRNAs and focal CNV in lung adenocarcinoma, we identify a proto-oncogenic circular RNA (circPRKCI) from the 3q26.2 amplicon, one of the most frequent genomic aberrations in multiple cancers. circPRKCI was overexpressed in lung adenocarcinoma tissues, in part due to amplification of the 3q26.2 locus, and promoted proliferation and tumorigenesis of lung adenocarcinoma. circPRKCI functioned as a sponge for both miR-545 and miR-589 and abrogated their suppression of the protumorigenic transcription factor E2F7. Intratumor injection of cholesterol-conjugated siRNA specifically targeting circPRKCI inhibited tumor growth in a patient-derived lung adenocarcinoma xenograft model. In summary, circPRKCI is crucial for tumorigenesis and may serve as a potential therapeutic target in patients with lung adenocarcinoma. Significance: These findings reveal high expression of the circular RNA circPRKCI drives lung adenocarcinoma tumorigenesis. Cancer Res; 78(11); 2839–51. ©2018 AACR.

Introduction: Poziotinib is a novel, pan-HER kinase inhibitor which showed potent anti-tumor activities through irreversible inhibition of HER family tyrosine kinases in preclinical and early clinical studies. Recent the open-label, multicenter phase II trial of poziotinib monotherapy evaluated that poziotinib is a new promising option for patients with HER2-positive metastatic BC who have failed more than two HER2 targeted therapy (NCT02418689). We evaluated genetic profiles of HER2-positive metastatic BC and investigated potential biomarkers of poziotinib for HER2-positive metastatic BC (MBC). Methods: All participants were diagnosed as HER2-positive BCs according to American Society of Clinical Oncology/College of American Pathologists HER2 guideline and provided tissue specimens that would be possible to extract DNA and RNA for next generation sequencing. We performed targeted deep sequencing with a customized 381 cancer gene panel (CancerSCAN™) and analyzed the relationship among the sequencing data, immunohistochemistry and clinical outcome. Results: From Apr 2015 to Feb 2016, 106 patients were enrolled in the trial from 7 institutes in Korea. Of 106 patients, biomarker data were available for 79 patients. TP53 was the most frequently mutated gene (70.8%) followed by PIK3CA (45.6%). HER2 single nucleotide variant (SNV) was detected in 13 BCs (16.5%) and HER3 SNV was in 9 (11.4%). The score of HER2 immunohistochemistry (IHC) was 3+ in 68 BCs and 2+ with positive in situ hybridization in 11 BCs. In copy number variant (CNV) analysis, HER2 amplification (86.1%) was most frequently observed and followed by CDK12 amplification (58.2%) and APOBEC3B deletion (30.4%). IHC score of HER2 was positively correlated to copy number (CN) of HER2 (P=0.001) but 11 breast cancer tissue did not have copy number amplification of HER2 (13.9%) (Six of HER2 IHC score 2+ and 5 of 3+). The median progression free survival (PFS) was 4.04 months (95% CI, 2.96 - 4.40) for patients who treated with poziotinib in this study. PIK3CA activating mutations were associated with short PFS compared to wild type (WT) and other SNVs (Median PFS of activating mutations vs. WT and others: 2.66 vs. 4.40 (months), P=0.009). HER2 CN amplification was positively correlated to duration of PFS (Median PFS of no amplification vs. 4 ≤ CN < 16 vs. 16 ≤ CN: 2.56 vs. 3.02 vs. 4.86 (months), P=0.032). HER2 SNVs prolonged duration of PFS without statistical significance (Median PFS of HER2 SNVs vs. WT: 4.24 vs. 3.19 (months), P=0.114), but 10 of 13 BCs with HER2 SNV (76.9%) had clinical benefit from poziotinib and 5 BCs (38.5%) had durable response more than 6 months. Conclusion: In this biomarker analysis, SNV of HER2 was frequently observed in HER2 positive MBCs and HER2 CN amplification was detected not in all. High CN amplification of HER2 derived longer PFS than those with low CN. To contrary to this, activating PIK3CA mutations shorten PFS compared to those with WT. In addition, HER2 SNVs might be a potential biomarker of poziotinib in HER2-positive MBC. Further functional study would be warranted. Citation Format: Kim J-Y, Lee E, Park K, Jung HH, Park W-Y, Lee K-H, Sohn JH, Lee KS, Jung KH, Kim J-H, Lee KH, Im S-A, Park YH. Molecular alterations and poziotinib, a pan-HER inhibitor efficacy in human epidermal growth factor receptor 2(HER2) positive breast cancers: Combined exploratory biomarker analysis from phase II clinical trial of poziotinib for refractory HER2 positive breast cancer(BC) patients [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 P2-09-21.

We are reconstructing the dynamics of tumour growth, treatment response, and origins of resistance of the Breast International Group (BIG) 1-98 study to determine the evolutionary paths of non-responders. We established the somatic alteration landscape of primary tumours collected at surgery from BIG 1-98 patients using targeted DNA sequencing of 287 genes of FFPE samples from 538 patients. Combined analysis of genetic and clinical data indicates increased genomic instability; TP53 mutations and copy number variations of 11q13 and 8p11 are associated with poor prognosis. We developed an in silico simulation framework to investigate if the non-responder outcome is due to (over)-treatment or intrinsic to evolutionary selection prior to diagnosis. To reconstruct the evolutionary history of tumours, we created a computational multitype branching process that tracks the expansion of diverse clonal lineages as they acquire driver and passenger mutations that alter their proliferation and mutation rates. To account for the heterogeneity between patients, we created a fitting procedure based on the Cramer-von Misses statistic to find the likelihood of parameters of our computational model that recreate the mutational landscape and clinicopathological factors observed in each patient. Once the fitting procedure is done, we simulate the course of treatment following the study arm scheme accounting for the cell-cycle action mechanism of Tamoxifen and Letrozole. Using our tool, we are characterising the range of tumour development scenarios covering different degrees of aggressiveness and genomic instability. Our computational model allows simulation of diverse adjuvant-schemes to predict optimised treatment regimes for validation in patient-derived tumour xenograft mouse models. Citation Format: Lara-Gonzalez LL, Loi S, Ferrari D, Papenfuss A, Goode DL. Reconstructing the evolutionary paths of BIG 1-98 [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 P5-08-01.

Liquid biopsy has become increasingly important in cancer diagnosis, personalized medicine, and disease progression monitoring. Conventional liquid biopsy relies on cancer gene panels which often contain fewer than 500 genes. Despite the revolutionary impact it has brought to the cancer research and cancer patient care, gene panels often miss many key mutations involved in cancer development and drug response. To fully capture the genetic variations embedded in cfDNA, we have developed Predicine-cfWES assay, which expands the features of our existing 152-gene PredicineCARE and 600-gene PredicineATLAS NGS panel with additional coverage of coding regions of the entire human genome. With 2500 x sequencing depth, Predicine-cfWES assay detects common cancer variants at sensitivity of 1% allele frequency and precisely measures copy number variations, such as loss of tumor suppressor genes (PTEN, BRCA2, p53) and gain of oncogenic driver genes (AR, Myc, ERBB2). Variants detected by Predicine-cfWES assay was confirmed by PredicineCARE panel based NGS assay and further used in tumor-agnostic personalized MRD assay. In conclusion, we have developed a comprehensive liquid biopsy solution to profile cfDNA with broad coverage of ~20,000 genes in the whole exome and higher sensitivity in a focused set of clinically actionable cancer variants for more personalized MRD monitoring in cancer patients. Citation Format: Amy Xiaohong Wang, Zhixin Zhao, Chao Dai, Kemin Zhou, Shidong Jia, Pan Du, Shujun Luo. Comparative analysis of cell free DNA with whole-exome sequencing (cfWES) and application to personalized minimal residual disease (MRD) monitoring [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 578.

Background: Gastric adenocarcinoma (GAC) is a deadly disease in need for new therapies. Recently, the most exciting developments in oncology are immune checkpoint inhibitors (ICIs) such as inhibitors of programmed cell death protein 1 (PD-1) which has shown response rates of 11-16% in unselected GAC patients. This study aimed to characterize the distributions of immunotherapy-related biomarkers in Chinese patients with gastric adenocarcinoma. Methods: Formalin-fixed, paraffin-embedded (FFPE) tumor samples and matched peripheral blood samples of 497 Chinese gastric adenocarcinoma patients were collected for YuansuTM NGS-panel assay (OrigiMed, Shanghai). Genomic alterations (GAs) including single nucleotide variations, short and long insertions/deletions, copy number variations, and gene rearrangements were assessed. Microsatellite instability (MSI), PD-L1 expression and tumor mutational burden (TMB) were assessed in 467(94%), 208(42%) and 467(94%) patients, respectively. Results: There were 334 males (67.2%) and 163 females (32.8%) with a median age of 60 (range 23-87) years in this cohort, including 440 primary tumors and 57 metastases. MSI-H was detected in 39 samples (7.8%). The rate of MSI-H in the ACVR2A mutation group was significantly higher than that in the ACVR2A wild-type group (82.1% vs 17.9%, p Conclusions: We analyzed the distribution of microsatellite status, PD-L1 expression and TMB in Chinese patients with GAC. Interestingly, our data showed that mutated ACVR2A was significantly associated with MSI-H and TMB-H. In addition, approximately 8% of patients were identified harboring hyperprogression-related gene alterations. Therefore, the application of comprehensive genomic profiling might be necessary in guiding ICI treatments in GAC. Citation Format: Qun Zhao, Yuntong Guo, XiaoLong Cui, LiangBiao Peng, Junying Li, Fei Wang, Lin Zhang. Comprehensive analysis of immunotherapy-related biomarkers in Chinese patients with gastric adenocarcinoma [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 2375.

Wilms tumor is the most common childhood kidney cancer. 15% of patients with Wilms tumor have germline pathogenic variants in genes or regions such as WT1 or the 11p15 region. Variants in these regions can include structural or copy number alterations or alterations in methylation. In the majority of cases of Wilms tumor no known pathogenic variant could be found using the state-of-the-art technologies, including comprehensive approaches such as Illumina whole exome sequencing. One explanation for this is that such technologies have difficulties in detecting structural variants (SVs) in areas associated with repeat or low complexity sequence. In addition, Illumina technology does not immediately support direct methylation detection. Therefore, we hypothesized that analysis of bilateral Wilms tumor of unknown etiology using long-read sequencing could reveal molecular events of potential clinical interest. We performed in-depth genomic analysis on a whole blood DNA sample from a patient with a bilateral Wilms tumor. This patient had no significant family history of cancer, and previously tested negative for Beckwith-Wiedemann syndrome by methylation testing of the 11p15 region; clinical exome sequencing of the patient's germline detected no variants associated with Wilms tumors. We sequenced the genome at 40x depth using PromethION Nanopore sequencing. 29180 SVs (deletions or insertions larger than 30 base pairs) were detected using Sniffles and 26480 were detected with SVIM. Only SVs that were detected by both methods were considered for downstream analysis. Variants were annotated and filtered using a short-read catalog of SVs (gnomAD-SV), a long-read catalog, the Database of Genomic Variants catalog, and by comparison to 11 in-house genomes. We focused on SVs, copy number variants, and methylation events affecting genes previously associated with Wilms tumor. Our long-read sequencing approach detected compound heterozygotes using phased variant calls. A heterozygous missense mutation was identified in haplotype one, while a 300 base pair insertion in an ALU element was present in haplotype two. These two compound heterozygous variants overlap an exon of the OVCH2 gene, and the ALU element was not detected by the prior Illumina analysis. Additionally, we determined the frequency of methylation in CpG sites genomewide using nanopolish. Using a normal blood sample from an unrelated individual as a control, we searched for extreme differences across large and gene promoter regions. Hypermethylation in the promoter regions of genes in the 11p15.5 locus was observed in the patient as compared to the control. Hypomethylation in this region is associated with Beckwith-Wiedemann syndrome. In conclusion, nanopore technology is able to detect variants missed by Illumina sequencing, and has the potential to yield new findings of interest in a case of a child with suspected cancer predisposition syndrome. Citation Format: Allison R. Cheney, Jean Monlong, Holly C. Beale, Hugh Olsen, Ellen Towle Kephart, Katrina Learned, Shanna White, Julian A. Martinez-Agosto, Noah Federman, Mark Akeson, Miten Jain, Vivian Y. Chang, Olena M. Vaske. Long-read sequencing characterization of a patient with bilateral Wilms tumor of unknown etiology [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 261.

Metastasis-Associated in Colon Cancer 1 (MACC1) is a strong prognostic marker for human colorectal cancers (CRC). The gene is a key regulator of the HGF/c-Met pathway and its overexpression is associated with cancer cell proliferation, migration, invasiveness and metastasis. The contexture of MACC1 overexpression in cancers is still poorly understood. Here we used the data released by The Cancer Genome Atlas (TCGA) and in silico approaches for a deep and integrative analysis of MACC1 expression modalities in CRCs. First, we curated COAD-READ TCGA datasets and obtained full clinical, genomic, and transcriptomic annotation for a total of 259 CRC samples. In tumors, we showed high heterogeneity of MACC1 mRNA expression levels with data varying from 5.58 to 13.07 U (RNAseq log transformed data - RSEM normalized). By correlating MACC1 expression to clinical data, we showed its levels were associated with anatomic regions of the disease (p=6.35.10-5). The tumors from descending colon and rectum having higher levels than those of the ascending colon and caecum (p=7.44-06). High MACC1 was also associated with higher tumor stages (p=0.01) and nodal invasion (p=0.03). At genomic level, MACC1 mutation was not frequent in CRCs ( normal in 155/259: 60%). Correlation analysis indicated strong positive association between MACC1 copy number variations and mRNA expression levels (p Citation Format: Vincent Vuaroqueaux, Hoor Al-Hasani, Dennis Kobelt, Thomas Risch, Susen Burock, Marie-Laure Yaspo, Heinz-Herbert Fiebig, Ulrike Stein. MACC1 overexpression in colorectal cancer is driven by chromosomal instability and is associated with molecular subtype [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 2879.

PARP7 (TIPARP) is a monoPARP which catalyzes the transfer of single units of ADP-ribose onto substrates to change their function. Its expression is upregulated during cellular stress, including viral infection or through the activation of the aryl hydrocarbon receptor after exposure to cigarette smoke. We and others have shown that PARP7 activity suppresses the Type I interferon (IFN) response following activation by cytosolic nucleic acid sensing pathways. RBN-2397 is a first-in class PARP7 inhibitor, which induces cancer cell autonomous and immune stimulatory effects in preclinical models through enhanced Type I IFN signaling in cancer cells. Here we describe the presence of PARP7 genomic amplification with corresponding increased mRNA expression across select cancers. Elevated PARP7 expression or amplification may identify cancer patients that could derive benefit from treatment with RBN-2397. In characterizing PARP7 copy number and mRNA expression from The Cancer Genome Atlas (TCGA) database, we found the presence of PARP7 copy number amplification in a subset of tumor types, particularly those of squamous histology, as well as ovarian cancer that corresponded to higher mRNA expression levels. High PARP7 expression correlated with poor survival in squamous cancers, while it had no effect on survival in ovarian cancer. Interestingly, tumor types with high PARP7 expression also expressed higher levels of baseline interferon stimulated genes (ISGs). This parallels our findings that cancer cell lines with higher ISGs at baseline are more responsive to PARP7 inhibition. To explore PARP7 copy number variations (CNVs) in advanced cancer patients, we queried the FoundationCore® (Foundation Medicine, Inc) database. Similar to TCGA, squamous cancers as well as ovarian, breast, and pancreatic ductal adenocarcinoma (PDAC) were among the tumor types with PARP7 amplifications. Moreover, PARP7 was found to be amplified both on the background of chromosome 3q arm-level amplifications as well as focally. Congruent to our analysis of PARP7 amplifications, we evaluated PARP7 mRNA expression in both squamous and non-squamous non-small cell lung carcinoma (NSCLC), breast cancer, and PDAC primary tumor samples. Using a validated RNAscope ISH probe set, we analyzed over 1,000 patient samples and found that PARP7 was more highly expressed in tumor cells relative to corresponding normal tissues. Overall, there were varying levels of PARP7 expression across samples with higher expression levels of PARP7 in tumor cells, compared to stroma, across all cancers examined. Our analyses describing the presence of PARP7 amplifications as well as high expression levels in several tumor types including NSCLC, breast, and PDAC provides evidence for the therapeutic relevance of PARP7 inhibition and highlights potential patient selection strategies to identify those patients more likely to benefit from RBN-2397 treatment. Citation Format: Jodie Wong, Kristy Kuplast-Barr, Ryan P. Abo, Anupriya S. Kulkarni, Linda T. Nieman, Azfar Neyaz, Katherine H. Xu, Radwa Sharaf, Lee A. Albacker, David T. Ting, Heike Keilhack, Kristen A. McEachern. Elevated PARP7 expression in select cancers identifies a target population for RBN-2397 therapy [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 381.

Background: Electronic cigarettes (e-cigs) are one of the most popular tobacco products in the US. Little is known regarding their pulmonary effects. E-cigs induce similar oxidant reactivity as cigarette smoke and promote oxidative damage/inflammation in airway cells. Given that mtDNA is more prone to oxidative stress than nuclear DNA because of a less effective proofreading system, mtDNA alterations may be important indicators of e-cigs' toxic effects. Clinically, mitochondrial DNA alteration is an emerging biomarker of respiratory diseases. Methods: We compared mtDNA copy number (mtDNA-CN) from lung brushings in a cross-sectional bronchoscopy study of healthy young adults, e-cig users (EC)(n=15), non-smokers, non-EC users (NS)(n=43), and smokers (SM)(n=26). We examined associations of mtDNA-CN with immune response (differential cell counts and cytokines in bronchoalveolar lavage), DNA methylation and gene expression brushings. Associations for: 1) EC vs NS vs SM, and 2) tobacco product users (EC+SM) vs NS for MtDNA-CN with immune response, methylation, and expression were made using linear regression. Further, significant features by group interactions were followed up by within-group tests. False Discovery Rate (FDR) at 0.1 was considered significant. Ingenuity pathway analysis was used to identify the most significantly enriched pathways/molecular functions/diseases. Results: MtDNA-CN was not significantly different among the three groups (P=0.06). MtDNA-CN was higher in SM than NS (P=0.02), and in tobacco product users than NS (P=0.02); EC mtDNA-CN tended to be intermediate between the 2 other groups. There were significantly positive associations of IL-2 and IL-4 with mtDNA-CN in EC, but not in SM or NS (Interaction FDR=0.06 for both). We found 147 transcripts (60 genes) and 1,153 CpGs (713 genes) to be significantly associated with mtDNA-CN in all three groups. The most common canonical pathway of the signatures for both expression and methylation were immune responses. The top molecular and cellular functions for both included cell death and survival. Ten transcripts (LINC01184, SNU13, RPL35A, COLCA1, HLA-DRB1, LOC105379655, TRIM9, TCIRG1, CLPB, MIR2114) and 3,929 CpGs (top: ULK4, STARD13, HLCS, FLT1, TMEM91, CYP2J2) were associated with mtDNA-CN in E-cig users only. Some of these genes are known to play a role in lung diseases, including cancer. For the signatures associated in all groups, we found many more significant signatures (236 vs 147 transcripts and 40,830 vs 1,153 CpGs) in the two group vs. three group comparisons, respectively. Conclusion: While the sample size was small, this study is the first to suggest that mtDNA-CN is a site of pulmonary toxic effects. We found associations of mtDNA-CN with inflammatory markers among EC users, and with a number of biological signatures, particularly genes related to immune response, in the lungs of EC, SM, and NS, but differently by groups for some. Citation Format: Kellie M. Mori, Joseph P. McElroy, Daniel Y. Weng, Sangwoon Chung, Sarah A. Reisinger, Kevin L. Ying, Quentin A. Nickerson, Theodore M. Brasky, Mark D. Wewers, Jo L. Freudenheim, Peter G. Shields, Min-Ae Song. Lung mitochondrial DNA copy number variations: E-cig users, smokers, and never-smokers [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2021; 2021 Apr 10-15 and May 17-21. Philadelphia (PA): AACR; Cancer Res 2021;81(13_Suppl):Abstract nr 751.

The presence of tumor-infiltrating lymphocytes (TIL) is a favorable prognostic factor in breast cancer, but what drives immune infiltration remains unknown. Here we examine if clonal heterogeneity, total mutation load, neoantigen load, copy number variations (CNV), gene- or pathway-level somatic mutations, or germline polymorphisms (SNP) are associated with immune metagene expression in breast cancer subtypes. Thirteen published immune metagenes correlated separately with genomic metrics in the three major breast cancer subtypes. We analyzed RNA-Seq, DNA copy number, mutation and germline SNP data of 627 ER+, 207 HER2+, and 191 triple-negative (TNBC) cancers from The Cancer Genome Atlas. P-values were adjusted for multiple comparisons, and permutation testing was used to assess false discovery rates. Increased immune metagene expression associated significantly with lower clonal heterogeneity estimated by MATH score in all subtypes and with a trend for lower overall mutation, neoantigen, and CNV loads in TNBC and HER2+ cancers. In ER+ cancers, mutation load, neoantigen load, and CNV load weakly but positively associated with immune infiltration, which reached significance for overall mutation load only. No highly recurrent single gene or pathway level mutations associated with immune infiltration. High immune gene expression and lower clonal heterogeneity in TNBC and HER2+ cancers suggest an immune pruning effect and equilibrium between immune surveillance and clonal expansion. Thus, immune checkpoint inhibitors may tip the balance in favor of immune surveillance in these cancers. Cancer Res; 77(12); 3317–24. ©2017 AACR.

Background: 24% of familial breast cancer (BC) and/or ovarian cancer (OC) cases analyzed within the framework of the German Consortium for Hereditary Breast and Ovarian Cancer (GC-HBOC) are due to pathogenic BRCA1/2 mutations. However, the mutation prevalence of non-BRCA1/2 genes associated with familial BC and/or BC/OC is largely unknown. Methods: Here, we present the first NGS data generated using the GC-HBOC-designed TruRisk® gene panel. In this study a cohort of 2028 BRCA1/2 and CHEK2 c.1100delC negative index cases was analyzed which comprises consecutive patients from BC families and BC/OC families complying the inclusion criteria of the GC-HBOC. Sequencing was performed on MiSeq, NextSeq, or HiSeq devices (Illumina) using customized SureSelect XT enrichment (Agilent). Data analysis was carried out using the SeqPilot software (version 4.2.2), SophiaDDM (Version 3.5.0.12-p5.0.0) as well as an in house bioinformatics pipeline (Cologne Center for Genomics, varpipe_v2.X). The analysis of copy number variations (CNV) based on NGS-data is currently in process and not yet included in the present mutation prevalence. Results: By focusing on 22 BC/OC associated genes (ATM, BARD1, BRIP1, CDH1, CHEK2, FAM175A, FANCM, MLH1, MRE11A, MSH2, MSH6, NBN, PALB2, PMS2, PTEN, RAD50, RAD51C, RAD51D, RINT1, STK11, TP53, XRCC2), we identified 71 different deleterious variants in 104 unrelated mutation carriers derived from 2028 BC and BC/OC families (8%). Interestingly, we identified a high prevalence of ATM mutations (n=29, 1.4%) in the familial cases. Additionally mutations in PALB2 (n=27), NBN (n=9), CHEK2 (n=14), BARD1 (n=9), BRIP1 (n=10), RAD51C (n=11) were frequently observed and we confirmed FANCM (n=17) as a novel BC predisposing gene. No mutations in MLH1, MRE11A, PTEN, RAD51D, STK11 and XRCC2 were identified in our collective. Conclusions: Due to the unexpectedly high mutation prevalence in familial cases, our study highlights the importance of these genes to be included in BC/OC routine diagnostics. In contrast we found low occurrence or absence of mutations for a subset of our gene selection which requires further investigation to optimize the gene panel for diagnostic purposes. Nevertheless this approach confirms the TruRisk® gene panel as a reliable tool for this comprehensive analysis. Citation Format: Keupp K, Richters L, Bülow L, Kröber S, Ernst C, Blümcke B, Versmold B, Waha A, Driesen J, Baasner A, Altmüller J, Thiele H, Nuernberg P, Wappenschmidt B, Neidhardt G, Rhiem K, Schmutzler R, Hahnen E, Hauke J. TruRisk® based next-generation sequencing in BRCA1/2-negative breast and ovarian cancer families reveal high mutation prevalence in additional risk genes [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 P3-08-01.

Background: Next generation sequencing (NGS) analysis of actionable molecular alterations has the potential to guide cancer treatment, especially for patients with advanced cancer who have progressed on standard treatment. In this study, we analyzed tumor biopsies and peripheral blood from 62 patients with advanced breast cancer by two different NGS clinical-grade assays for molecular alterations in tumor tissue or in circulating tumor DNA (ctDNA). We used these results to determine if these specimens have potentially “actionable” alterations that could guide cancer therapy. Methods: From 2014 to 2016, 62 patients with advanced breast cancer had plasma sent for ctDNA analysis (Guardant360 assay; 54 to 70 genes) Thirty-eight of these patients (61%) also had tumor biopsies evaluated by NGS (FoundationOne®; 182 to 315 genes). Alterations were defined as mutations, insertions, deletions, truncations, or rearrangements or amplifications/copy number variations. Patients that harbored multiple alterations in the same gene were not counted as having separate alterations; however, if a gene amplification and an alteration were found in the same gene these were counted as separate events. Variants of unknown significance (VUS) and synonymous mutations were excluded from both assays. Data were collected and analyzed according to a UCSD Institutional Review Board approved protocol. Results: The median age of our patients at the time of ctDNA analysis was 55 years (range, 44 to 84 years); the median age at the time of tissue biopsy for NGS was 52 years (range, 39 to 82 years). One patient was male. The most common receptor status was estrogen receptor (ER) and progesterone receptor (PR) positive, human epidermal growth factor receptor 2 (HER2) non-amplified or negative (neg)(N=44; 71%), followed by triple negative breast cancer (ERnegPRnegHer2neg) (N=10, 16%), triple positive (N=6, 1%) and finally ERnegPRnegHER2positive (N=2, 0.03%). One patient of 38 (2%) had no tumor alteration detected and 19 of 62 had no ctDNA alterations (31%). In 38 breast cancer patients with tumor NGS results, alterations were detected in 79 unique genes, with the most frequent being TP53 (37% of patients), PIK3CA (24%) and GATA3 (24%) genes. In the 62 patients with ctDNA analysis, 31 unique genes had at least one alteration, with the most frequent being TP53 (36% of patients) and PIK3CA (23%) and EGFR amplification (11%) (GATA3 was not analyzed in the ctDNA assay). Both assays had a high rate of detection for potentially actionable mutations: 41 out of 62 patients (66%) by ctDNA and 34 out of 38 (89%) by tumor NGS. No two patients harbored identical genomic profiles by either tumor NGS or plasma ctDNA analysis except for 1 patient who had no alterations detected by either assay. Conclusions: Plasma and tissue NGS analysis appear to be complementary assays that yield a high percentage of potentially actionable alterations in patients with advanced breast cancer. Studies of the clinical impact of NGS-guided therapy in breast cancer are warranted. Citation Format: Shatsky RA, Parker BA, Schwab R, Helsten T, Boles SG, Subramanian R, Piccioni D, Kurzrock R. Detection of molecular alterations in breast cancer through next generation sequencing of both tumor tissue and circulating tumor DNA: The UC San Diego Moores Cancer Center experience [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 P1-07-15.

Background: Result of recent advances in genetics, guanine nucleotide binding protein, alpha stimulating (GNAS), transducer of signals from G-protein coupled receptors, has been noted that this factor is related with the onset and progression of tumor in various cancers. We aimed to analyze gene amplification, mRNA and protein expression of GNAS and their potential association with clinicopathological factors and prognosis in primary breast cancer. Methods: The cohort of this study included 432 primary invasive breast cancer patients treated with standard care at Kumamoto University Hospital between June 2000 and January 2011. We performed a comprehensive analysis of GNAS at the levels of gene copy number, mRNA and GNAS protein expression analyzed by qPCR, qRT-PCR and immunohistochemistry (IHC), respectively. In the IHC assessment of GNAS protein expression, an H-score Results: The median age at diagnosis was 60 (range 27-93). Three hundred fifteen (72.9%) of these were postmenopausal women. One hundred forty two patients (32.9%) had axillary lymph node metastasis. The median Ki67 labeling index was 23.6 (range 0.5-97.0). The subtypes were 321 ER+/HER2-, 24 ER+/HER2+, 34 ER-/HER2+ and 53 ER-/HER2-. Three hundred twenty one patients (75.9%) were treated with endocrine therapy and 146 patients (34.6%) chemotherapy. Most notably, a low levels of GNAS protein expression was observed in 191(44.2%) patients, and was positivity associated with Ki67 (P=0.028). Furthermore, univariate and multivariate analysis revealed that low GNAS protein expression was significantly related with poor relapse-free survival rate (Log-rank test; P=0.0013, OR:0.40, 05%CI:0.22-0.70) and breast cancer specific survival rate (Log-rank-test; P=0.041, OR:0.43, 95%CI:0.19-0.97). GNAS amplification and mRNA expression were not correlated with prognoses. Conclusion: Contrary to expectations, GNAS expression was positively related with favorable tumor characteristics. Expression levels of GNAS protein may be an independent prognostic factor for primary breast cancer. Citation Format: Tomiguchi M, Yamamoto Y, Yamamoto-Ibusuki M, Goto-Yamaguchi L, Fujiki Y, Sueta A, Takeshita T, Iwase H. A comprehensive analysis of GNAS DNA copy number, levels of mRNA and protein expression in primary breast cancer [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 P6-09-46.

Background: Invasive lobular carcinoma (ILC) is the most commonly diagnosed special histological type of breast cancer, accounting for up to 15% of all cases. Accumulating data suggests the biology and clinical features of ILC differ to those of the more commonly diagnosed Invasive Carcinoma of No Special Type (IC-NST), including evidence of there being an overall worse long term outcome associated with ILC. Several large-scale molecular profiling studies have recently been published regarding ILC, highlighting molecular heterogeneity and important drivers of tumour behaviour. We hypothesize that an integrative analysis of gene expression and DNA copy number data will identify novel drivers, and prognostic and predictive biomarkers of lobular phenotype. Methods: Gene expression and copy number data from ILC tumours profiled in-house (n=25), METABRIC (n=125) and TCGA (n=145) were assembled and analysed. Integration of genome and transcriptome data was performed using two methods. Firstly, by Spearman correlation with a meta-analysis by combining gene level correlation coefficients from each study using a random effects model, weighting each study with inverse variance. Secondly using ANOVA, followed by combining gene level p-values across studies using Stouffers Z-score. Disease specific survival was examined at an individual gene level and using all genes simultaneously (via a Cox Boost analysis) to identify gene expression changes that are associated with poor outcome. Results: DNA copy number profiling identified recurrent gains (1q, 8q, 16p), losses (11q, 16q) and amplifications (1q32, 8p12-p11.2, 11q13). 11q13 amplifications were prevalent at a higher frequency in patients with poor outcome compared to patients with better outcome. The integrative analysis identified 1928 candidate genes whose expression was associated with gene copy number; as expected being enriched from genome regions highlighted above. One hundred and sixty of these genes were of prognostic significance in the METABRIC ILC cohort, several are known cancer drivers (e.g. PBX1, CCND1) and five new candidates are being investigated as novel prognostic biomarkers by immunohistochemistry. Conclusions: An integrated molecular analysis of ILC has identified a large number of gene expression changes that are dictated by gene copy number in ILC. Some of these are expected to influence tumour behavior and to highlight aggressive cancers with poor prognosis. Citation Format: Lal S, McCart Reed A, Nones K, Wockner L, Song S, Lakhani S, Simpson P. An integrated molecular analysis of invasive lobular carcinoma [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 P1-05-16.

Background: FGFR alterations are a known mechanism of resistance against breast cancer systemic therapies with higher FGFR RNA expression in breast cancer brain metastases compared to their primary tumors. While the genomic characterization of FGFR alterations has occurred from bulk breast tumors, the frequency and type of FGFR alterations in metastatic breast cancer (MBC) have not been fully characterized from cfDNA. The purpose of this study was to identify the incidence of FGFR1, FGFR2, and FGFR3 genomic alterations in cfDNA from patients with MBC and elucidate which FGFR alterations may increase FGFR kinase activity or may function as mechanisms of resistance against the FDA approved FGFR inhibitor, erdafitinib. Methods: We queried 16,053 reports from Guardant Health between June 2015 - October 2019 to identify the incidence of FGFR1, FGFR2, and FGFR3 alterations detected in cfDNA from MBC. We classified each alteration type into the following categories: copy number variation (CNV), fusion, indel, or single nucleotide variant (SNV). Focus was placed on characterizing FGFR1, FGFR2, and FGFR3 SNVs. We compared known activating mutations in EGFR, ERBB2, and BRAF with homologous regions in FGFR1, FGFR2, and FGFR3. In silico modeling with PyRx was used to dock erdafitinib onto FGFR1 (PDB 4V05), FGFR2 (PDB 5B7V), and FGFR3 (PDB 6LVM) kinases. Three-dimensional in silico analyses with ChimeraX was utilized to further determine which alterations may increase FGFR1, FGFR2, and FGFR3 kinase activity or may induce resistance against erdafitinib. Results: The incidence of nonsynonymous alterations occurring in FGFR1, 2213 (13.8%); FGFR2, 1017 (6.3%); and FGFR3, 144 (0.9%) were identified in the Guardant Health MBC database. FGFR1, FGFR2, and FGFR3 alterations are detailed in Table 1. We identified 40 (9.15%) and 167 (38.2%), 55 (7.2%) and 310 (40.4%), and 31 (27.4%) and 32 (28.3%) mutations occurring in the transmembrane/juxtamembrane and kinase domains of FGFR1, FGFR2, and FGFR3, respectively. Hotspot mutations were observed at R54C/S/G/R and N546K/D/S in FGFR1, D304N and N549K/D/S in FGFR2, and at S408/F/C/L/Y in FGFR3. We aligned FGFR1, FGFR2, and FGFR3 with homologous kinases and found 12 unique samples with mutations in our FGFR1 dataset corresponded to known activating mutations in EGFR and ERBB2; 18 samples with mutations in our FGFR2 dataset corresponded to known activating mutations in EGFR, ERBB2, and BRAF; and 1 mutation in our FGFR3 dataset corresponded to a known activating mutation in ERBB2. FGFR1 mutations C488S, V492M, M535I, L569V and FGFR2 mutations L550V, E565A, Y566N, S568C, G570R, G685E are postulated to induce resistance against erdafitinib. Conclusions: We found that 21% of MBC in this dataset harbor FGFR genomic alterations detected from cfDNA. Novel somatic alterations in FGFR1, FGFR2, and FGFR3 were identified from Guardant Health that were not detected in the public domain. A portion of FGFR SNVs occurred at known homologous kinase activating mutations in EGFR, ERBB2, and BRAF suggesting these specific FGFR mutations may increase FGFR kinase activity and might be actionable therapeutic targets in breast cancers harboring these mutations. Three dimensional analyses of the FGFR protein tyrosine kinase further illustrate which specific alterations may increase FGFR kinase activity or induce resistance against erdafitinib. Table 1. Incidence and type of FGFR alterations in Guardant Health MBC databaseGeneCNV (%)Fusion (%)Indel (%)SNV (%)FGFR11770 (80%)-6 (0.3%)437 (19.7%)FGFR2224 (22%)15 (1.5%)11 (1.1%)767 (75.4%)FGFR3-24 (16.7%)7 (4.9%)113 (78.4%) Citation Format: Mary Love Taylor, Benjamin Mayro, Leylah Drusbosky, Robin Batchelder, P. Kelly Marcom, Carey Anders, Jeremy Force. Identification of pathogenic FGFR1, FGFR2, and FGFR3 alterations in cell-free DNA (cfDNA) from patients with metastatic breast cancer [abstract]. In: Proceedings of the 2020 San Antonio Breast Cancer Virtual Symposium; 2020 Dec 8-11; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2021;81(4 Suppl):Abstract nr PS16-06.

Background: Mutations or epigenetic silencing of BRCA1/2 genes result in DNA repair deficiency in a large proportion of TNBC cases. Yet it is unclear whether this deficiency is associated with increased chemosensitivity and improved benefit from standard-of-care chemotherapy. We systematically evaluated BRCA deficiency in TNBC using integrated DNA and RNA sequencing data and its association clinical outcome, exploring the potential role of tumor immune response. Patients and Methods: Whole-exome, DNA methylation, copy number variation, and RNA sequencing data from 102 stage I-III TNBCs were retrieved from The Cancer Genome Atlas (TCGA). Almost all patients received adjuvant taxane-anthracycline-cyclophosphamide (T-AC) chemotherapy and had >30 days of follow-up. The number of predicted neoantigens and an estimate of the level of immune cell activity for 77 of these tumors were previously published. Deleterious germline or somatic BRCA1/2 mutations were identified by majority voting on predictions of 5 variant scoring algorithms. Definition of BRCA1/2 deficiency (BRCA-D) included carrier of deleterious BRCA1/2 mutations or BRACA1/2 normal (BRCA-N) with wild type BRCA1/2 expression less than the maximum observed in mutation carriers. Normalized genomic mutation rate and mutant allele tumor heterogeneity (MATH) were computed as broad measures of genomic instability. Characteristics of BRCA-D vs N tumors were compared using the Wilcoxon rank test. Results: Twenty tumors (19.6%) had mutations in BRCA1, 6 (5.8%) in BRCA2, and 2 (1.9%) in both. Based on the expanded definition, 39 cases (38%) were characterized as BRCA1 deficient, 5 (4.9%) as BRCA2 deficient and 4 (3.9%) as deficient in both. BRCA-D tumors (47%) were associated with a significantly higher mutation rate (P=8x10-4) but had similar clonal heterogeneity (P=0.55) as BRCA-N tumors. BRCA-D tumors had excellent 4-year overall survival (100%) compared to 79.5% (95%CI: 66.6- 94.9) for BRCA-N tumors (log-rank P=0.02). BRCA-D tumors also presented a significantly higher number of predicted neoantigents (P=0.003), which resulted in increased level of immune cell activity. In contrast, low immunogenic TNBC tumors were underrepresented in BRCA-D (p=0.05) and showed potential signs of immunoediting (observed/expected number of predicted neoantigens < 1; p=0.07). Conclusions: Deleterious mutations in BRCA1/2 genes occur in 25% of TNBC tumors, but parallel quantification of wild-type BRCA1/2 expression identifies 47% of TNBC samples with double strand break DNA repair deficiency. These BRCA-D TNBC tumors are characterized by a significantly higher mutation rate and present a significantly greater number of neoantigens that result in increased immune cell activity. Our analysis suggested that enhanced immune activation could explain to a large extent the excellent clinical outcome in patients with BRCA-D tumors treated with standard-of-care T-AC chemotherapy. Citation Format: Jiang T, Safonov A, Bianchini G, Shi W, Wali VB, Pusztai L, Hatzis C. DNA repair deficiency enhances immune response and correlates with excellent clinical outcome in triple negative breast cancer. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P4-07-01.

Introduction: Malignant phyllodes tumors are rare breast malignancies (0.1% of all breast tumors) with limited effective treatment options for recurrent and metastatic disease. Recent trials indicated a potential for anti-angiogenic therapy in soft tissue sarcomas, which led us to investigate these pathways. Materials and Methods: Thirty-five malignant phyllodes tumors (including two cases with matched primary and metastatic tumors) were profiled using gene sequencing (Next-generation and Sanger), gene copy number analysis (in-situ hybridization), whole genome RNA expression, and protein expression (immunohistochemical assay). Results: RNA microarray assay showed consistent over-expression of genes involved in angiogenesis including VEGFA, Angiopoietin2, VCAM1, PDGFRA, PTTG1, and CYP3A5 in all cases analyzed (n=5). No mutations in KDR (VEGFR2) were detected (0/26). EGFR protein overexpression was observed in 25/26 (96%) of cases with amplification of the EGFR gene in 8 cases (33%). EGFR gene mutations were identified in 2 cases (8%) including one case with presumed pathogenic V774M mutation and one case with EGFRvIII mutation. The most common mutations included those of TP53 (50%) and PIK3CA (15%) while other mutations (BRCA1, BRCA2, RET, CDH1, MLH1, ATM) were rare affecting single phyllodes cases. Two cases with matched primary and metastatic cancers harbored the same mutations in both sites (PIK3CA/KRAS and RB1 gene mutations, respectively). Conclusions: Comprehensive multiplatform profiling approach to phyllodes tumors identifies various molecular alterations of which some are potentially actionable. Our data suggests that anti-angiogenic therapy may also be effective in patients with malignant phyllodes tumor. Evaluation of EGFR pathway discovered consistent protein over-expression but rare activating mutations, which necessitates refinement in patient selection targeting these pathways. Citation Format: Gatalica Z, Vranic S, Ghazalpour A, Xiu J, Ocal I, McGill J, Bender R, Discianno E, Sanati S, Reddy S, Pockaj B. Comprehensive multiplatform molecular profiling identifies potentially targetable biomarkers in malignant phyllodes tumors of the breast. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P4-09-19.

Background: The Cancer Genome Atlas (TCGA) showed that copy number gain/amplification of FGFR1 was around 10% in primary breast cancer. FGFR1 gene amplification in breast cancer has been reported in some studies, more likely seen in ER-positive subtype. Several preclinical and clinical studies demonstrated that FGFR1 was one of novel targets of therapy for metastatic breast cancer. Previous studies suggested that aberrant FGFR1 expression was associated with poor prognosis, while there was no report that compared copy number aberration, mRNA and protein expression. The aim of this study is to analyze FGFR1 gene copy number, expression levels of FGFR1 mRNA and FGFR1 protein in ER-positive/HER2-negative primary breast cancer, and to examine the relationship between FGFR1 status and clinicopathological parameters including prognosis. Methods: The cohort of this study included 307 ER-positive/HER2-negative primary invasive breast cancer patients treated with standard care at Kumamoto University Hospital between June 2000 and January 2011. We performed a comprehensive analysis of FGFR1 at the levels of gene copy number, mRNA and FGFR1 protein expression analyzed by qPCR, qRT-PCR and immunohistochemistry, respectively. Results: FGFR1 gain/amplification was identified in 43 (14.0%) out of 307 patients. FGFR1 gain/amplification had significantly associated with higher nuclear grade (p=0.010). No correlations between FGFR1 mRNA expression levels and any clinicopathological factors were found. Expression levels of FGFR1 protein was positively associated with invasive tumor size (p=0.039). Modest positive correlations between these three (FGFR1 gene gain/amplification, expression levels of FGFR1 mRNA and FGFR1 protein) were found. The univariate analysis revealed that high FGFR1 protein expression was significantly related to poor prognosis (p=0.0019, HR: 2.63, 95%CI: 1.17-5.98) in terms of relapse-free survival (RFS) but not breast cancer-specific survival. The univariate analysis did not show that any factors except FGFR1 protein expression were significantly associated with RFS in this cohort. Conclusion: Expression levels of FGFR1 protein may be an independent prognostic factor in terms of RFS for ER-positive/HER2-negative breast cancer patients receiving standard care. Citation Format: Tomiguchi M, Yamamoto Y, Yamamoto-Ibusuki M, Yamaguchi R, Fujiki Y, Fujiwara S, Sueta A, Takeshita T, Inao T, Iwase H. FGFR1 protein expression is associated with prognosis in primary breast cancer: A comprehensive analysis of gene copy number, mRNA and protein expression. [abstract]. In: Proceedings of the Thirty-Eighth Annual CTRC-AACR San Antonio Breast Cancer Symposium: 2015 Dec 8-12; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2016;76(4 Suppl):Abstract nr P4-09-01.

Background: Poor prognosis in TNBC can be predicted in the significant fraction of patients with large volume residual cancer burden (RCB-II/III) after neoadjuvant chemotherapy (NACT). Whilst residual disease has been characterised to identify “driver” mutations and copy number variations, the contribution of the immune response within its tumour microenvironment remains unclear. Here we aimed to: 1) assess the potential spatial heterogeneity of immune transcript related gene expression between areas of tumour approximately 1cm apart as might still occur with a radiologically guided biopsy through the residual disease; and 2) assess the immune stroma composition of the TNBC high metastatic risk RCB II/III disease. Method: 12 TNBC post NACT RCB II/III residual cases were identified from the KHP biobank. H&E sections were reviewed and areas of tumor 1cm apart within a residual resection specimen marked as area A and area B. HistoQuest analysis software was used to quantify the proportion of tumor infiltrating lymphocytes (per total cell count) within both areas. RNA was extracted from both areas and immune gene expression profiling performed using a Nanostring nCounter® on all 24 samples. The immune PanCancer panel consisted of 770 genes combining markers for different immune cell populations. Differential genes between paired samples were compared and unsupervised hierarchical clustering using 770 genes and immune cell types performed. Results:Quantitative comparison of the tumour infiltrating lymphocytes (TILS) between area A and B revealed that 73% (8/11) of the cases had a 2-fold (range 2.03-3.16) difference in the TILS. When comparing the 770 gene expression profiles between sampling areas in the same tumour, we found little spatial heterogeneity with areas A/B clustering together in 10 out of 12 cases. Interestingly, the two cases that revealed spatial heterogeneity within the paired samples displayed little immune cell heterogeneity histologically (i.e. Conclusion: The findings that high metastatic risk residual disease can be further characterized as either “immunologically inert” or “immunologically enriched” at the level of extensive immunological transcript gene expression and by histological assessment of TILS requires further investigation; and is being validated in a larger sample set. Citation Format: Irshad S, Cheang M, Gazinka P, Naidoo K, Buus R, Pinder S, Dowsett M, Tutt A. Immune profiling of post neoadjuvant high metastatic risk (RCB-II/III) residual disease in patients with early triple negative breast cancers [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 P2-04-07.

The standard treatment for locally advanced rectal cancer is chemoradiotherapy followed by surgery. However, patient response to pre-operative chemoradiotherapy is highly heterogeneous, ranging from complete tumor regression to no response. Over the course of disease, up to 40% of patients suffer from local relapse or the development of metachronous metastatic disease, compromising survival. It is not yet known how treatment resistance and disease progression in these patients evolve – whether from the selective outgrowth of pre-existing resistant clones during treatment or through the acquisition of additional, resistance conferring genomic alterations. We aim to delineate the genomic evolution and clonal architecture underlying treatment resistance and metastatic progression in rectal cancer based on a cohort of 42 patients, from which 98 tumor samples were collected longitudinally over the course of treatment and metastatic progression. Tumor diagnosis and treatment response assessment are routinely based on the histopathologic evaluation of formalin-fixed paraffin-embedded (FFPE) patient tissues. While formalin fixation and paraffin embedding optimally preserve histomorphology and allow long-term tissue storage at ambient temperature, they, on the other hand, lead to DNA crosslinks posing a challenge for sequencing analyses. To overcome this limitation, we have successfully developed a protocol for dissociation of FFPE tissues into single cells that are suitable for immunolabeling, flow sorting and genomic analysis by whole exome sequencing of pure tumor cell populations, single cell copy number variation (CNV) sequencing and multiplex interphase fluorescence in situ hybridization (miFISH). Our genomic analyses at the population as well as individual cell level revealed divergent aberration patterns and distinct levels of intra-tumor heterogeneity. Major tumor clones were characterized by gains of EGFR (7p11.2), MYC (8q24.21), CDX2 (13q12.2) and ZNF217 (20q13.2) along with losses of SMAD4 (18q21.2) and TP53 (17p13.1). While in some patients the clonal composition remained largely stable throughout treatment, in others major clonal shifts occurred favoring clones with TP53 loss. Our preliminary results indicate different propensities of genetically distinct tumor cell clones in therapy response and metastatic progression. Furthermore, our data show the feasibility of high-resolution clonal reconstruction from whole exome sequencing and single cell CNV sequencing data of FFPE cells. Citation Format: Daniela Hirsch, Judith Lieberich, Kerstin Heselmeyer-Haddad, Yonca Ceribas, Michael Kelly, Keyur Talsania, Yongmei Zhao, Timo Gaiser, Thomas Ried. Genomic heterogeneity and clonal dynamics of resistance evolution and metastatic progression in rectal cancer [abstract]. In: Proceedings of the AACR Virtual Special Conference on Tumor Heterogeneity: From Single Cells to Clinical Impact; 2020 Sep 17-18. Philadelphia (PA): AACR; Cancer Res 2020;80(21 Suppl):Abstract nr PO-106.

Tumor evolution is a key feature intrinsic to tumor biology and contributes to intratumor heterogeneity, escape of immune surveillance, treatment failure, and patients’ prognosis. The evolutionary process of tumor is driven by selecting favorable phenotypes in terms of their fitness and survival in a tumor ecosystem. While genomic alterations provide rich materials for tumor evolution, only a few can induce a recognizable phenotypic change with even fewer for a fitness advantage. Thus, transcriptomics, a major molecular feature reflecting functional activities, will be informative in modeling tumor heterogeneity and crucial in understanding tumor evolution. Here, we aim to study tumor clonal evolution by single-cell transcriptomic profiling of hepatocellular carcinoma and intrahepatic cholangiocarcinoma from 37 patients participating the immune checkpoint inhibition trials. By analyzing core biopsies before or after treatment, we determined the single-cell atlas of liver tumors and confidently separated malignant cells and non-malignant cells by inferring chromosomal copy number variations. We developed a consensus clustering model based on machine learning algorithms and statistical methods to identify functional clones from malignant cells within each tumor. We further determined the clonal relationship by constructing the phylogenetic tree of the clones from all tumors. The clonal relationship within each tumor was independently assessed by manifold based single-cell trajectory and RNA-velocity based cell lineage. The analyses revealed a tumor branching evolutionary architecture of the clones. Noticeably, tumor branching evolution was associated with patient outcomes, which was also validated by using bulk transcriptomic data from 765 liver tumors. We found tumors in the poor prognosis branch were enriched in the pathways of hypoxia, epithelial-mesenchymal transition and angiogenesis. Remarkably, the functional role of the clones within a tumor varied, indicating a cooperative tumor cell community. We found a polarization of immune landscape associated with tumor branching evolution driven by tumor cell-specific cytokines. Our results offer insight into the collective behavior of tumor cell communities in liver cancer as well as potential drivers for tumor evolution in response to immunotherapy. Citation Format: Lichun Ma, Limin Wang, Ching-Wen Chang, Sophia Franck, Dana Dominguez, Marshonna Forgues, Julian Candia, Maria O. Hernandez, Michael Kelly, Yongmei Zhao, Bao Tran, Jonathan M. Hernandez, Jeremy L. Davis, David E. Kleiner, Bradford J. Wood, Tim F. Greten, Xin Wei Wang. Understanding tumor clonal evolution by single-cell transcriptomic analysis in liver cancer [abstract]. In: Proceedings of the AACR Virtual Special Conference on Tumor Heterogeneity: From Single Cells to Clinical Impact; 2020 Sep 17-18. Philadelphia (PA): AACR; Cancer Res 2020;80(21 Suppl):Abstract nr PR04.

Introduction: We describe the development and performance of a new sample-to-report targeted sequencing solution for testing solid tissue cancers using the Genexus Integrated Sequencing System and accompanying software. The assay is designed for research applications from either formalin fixed paraffin embedded (FFPE) solid tumor samples or cell-free total nucleic acid (cfTNA) from liquid biopsy samples. The Genexus Integrated Sequencer is a fully automated system requiring minimal touch points and hands on time allowing a novice user to go from nucleic acid to variant calls for somatic variant testing across multiple cancer types in less than two days. Methods: The Oncomine Precision Assay is a new amplicon-based assay targeting specific somatic variants in 50 genes with coverage for multiple cancer types. The assay uses AmpliSeq HD chemistry capable of distinguishing true sample biological variants from errors generated during library preparation, templating, and sequencing through incorporation of molecular tags during target amplification. With about 15 minutes of hands on time, a run is set-up using pre-filled reagent strips for a fully automated run that includes library prep, templating, sequencing, variant calling, and a final report if desired. Results: Reported here are the results generated from an early external test site along with development data. The Oncomine Precision Assay is designed to detect somatic variants in 50 unique genes testing all major variant types important in the oncology research. The content was selected based on published accounts of target actionability and prevalence across multiple cancer types. All major variant types are targeted including SNVs, insertions, deletions, copy number variation, fusion transcripts and alternate splice forms. Data is shared from an external test lab using the Oncomine Precision Assay on the Genexus Integrated Sequencer with control and research samples. Results from both multiplexed FFPE and liquid biopsy runs are presented. Data demonstrates use of a single assay and system to effectively call variants from both FFPE and liquid biopsy sample types with a turn-around time of less than 30 hours. Conclusion: The Oncomine Precision Assay and Genexus Integrated Sequencer enable detection of key oncology variants in 50 genes using either solid tissue or liquid biopsy samples as input. This fully automated solution for oncology research generates variant calls from nucleic acid input in less than 2 days with minimal hands-on time and touch-points from the user. Many features of the automated system increase success rates by ensuring the appropriate reagents are properly installed before a run. The user friendly, highly automated, and fully optimized sample to answer system described here has great potential for targeted oncology sequencing in the research setting. Citation Format: Jian Gu, Ru Cao, Jeff Schageman, Kris Lea, Priyanka Kshatriya, Amir Marcovitz, Paul Williams, Rasika Sunnadeniya, Varun Bagai, Khalid Hanif, Jose L. Costa, Kelli Bramlett. Demonstration of the genexus integrated sequencing system with the oncomine precision assay [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 213.

Background: The current standard of genomic profiling of cancer tissues relies upon multiple separate technologies to aid in tumor characterization. Here we describe an NGS panel and analysis workflow for cancer samples that can detect single nucleotide polymorphisms (SNPs), insertions and deletions (INDEL), somatic copy number alterations (SCNAs), and translocations (TLs), Tumor Mutation Burden (TMB) and Microsatellite Instability (MSI) in a single assay. Methods: We have developed a SureSelect target enrichment sequencing panels for comprehensive profiling of single nucleotide variations (SNP/INDEL), gene copy number variations (CNV), and DNA translocations (TL) and relevant microsatellite regions in a single assay. The TMB panel gene list was curated based on genes found in cancer gene databases such as Clinical Interpretations of Variants in Cancer (CIViC), Cancer Genome Interpreter (CGIdb), Catalog of Somatic Mutations in Cancer Census (COSMIC), Precision Oncology Knowledge Base (OncoKB). An accompanying data analysis pipeline will provide highly sensitive and accurate detection of variants and allows for the determination TMB and MSI status. Probes have been selected for over 500 genes to detect SCNAs and translocation in several cancer subtypes. The probes are optimized with the SureSelect XTHS protocol, which enables sensitive and accurate detection of rare mutation events within a heterogeneous sample with molecular barcode-mediated error correction. The SCNAs probes target regions in the genome that help in decreasing the noise for SCNAs calling. The data analysis utilizes either a matched normal or standard control DNA to compute read depth log ratios. Aberrant regions of constant ploidy are first identified followed by the application of a variational streaming algorithm on log ratio and SNP allele frequencies to determine the major clones present in the tumor. Translocation detection was verified on lung cancer samples by analyzing split reads across the fusion breakpoints. Results: By diluting samples with known copy-number aberrations into normal DNA, a limit of detection of ⇐2.3 copies/cell was demonstrated, equivalent to the detection of 3 copies at 30% tumor fraction. A similar approach demonstrated an ability to detect EML4-ALK and SLC24A2-ROS1 translocations at 3% allele frequency. Performance of the assay was further evaluated using 100+ FFPE samples harboring rearrangements, gene amplifications, and SNP/Indels. Concordance to the reference methods (FISH, WES, WGS) was >90%. TMB and MSI algorithms were benchmarked using in silico analysis of TCGA data and performance of the assay was evaluated on FFPE samples with orthogonally-determined MSI/TMB status. Conclusions: This work represents an important advancement in the development of a single assay to detect copy number variation, DNA rearrangement and mutations, TMB and MSI status of a FFPE sample. Citation Format: Arjun Vadapalli, Akanksha Khare, Hanjun Shin, Ashutosh Ashutosh, Linus Forsmark, Anne Lucas, Scott Happe, Gilbert Amparo, Carlos Pabon, Jayati Ghosh, Tracy Liu, Jimmy Jin, Mike Ruvolo, Douglas Roberts. SureSelect sequencing panels and algorithms to detect copy number variations (CNVs), DNA rearrangements, microsatellite Instability and tumor mutational burden (TMB) in FFPE specimens [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 184.

Background: Cancer immunotherapy, especially immune checkpoint inhibitors (ICIs), has a remarkable antitumor response in non-small cell lung cancer (NSCLC), but only a limited number of patients can derive durable benefits from it. Therefore, more accurate biomarkers are highly needed. STK11 is a well-known negative regulator for response to ICIs, but the mutation frequency of STK11 and its correlation with recognized immune biomarkers in Chinese patients with NSCLC remains unclear. Methods: Formalin-fixed, paraffin-embedded (FFPE) tumor and matched blood samples of 637 NSCLC patients from the OrigiMed were collected for targeted next-generation (NGS) panel sequencing from December 2017 to January 2019. Genomic alterations (GAs) including single nucleotide variations, short and long insertions/deletions, copy number variations, and gene rearrangements were assessed. Tumor mutational burden high (TMB-H) was defined as ≥10 muts/Mb. PD-L1 expression positivity was defined as ≥1% of tumor cells with membranous staining (22C3, DAKO). Genomic data and ICIs treatment outcome of one cohort of NSCLC patients were derived from cBioPortal database. Results: STK11 GAs were found in 6.9% (44/637) of patients in the OrigiMed database. No significant difference was observed in the frequency of STK11 GAs within the histologic subtypes. Sites of the STK11 mutations were found scattered throughout the gene. STK11 GAs were associated with a significantly higher TMB compared with wild-type (WT) STK11 (11.95 vs. 4.6 muts/Mb, respectively, p Conclusion: The results showed that in our cohort, Chinese NSCLC patients with STK11 GAs were enriched in PD-L1-neg/TMB-H subgroup and associated with a poor response to ICIs in Chinese NSCLC patients. We also found that co-mutations of STK11 and KEAP1 may have implications as a negative biomarker for guiding ICI treatment. Keywords: non-small cell lung cancer, FGFR4, tumor mutational burden, immunotherapy Citation Format: Youhua Jiang, Kaiyi Tao, Liang Wang, Yuqian Hu, Shiyue Zhang, Lin Zhang, Ming Yao. Mutational landscape of STK11, PD-L1 expression, and TMB in patients with non-small cell lung cancer [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 2021.

Introduction: NRF2/KEAP1 copy number variations and somatic mutations lead to NRF2 activation in a variety of human cancers. Though NRF2 plays a role in activating KEAP1 expression in turn, exact mechanisms and clinical implications remain unknown. Methods: We examined expression correlation between NRF2 and KEAP1 in RNA profiling databases (cBioportal, GEO) with focus on lung cancer. We performed qPCR and western blotting to measure KEAP1 mRNA and protein alterations during NRF2 activator or knockdown treatments. We also queried ENCODE and SRA for ChIP-seq datasets to identify NRF2 binding within KEAP1 promoters in human cells. To demonstrate the regulatory role of NRF2 binding on KEAP1 expression, we applied luciferase reporter assay and CRISPR editing to assess the ARE-related KEAP1 promoter activity and endogenous mRNA abundance change. To determine specimen-level implication of the feedback pattern, we tested gene expression ratio for predicting NRF2/KEAP1 copy number variations and somatic mutations. Results: RNA profiling datasets analysis showed that mRNA expression of KEAP1 was consistently in positive correlation with NRF2 in multiple primary squamous cell cancers, including LUSC (Pearson r=0.50, p Conclusion: NRF2-driven KEAP1 transcription suggested a calibrated NRF2 signaling transduction. Novel cancer prevention and therapy targeting NRF2 signaling will benefit from understanding this unique pattern. Citation Format: Yijun Tian, Qian Liu, Shengnan Yu, Qian Chu, Yuan Chen, Kongming Wu, Liang Wang. NRF2-driven KEAP1 transcription in human lung cancers [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 3587.

Background: Salivary adenoid cystic carcinoma (ACC) and oral tongue squamous cell carcinoma (OTSCC) are rare malignancies with low incidence and dismal prognosis. The molecular underpinnings of these remain poorly understood, particularly in Asian population. Methods: Through a whole-exome sequencing analysis of 13 OTSCC and 30 ACC patients, a systematic overview of the mutational features in single-nucleotide variations, copy number variations and chromosomal variations were illustrated. Results: In addition to the well-characterized alterations previously reported in Caucasian patients, such as TP53 mutations (11/13, 84.6%) in OTSCC and MYB-NFIB fusion in ACC (11/30, 36.7%), we also observed several novel genetic changes, including FOXP1-TEX261 fusion (1/13, 7.7%) and MYB-FMO5 fusion (1/30, 3.3%) in OTSCC and ACC respectively, which might play a role in facilitating tumor progression. The data exhibited promising translational value as numerous actionable genes were identified with mutations. 76.9% OTSCC and 20% salivary ACC harbored at least one alteration in actionable genes, mutations in CDKN2A (4/13, 30.8%), NTRK3 (3/13, 23.1%) in OTSCC and FGFR2 (2/30, 6.7%), BRAF (2/30, 6.7%) in salivary gland ACC, suggesting targeted immunotherapy might be an effective intervention. Instead of MYB-independent NFIB fusions previously reported in ACC, only NFIB-independent MYB fusions were found in our cohort, indicating that alterations in NFIB rather than MYB might be essential to ACC tumorigenesis. Conclusions: This study reveals the genetic underpinnings of two rare malignancies in Chinese population, illustrating consistent mutational features with published literature as well as demonstrating several distinct genetic alterations, warranting further investigations. Citation Format: Ning Li, Shuhang Wang, Yue Yu, Yuan Fang, Huiyao Huang, Dawei Wu, Hong Fang, Chao Sun, Anqi Yu, Qi Fan, Shi-Yuan Cheng. Whole-exome sequencing reveals genetic underpinnings of salivary adenoid cystic carcinoma and tongue carcinoma in Chinese population [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr LB-310.

Background: Hereditary cancer accounts for 5% to 10% of all cancers. In addition to sequence variants, copy number variants (CNVs) are a cause of inherited cancer syndromes. Next-generation sequencing (NGS) technology enables simultaneous interrogation of multiple genes in a cost-effective manner. However, accurate CNV detection by NGS remains challenging owing to bias and noise that distort the association between copy number and read-depth. We previously developed and validated an NGS-based CNV detection algorithm using 18 CNV-positive and 85 CNV-negative specimens; the algorithm yielded 100% sensitivity and 98% specificity. In this study, we evaluated the performance of the algorithm on a large group of clinical specimens. Methods: Retrospective CNV analysis was performed on 19,722 specimens submitted for a 34-gene cancer predisposition panel (n=5,726) or 4 sub-panels: BRCA1/2 (n=13,156), a 5-gene Lynch syndrome panel (n=553), a 13-gene colorectal cancer panel (n=205), or a 14-gene women's health panel (n=82). For NGS, target genes were captured using an RNA-bait hybridization method and sequenced on an Illumina NextSeq 500 instrument. CNV-positive specimens were identified using a CNV-flagging algorithm developed in-house that is based on the well-established NGS read-depth approach. All specimens flagged as having CNVs by NGS were reflexed to array comparative genomic hybridization (aCGH) using a custom-designed Agilent array for confirmation. Although highly reliable and widely used, aCGH can only simultaneously process a limited number of specimens. Results: Among the 19,722 clinical specimens, the NGS CNV flagging algorithm identified 485 CNVs; 235 were positive by aCGH and 250 were negative; thus, specificity of the algorithm alone was 98.7% (19,472/19,722) and improved to 100% when aCGH was incorporated. Of 235 aCGH-confirmed CNVs, 224 (95.3%) were pathogenic or likely pathogenic. Approximately 1% of specimens tested by a 34-gene panel (60/5,726) was CNV-positive. BRCA1 (20% of positives) was the most frequently mutated gene followed by CHEK2 (15%), ATM (10%), and PALB2 (10%). From the 4 sub-panels, the 5-gene lynch panel had highest positive CNV rate (4%, 22/553), followed by 13-gene colorectal panel (2%, 4/205), and BRCA1/2 test (1%, 149/13,156). Conclusions: The developed CNV-flagging algorithm enabled CNV detection of clinical specimens, in a variety of panels, with high specificity. Supplementation of aCGH confirmation improved CNV detection specificity to 100%. Our data underscore the important clinical utility of NGS-based CNV detection in the molecular diagnosis of hereditary cancer. Citation Format: Sun Hee Rosenthal, Ke Zhang, Yan Liu, Renius Owen, Alla Smolgovsky, Arlene Buller-Burckle, Felicitas Lacbawan. Highly specific copy number variant-flagging algorithm using next-generation sequencing in cancer-predisposition genes [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 5741.

Introduction POLE-mutated tumor is a rare subtype of colorectal cancer (POLE-CRC), whose pathogenesis has been poorly understood because of its rarity. In this study, we have investigated the clinical and genetic features of POLE-CRC in the largest cohort of this unique subtype of CRC ever analyzed. Methods We included a total of 3,265 patients who had been treated at the Cancer Institute Hospital of Japanese Foundation for Cancer Research between 2004 and 2017, which were screened for POLE mutations by targeted-panel sequencing. The POLE-mutated samples were further analyzed by whole exome sequencing. All samples were well-annotated for clinical information. Results In total, 44 samples showed prominent hypermutation with a median of 5,346 (range: 837-16,990) with the predominance of COSMIC signature 10 associated with defective POLE functions. By contrast, copy number variations (CNV) in POLE-CRC was much less common compared to non-hypermutated CRC. In accordance with previous reports, mutational hotspots in the POLE gene were found at codons 411, 286, 459, and 456. Patients with POLE-CRC were significantly younger (median of 50 years), compared with non-hypermutated CRC (median of 65 years, p Conclusions POLE-CRC is a CRC subtype with a favorable prognosis compared with other CRC subtypes and is characterized by high frequency of genetic alterations affecting APM, which is implicated in immune evasion of this unique subtype of CRC with an extremely high mutational burden. Citation Format: Yoshikage Inoue, Nobuyuki Kakiuchi, Kenichi Yoshida, Yasuhide Takeuchi, Yusuke Shiozawa, Yuichi Shiraishi, Kenichi Chiba, Tetsuichi Yoshizato, Hiroko Tanaka, Ai Okada, Satoshi Nagayama, Satoru Miyano, Yoshiharu Sakai, Seishi Ogawa. Frequent genomic alterations to evade the immune system in colorectal cancer with POLE gene mutation [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 3547.

Background: Prior molecular characterization efforts for high-grade serous ovarian cancer (HGSOC) were restricted to those who had upfront surgical debulking with variable treatment paradigms. Thus, we sought to examine molecular and cellular differences between clinically defined groups (tumor tissues from patients who had complete gross resection (CGR) versus those who were triaged to neoadjuvant chemotherapy (NACT) and experienced either excellent or poor response. Methods: Tumor biopsies were collected from three patient groups managed under a systematic surgical algorithm: CGR after primary surgery (R0, n=10); poor tumor response to NACT (NACT-PR, n=10); excellent tumor response to NACT (NACT-ER, n=10). Primary and multiple metastatic tumor sites per each patient were obtained pre-treatment and subjected to comprehensive omics analyses including high-pass whole-genome (WGS) and targeted deep DNA sequencing. Results: An average of 71 nonsynonymous somatic mutations from each sample for 75 samples with high-purity tumors (≥75%) by WGS were identified from each sample for the entire cohort. Fourteen ovarian-cancer-associated genes were found mutated in our patient cohort and, as expected, the most frequently mutated gene was TP53 in both primary and metastatic sites in all three groups. TP53 nonsense mutations were exclusively identified in the NACT-ER (36.0%) and NACT-PR groups (15.4%), while in the R0 group most TP53 mutations were missense mutations (62.5%). Nonsense mutations in CSMD3 and PIK3CA were exclusively identified in both primary and metastatic sites in the NACT-PR group. The most frequent copy number variations (CNVs) in the R0 were copy number gain/loss of CSMD3 (67%) and copy number loss of NF1 (54%) and CDK12 (50%) in both primary and metastatic sites. Interestingly, copy number losses of NF1 were significantly lower in the NACT groups (18%, p=0.002), especially in the NACT-PR (8%, p=0.0004), when compared to the R0 group (54%). We also identified significant less observation of chromothripsis-like patterns, and a significantly higher level of strong-binding neoantigens in the R0 than in the NACT groups. Conclusions: Our findings using HGSOC samples obtained from patients treated on a prospective algorithm identified distinct molecular abnormalities, and could have prognostic and therapeutic implications for patients with HGSOC. Citation Format: Sanghoon Lee, Li Zhao, Shannon N. Westin, Nicholas W. Bateman, Amir A. Jazaeri, Nicole D. Fleming, Karen H. Lu, Robert L. Coleman, Gordon B. Mills, Jianhua Zhang, Thomas P. Conrads, George L. Maxwell, P. Andrew Futreal, Anil K. Sood. High-depth whole genome sequencing of clinically-annotated high-grade serous ovarian cancers [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 2510.

Introduction: Many tumors display phenotypic similarity across tissue types, characterized by defects in biological pathways leading to susceptibility to particular therapies. Deficiencies in one such pathway, homologous recombination DNA damage repair (HR DDR), affect the response of breast and ovarian cancers to agents that inhibit PARP function. The Oncomine BRCA Expanded panel targets 15 genes (BRCA1 and 2, ATM, BARD1, BRIP1, CDK12, CHEK2, FANCD2, MRE11, NBN, PALB2, PPP2R2A, RAD51B, RAD54L, and TP53) that play various roles in the HR DDR pathway. The assay includes 1011 amplicons covering both hotspot variants and coding regions of these genes. An additional set of pre-qualified genes is available for assay customization, including a recommended set of 25 genes of particular relevance to HR DDR pathway. Methods: Targeted regions were selected, and amplicons were designed, optimized, and experimentally verified to perform well with semi-conductor based sequencing. Measures used included gene-base and hotspot-base coverage uniformity, variant calling accuracy, and detection of known CNV gains in both FFPE and cell line DNA samples. Large gene rearrangements in BRCA1 and BRCA2, including exon deletions and duplications, could be detected by using a custom exon deletion pipeline. Analysis was performed on an integrated bioinformatics platform, Ion Reporter, which includes variant calling, identification and annotation of the likely deleterious mutations, and production of configurable reports, including visualization of the extent of copy number changes. Results: The assay's performance was evaluated using a variety of sample types, including commercially available controls and FFPE samples for breast, ovarian, prostate, and other tumor types. Sensitivity and positive predictive value for small variant detection was >95% in Acrometrix Hotspot Controls, and >90% of CNV gains greater than 6 were detected in cell line samples. Examples of all detected variant classes were displayed based on the Ion Reporter user interface. Conclusions: We introduce an expanded BRCA panel for researching genetic changes influencing homologous recombination DNA damage repair. The option to customize the gene content of the panel is enabled via a web-based interface to inventoried, pre-qualified targets. This approach allows identifying a broad set of mutation types in a single assay with high accuracy, enabling interrogation of HR DDR genes in FFPE preparations. For research use only. Not for use in diagnostic procedures. Citation Format: Charles Scafe, Yun Zhu, Chenchen Yang, Yu-Ting Tseng, Brooke McKnight, Fernando Farfan, Santoshi Bandla, Seth Sadis, Steven Roman, Fiona C. Hyland. The Oncomine BRCA expanded next-generation sequencing assay: Development and analytical validation of a new panel for detection of SNV, insertions, deletions and copy number variants in a panel of 15 genes involved in homologous recombination repair of double-strand break DNA damage [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 2505.

Purpose: Luminal type breast cancers account for the most common subtypes of breast cancers, among which luminal B subtype is relatively faster tumor growing and confers substantial risk of recurrence after primary treatment despite adjuvant endocrine therapy. Understanding the molecular alterations underlying the poorer survival outcome for luminal B tumors represents an unmet medical need. Experimental design: The archival samples of 16 recurrences within 5 years, 3 recurrences between 5-10 years and 24 non-recurrence tumor tissues from luminal B breast cancer patients were retrieved. Genomic alterations of these 43 breast cancer samples were detected using deep next-generation sequencing of a 440-gene panel. The mRNA levels of candidate genes in breast cancer patients were analyzed from The Cancer Genome Atlas (TCGA) database. Results: Overall, genomic alterations were found in 340 genes, including single nucleotide variants and copy number variation. Comparing to the non-recurrence tumors, NUP98 (31.3%), MAPK4 (18.8%), PTEN (18.8%), PER1 (18.8%) and TRIP11 (18.8%) showed higher population frequency in the recurrence tumors, and IGF2R (50.0%), PIM1 (45.8%) and ROS1 (37.5%) were dominate in the non-recurrence tumors. After a validation with the public clinical dataset, we identified three potential prognostic biomarkers, including NUP98, PER1, and TRIP11. We observed missense mutation or copy number deletion of NUP98, PER1 and TRIP11 in early recurrence tumors. Data from TCGA database exhibited patients with breast cancer harbored lower NUP98, PER1 and TRIP11 transcripts levels. Notably, lower NUP98, PER1 and TRIP11 expression levels correlated with worse recurrence-free survival in luminal B breast cancer. Interestingly, based on algorithm for visualization of protein interaction network, we found these 3 genes were all involved in androgen receptor signaling network. Conclusion: Our study identified NUP98, PER1, and TRIP11 which potentially served as biomarkers to predict recurrence in luminal B breast cancer. Further research is required to understand the association between genomic alterations and androgen receptor signaling as well as its clinical impact. Citation Format: Chun-Yu Liu, Ji-Lin Chen, Yi-Ting Yang, Yi-Fang Tsai, Ta-Chung Chao, Kien Thiam Tan, Wan-Lun Wang, Pei-Ju Lien, Shu-Jen Chen, Ling-Ming Tseng. Prognostic genomic alterations associated with recurrence after primary therapy for patients with luminal B breast cancer [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 6468.

Introduction Despite many advances, the treatment for most patients with advanced solid tumors in the abdomen continues to be a clinical challenge. Copy number variants (CNVs) are an important part of the genomic landscape of these patients. Next-generation sequencing (NGS) tests used in the clinical setting are predominantly multi-gene targeted panels (FM, Guardant, NeoGenomics). Although this is the current standard, there is still doubt about the reliability and accuracy of such panels (PMID: 28472276), and panel based CNV-calling in clinical labs use stringent thresholds that may filter out many CNVs. The purpose of our study was to evaluate the clinical utility of a 125-gene targeted sequencing panel in detecting actionable CNVs in a cohort of colorectal cancer patients, with the eventual goal of identifying appropriate genome-guided therapy options. Materials and Methods We performed (i) targeted (350X) NGS based on the IDT xGen 125-gene Pan-Cancer panel on tumor and matched normal blood samples of 54 patients as well as (ii) very low-coverage (LC-WGS, 0.1X) whole genome sequencing on tumor and matched blood samples for the same set of patients. Bioinformatics methods included adapter trimming with bbduk, alignment with bwa-mem, quality control with FASTQC and Qualimap. CNVkit, ichorNCA, and Control-FREEC were used to call CNVs. For CNVkit and ichorNCA, the normal samples were combined to create a ‘reference normal'. Multiple window sizes were utilized to call CNVs at different scales. Results We compared the results of CNV calls from CNVkit, ichorCNA and Control-FREEC from targeted and LC-WGS data. This analysis indicates that targeted panels are at least currently not capable of reliably identifying the copy number landscape of clinical colorectal samples when matched normal samples are utilized. While within the targeted region we identified a median of five clinically actionable CNVs per sample for both targeted and whole genome sequencing (CNV amplification threshold of 5 copies), discordance of called CNVs on average was 56 percent. Numerical copy number calls tend to be higher for the targeted panel (mean: 10) vs LC-WGS (mean: 6) Beyond the targeted regions, LC-WGS detected a median of one additional clinical actionable CNV in genes such as PD-L1, JAK2, LRP1B, VEGFA, BCL2, and CCND3 which are not part of the targeted gene panel. Conclusions Our analysis suggests that CNV calling from targeted panels can be improved upon by incorporating very low-coverage whole genome sequencing. Citation Format: Anu Amallraja, Janina Fuß, Casey B. Williams, Michael Forster, Sebastian Hinz, Tobias Meißner, Alexander Hendricks. Very low coverage whole genome sequencing improves clinically relevant copy number variation calling compared to targeted sequencing [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 3617.

Background: Tumors of unknown origin account for up to 5% of newly diagnosed cancers and the average survival time is 9 to 12 months from diagnosis. Establishing tumor type and subtype guides standard of care treatment for several NCCN targeted therapy guidelines. Methods: Targeted DNA sequencing for more than 500 cancer-associated genes and exome-capture RNA sequencing was carried out in more than 25,000 fresh frozen or paraffin embedded tumor samples, including both primary and metastatic tumors. Mutations, copy number variants, and viral sequences were detected from DNA sequencing while gene expression and fusion events were determined from RNA sequencing. We aimed to predict cancer type by utilizing multiple machine learning models trained on individual data types and harmonize predictions across multiple data types. Results: The transcriptome model predicts more than 60 unique diagnoses covering both solid and hematological cancers with >90% overall accuracy on a held-out test set. Of note, the model can accurately predict 10 subtypes of sarcoma and 6 subtypes of neuroendocrine tumors. Gene expression and splicing were the most informative data types, but a performant DNA-only model was also evaluated for application when only DNA data is available. Finally, we evaluated the model on an unlabeled cohort of poorly differentiated samples with inconclusive diagnosis. Conclusions: The incorporation of multiple modes of omics data can improve the interpretability and robustness of machine learning models to predict cancer diagnosis. Citation Format: Jackson Michuda, Benjamin Leibowitz, Shlomit Amar-Farkash, Crystal Bevis, Alessandra Breschi, Joshuah Kapilivsky, Catherine Igartua, Joshua S. Bell, Kyle A. Beauchamp, Kevin White, Martin Stumpe, Nike Beaubier, Timothy Taxter. Multimodal prediction of diagnosis for cancers of unknown primary [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 5423.

Patients with aggressive osteosarcoma (OS) have poor prognosis due in part to copy number variations (CNVs) that contribute to dysregulation of gene expression (GE) and therapeutic resistance. The objective of the present study was to utilize the TARGET database to integrate CNV and corresponding GE with poor prognosis in pediatric OS (n=85) followed by functional validation of prioritized targets. Cox regression analysis indicated that CNVs in 2642 genes correlated with relapse risk in pediatric OS. Furthermore, the top 10 genes with CNVs significantly associated with increased risk for relapse were present on chromosome 8. The MYC and RAD21 copy number gain (MYC-RAD21 CNV+) located on chromosome 8q correlated with increased GE and poor survival in >90% of the relapsed patients. Based on network analysis, the MYC-RAD21 CNV+ was prioritized for development of targeted therapy. MYC, an oncogenic driver of OS growth, can be indirectly inhibited by bromodomain and extra-terminal domain inhibitors (BETi). RAD21 expression has been associated with increased sensitivity to cell cycle checkpoint kinase 1 inhibitors (CHK1i) in melanoma. Additionally, mechanistic links exist between MYC and CHK1, especially during replication stress. Our hypothesis was that the MYC-RAD21 CNV+ serves as a biomarker of poor prognosis and therapeutic response to BETi+CHK1i therapy. Cell growth response to BETi and CHK1i was evaluated in MYC-RAD21 CNV+ pediatric OS cell lines and a patient-derived xenograft (PDX)-derived xenoline (TT2-77). OS lines (G292, MG63, U2OS, and TT2-77) were highly sensitive to single agent BETi/OTX-015, CHK1i/ SRA737 or CHK1i/LY2606368 at clinically relevant concentrations. Combination index and Bliss independence analysis indicated that BETi+CHK1i did not result in synergistic or additive inhibition of growth at clinically relevant concentrations. However, in OS lines Saos2 and Saos2-LM7 BETi+CHK1i resulted in additive to synergistic inhibition of growth at multiple dose-ratios and at clinically relevant concentrations. In the TT2-77 PDX, whole genome sequencing indicated that the original OS biopsy and the TT2-77 PDX generated from a resection sample harbor the MYC-RAD21 CNV+ (4 copies/amplicon). PDX tumor fragments were implanted into the flank of immunodeficient NOD/SCID/IL2Rγ mice. Once tumor volumes reached 100-150 mm3, mice were randomized and treated with four 5-day cycles of BETi/OTX-015 and/or CHK1i/SRA737. BETi+CHK1i significantly decreased TT2-77 growth, increased probability of survival, and was well tolerated. BETi+CHK1i is a promising therapeutic approach for treatment of relapsed pediatric MYC-RAD21 CNV+ OS. It is possible that MYC, BETs, RAD21 and CHK1 protein levels could dictate sensitivity to combination BETi+CHK1i independent of MYC-RAD2 CNV+ status. Studies are in progress to identify responder versus non-responder signatures in OS. Citation Format: Khadijeh Bijangi-Vishehsaraei, Pankita Pandya, Cheng Lijun, Tang Shan, Anthony Sinn, Melissa Trowbridge, Kathy Coy, Courtney Hemenway, Barbara Bailey, Harlan Shannon, Jixin Ding, Erika Dobrota, M. Reza Saadatzadeh, Adily Elmi, Jeremiah Shultz, Mary Murray, Mark Marshall, Michael Ferguson, Todd Bertrand, L. Daniel Wurtz, Sandeep Batra, Lang Li, Jamie Renbarger, Karen Pollok. Systems biology approach provides rationale for dual-targeted inhibition of BET and CHK1 in aggressive pediatric osteosarcoma [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 450.

N6-methyladenosine (m6A) is a common RNA modification on eukaryotic mRNA and several m6a regulatory proteins, which plays a crucial part in breast cancer. But the copy number variations (CNVs) for m6a regulatory proteins and their role in pathogenesis and progression in breast cancer is still unclear. By analyzing the Molecular Taxonomy of Breast Cancer International Consortium (METABRIC) and the cancer genome atlas (TCGA database), we screened CNVs of ten m6A regulatory genes, including YTHDF3, YTHDF1, FTO, YTHDC1, ALKBH5, METTL3, WTAP, METTL14, YTHDF2, YTHDC2. Among them, YTHDF3 and YTHDF1 amplification had higher alteration rates among ten m6A regulatory genes. YTHDF1 and YTHDF3 amplification resulted in higher mRNA expression (P Univariate and multivariate Cox proportional hazard analysis for 10-year overall survivalUnivariateUnivariateHR95% CIP valueHR95% CIP valueNode status1.7181.566-1.885 Citation Format: Cheukfai Li, Guochun Zhang, Yulei Wang, Bo Chen, Kai Li, Li Cao, Chongyang Ren, Lingzhu Wen, Minghan Jia, Hsiaopei Mok, Jianguo Lai, Weikai Xiao, Xuerui Li, Ning Liao. YTHDF1 amplification associates with clinicopathological featuresand predicts worse survival in breast cancer [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 5912.

The analysis of Next Generation Sequencing (NGS) data relies on workflows that integrate different bioinformatics tools. Designing workflows that are assay agnostic and have a robust computational framework for result generation in a timely manner is particularly important for clinical applications of NGS. These technologies play an important role in classifying a patient's tumor at a molecular level however, there is still a need to develop workflows that generate consistent mutation profiles for oncology. We developed a bioinformatics tool to analyze DNA and RNA sequencing data in a single pipeline and to specifically detect somatic mutations in oncology clinical workflows. Our automated pipeline report contains extensive quality metrics for each sample and generates a robust mutation annotation and filtration module vital for clinical use. We use data driven filters and algorithms that predict the effectiveness of mutations to filter out mutations of poor quality or mutations without clinical significance. We applied our pipeline to standard cell lines and various cancer FFPE samples acquired from the Intermountain Healthcare Biorepository. Our pipeline successfully processes different mutation classes including Single Nucleotide Variants (SNVs), small (21 bp) insertions and deletions (InDels), gene fusions, and exon skipping mutations. In addition, we established a reference set that allows us to call Copy Number Variants (CNVs) in a gene panel designed to detect clinically relevant mutations in lung cancers. Comparison of this pipeline to a validated external workflow yield high concordance of clinically significant mutations. We conclude that our pipeline offers a sensitive and specific methodology to analyze DNA and RNA sequencing data to clinically relevant results from low input amplicon-based assays for various solid tumor cancer types. Citation Format: Katherine Shortt, Christopher Johnson, Archana Ramesh, Tom Neuwerth, Chris Giauque, Sharanya Raghunath. A data driven pipeline to detect somatic mutations in oncology [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 5454.

Background Individual carrying germline mutations within BRCA1 or BRCA2 gene confers well-established increased risk of developing breast and ovarian cancers in the life-time than the general population. Recent advancement in clinical oncology and pharmacology discovered that poly ADP ribose polymerase (PARP) inhibitors are particularly effective in treating cancer that is homologous recombination deficient (HRD) through synthetic lethality, irrespective of the primary tumor site or the type of the BRCA1/2 inactivation variant. Methods A dedicated next-generation sequencing (NGS) panel was designed to interrogate the mutational status of BRCA1/2 and 18 other HRD-associated genes. Using a cohort of 522 clinical samples and artificially engineered cell line samples, we evaluated the mutation detection proficiency for single nucleotide variant (SNV) and small insertion or deletion (Indel) variants. BRCA1/2 inactivation is often the consequence of large genomic rearrangement (LGR) within the gene and phenotypically shown as exon-level copy number variant (CNV). Given the technical challenge of accurate exon-level CNV detection within a small panel, we developed a noise-reduction model based on a pool of normal (PON) samples for CNV result analysis. We further designed droplet digital polymerase chain reaction (ddPCR) for each exon of BRCA1/2 to verify the exonic CNV findings. Results Using a comparator assay as the gold standard, the BRCA1/2 panel achieved an accuracy of 96.7%, a precision of 98.9%, and a lower limit of detection of 3% variant allelic frequency (VAF) for SNV and Indel. The CNV pipeline achieved 97.8% sensitivity and 99.7% specificity for exonic CNV events that were supported by at least one piece of secondary evidence, out-performing publicly available software. The ddPCR verification further enhances assay accuracy and provides a significantly improved statistically confidence for the final CNV calls unparalleled by any other experimental methodologies. Conclusions Using a large cohort of clinical samples, the authors demonstrated the proficiency of a targeted NGS approach for variant and LGR detection of BRCA1/2 and the other genes of the DNA damage repair pathway. An effective method was developed to suppress the noise within the NGS-derived CNV profile, which is exaggerated by the small panel size and short targeted regions. The authors would also like to call the researcher's attention to a critical caveat within the CNV testing results generated from NGS-based methods. Citation Format: Xiangyuan Ma, Hua Bao, Zhili Chang, Yong Wu, Changwan Du, Rong Gao, Yu Song, Ryan S. Robetyore, Xue Wu, Yang W. Shao. Technical validation of a NGS-based analysis pipeline for BRCA1/2 variants and large genomic rearrangements detection [abstract]. In: Proceedings of the Annual Meeting of the American Association for Cancer Research 2020; 2020 Apr 27-28 and Jun 22-24. Philadelphia (PA): AACR; Cancer Res 2020;80(16 Suppl):Abstract nr 4268.