MSK-IMPACT Assay Research Articles

Microsatellite instability (MSI-H) in metastatic urothelial carcinoma (mUC): A biomarker of divergent responses to systemic therapy.

566 Background: MSI frequently arises from deficiencies in mismatch repair mechanisms (dMMR) and are associated with higher response rates to anti-PD1/PDL1 checkpoint inhibitors (CPI) in different cancer types, including mUC (Iyer et al, ASCO 2017). Here, we sought to (1) assess the platinum-based chemotherapy response in MSI-H mUC, and (2) present an updated analysis of CPI response in MSI-H mUC from a database of > 1000 sequenced UC. Methods: From an institutional database of UC tumors genomically characterized using the MSK-IMPACT assay, we identified tumors with high MSIsensor score > 10 (MSI-H) for in-depth analysis. MSIsensor has been shown to correlate strongly with MSI status (Middha et al, JCO Precis Oncol 2017). Patient and disease characteristics, as well as response to (1) platinum-based chemotherapy and (2) CPI were analyzed. Results: A total of 1,333 UC from 1,194 pts were sequenced. 26 pts (2%) had an MSI score > 10. Of these, 20 (77%) were male, 17 (65%) were current or former smokers, and 20 (77%) had upper tract UC. 23/26 pts had alterations in MMR genes (MSH2, n = 15; MLH1, n = 6; MSH6 and PMS2, n = 1), 18 of which underwent germline testing which was positive in 14 cases (78%). 10/26 MSI-H patients developed metastatic UC. Six received first-line platinum-based chemotherapy with median PFS of 2.6 months, four with primary progressive disease. Ten patients received CPI in the mUC setting, two of them in combination with non-CTLA4 agents, six as second or third-line therapies, two following recurrence of disease within 12 months of peri-operative chemotherapy, and two as first line therapy. With a median follow-up of 24.3 months, PFS rate was 90% and 77% at 12 and 24 months, respectively, with near-complete or complete response in 9 patients. Conclusions: MSI-H status in UC predicts deep and durable responses to CPI and is associated with inferior chemotherapy responses. CPI should be considered for first-line treatment in this subset of patients. Pts with somatic MSI-H bladder or upper tract UC should be offered genetic testing for Lynch syndrome.

Abstract P2-10-01: The genomic landscape of breast cancer in African American women

Abstract Background: According to several estimates, African American women are 40% more likely to die from breast cancer than white women in the United States. Proposed explanations for this disparity include differential socioeconomic factors and access to care, however some studies have raised the possibility that variation in tumor biology is a contributing factor. Methods: We prospectively sequenced primary and metastatic breast cancer tumors and their matched normal DNA using the MSK-IMPACT assay. We performed gene enrichment analyses to identify the oncogenic mutations and copy number alterations that were more frequent in patients who self-identified as African American or black (AA/B) compared with patients who self-identified as non-Hispanic white (NHW). Detailed clinicopathologic variables were collected for the full cohort. Mann-Whitney U test, Fisher’s exact test, and multivariate binomial regression models were used for statistical analyses. Results: Genomic profiling was performed on 339 tumors from 301 AA/B patients (44.6% metastatic), and 2,607 tumors from 2,248 NHW patients (48.5% metastatic). Age of AA/B and NHW patients at diagnosis was similar (51.5 vs. 52.8 years, p = 0.065). However, AA/B patients were more likely to be pre- or peri-menopausal at diagnosis (51.6 vs. 44%, p = 0.013), have triple-negative disease (26.5 vs. 13%, p < 0.001), and have higher stage at diagnosis (p = 0.024). Of note, invasive lobular carcinomas were significantly less frequent in AA/B patients compared to NHW (5.9 vs. 14.5%; p < 0.001), a trend that persisted when controlling for receptor subtype. In the unadjusted analysis, AA/B women were more likely to have TP53 mutations (53.4 vs. 36.5%; q < 0.01) and NF1 loss of function mutations (9.9% vs. 3.7%; q < 0.01), and less likely to have CDH1 mutations (6.3% vs. 15.4%; q < 0.01) and PIK3CA mutations (25.7 vs. 35.6%, q = 0.017). However, in a multivariate analysis adjusted for receptor subtype, histology, and sample type (primary vs. metastatic), NF1 was the only gene found to be more commonly mutated in AA/B women (odds ratio: 2.84; 95% CI: 1.73 - 4.08, q < 0.01). Focusing specifically on ER+/HER2- disease, AA/B women were more likely to have mutations in TP53 (31.6% vs. 24.6%), NF1 (7.7% vs. 3.1%) and FGFR1 amplification (21.4% vs. 13.1%), and less likely to have mutations in CDH1 (9.2% vs. 18.8%) or PIK3CA (29.6% vs. 39.6%), however these results did not retain statistical significance when adjusted for multiple comparisons. In triple-negative breast cancer patients, TP53 was mutated at equal rates in AA/B and NHW patients (88.9% vs. 85.9%; p = 0.59), and there was a numerically higher frequency of NF1 mutations in AA/B patients (11.1% vs. 5%). There was no significant difference in tumor mutational burden between AA/B and NHW women (4.24 vs. 4.87; p = 0.19), and no difference in the frequency of microsatellite instability (defined as MSISensor score > 10, 0.9% vs. 0.5%; p = 0.44). Conclusions: In this large clinico-genomic analysis, as previously reported, AA/B patients were more likely to have the clinical hallmarks of aggressive disease, as defined by triple-negative subtype, higher stage, and premenopausal status at diagnosis. Our analysis demonstrated trends towards enrichment for some of the genomic alterations previously identified to be associated with poor outcome in the AA/B population, however, after controlling for the aforementioned clinical factors, breast cancer in AA/B did not differ significantly from breast cancer in NHW in terms of their driver genomic alterations, MSI or tumor mutation burden. Further studies are required to fully characterize the genomics of breast cancer in AA/B women, which may play a role in larger efforts to equalize the disparities observed in this population. Citation Format: Cristian Serna-Tamayo, Joshua Z Drago, Carlos Dos Anjos, Joshua Herbert, Fresia Pareja, Shanu Modi, Komal Jhaveri, Chau Dang, David B Solit, Larry Norton, Mauricio Scaltriti, Jorge S Reis-Filho, Sarat Chandarlapaty, Mark E. Robson, Pedram Razavi. The genomic landscape of breast cancer in African American women [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr P2-10-01.

P1.14-12 A Novel Activating MAP2K1 In-Frame Deletion Mediates Acquired Resistance to ROS1 TKIs in a Patient with ROS1 Fusion-Positive NSCLC

ROS1 tyrosine kinase inhibitors (TKIs) such as crizotinib, entrectinib and lorlatinib provide significant benefit in non-small cell lung cancer (NSCLC) patients with ROS1 fusions. As observed with all targeted therapies however, resistance arises. With the widespread adoption of large panel next generation sequencing (NGS) at the time of acquired resistance (AR), our appreciation of novel off-target mechanisms continues to grow. Detecting additional mechanisms of acquired resistance (AR) is crucial to find novel therapies and improve patient outcomes. We reviewed targeted large-panel sequencing data (using the MSK-IMPACT assay) of paired pre-treatment and post-progression samples from patients treated with ROS1 TKIs. Genetic alterations hypothesized to confer AR were modeled in a patient-derived cell line (LUAD-0003, expressing EZR/ROS1) as well as isogenic human (HBEC) and murine (NIH-3T3) cell lines. ROS1 fusions were expressed in these cells either by cDNA overexpression (CD74/ROS1, SLC34A2/ROS1) or CRISPR-Cas9-mediated genomic engineering (EZR/ROS1). Using these cell line models, alterations in drug sensitivity and downstream signal pathways were examined. We also explored possible therapeutic strategies to overcome the drug resistance caused by the novel AR mechanisms examined in this study. We identified a patient with NSCLC harboring a MAP2K1 (MEK1) variant encoding an in-frame deletion of amino acids E41-L54 (MEK1del) in a sample taken at the time of resistance to lorlatinib (after 9 months’ treatment). This mutation was not detected in the pre-TKI sample. Induction of ROS1 fusions in HBEC and NIH-3T3 cells increased the sensitivity of these cells to ROS1 TKIs and stimulated activation of MEK/ERK signaling in comparison with AKT signaling, suggesting the importance of the RAS-MAPK pathway in driving ROS1 fusion-positive cancers. Underscoring the importance of the RAS-MAPK pathway in ROS1-mediated tumorigenesis, we identified three patients (pancreatic, salivary, and breast cancer) with a ROS1 fusion and NF1 loss-of-function mutation concurrently, in TKI-naïve samples. Expression of MEK1del in HBEC and NIH-3T3 cells harboring ROS1 fusions, and knockdown of NF1 in LUAD-0003, activated ERK signaling and conferred resistance to ROS1 TKIs. Combined targeting of ROS1 (crizotinib, lorlatinib) and MEK (selumetinib, trametinib) inhibited growth of cells expressing both ROS1 fusion and MEK1del. Our results suggest that the activation of the RAS-MAPK pathway plays a critical role in tumorigenesis mediated by ROS1 fusions, and that activating mutations in this pathway can drive AR to ROS1 TKIs. Combined inhibition of ROS1 and MEK is a potential therapeutic strategy that should be explored clinically.

Genomic Assessment of a High Grade Low PSA Prostate Cancer Cohort

High grade prostate cancer with a paradoxically low PSA has been associated with a more aggressive phenotype compared with high grade disease with a high PSA. We sought to examine the genomic landscape of these patients using next generation DNA sequencing to identify a characteristic biology that could explain their unique clinical presentation. A database search identified 19 patients with a Gleason score (GS) of 8 or higher and a PSA less than 4 who had prostate tissue that had undergone next-generation DNA sequencing of targeted gene panels using the MSK-IMPACT assay. A comparison cohort was identified of 146 patients with a GS of 8 or higher, PSA greater than 4, and sequenced prostate tissue. Chi square tests were used to assess the significance of different characteristics between the two groups. Genomic outcomes examined included tumor mutational burden (TMB), the fraction of the genome that is copy number altered (FCNA), and specific genomic alterations, including somatic and germline mutations. Median PSA was 2.35 in the low PSA group and 9.27 in the comparison group. All patients had Gleason 8 or 9 disease. All prostate samples were radical prostatectomy specimens, except for one which was a biopsy specimen. A minority of patients received androgen deprivation therapy prior to tissue collection: 21% in the low PSA group and 13% in the high PSA group (p=0.341). The 2 groups did not differ significantly in pathologic T stage (pT2, pT3, and pT4 were 5%, 68%, and 21% in the low PSA group vs 6%, 73% and 21%, p=0.996), pathologic N stage (pN1 37% vs 51%, p=0.233), and clinical M stage at diagnosis (M1 21% vs 18%, p=0.762). While the 2 groups did not differ with respect to TMB (2.1 vs 2.0) or FCNA (0.07 vs 0.07), there were some notable genomic differences. The low PSA cohort had more germline mutations in the DNA damage response (DDR) pathway, with 26% of patients harboring a germline DDR alteration versus 8% of patients in the high PSA cohort, p=0.0097. The difference in the incidence of both germline and somatic DDR alterations approached significance: 32% of low PSA patients versus 15% of high PSA patients, p=0.059. Genes considered as part of the DDR pathway included BRCA2, BRCA1, ATM, RAD50, CHEK2, FANCA, PALB2, and MUTYH. Notably, the low PSA cohort did not harbor an increased number of alterations in genes typically associated with small cell neuroendocrine prostate cancer, including TP53 (26% vs 32%), RB1 (0% vs 2%) and PTEN (5% vs 20%). Other genes with recurring alterations in these populations included APC (16% and 6%) and TMPRSS2-ERG fusions (26% and 23%). Patients presenting with high grade prostate cancer and a PSA <4 should undergo genomic testing since about one third harbor DDR mutations and may be future candidates for PARP inhibitor therapy. Alterations commonly associated with small cell neuroendocrine prostate cancer were not seen with increased frequency.

Detection of subthreshold microsatellite instability in breast cancer: An ongoing investigation.

1089 Background: Microsatellite instability (MSI) and mismatch repair deficiency (MMRd) can occur sporadically in solid tumors across histologic subtype and predict clinical response to immune checkpoint blockade (ICB). MSIsensor uses next generation sequencing to report the percentage of unstable microsatellites in a given tumor sample as a score. While scores of ≥10 and &lt; 3 have excellent sensitivity/specificity for MSI/MMRd, intermediate scores of 3-10 are of unknown clinical significance, and may include patients with MSI-like genomic signatures who would be overlooked by historical assays. Methods: MSIsensor testing was performed on primary or metastatic tumor specimens from 2,371 patients with breast cancer. Scores of 3-10 were considered MSI-Indeterminate (MSI-I). Concurrent somatic mutations were determined using the MSK-IMPACT assay. Results: Of the 2,371 patients tested, 147 were found to have MSI-I tumors. Ninety-nine (67.3%) were metastatic specimens, and 48 (32.7%) were primary specimens. Of the MSI-I patients, 64 (43.5%) were estrogen receptor positive, 57 (38.8%) were triple negative, and 24 (16.3%) were HER-2 amplified. Somatic TP53 mutations were detected in 109 (74.1%) MSI-I patients, including 54 (94.7%) triple negative patients. Fourteen MSI-I patients were found to have somatic mutations in MMR genes ( MLH1, MSH2, MSH6, PMS2 or EPCAM), but no germline mutations were found. Of the 99 patients with MSI-I metastatic disease, 15 received ICB as part of their clinical care, exhibiting a median progression-free-survival (PFS) of 3 months. Of the 4 patients who demonstrated a PFS of &gt; 6 months on ICB, 2 were found to have somatic mutations in MMR genes. Conclusions: We introduce a subpopulation of breast cancer patients whose tumors exhibit genomic signs of microsatellite instability without meeting the classic criteria for MSI/MMRd as previously defined in colon and endometrial cancers. A subset of these patients may exhibit the properties of MSI/MMRd at the epigenetic and protein-expression levels, and thus potentially benefit from ICB.

A KRAS mutation is associated with an immunosuppressive tumor microenvironment in mismatch-repair proficient colorectal cancer.

609 Background: KRAS-mutant (KRASmut) colorectal cancers (CRCs) are associated with worse prognosis and resistance to therapy. We have previously shown that KRASmut CRCs have different transcriptomic signature of stromal and immune-related genes compared to KRAS-wild type (KRASwt) tumors. Here, we validated the immune-related changes in the tumor microenvironment associated with the KRAS mutation in CRC to guide the design of novel immunotherapy strategies. Methods: The expression of different immune markers (T cells, B cells, macrophages, natural killer cells, and immune check ligands) were assessed using multiplex immunofluorescence (M-IF) technique in both tumor core (TC) and invasive margin (IM). Sequential slides were cut from paraffin blocks of CRC resected at our institute. Each slide was stained with 4 immune markers using M-IF technique. The stained slides were scanned, and quantification of immune cells was done using ImageJ software. Student’s t test was used for statistical analyses. DNA was extracted from each tumor and profiled for 420 cancer genes using targeted exome-capture sequencing (MSK-IMPACT assay). DNA mismatch-repair (MMR) proteins deficiency were analyzed by immunohistochemistry. Only MMR-proficient (pMMR) tumors were included. Results: A total of 39 patients with pMMR CRC were included. AJCC stages (I-III) were not different between KRASmut (n = 15) and KRASwt (n = 25) tumors. M2-macrophages (CD68+CD163+ cells) and IL-17-producing cells (IL17+ cells) were significantly higher (p = 0.002, and 2.9e-6 respectively), while T-helper cells (CD3+CD4+cells) were significantly lower (p = 3.9e-4) in TC of KRASmut tumors compared to KRASwt. Treg (CD3+CD4+FOXP3+cells) were significantly higher in IM of KRASmut tumors (p = 0.01). KRASmut tumors had significantly higher ratios of Treg:T-helper cells, and Treg:T cytotoxic cell (p = 0.008, and p = 0.04; respectively). Conclusions: KRAS oncogene is associated with more pro-tumorigenic (M2 macrophages, IL17 and Treg) and less anti-tumorigenic (CD4 T-helper) immune cells in CRC. These results can be used to guide further research to design novel immunotherapy strategies against KRASmut CRC.

Tumor Mutation Burden and Efficacy of EGFR-Tyrosine Kinase Inhibitors in Patients with EGFR-Mutant Lung Cancers

Tumor mutation burden (TMB) is a biomarker of response to immune checkpoint blockade (ICB). The impact of TMB on outcomes with targeted therapies has not been explored. We identified all patients with metastatic EGFR exon19del or L858R-mutant lung cancers treated with first/second-generation EGFR tyrosine kinase inhibitors (TKIs) with pretreatment next-generation sequencing data (MSK-IMPACT assay). The effect of TMB on time-to-treatment discontinuation (TTD) and overall survival (OS) were evaluated in univariate and multivariate analyses. EGFR wild-type lung adenocarcinoma samples were used for comparison. Among 153 patients with EGFR-mutant lung cancer, TMB was lower compared with EGFR wild-type (n = 1,849; median 3.77 vs. 6.12 mutations/Mb; P < 0.0001) with a broad range (0.82-17.9 mutations/Mb). Patients with EGFR-mutant lung cancer whose tumors had TMB in the high tertile had shorter TTD (HR, 0.46; P = 0.0008) and OS (HR, 0.40; P = 0.006) compared with patients with low/intermediate TMB. Evaluating by median TMB, there was significantly shorter TTD and OS for patients with higher TMB (TTD, P = 0.006; OS, P = 0.03). In multivariate analysis, TTD and OS remained significantly longer in the low/intermediate tertile compared with high TMB (HR = 0.57, P = 0.01; HR = 0.50, P = 0.02, respectively). In paired pretreatment and postprogression samples, TMB was increased at resistance (median 3.42 vs. 6.56 mutations/Mb; P = 0.008). TMB is negatively associated with clinical outcomes in metastatic patients with EGFR-mutant lung cancer treated with EGFR-TKI. This relationship contrasts with that seen in lung cancers treated with immunotherapy.See related commentary by Cheng and Oxnard, p. 899.

MP18-09 GENOMIC PREDICTORS OF RESPONSE TO NEOADJUVANT CISPLATIN-BASED CHEMOTHERAPY IN UPPER TRACT UROTHELIAL CARCINOMA (UTUC)

You have accessJournal of UrologyBladder Cancer: Upper Tract Transitional Cell Carcinoma I1 Apr 2018MP18-09 GENOMIC PREDICTORS OF RESPONSE TO NEOADJUVANT CISPLATIN-BASED CHEMOTHERAPY IN UPPER TRACT UROTHELIAL CARCINOMA (UTUC) NIRMAL THAMPI JOHN, AMY TIN, GOPAKUMAR IYER, RENZO DINATALE, NICOLE BENFANTE, DANIEL SJOBERG, A ARI HAKIMI, PAUL RUSSO, and JONATHAN COLEMAN NIRMAL THAMPI JOHNNIRMAL THAMPI JOHN More articles by this author , AMY TINAMY TIN More articles by this author , GOPAKUMAR IYERGOPAKUMAR IYER More articles by this author , RENZO DINATALERENZO DINATALE More articles by this author , NICOLE BENFANTENICOLE BENFANTE More articles by this author , DANIEL SJOBERGDANIEL SJOBERG More articles by this author , A ARI HAKIMIA ARI HAKIMI More articles by this author , PAUL RUSSOPAUL RUSSO More articles by this author , and JONATHAN COLEMANJONATHAN COLEMAN More articles by this author View All Author Informationhttps://doi.org/10.1016/j.juro.2018.02.585AboutPDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareFacebookTwitterLinked InEmail INTRODUCTION AND OBJECTIVES Use of neoadjuvant cisplatin-based chemotherapy (NAC) in UTUC is based on evidence of survival benefit in urothelial carcinoma (UC) of the bladder. However, concerns regarding toxicity and lack of efficacy have prevented its widespread adoption. We aim to analyze the genomic profiling data of UTUCs from patients who received NAC to identify predictors of chemo-sensitivity. METHODS We reviewed data on all >=cT2anyNM0 high-grade UTUC patients who underwent genomic profiling prior to cisplatin-based NAC to identify somatic mutations that correlated with response, defined as <pT2 tumor stage at nephroureterectomy (NU). We evaluated 14 of the most commonly altered genes in UTUC (>=10%) identified by the MSK-IMPACT assay in a recently published study (FGFR3, KMT2D, KDM6A, KMT2C, STAG2, CDKN2A, TP53, CDKN2B, CREBBP, TSC1, PIK3CA, ARID1A, CCND1 and, HRAS). We also looked at ERCC2 and BCL2, which have been linked with response to chemotherapy in bladder UC. We assessed the association between alterations and pathologic response using univariate logistic regression. RESULTS In 678 patients who underwent NU, 62 (9.1%) received cisplatin-based NAC and 22 underwent MSK-IMPACT sequencing of pre-treatment tumor tissue. Of these patients, 16 (73%; 95% CI 50%, 89%) achieved <pT2 response to NAC. Three patients had CCND1 amplification, one of whom had pathological NAC response. CCND1 amplification was associated with non-significant lower odds of NAC response on pathology (OR 0.13; 95% CI 0.01, 1.87). In particular, 15/19 (79%) patients without CCND1 amplification who had NAC response on pathology compared to 1/3 (33%) patients with CCND1 amplification had NAC response on pathology (diff 46%, 95%CI -11%, 100%). Only one patient was indicated as having BCL2 deletion, and he achieved a pathologic response, compared to 15 of the remaining 21 patients without BCL2 deletion and pathologic response. No patients in our cohort had genetic alterations in any of the other markers assessed. CONCLUSIONS CCND1 amplification was found to be non-significantly associated with lack of response following cisplatin-based NAC. Similar findings in cancers of the head and neck, lung, breast and bladder suggest the importance of CCND1 as a prognostic marker and potential actionable target in cisplatin-resistant high risk UTUC; however larger studies are needed for confirmation. © 2018FiguresReferencesRelatedDetails Volume 199Issue 4SApril 2018Page: e216 Advertisement Copyright & Permissions© 2018MetricsAuthor Information NIRMAL THAMPI JOHN More articles by this author AMY TIN More articles by this author GOPAKUMAR IYER More articles by this author RENZO DINATALE More articles by this author NICOLE BENFANTE More articles by this author DANIEL SJOBERG More articles by this author A ARI HAKIMI More articles by this author PAUL RUSSO More articles by this author JONATHAN COLEMAN More articles by this author Expand All Advertisement Advertisement PDF downloadLoading ...

Defining the DNA damage repair (DDR) genomic landscape of urothelial carcinoma of the bladder (UCB).

502 Background: Impaired DDR is associated with response to platinum-based chemotherapy. Several genomically directed trials have been proposed using DDR alteration status for enrolment. In this study, we characterized the deleterious DDR alteration (DDRa) landscape of UCB across various clinical states. Methods: Targeted exon capture and sequencing of at least 341 cancer-associated genes was performed prospectively on 451 UCB specimens (MSK-IMPACT assay). We assessed sequencing data for deleterious alterations in 34 genes representing canonical DDR pathways. Deleterious alterations included truncating mutations, homozygous deletions, and functionally validated missense mutations. Results: Table 1. In NMIBC, deleterious DDRa were enriched in high-grade disease ([39/136, 28.7%] high grade vs. low-grade [2/28, 7.1%]; p=0.02). The frequency of deleterious DDRa in chemo-naïve MIBC was enriched relative to unmatched post-neoadjuvant chemotherapy (NAC) residual MIBC ([33/112, 29.5%] vs. [8/55, 14.5%]; p=0.01). Patients with metastatic disease had similar rates of deleterious DDRa to MIBC ([31/116, 26.7%] vs. [33/112, 29.5%]). The percentage of patients having any type of DDR alteration was similar across states. The proportion of patients with a deleterious DDRa relative to any DDRa was 41/77 (53.2%) in NMIBC, 33/65 (50.8%) in chemo-naïve MIBC, 8/27(29.6%) in post-NAC residual MIBC, and 31/68 (45.6%) in metastatic disease. Conclusions: DDRa are found across the UCB disease spectrum. ERCC2 and ATM are the most common DDRa although alterations were seen in most other DDR genes. Many alterations are of unknown significance and further characterization is needed to develop genomically directed treatment. [Table: see text]

Abstract 3379: Massively parallel sequencing analysis of breast adenomyoepitheliomas reveals the heterogeneity of the disease and identifies a subset driven by HRAS hotspot mutations

Abstract Adenomyoepithelioma (AME) is a rare biphasic proliferative breast lesion, which may resemble salivary gland epithelial-myoepithelial carcinomas (EMCs). Most AMEs have an indolent clinical course, but malignant transformation and local and distant recurrences have been reported. We sought to define the mutational landscape of AMEs and investigate the functional impact of recurrent likely pathogenic mutations identified in these tumors. Nineteen AMEs were subjected to whole-exome massively parallel sequencing (MPS, n=7) or targeted capture MPS using MSK-IMPACT assay (n=12). Somatic genetic alterations and the cancer cell fraction of mutations were defined using state-of-the-art bioinformatics algorithms. Selected genes (i.e. HRAS and PIK3CA) were subjected to Sanger sequencing in a series of 17 additional AMEs (total n=36). Non-tumorigenic mammary epithelial cells (i.e. MCF10A, MCF10A with the PIK3CAH1047R mutation and MCF12A), which are estrogen receptor (ER)-negative, were utilized for 2D and 3D functional studies. Of 36 cases, 22 were ER-positive and 14 were ER-negative. MPS analysis revealed a low mutation burden and HRASQ61 and PIK3CA hotspot mutations in 6/19 (32%) and 11/19 (58%) AMEs, respectively. All HRASQ61 and all but one PIK3CA mutations were clonal. ER-positive and ER-negative AMEs were fundamentally histologically and genetically distinct. Whilst ER-positive AMEs displayed recurrent PIK3CA mutations (50%, 11/22) but lacked HRAS mutations, ER-negative AMEs displayed, in addition to PIK3CA mutations (57%, 8/14), recurrent HRASQ61 mutations (57%, 8/14). HRASQ61 mutations co-occurred with PIK3CA mutations (50%, 4/8), PIK3R1 deletions (12.5%, 1/8) and/or CDKN2A homozygous deletions (25%, 2/8). HRASQ61 mutations, but not PIK3CA mutations, were significantly associated with ER-negativity (100% vs 21%), concurrent carcinoma (50% vs 7%), axillary metastases (38% vs 0%), high proliferation (63% vs 4%), necrosis (63% vs 11%) and nuclear pleomorphism (75% vs 29%). In vitro forced HRASQ61R expression in MCF10A and MCF12A cells resulted in increased proliferation and transformation. In 3D organotypic cell cultures, forced HRASQ61R resulted in a highly disorganized growth pattern, a partial loss of epithelial phenotype and acquisition of aberrant myoepithelial differentiation, which was more overt in PIK3CA-mutant MCF10A cells. In conclusion, AMEs are phenotypically and genetically heterogeneous. Whilst PIK3CA hotspot mutations occur across the spectrum of lesions, HRASQ61 hotspot mutations are restricted to ER-negative AMEs, which should arguably be classified as breast EMCs. Our genomic and functional analyses are consistent with the notion that HRASQ61 mutations are driver events in the pathogenesis of ER-negative AMEs and may be sufficient for the acquisition of myoepithelial differentiation in breast cells. Citation Format: Felipe C. Geyer, Kathleen A. Burke, Anqi Li, Anastasios D. Papanastatiou, Fresia Pareja, Anne S. Schulteis, Charlotte K. Ng, Salvatore Piscuoglio, Marcia Edelweiss, Luciano G. Martelotto, Pier Selenica, Maria R. Filippo, Gabriel S. Macedo, Achim Jungbluth, Hannah Y. Wen, Juan Palazzo, Zsuzsanna Varga, Emad Rakha, Ian O. Ellis, Brian Rubin, Britta Weigelt, Jorge S. Reis-Filho. Massively parallel sequencing analysis of breast adenomyoepitheliomas reveals the heterogeneity of the disease and identifies a subset driven by HRAS hotspot mutations [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2017; 2017 Apr 1-5; Washington, DC. Philadelphia (PA): AACR; Cancer Res 2017;77(13 Suppl):Abstract nr 3379. doi:10.1158/1538-7445.AM2017-3379

Performance of a high-intensity 508-gene circulating-tumor DNA (ctDNA) assay in patients with metastatic breast, lung, and prostate cancer.

LBA11516 Background: ctDNA assays can noninvasively assess tumor burden and biology by identifying tumor-derived somatic alterations. For broad applicability, including early cancer detection, an unprecedented high-intensity approach (ultra-deep sequencing of plasma cell-free DNA (cfDNA) with broad genomic coverage) is needed to address intra-patient and population-level heterogeneity. We present initial results with this approach in patients (pts) with metastatic breast (BC), non-small cell lung (NSCLC), and castration-resistant prostate cancer (CRPC). Methods: Blood and tissue were prospectively collected w/in 6 wks with no intervening therapy change from pts with de novo or progressive cancer. cfDNA and white blood cell (WBC) genomic DNA from each pt were sequenced with a 508-gene panel (2 Mb; &gt;60,000X raw depth). cfDNA variant calling used molecular barcoding for error correction and filtering for WBC variants. Tissue was sequenced using the MSK-IMPACT assay (410 genes, 1.4 Mb, &gt;500X depth) blinded to plasma/WBC sequencing. Variants were classified as clonal or subclonal based on tumor sequencing in BC and NSCLC. Results: Of 161 eligible pts, 124 (39 BC, 41 NSCLC, and 44 CRPC) were evaluable for concordance. In tissue, 864 variants were detected across the 3 tumor types, with 627 (73%) also detected in plasma: single nucleotide variants/indels - 75%, fusions - 67%, and copy number alterations - 58%. In 90% of pts, at least 1 of the variants detected in tumor tissue was also detected in plasma: BC - 97%, NSCLC - 85%, CRPC - 84%. Most actionable mutations detected in tissue were also detected in plasma (54/71, 76%; SNVs only: 28/31, 90%). A subset of driver mutations (eg. in ESR1, PIK3CA, ERBB2, EGFR) were observed in plasma but not tissue. Clonal variants in tissue were more likely to be detected in plasma than subclonal variants (p&lt;.001). Conclusions: This novel, high-intensity ctDNA assay enabled broad detection of genomic variants in plasma at high rates of concordance with corresponding tumor tissue, providing strong evidence for tumor-derivation of these signals. This study will inform development of a high-intensity sequencing approach for early cancer detection.

838 - Targeting ERBB2 mutations in urothelial carcinoma

354Background: ERBB2 encodes human epidermal growth factor receptor 2 (HER2), a member of the EGFR family of receptor tyrosine kinases that signal through the pro-oncogenic MAP- and PI3K-kinase pathways. ERBB2 is altered by amplification and/or overexpression in various cancers, including urothelial carcinoma (UC). These alterations could confer sensitivity to ERBB2 kinase inhibitors in selected patients with UC. Methods: Patients diagnosed with UC were enrolled onto an institutional review board approved prospective sequencing protocol. Tumor and matched germline DNA were analyzed using the MSK-IMPACT assay that detects alterations in 410 oncogenes and tumor suppressor genes, including ERBB2. Results: Overall, 449 samples from 429 patients were sequenced from 2013 to August 2016. Genetic alterations in ERBB2 were found in 78 samples (17%). At the time of sample collection, 24 patients (31%) had non-muscle invasive bladder cancer (NMIBC), 30 patients (38%) had muscle-invasive bladder cancer (MIBC), 18 (23...

Patient derived xenografts of upper tract urothelial carcinoma: A potential tool for personalized medicine.

344 Background: Historically, upper tract urothelial carcinomas (UTUC) are treated similarly to urothelial carcinoma of the bladder (UCB). However, compared to UCB, UTUC demonstrates a more aggressive clinical course which may be explained by significant differences in mutational frequencies between UTUC and UCB that were reported using a customized exon capture sequencing assay (the MSK-IMPACT assay). A major limitation in the advancement of the UTUC field is the lack of appropriate models specific to the disease. The objective of this study was to develop and evaluate preclinical models that would recapitulate treatment responses observed in patients. Methods: 35 surgical specimens from nephroureterectomy of patients with UTUC were implanted into immunocompromised NOD-SCID IL2Rg−/− (NSG) mice. The histological and the genomic characterization of patient tumors and PDXs were examined by a board certified pathologist and MSK-IMPACT assay, respectively. Cell lines were also established to assess histologic and genetic fidelity. Chemosensitivity of PDX models was assessed using a 4 week cycle of gemcitabine/cisplatin (or carboplatin) administration and analysis of tumor growth was performed using a Two-way ANOVA test. Results: 12 patient-derived xenograft (PDX) models were established with a success rate of 34% (12/35) and a 14% (3/21) success in developing cell lines. Both models were highly reflective of their original tumors in terms of histology and genomic characteristics in mutation status and copy number alterations. For a representative PDX, chemosensitivity experiments identified gemcitabine/carboplatin as a potentially effective combination, which was also used in the clinical scenario with a therapeutic response. Chemosensitivity of additional PDXs is continuing to be investigated to potentially guide post-operative decision making for treatment of the patient’s clinically. Conclusions: We developed a cohort of stable PDX models and cell lines for UTUC that maintains the genetic characteristics of the patient’s initial tumor. The continued development of these models may facilitate personalized medicine strategies in the treatment of UTUC.

Characterization of POLE-mutated (POLE+) urothelial carcinoma of the bladder and urinary tract (UC).

387 Background: The POLE gene encodes for the catalytic and proofreading unit of DNA polymerase epsilon. Mutations in POLE gene, recently observed in endometrial and colorectal cancers, are associated with unique clinical features and an ultramutated phenotype. We sought to quantify and characterize POLE+ in UC. Methods: POLE+ cases were identified from a database of 532 UC tumors sequenced with the MSK-IMPACT assay (Cheng et al J Mol Diagn 2015). Tumors were classified based on total mutation: hypermutated (20 - 149 mutations) and ultramutated ( ≥ 150 mutations) as previously described in colon cancer (Stadler et al J Clin Oncol 2016). An institutional cohort of patients with metastatic UC was used as historical POLE- control (n = 94). Results: Thirty two POLE+ patients (6%) and 42 mutations were identified. Median age of diagnosis was 65 years (range: 31 - 94) and 81% were male. Seven patients were found to have Lynch’s syndrome, 8 with ERCC2 mutations which were hotspots or within the helicase domains and 1 with germline NBN mutation. These were excluded from further analysis. Of the remaining 16 patients (3%) and 22 mutations, four were found within the proofreading domain (positions 268 – 471), one was a previously reported hotspot (P286R) and one was a truncating mutation (R114*). Four cases were considered hypermutated and two ultramutated based on mutation counts. Clinically, four had non-invasive disease and 4 had metastatic disease at diagnosis. Eight patients had muscle-invasive disease, five of whom received neoadjuvant chemotherapy. All 8 patients developed metastatic disease with a median time interval of 10.3 months (range: 3.3 – 29.6). Nine of the 12 metastatic patients had nodal metastases while only one had visceral metastasis. With median follow-up of 25.4 months, median overall survival from metastatic diagnosis was not reached and 83% remained alive at 2 years, which was superior to survival for POLE- metastatic patients (HR 0.36, p = 0.04) and POLE- nonvisceral, node only metastatic patients (HR 0.33, p = 0.03). Conclusions: POLE+ prevalence is low in UC but might exhibit superior overall survival. It remains to be investigated if this is strictly the effect of POLE+ or other as yet unidentified genomic events.

Targeting ERBB2 mutations in urothelial carcinoma.

354 Background: ERBB2 encodes human epidermal growth factor receptor 2 (HER2), a member of the EGFR family of receptor tyrosine kinases that signal through the pro-oncogenic MAP- and PI3K-kinase pathways. ERBB2 is altered by amplification and/or overexpression in various cancers, including urothelial carcinoma (UC). These alterations could confer sensitivity to ERBB2 kinase inhibitors in selected patients with UC. Methods: Patients diagnosed with UC were enrolled onto an institutional review board approved prospective sequencing protocol. Tumor and matched germline DNA were analyzed using the MSK-IMPACT assay that detects alterations in 410 oncogenes and tumor suppressor genes, including ERBB2. Results: Overall, 449 samples from 429 patients were sequenced from 2013 to August 2016. Genetic alterations in ERBB2 were found in 78 samples (17%). At the time of sample collection, 24 patients (31%) had non-muscle invasive bladder cancer (NMIBC), 30 patients (38%) had muscle-invasive bladder cancer (MIBC), 18 (23%) had metastatic disease and 7 (8%) had upper tract urothelial carcinoma (UTUC). Sixteen samples (21%) came from metastatic specimens. Of the observed 78 ERBB2 alterations, 20 samples had amplifications (26%) and 58 samples had mutations (74%). Seven samples (9%) had both. We identified 71 missense, 2 inframe and 1 fusion alteration, corresponding to a somatic mutation rate of 13.1%. Thirty-seven mutations (50%) were functionally characterized hot spots that are potentially actionable. Of note, the hot spot mutation S310F/Y was found in 29 samples (37%). Its mutation allele frequency varied significantly. There was a trend towards a higher mutant allele frequency of S310F/Y in higher stage disease without reaching statistical significance (Table). In our cohort, 3 patients were enrolled in clinical trials with ERBB2 kinase inhibitors based upon the presence of an ERBB2alteration. Conclusions: ERBB2is a frequent mutation at different stages of UC. In higher stage disease, clonal selection of the S310F/Y hot-spot mutation may occur and requires further study. [Table: see text]

Abstract S2-02: The landscape of somatic genetic alterations in BRCA1 and BRCA2 breast cancers

Abstract Introduction: Women carrying BRCA1 or BRCA2 germline mutations have a 45-80% lifetime risk of developing breast cancer (BC). BRCA1 and BRCA2 are perceived as bona fide tumor suppressor genes, whereby bi-allelic inactivation in tumor cells is required for tumorigenesis. Recent studies have indicated that loss of heterozygosity (LOH) of the wild-type allele of BRCA1 may be heterogeneous and constitute a late event. Therefore, additional somatic events prior to full BRCA1/2 inactivation may be required for tumorigenesis. Given that the somatic events that result in the development of BRCA1/2-BCs and their chronology are not understood, here we sought to define the genomic landscape of BRCA1/2-BCs and whether LOH of BRCA1/2 wild-type allele and/or mutations affecting additional tumor suppressor genes would be clonal or subclonal in these cancers. Methods: We retrieved 29 BRCA1-BCs and 10 BRCA2-BCs from the Pathology Departments of the authors' institutions. DNA extracted from microdissected tumor and normal breast samples was subjected to targeted capture massively parallel sequencing using either the MSK-IMPACT assay or an assay targeting all exons of 254 genes recurrently mutated in BC or related to DNA repair. Somatic single nucleotide variants, small insertions and deletions and copy number alterations affecting genes present in both sequencing assays (111 genes) were defined using state-of-the-art bioinformatics algorithms. ABSOLUTE and FACETS were employed to define clonal (i.e. present virtually in 100% of the cancer cells of a given case) and subclonal mutations and the presence of LOH of the BRCA1 and BRCA2 wild-type alleles. Results: Our analysis revealed bi-allelic inactivation of BRCA1 in 28 of 29 BRCA1-BCs (93% harbored LOH of the BRCA1 wild-type allele and 3% harbored a second somatic BRCA1 pathogenic mutation). The only BRCA1-BC lacking bi-allelic inactivation of BRCA1 was an estrogen receptor-positive lobular carcinoma, lacking genomic features consistent with homologous recombination DNA repair defects, diagnosed at 62 years of age. Bi-allelic inactivation of BRCA2 was found in all cases (100% of harbored LOH of the BRCA2 wild-type allele). A clonal somatic 'second hit' resulting in bi-allelic inactivation of BRCA1 or BRCA2 was detected in 76% and 100% of BRCA1-BCs and BRCA2-BCs, respectively. In BRCA1-BCs, TP53 mutations were detected in 76% of cases, and these mutations were found to be clonal in 58% of cases. The repertoire of somatic mutations affecting BRCA1-BCs included clonal somatic mutations or homozygous deletions of known tumor suppressor genes, such as PTEN, RB1, CDKN2A and NF1. In contrast, only 10% of the BRCA2-BCs harbored TP53 somatic mutations. Though clonal somatic mutations in several cancer genes were detected, 40% of BRCA2-BCs had no mutations affecting the cancer genes analyzed. Conclusions: Bi-allelic inactivation of BRCA1 and BRCA2 are frequent events in BRCA1-BCs and BRCA2-BCs, respectively. In a subset of BRCA1-BCs, however, the second 'hit' appeared to be subclonal, whereas mutations affecting TP53 and other tumor suppressor genes were clonal, supporting the notion that at least in a subset of these tumors, loss of the wild-type allele of BRCA1 may be preceded by inactivation of another tumor suppressor gene. Citation Format: Geyer FC, Burke KA, Macedo GS, Piscuoglio S, Ng CK, Martelotto LG, Papanastatiou AD, De Filippo MR, Schultheis AM, Brogi E, Robson M, Wen YH, Weigelt B, Schnitt SJ, Tung N, Reis-Filho JS. The landscape of somatic genetic alterations in BRCA1 and BRCA2 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 S2-02.

Abstract 4078: KRAS mutation status is associated with stromal inactivation in colorectal cancer and predicts poor response to neoadjuvant chemoradiotherapy

Abstract Background: Treatment for locally advanced rectal cancer (LARC) consists of neoadjuvant chemoradiotherapy (NCRT) followed by radical excision. Patients with tumors carrying a mutant KRAS are less likely to respond to NCRT compared to KRAS wild type tumors. We hypothesized that an RNA-based signature differentiating KRAS mutant and wild type patients could serve as an indicator of the biological process associated with response to NCRT. We found that the RNA-based signature is enriched for stromal and immune genes. Furthermore, the stromal component of the signature is a predictor of response to NCRT. Methods: Tumors from 120 LARC patients enrolled in a multicenter phase 2 trial studying response to NCRT were tested for KRAS status by Sanger Sequencing or Memorial Sloan Kettering (MSK)-IMPACT assay and gene expression was quantified by sequencing. Colorectal cancer (CRC) patients from MSK (n = 95) and TCGA (n = 261), previously annotated for KRAS mutation status and gene expression, were used for validation. A KRAS-inducible mouse model and CRC patient-derived xenografts (PDXs) were utilized to determine the cell of origin for the gene expression signature. Stromal enrichment was assessed with the ESTIMATE stromal gene signature. Immunohistochemistry (IHC) was completed for Periostin (POSTN), a stromal marker from the RNA-signature. Variant Allele Frequency (VAF) was used to measure the abundance of KRAS, TP53 and Adenomatous Polyposis Coli (APC) mutant alleles in tumors, and was quantified by targeted exome sequencing with the MSK-IMPACT assay. Results: Analysis of the KRAS-associated gene signature showed significant stromal inactivation in KRAS mutant patients. The signature was validated in the MSK and TCGA cohorts. The stromal signature was recapitulated in a KRAS inducible mouse model. Human CRC PDXs in mouse indicated that the signature arose from murine stroma and not human epithelium. Consistent with the stromal signature, IHC for POSTN, a stromal marker, was significantly lower in the KRAS mutant tumors compared with the KRAS wild type tumors (p&amp;lt;0.05) and was absent from the epithelium. The stromal enrichment in mutant KRAS tumors was inversely correlated with the KRAS VAF (p&amp;lt;0.01). This finding was not observed for TP53 or APC VAF indicating specificity for stromal inactivation in KRAS mutant tumors. Furthermore, the stromal component of the signature is associated with poor response to NCRT in LARC. Conclusions: This study shows that a KRAS mutation in CRC is associated with a lower expression of a stromal signature and that this signature is derived from the tumor microenvironment. This study indicates that CRC KRAS mutant tumors and a stromal subtype are closely related. Understanding this relationship may play a key role in elucidating the mechanism by which a KRAS mutant tumor is resistant to standard therapy. Citation Format: Raphael Pelossof, Moshe Elkatebts, Oliver Chow, Lauren Fairchild, Kevin O’Rourke, Jesse J. Smith, Chin-Tung Chen, Samuel Brook, Maurizio Scaltriti, Jinru Shia, Philip Paty, Christina Leslie, Scott Lowe, Jose Baselga, Julio Garcia-Aguilar. KRAS mutation status is associated with stromal inactivation in colorectal cancer and predicts poor response to neoadjuvant chemoradiotherapy. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 4078.

Abstract 119: Genomic analysis of remnant gastric adenocarcinoma

Abstract Introduction: Prior to pharmaceutical treatment of gastroduodenal ulcer disease, partial gastrectomy was performed for the management of ulcer disease complicated by bleeding (PUD). The proposed mechanism for the development of gastric cancer in this setting is the chronic irritation of the mucosa by duodenal reflux into the gastric remnant. The aim of this study is to characterize the genetic profile of remnant gastric carcinoma (RGC). Material and Methods: Thirty cases of RGC developed in the setting of prior partial gastrectomy for PUD and without neoadjuvant chemotherapy were retrospectively collected from the institution database. Histopathological evaluation was performed to ensure &amp;gt;80% tumor cellularity with matching normal gastric mucosa prior to DNA extraction. Ultra-deep targeted sequencing of the thirty matched pairs was performed using the MSK-IMPACT assay (Integrated Mutation Profiling of Actionable Cancer Targets) to identify mutations and copy number alterations in a common set of 410 cancer-associated genes. MSK-IMPACT data from previously sequenced 45 sporadic gastric carcinomas (GC) was used as comparison. Results: The cohort consisted of 22 (73%) males and 8 (27%) females with the mean age of 73 years. The median interval between PUD and RGC resections was 34.8 years (7-60 years). By Lauren's classification, 60% of the tumors were intestinal, 23% diffuse and 17% mixed. Most of the tumors (47%) were poorly differentiated and stage II (40%). MSK-IMPACT sequencing revealed 299 total somatic alterations occurring in 148 of 410 targeted genes with a median number of non-synonymous somatic mutations and copy number events of 3 and 1 per patient, respectively. Of the 207 total mutations, 27.5% were hotspot COSMIC and 23.2% were loss-of-function mutations. Recurrent genomic events included mutually exclusive mutations in TP53 (33%) and amplifications in MDM2 (26%), alterations in the ERBB family genes (37%), ARID1A (16%), KMT2D (13%), CDH1 (13%), and CDK4 (13%). In addition, actionable alterations in PIK3CA (10%), FGFR1 (4%), PTEN (4%), and concurrent hotspot KRAS mutations and inactivation of p15/p16 (10%) were observed. Compared to the sporadic GC cases, RGC cases presented differences in the landscape and recurrence of alterations. Conclusion: The pathogenesis of RGC is likely associated with the prolonged exposure of the gastric mucosa to gastroduodenal contents. The mean age of patients may also exert effects on the development of the RGC. Based on our sequencing results, RGC may present a genetic profile distinct from that of sporadic gastric tumors. Assessment of these alterations identified 30% of patients with potential therapeutic targets for the treatment of this disease and deserves further investigation. Citation Format: Edaise M. da Silva, Helen H. Won, Dianne Torrence, Daniel Coit, Michael F. Berger, Laura Tang. Genomic analysis of remnant gastric adenocarcinoma. [abstract]. In: Proceedings of the 107th Annual Meeting of the American Association for Cancer Research; 2016 Apr 16-20; New Orleans, LA. Philadelphia (PA): AACR; Cancer Res 2016;76(14 Suppl):Abstract nr 119.

Next-generation sequencing (NGS) in advanced well differentiated pancreatic neuroendocrine tumors (WD pNETs): A study using MSK-IMPACT.

246 Background: Whole exome sequencing in WD pNETs demonstrated an increased number of mutations in chromatin modeling genes (MEN1, DAXX, and ATRX), and in the mTOR pathway. NGS brings this technology to the clinic but its relevance in practice has been questioned. In this prospective, IRB-approved study (NCT01775072), we used the MSK-IMPACT assay to perform NGS on WD pNETs in a routine practice setting. MSK-IMPACT, performed in a CLIA-compliant laboratory, is a multiplexed assay (Illumina HiSeq) providing full exon coverage of 410 cancer related genes, detecting base substitutions, small indels, copy number and select gene rearrangements. Methods: After written consent, tumor and germline DNA were analyzed. Genomic alterations were catalogued. Results: MSK-IMPACT results are available in 39 patients (pts). Actionable alterations were identified in 13/39 pts (33.3%). These actionable alterations included BRAF V600E (2 pts; 5.1%), and mutations in TSC1 (1 pt; 2.6%), TSC2 (5 pts; 12.8%), PTEN (4 pts; 10.3%), CDKN1B (2 pts; 5.1%), CDKN2A (2 pts; 5.1%), CDKN2B (2 pts; 5.1%), CDKN2C (1 pt; 2.6%), and ARID1A (5 pts; 12.8%). Other recurrently altered genes included MEN1 (23 pts, 59.0%), DAXX (13 pts, 33.3%), and ATRX (10 pts, 25.6%). Notably, 7 pts (17.9%) had alterations in the histone methyltransferase SETD2. All 7 pts presented with metastatic liver disease. No tumors were low grade; 6/7 (85.7%) were intermediate grade and 1/7 (14.3%) were high grade. All 7 pts received capecitabine/temozolomide (cape/tem); 6 pts (85.7%) had disease shrinkage and 1 pt (14.3%) had stable disease. Conclusions: The mutational landscape in our study was in line with prior work in pNET whole exome sequencing. In one-third of our cohort, potentially actionable alterations were identified through NGS but have not yet been shown to be therapeutically relevant. An exploratory analysis to identify responses to different therapies is ongoing. In addition, we report a novel finding of SETD2 alterations in pNETs. All pts with SETD2 alterations had intermediate to high grade tumors that responded to cape/tem. The clinical significance of SETD2 in pNETs is unknown, and evaluation using tissue microarrays is ongoing.

Abstract B14: Molecular characterization of mucinous ovarian carcinoma.

Abstract Mucinous ovarian carcinoma (MOC) is a rare, chemoresistant tumor known to share pathologic features with tumors of the gastrointestinal and pancreaticobiliary tracts. To better understand the genomic and proteomic landscapes of invasive MOCs, we identified somatic mutations and proteins expression in a single institution cohort and compared these results with data from TCGA tumor projects. Twenty-six tumors consistent with primary invasive MOC after expert pathology review and with available paired tumor and normal tissue were identified from institutional databases between July 2001 and July 2012. DNA extracted from FFPE or fresh frozen samples underwent next generation sequencing with a combination of a candidate gene assay (37 genes), the MSK-IMPACT assay (341 genes), transcriptome sequencing, and whole exome sequencing. Copy number alterations were identified using data from the MSK-IMPACT assay, whole exome sequencing or Affymetrix SNP 6.0 arrays. Immunohistochemistry (IHC) was performed using optimized antibodies for six proteins to confirm the diagnosis of MOC (ER, PR, CK7, CK20, CDX-2, PAX8) and seven proteins to correlate with mutation status and copy number alterations (p53, ARID1A[Baf250a], PTEN, PMS2, MSH6, HER2, p16). Mutation data for other TCGA tumor types was obtained from the cBio Cancer Genomics Portal (cbioportal.org). The median age of the cohort was 58 years (range 20-86 years). Most (19/26, 73%) of the tumors were stage I. Somatic TP53 and KRAS mutations were the most common seen and were identified in 18 (69%) cases each, with a co-mutation rate of 50% (13/26). Other commonly mutated genes include ARID1A, PTEN, and PIK3CA. Homozygous deletions of CDKN2A were found in 27% (7/26). ERBB2 alterations were identified in 19% (5/26) and consisted of three amplifications and two mutations. Mutations in at least one potentially targetable gene were identified in 42% (11/26) of tumors. IHC was concordant with sequencing results in 154/182 (85%) of stained cases. Pancreatic, colorectal, lung adenocarcinoma, endometrial, and stomach cancers have the highest frequency of KRAS mutations. Co-mutations of KRAS and TP53 occur most commonly in pancreatic (59%) and colorectal (21%) carcinomas. CDKN2A homozygous deletions are also found at a similar frequency in pancreatic adenocarcinomas (28%). When evaluating the mutation rates of the five most commonly mutated genes in our MOC cohort with colorectal, pancreatic, gastric and high-grade serous ovarian carcinomas (HGSOC) in the TCGA datasets, pancreatic adenocarcinoma showed the most similarity. HGSOC showed little similarity to the MOCs. KRAS and TP53 co-mutation are common in invasive MOCs. Other commonly mutated genes include ARID1A, PTEN, and PIK3CA. Potentially targetable ERBB2 alterations were identified in several cases. Despite anatomic distinctions, the mutational landscape of MOC shares similarities with that of pancreatic adenocarcinoma including frequent CDKN2A deletions and KRAS/TP53 co-mutation. The suggested shared molecular pattern with pancreatic adenocarcinoma offers potential to guide future developmental therapeutics. Citation Format: Brooke Schlappe, Jennifer J. Mueller, Narciso Olvera, Fanny Dao, Cyriac Kandoth, Faina Bogomolniy, Agnes Viale, Kety Huberman, Gouri Nanjangud, Yaser Hussein, Barry Taylor, Robert Soslow, Douglas A. Levine. Molecular characterization of mucinous ovarian carcinoma. [abstract]. In: Proceedings of the AACR Special Conference on Advances in Ovarian Cancer Research: Exploiting Vulnerabilities; Oct 17-20, 2015; Orlando, FL. Philadelphia (PA): AACR; Clin Cancer Res 2016;22(2 Suppl):Abstract nr B14.