Search

Background Estrogen receptor-positive (ER-positive) metastatic breast cancer is often intractable due to endocrine therapy resistance. Although ESR1 promoter switching events have been associated with endocrine-therapy resistance, recurrent ESR1 fusion proteins have yet to be identified in advanced breast cancer. Patients and methods To identify genomic structural rearrangements (REs) including gene fusions in acquired resistance, we undertook a multimodal sequencing effort in three breast cancer patient cohorts: (i) mate-pair and/or RNAseq in 6 patient-matched primary-metastatic tumors and 51 metastases, (ii) high coverage (>500×) comprehensive genomic profiling of 287–395 cancer-related genes across 9542 solid tumors (5216 from metastatic disease), and (iii) ultra-high coverage (>5000×) genomic profiling of 62 cancer-related genes in 254 ctDNA samples. In addition to traditional gene fusion detection methods (i.e. discordant reads, split reads), ESR1 REs were detected from targeted sequencing data by applying a novel algorithm (copyshift) that identifies major copy number shifts at rearrangement hotspots. Results We identify 88 ESR1 REs across 83 unique patients with direct confirmation of 9 ESR1 fusion proteins (including 2 via immunoblot). ESR1 REs are highly enriched in ER-positive, metastatic disease and co-occur with known ESR1 missense alterations, suggestive of polyclonal resistance. Importantly, all fusions result from a breakpoint in or near ESR1 intron 6 and therefore lack an intact ligand binding domain (LBD). In vitro characterization of three fusions reveals ligand-independence and hyperactivity dependent upon the 3′ partner gene. Our lower-bound estimate of ESR1 fusions is at least 1% of metastatic solid breast cancers, the prevalence in ctDNA is at least 10× enriched. We postulate this enrichment may represent secondary resistance to more aggressive endocrine therapies applied to patients with ESR1 LBD missense alterations. Conclusions Collectively, these data indicate that N-terminal ESR1 fusions involving exons 6–7 are a recurrent driver of endocrine therapy resistance and are impervious to ER-targeted therapies.

Background We sought to identify genomic alterations (GAs) in salivary mucoepidermoid carcinomas. Patients and methods DNA was extracted from 48 mucoepidermoid carcinomas. Comprehensive genomic profiling (CGP) including the calculation to tumor mutational burden (TMB) was performed on hybridization-captured adaptor ligation-based libraries of 315 cancer-related genes plus introns from 28 genes frequently rearranged for cancer and evaluated for all classes of GAs. Results A total of 183 GAs were found in 80 unique genes. High-grade tumors had more GAs (mean 5 ± 3.8) compared with low (2.3 ± 1.4) or intermediate (2.6 ± 1.5) (P = 0.019).TP53 GAs were seen in all tumor grades (41.7%) but were most common in high-grade malignancies (56%) (P = 0.047).CDKN2A GAs were seen in 41.6% of tumors. PI3K/mTOR pathway activation, includingPI3KCA mutations, were more common in high grade (52%) than in low- and intermediate-grade tumors (4.3%) (P = 0.007).BAP1 GAs were observed in 20.8% of tumors andBRCA1/2 GAs present in 10.5% of specimens.ERBB2 amplifications were seen in only 8.3% of tumors. The TMB for this patient group was relatively low with only 5 (10%) of cases having greater than 10 mutations/megabase of sequenced DNA. Conclusion CGP of salivary mucoepidermoid carcinomas revealed diverse GAs that may lead to customized treatment options for patients with these rare tumors.

Background The PURE01 study (NCT02736266) evaluates preoperative pembrolizumab before radical cystectomy (RC). Here, the possibility to predict the pathologic complete response (pT0) after neoadjuvant immunotherapy could have paradigm-shifting implications for the management of patients with MIBC. Methods The reported analyses include comprehensive genomic profiling with FoundationONE CDx assay and programmed cell-death-ligand-1 (PD-L1) combined positive score (CPS, Dako 22C3 antibody) on baseline TURB samples. Multivariable logistic regression analyses (MVA) evaluated clinical T-stage and biomarkers (tumor mutational burden [TMB], and CPS) in association with pT0 response. Multivariable-derived coefficients were used to develop a novel risk calculator. Decision-curve analysis was used to evaluate the net benefit of the predictive model. Results From 02/2017 to 06/2019, 112 patients with full biomarker data were enrolled. At MVA, only CPS showed a significant association with pT0 (p = 0.005). Increasing TMB and CPS values featured a linear association with logistic pT0 probabilities (p = 0.02 and p = 0.004). Very high TMB values, associated with a predicted probability of pT0 > = 55%, were independent from CPS contribution. The coefficients of the predictive model were used to develop a risk calculator. The c-index of the model was 0.77. At decision-curve analysis, the net-benefit of the model was higher than the “treat-all” option from more than 10% threshold probabilities. Conclusion We presented the first biomarker-based risk stratification tool that can be used to recommend use of neoadjuvant pembrolizumab in those patients who have equal or greater chances to achieve a pT0 status compared to the literature on neoadjuvant chemotherapy or to RC alone, depending on cisplatin eligibility. Clinical trial identification PURE01 study (NCT02736266). Legal entity responsible for the study The authors. Funding Merck & Co., Inc., Kenilworth, NJ, USA; Associazione Italiana per la Ricerca sul Cancro (AIRC). Disclosure R. Madison: Shareholder / Stockholder / Stock options, Full / Part-time employment, employee and stock owner of Foundation Medicine Inc: Foundation Medicine Inc. M. Colecchia: Speaker Bureau / Expert testimony: Roche Diagnostics. S.M. Ali: Full / Part-time employment: Foundation Medicine. J.H. Chung: Full / Part-time employment: Foundation Medicine. J. Ross: Leadership role, Research grant / Funding (institution), Shareholder / Stockholder / Stock options, Full / Part-time employment: Foundation Medicine. A. Salonia: Speaker Bureau / Expert testimony: Astellas Pharma; Travel / Accommodation / Expenses: Konpharma. A. Briganti: Advisory / Consultancy: Astellas Pharma; Advisory / Consultancy: Hanssen-Cilag; Advisory / Consultancy: OPKO Health; Advisory / Consultancy: MDxHealth; Advisory / Consultancy: Ferring; Research grant / Funding (institution): Novartis. A. Necchi: Advisory / Consultancy: Merck, Incyte, AstraZeneca, Roche, Rainier Therapeutics, Clovis Oncology, Bristol-Myers Squibb, Bayer, Basilea Pharmaceutica; Research grant / Funding (institution): AstraZeneca, Ipsen, Merck and Incyte; Spouse / Financial dependant: Bayer. All other authors have declared no conflicts of interest.

Introduction Environ 7 % des melanomes surviennent dans un contexte evocateur d’une predisposition hereditaire, or seuls 20 % d’entre eux ont une cause moleculaire clairement identifiee a ce jour. Au decours de l’analyse somatique d’un melanome, il arrive de detecter des mutations constitutionnelles. Materiel et methodes Le test NGS FoundationOne® permettant le sequencage de 315 genes a ete propose a un patient de 23 ans atteint de melanome stade III dans un contexte de multiples antecedents familiaux de melanomes (mere et 2 sœurs, l’une decedee de son melanome a 33 ans). Une CGH-array a ete realisee a partir d’ADN tumoral. L’immunohistochimie (IHC) des proteines impliquees dans la voie Wnt/β-catenine a ete realisee sur le melanome du patient et une serie de melanomes controles. Les caracteristiques cliniques de la famille ont ete recueillies. Resultats Le test FoundationOne® sur ADN tumoral a mis en evidence une mutation BRAFV600E avec une frequence allelique (FA) de 3 % associee a une mutation CTNNA1 F611fs*10 (gene suppresseur de tumeur codant pour l’α-catenine). Cette mutation, responsable d’un decalage du cadre de lecture induisant un codon stop, est trouvee avec une FA de 41 %. Cette discordance de FA pouvait faire evoquer une mutation constitutionnelle de CTNNA1, confirmee dans le sang de sa mere et sa sœur. La CGH sur l’ADN tumoral a identifie une perte segmentaire du chromosome 5 en 5q31.2. Cette perte bi-allelique de CTNNA1 se traduit par une perte d’expression de CTNNA1 en IHC dans le melanome du patient. Sur le plan clinique, les 7 melanomes familiaux ont ete operes a un âge moyen de 29 ans (18–56), 2 etaient avances (Breslow 6,7 mm/metastatique d’emblee), et les autres etaient des SSM in situ (n = 4) ou de Breslow 0,56 mm (n = 1). Le patient index presente une agenesie dentaire, une fente labiopalatine et un goitre multinodulaire. Enfin, le sequencage haut debit FoundationOne® sur une serie de 4889 melanomes successifs a retrouve CTNNA1 mute dans 145 cas (3 %). Les mutations perte de fonction sont frequemment associees a des mutations activatrices de BRAF ou NRAS. Discussion CTNNA1 est un frein de la voie Wnt/β-catenine. In vitro, sa perte de fonction favorise l’evolution metastatique par hyperproliferation et perte d’adhesion cellulaire. Des mutations germinales CTNNA1 ont ete mises en cause dans les cancers du sein lobulaires et les cancers gastriques diffus familiaux. Nous decrivons ici la 1ere mutation CTNNA1 dans un contexte de melanomes familiaux. Conclusion CTNNA1 est un bon candidat pour expliquer les multiples melanomes de cette famille, avec selon le modele des 2 hits de Knudson, une mutation constitutionnelle tronquante de CTNNA1 et une perte somatique d’un segment du chromosome 5. Outre des perspectives de prevention/depistage precoce du melanome chez les porteurs de mutation CTNNA1, une meilleure connaissance de cette voie oncogenique pourrait aider a concevoir de nouvelles strategies de combinaison de therapies ciblees dans le traitement du melanome.

Background BRAF genomic alterations (GA) occur in multiple tumor types and BRAF/MEK targeted therapies are approved in melanoma and NSCLC. Diverse mechanisms of AR to these therapies have been proposed but have not been comprehensively assessed. Methods Hybrid-capture based comprehensive genomic profiling (CGP) was performed on FFPE (n = 228,629) or blood-based cell free DNA (cfDNA, n = 15,069) samples for 222,952 patients (pts). Tumor mutational burden (TMB) was determined on 0.8-1.1 Mbp of sequenced DNA. Samples without evidence of tumor DNA or known to have not received RAF/MEK inhibitors were excluded. Paired samples were collected >60 days apart (median 523, range 71-5571). Results Paired samples with BRAF V600E (64%) or other activating BRAF GA (36%) were available for 154 pts with NSCLC (20%), melanoma (19%), CRC (15%) myeloma (8.4%) glioma (7.1%) or other (30%) cancers. Acquired GA previously described preclinically or clinically including in BRAF, KRAS, NRAS, MEK1, PIK3CA, PTEN, MET, and CCND1 occurred in 34 cases (Table). 56 additional cases had reportable acquired GA in other genes (eg. STK11, NF1). Median TMB was 4.0 vs 5.2 mut/Mb in the first vs second sample (p = 0.23). In 12% of cases (9 tissue, 9 cfDNA) a BRAF GA was not detected in the second sample. Most AR mechanisms (MET amp, KRAS mut, secondary BRAF GA) were tumor agnostic, but PIK3CA and PTEN GA were enriched in brain samples and absent in CRC, and NRAS mut were exclusive to melanoma (Table). Treatment status was available for a subset of cases. Notably V600E CRC, NSCLC and melanoma each had acquired MET amp post-dabrafenib + trametinib, and a V600E myeloma had acquired MEK C121S post-trametinib + vemurafenib. Additional clinical data will be presented. Table: 1878PD . Potential AR mechanism No. cases# AR subtypes Disease Histologies Associated Primary BRAF GA Biopsy location * KRAS mut 7 G12D (2), G12R, G12V, G13D, Q61H, K117N CRC (2), NSCLC (2), cholangiocarcinoma, multiple myeloma, CLL V600E (6), G466A omentum (2), liver NRAS mut 4 G12C, G13R, G13R/Q61H, Q61H/K melanoma (4) V600E (2), V600R, G469A brain (1), lymph node (1), soft tissue (1) NRAS amp 1 amp estimated copies: 41 NSCLC V600E pericardial fluid Secondary BRAF GA 10 N-terminal deletion exons 2-8 (6), duplications exons 10-18, L505H, N581I/D594G, amp estimated copies: 6 NSCLC (4), CRC (2), melanoma (2), multiple myeloma, pancreatic V600E (9), G466A liver (3), lymph node (2), lung, abdominal wall, brain MEK1 mut 1 C121S multiple myeloma V600E NA PIK3CA mut 5 H1047R (2), G1049R, R88Q, S405F glioma (3), NSCLC, thyroid V600E (3), N486_T491>K, R506_K507insVLR brain (4), lung PTEN GA 5 E7fs * , R130 * , G129R, splice site 165-1G>A, loss melanoma (2), glioma, NSCLC, UP neuroendocrine V600E, V600K, R506_K507insVLR, KHDRBS2-BRAF fusion brain (2), abdomen, soft tissue CCND1 amp 2 amp estimated copies: 9, 10 NSCLC, thyroid V600E, G464V brain, pleural fluid MET amp 4 amp estimated copies: 12, 14, 15, 56 NSCLC, CRC, melanoma, UP adenocarcinoma V600E (4) lymph node, colon, brain, liver * Indicated for tissue samples only (NA= not applicable); #5 cases had AR alterations in multiple genes included here; NSCLC: non-small cell lung cancer, CRC: colorectal carcinoma; CLL: chronic lymphocytic leukemia; UP: unknown primary; AR: acquired resistance; mut: mutation; amp: amplification. Conclusions Novel and previously observed potential AR alterations in paired BRAF altered clinical samples were detected using CGP. Most AR mechanisms appeared independent of tumor type and biopsy site. Additional clinical studies to explore effective treatments for these AR subsets are needed. Legal entity responsible for the study The authors. Funding Foundation Medicine. Disclosure F. Pietrantonio: Advisory / Consultancy: Roche; Advisory / Consultancy: Amgen; Advisory / Consultancy: Eli-Lily; Advisory / Consultancy: Bayer; Advisory / Consultancy: Sanofi; Advisory / Consultancy: Servier; Advisory / Consultancy: Merck Serono. J. Lee: Shareholder / Stockholder / Stock options, Full / Part-time employment: Foundation Medicine; Shareholder / Stockholder / Stock options: Roche. L. Boussemart: Advisory / Consultancy: Novartis; Advisory / Consultancy: Pierre Fabre. G. Srkalovic: Speaker Bureau / Expert testimony: Foundation Medicine. R. Madison: Shareholder / Stockholder / Stock options, Full / Part-time employment: Foundation Medicine; Shareholder / Stockholder / Stock options: Roche. J.S. Ross: Leadership role, Shareholder / Stockholder / Stock options, Full / Part-time employment: Foundation Medicine; Shareholder / Stockholder / Stock options: Roche. V.A. Miller: Leadership role, Shareholder / Stockholder / Stock options, Full / Part-time employment: Foundation Medicine; Shareholder / Stockholder / Stock options: Roche; Advisory / Consultancy: Revolution Medicines. B.M. Alexander: Leadership role, Full / Part-time employment: Foundation Medicine; Shareholder / Stockholder / Stock options: Roche. S.M. Ali: Shareholder / Stockholder / Stock options, Full / Part-time employment: Foundation Medicine; Shareholder / Stockholder / Stock options: Roche. A.B. Schrock: Shareholder / Stockholder / Stock options, Full / Part-time employment: Foundation Medicine; Shareholder / Stockholder / Stock options: Roche. All other authors have declared no conflicts of interest.

Background The CUPISCO trial (NCT03498521) is an ongoing, phase II, randomised, multicentre study comparing molecularly-guided therapy with standard platinum-based chemotherapy in newly diagnosed poor-risk CUP patients. Methods Eligible patients have poor-risk adeno- or undifferentiated CUP as defined by ESMO 2015 guidelines and tissue for molecular sequencing. Local sites initiate the screening process with potentially eligible patients. Patients then undergo central Eligibility Review (ER), a cooperative effort between a central pathology laboratory, external referent oncologists and each site’s investigator and pathology laboratory to confirm the diagnosis. Patients with favourable prognostic subsets or with a strong suspicion of an existing primary site of origin based on immunohistochemistry (IHC) signature and clinical picture are excluded. Results As of 19 March 2019, 157 patients had been screened, of whom 91 (58%) failed screening. Three patients were successfully re-screened. Of the 88 patients who permanently failed screening, 23 were due to technical reasons (e.g. insufficient quality/quantity of tissue for sequencing), 20 for failure to meet inclusion/exclusion criteria not directly related to CUP diagnosis, and 14 for other reasons (e.g. declining health status). A set of 31 patients were not enrolled because the CUP diagnosis could not be confirmed at the IHC level, 19 of those after ER review. Central IHC review results included pathological signatures more typical of specific primary tumours (e.g. prostate cancer or melanoma), or marker combinations typically positive in favourable CUP subsets or rare tumour entities. Conclusions Experience with the CUPISCO study has highlighted challenges with standardised screening and diagnostic processes in an international clinical trial and the difficulties inherent in accurate diagnosis of poor-risk CUP. Confirming a CUP diagnosis for a clinical trial with multiple review checkpoints can result in many reasons for screen failures. By sharing this experience, we aim to foster understanding and to improve diagnostic algorithms for CUP. Clinical trial identification NCT03498521. Editorial acknowledgement Medical writing assistance was provided by Ian Leighton, PhD, Nspm Ltd, Meggen, Switzerland, and supported by F. Hoffmann-La Roche Ltd, Basel, Switzerland. Legal entity responsible for the study F. Hoffmann-La Roche Ltd. Funding F. Hoffmann-La Roche Ltd. Disclosure C. Pauli: Honoraria (self), Research grant / Funding (institution), Travel / Accommodation / Expenses: Roche. T. Bochtler: Honoraria (self), Research grant / Funding (institution), Travel / Accommodation / Expenses: Roche. L. Mileshkin: Travel / Accommodation / Expenses: Roche; Travel / Accommodation / Expenses: Beigene. G. Baciarello: Advisory / Consultancy, Travel / Accommodation / Expenses: Amgen; Advisory / Consultancy, Travel / Accommodation / Expenses: Janssen Oncology; Advisory / Consultancy, Travel / Accommodation / Expenses: Sanofi; Advisory / Consultancy, Travel / Accommodation / Expenses: Astellas-Pharma; Advisory / Consultancy: Roche; Travel / Accommodation / Expenses: AstraZeneca; Travel / Accommodation / Expenses: Ipsen. F. Losa: Advisory / Consultancy, Speaker Bureau / Expert testimony, Research grant / Funding (institution), Travel / Accommodation / Expenses: Roche; Advisory / Consultancy, Research grant / Funding (institution): Amgen; Advisory / Consultancy, Research grant / Funding (institution), Travel / Accommodation / Expenses: Merck; Advisory / Consultancy, Speaker Bureau / Expert testimony: Sanofi; Advisory / Consultancy: Servier. J.S. Ross: Shareholder / Stockholder / Stock options, Full / Part-time employment: Foundation Medicine Inc. S. Yalcin: Honoraria (self): Roche; Honoraria (self): Sanofi; Honoraria (self): Amgen; Honoraria (self): Novartis; Honoraria (self): Lilly; Honoraria (self): MSD; Honoraria (self): Merck Serono. A. Beringer: Full / Part-time employment: F. Hoffmann-La Roche Ltd.. S. Foser: Full / Part-time employment: F. Hoffmann-La Roche Ltd. J. Scarato: Full / Part-time employment: F. Hoffmann-La Roche Ltd.. M. Mueller-Ohldach: Full / Part-time employment: Hoffmann-La Roche Ltd. H. Moch: Honoraria (self), Research grant / Funding (institution), Travel / Accommodation / Expenses: Roche. A. Kramer: Honoraria (self), Research grant / Funding (institution), Travel / Accommodation / Expenses: Roche. All other authors have declared no conflicts of interest.

Background cHCC-CCA is a rare primary liver carcinoma, with histologic features of both hepatocellular carcinomas (HCC) and liver cholangiocarcinomas (CCA). In order to elucidate their shared and distinctive biology, we used CGP to compare the genomic alterations (GA) of cHCC-CCA with those of HCC and CCA. Methods 1269 HCC, 3965 CCA and 44 cHCC-CCA were assayed by hybrid capture based CGP on 0.8-1.1 Mb of the coding genome to identify GAs in exons and select introns in up to 404 genes, tumor mutational burden (TMB), microsatellite status (MSI) and % monoallelic genome (gLOH). Utilizing the significant differences in GAs, biomarkers, and demographics between HCC and CCAs (Fischer exact test, FDR 10% shown, Table), we built a random forest-based machine learning model to rank a cHCC-CCA specimen in the CCA-HCC spectrum (Out of Bag error rate = 12.6%, AUC = 0.94). Results Biomarkers exclusively associated with one disease type included IDH1 (15% in CCA vs 1% in HCC, p = 1e-57), TERT (4% in CCA vs 47% in HCC, p = 2e-273) and Hepatitis B virus (HBV, 0.8% in CCA vs 8.5% in HCC, p = 9e-42). 10/44 cHCC-CCA were classified as CCA-like (IDH1 = 4/10, HBV = 0/10, TERT = 0/10, median TMB = 0.9) and 21/44 cases as HCC-like (IDH1=0/21, HBV=5/21, TERT = 15/21, median TMB = 4). 13/44 cHCC-CCA featured ambiguous/conflicting CGP findings (IDH1=0/13, HBV=1/13, TERT = 0/13, median TMB = 4). 5/13 cHCC-CCA had a computational tumor purity Table . 677PD HCC CCA GAs TERT, CTNNB1, MYC CDKN2A/B, KRAS, ARID1A, IDH1, BAP1, PBRM1, FGFR2 Biomarkers HBV +, low gLOH, intermediate TMB HBV -, high gLOH, low TMB Demographics Male, Younger patients, African & East Asian ancestry Female, Older patients, European ancestry Conclusions Using a machine learning model, the prevalence and mutual exclusivity of IDH1/2, FGFR2, TERT, CTNNB1 and other key GAs, suggest that the majority of cHCC-CCA cases fall into molecular subgroups either similar to CCA or HCC. These findings suggest that cHCC-CCA is not a true disease entity but may be a conglomerate of transformed CCA and HCC. Legal entity responsible for the study Foundation Medicine. Funding Foundation Medicine. Disclosure K. Murugesan: Full / Part-time employment: Foundation Medicine; Shareholder / Stockholder / Stock options: Roche. M. Javle: Honoraria (self), Honoraria (institution), Advisory / Consultancy, Speaker Bureau / Expert testimony, Leadership role, Research grant / Funding (self), Research grant / Funding (institution), Travel / Accommodation / Expenses, Shareholder / Stockholder /: Rafael Pharmaceuticals; Honoraria (self), Honoraria (institution), Advisory / Consultancy, Speaker Bureau / Expert testimony, Leadership role, Research grant / Funding (self), Research grant / Funding (institution), Travel / Accommodation / Expenses, Shareholder / Stockholder /: Incyte; Honoraria (self), Honoraria (institution), Advisory / Consultancy, Speaker Bureau / Expert testimony, Leadership role, Research grant / Funding (self), Research grant / Funding (institution), Travel / Accommodation / Expenses, Shareholder / Stockholder /: Pieris Pharmaceuticals; Honoraria (self), Honoraria (institution), Advisory / Consultancy, Speaker Bureau / Expert testimony, Leadership role, Research grant / Funding (self), Research grant / Funding (institution), Travel / Accommodation / Expenses, Shareholder / Stockholder /: Merck; Honoraria (self), Honoraria (institution), Advisory / Consultancy, Speaker Bureau / Expert testimony, Leadership role, Research grant / Funding (self), Research grant / Funding (institution), Travel / Accommodation / Expenses, Shareholder / Stockholder /: Merck Serono; Honoraria (self), Honoraria (institution), Advisory / Consultancy, Speaker Bureau / Expert testimony, Leadership role, Research grant / Funding (self), Research grant / Funding (institution), Travel / Accommodation / Expenses, Shareholder / Stockholder /: Novartis; Honoraria (self), Honoraria (institution), Advisory / Consultancy, Speaker Bureau / Expert testimony, Leadership role, Research grant / Funding (self), Research grant / Funding (institution), Travel / Accommodation / Expenses, Shareholder / Stockholder /: Seattle Genetics; Honoraria (self), Honoraria (institution), Advisory / Consultancy, Speaker Bureau / Expert testimony, Leadership role, Research grant / Funding (self), Research grant / Funding (institution), Travel / Accommodation / Expenses, Shareholder / Stockholder /: QED Therapeutics; Honoraria (self), Honoraria (institution), Advisory / Consultancy, Speaker Bureau / Expert testimony, Leadership role, Research grant / Funding (self), Research grant / Funding (institution), Travel / Accommodation / Expenses, Shareholder / Stockholder /: Bayer; Honoraria (self), Honoraria (institution), Advisory / Consultancy, Speaker Bureau / Expert testimony, Leadership role, Research grant / Funding (self), Research grant / Funding (institution), Travel / Accommodation / Expenses, Shareholder / Stockholder /: BeiGene. A.B. Schrock: Full / Part-time employment: Foundation Medicine; Shareholder / Stockholder / Stock options: Roche. N. Ngo: Full / Part-time employment: Foundation Medicine; Shareholder / Stockholder / Stock options: Roche. G.M. Frampton: Full / Part-time employment: Foundation Medicine; Shareholder / Stockholder / Stock options: Roche. B.M. Alexander: Full / Part-time employment: Foundation Medicine; Shareholder / Stockholder / Stock options: Roche. V.A. Miller: Full / Part-time employment: Foundation Medicine. T. Bekaii-Saab: Advisory / Consultancy: Amgen; Advisory / Consultancy: Bristol-Myers Squibb; Advisory / Consultancy: Lilly; Advisory / Consultancy: Regeneron; Advisory / Consultancy: AstraZeneca; Advisory / Consultancy: Celgene; Advisory / Consultancy: NCCN; Advisory / Consultancy: Genentech/Roche; Advisory / Consultancy: Bayer; Advisory / Consultancy: Taiho Pharmaceutical; Advisory / Consultancy: Boehringer Ingelheim; Advisory / Consultancy: Sirtex Medical; Advisory / Consultancy: Pfizer; Advisory / Consultancy: Glenmark; Advisory / Consultancy: Ipsen; Advisory / Consultancy: AbbVie; Advisory / Consultancy: Incyte; Advisory / Consultancy: Imugene; Advisory / Consultancy: Immuneering. L.A. Albacker: Full / Part-time employment: Foundation Medicine; Shareholder / Stockholder / Stock options: Roche. J.S. Ross: Full / Part-time employment: Foundation Medicine; Shareholder / Stockholder / Stock options: Foundation Medicine. S.M. Ali: Shareholder / Stockholder / Stock options, Licensing / Royalties, Full / Part-time employment: Foundation Medicine; Advisory / Consultancy: Revolution Medicines; Leadership role: Incysus; Shareholder / Stockholder / Stock options: Exelixis; Shareholder / Stockholder / Stock options: Blueprint Medicines; Shareholder / Stockholder / Stock options: Agios; Shareholder / Stockholder / Stock options: Genocea Biosciences.

Background ENB is a rare skull base tumor arising from the olfactory neuroepithelium with variable prognosis and no consensus guidelines for treatment. Previous genomic studies (Classe 2018, Capper 2018) have reported scant genomic alterations (GA) in advanced and metastatic ENB (mENB), including mutation of IDH2. Here, we explore the cytogenetic landscape of in 94 cases of mENB as assessed by NGS-based comprehensive genomic profiling (CGP). Methods Tumour specimens from 94 patients with mENB were assayed using hybrid capture-based CGP to delineate all classes of GA as point mutations, insertions deletions (indels), rearrangements and focal ( Results mENB cases had 1 or fewer GA detected in 80% (75/94) of cases with a corresponding median tumor mutational burden of 1.5 mut/mb. TP53 was the most frequently altered gene in the cohort, while IDH2 mutations were found in only 2 cases (2.1%). Focal and homozygous deletions were identified in 11.7% (11/94) and 14.9% (14/94) of cases respectively. We then assessed chr arm gain and LOH1/X and identified a median of 11 events per case. Based on chr arm level changes, we identified 3 distinct subtypes of mENB: type IA (n = 53), defined by wide spread LOHx, type IB (n = 22), defined by wide spread LOH1, and type II (n = 19) generally lacking arm CNA. Type IB lacked chr5 or chr20 gain, which were both seen in 36% of type IA (p = 0.001). Patients with IA/B types were older than type II (mean 54.9 v 46.5 years, p = 0.015) and both IDH2 mutated cases segregated into type II. Conclusions Given the pauci-mutational genomic landscape of mENB, clustering by chr arm level changes identifies distinct classes of mENB, which were associated with biologic factors including age and IDH2 mutation. Further studies to correlate clinico-pathologic characteristics with the cytogenetic characteristics of these newly defined subtypes of mENB are needed. Legal entity responsible for the study Foundation Medicine Inc. Funding Foundation Medicine Inc. Disclosure R. Madison: Full / Part-time employment: Foundation Medicine Inc; Shareholder / Stockholder / Stock options: Roche. D.C. Pavlick: Full / Part-time employment: Foundation Medicine Inc; Shareholder / Stockholder / Stock options: Roche. S. Khan: Honoraria (self): Ariad; Honoraria (self): Genentech; Honoraria (self): Foundation Medicine; Honoraria (self): EMD Serono; Honoraria (self): Genzyme; Honoraria (self): Guardant; Honoraria (self): Takeda; Honoraria (self): Oak Ridge Universities; Honoraria (self): Onyx Pharmaceuticals. J. Lee: Full / Part-time employment: Foundation Medicine Inc; Shareholder / Stockholder / Stock options: Roche. J.S. Ross: Leadership role, Full / Part-time employment: Foundation Medicine Inc; Shareholder / Stockholder / Stock options: Roche; Officer / Board of Directors: Celcius Therapeutics. V.A. Miller: Leadership role, Full / Part-time employment: Foundation Medicine Inc; Shareholder / Stockholder / Stock options: Roche; Advisory / Consultancy, Scientific Advisory Board: Revolution Medicines. B.M. Alexander: Full / Part-time employment, Officer / Board of Directors: Foundation Medicine Inc; Shareholder / Stockholder / Stock options: Roche; Advisory / Consultancy, Personal Fees: AbbVie; Advisory / Consultancy, Personal Fees: Schlesinger Associates; Advisory / Consultancy, Personal Fees: Bristol-Myers Squibb; Advisory / Consultancy, Personal Fees: Precision Health Economics; Research grant / Funding (self), grants outside submitted work: Puma Biotechnology; Research grant / Funding (self), grants outside submitted work: Celgene; Research grant / Funding (self), grants outside submitted work: Eli Lily. J. Chung: Full / Part-time employment: Foundation Medicine Inc; Shareholder / Stockholder / Stock options: Roche. All other authors have declared no conflicts of interest. A.B. Schrock: Full / Part-time employment: Foundation Medicine Inc; Shareholder / Stockholder / Stock options: Roche. S.M. Ali: Full / Part-time employment: Foundation Medicine Inc; Shareholder / Stockholder / Stock options: Roche.

Background In PURE01 study (NCT02736266), neoadjuvant pembro resulted in 42% pT0 in patients (pts) with MIBC. pT0 pts had features suggesting pre-existing immunity or higher tumor mutational burden (TMB) may promote response. In this study, we investigated potential mechanisms for resistance to pembro, including FGFR3 genomic alterations (GA). Methods Pts enrolled in the PURE-01, which is still recruiting pts in its amended design, had predominant urothelial carcinoma histology and stage cT ≤ 4N0 MIBC. Biomarker analyses in the expanded cohort of 96 pts included PD-L1 combined positive score and comprehensive genomic profiling (FoundationCDx assay). In addition, TCGA MIBC (n = 405) and a prospective commercial cohort (PCC) of 415 MIBC pts from the clinical use of the Decipher Bladder TURB test from the GRID registry (NCT02609269) were analyzed. A single-sample genomic classifier (GC) was trained to identify FGFR3-active tumors (FGFR3+). Results In PURE01 cohort, despite a linear association of TMB with ypT0 in multivariable models (p = 0.017), there was no association between FGFR3 GA and pathological response nor with PD-L1 expression. FGFR3 GA were found in 17.6% ypT0 vs 20.7% in ypT3-4 disease (corrected p = 0.82). Applying the GC to the TCGA MIBC cohort, we found 45% of FGFR3+ cases had FGFR3 mutations vs 9% for the rest of the cohort. Applying the GC to the PCC, we found that the FGFR3+ tumors were Luminal (n = 14), Luminal Infiltrated (n = 10) or Luminal Papillary (n = 30). FGFR3+ pts showed significantly lower PD-L1 (-0.03 vs. 0.109, p Conclusions In our combined cohorts, FGFR3 GA did not correlate with response to pembro. FGFR3 GA are enriched in FGFR3+ tumors, which in turn tend to have lower immune activity, PD-L1 and PD-L2, suggesting FGFR3 activity may provide a potential tool for further discriminating the mechanisms underlying response and resistance to neoadjuvant pembro in MIBC. Clinical trial identification NCT02736266 (PURE01) NCT02609269 (GRID Registry). Legal entity responsible for the study The authors. Funding Decipher Bioscience. Disclosure A. Necchi: Honoraria (institution), Advisory / Consultancy, Research grant / Funding (institution): Merck; Honoraria (institution), Advisory / Consultancy: Roche; Advisory / Consultancy: AstraZeneca; Advisory / Consultancy: BMS; Advisory / Consultancy, Research grant / Funding (institution): Incyte; Advisory / Consultancy: Clovis; Advisory / Consultancy: Janssen. R. Madison: Full / Part-time employment: Foundation Medicine. J. Chung: Full / Part-time employment: Foundation Medicine. Y. Liu: Full / Part-time employment: Decipher Bioscience. J. Chung: Full / Part-time employment: Foundation Medicine. J.S. Ross: Full / Part-time employment: Foundation Medicine. S.M. Ali: Full / Part-time employment: Foundation Medicine. E. Davicioni: Full / Part-time employment: Decipher Bioscience. E.A. Gibb: Full / Part-time employment: Decipher Bioscience. All other authors have declared no conflicts of interest.

Background Standard CUP therapy has not changed for decades. CUPISCO (NCT03498521) is an ongoing randomised prospective trial using CGP to assign CUP pts to individualised targeted or immunotherapy arms (molecularly guided therapy; MGT). Methods Archival tissue from 303 centrally reviewed undifferentiated- and adeno-CUP cases in the FoundationCore™ database underwent hybrid capture-based CGP (FoundationOne® CDx). Microsatellite instability (MSI), tumour mutational burden (TMB) and genomic loss of heterozygosity (gLOH) were calculated (Frampton Nat Biotechnol 2013; Swisher Lancet Oncol; Chambers Genome Med 2017). PD-L1 expression was measured by DAKO 22C3 immunohistochemistry. Pts were classified by whether CGP results could have informed assignment to CUPISCO arms; TTF-1+, CK7–/CK20+/CDX2+ or TMPRSS2:ERG+ cases were excluded. Results The sex ratio was 1:1; median age was 67 yrs (range 22–89+). CGP revealed 96 pts (32%) matched to a CUPISCO arm (Table). Table . 1983PD_PR CUPISCO arm % Genomic alterations Alectinib 0.66 Vismodegib 1.32 Ipatasertib 8.25 Olaparib 5.61 Erlotinib + bevacizumab 2.31 Vemurafenib + cobimetinib 2.97 Subcutaneous trastuzumab + pertuzumab + chemo 9.24 Atezolizumab 9.24 Entrectinib (in development) 0.33 Mean TMB was 8 mut/Mb; 23% had ≥10 (low), 12%, ≥16 (int) and 9%, ≥20 (high). 3 (1%) had high MSI and 20%, gLOH ≥16. 42 (14%) had high PD-L1 (tumour proportion score ≥50%). Key genomic alterations included HER2 (7%), PIK3CA, NF1 (6% each), NF2 (5%), BRAF, PTEN, FGFR2, EGFR, MET (all 4%), CDK6 (3%), FBXW7 and CDK4 (2% each). Key gene fusions involved ALK, RET and ROS1 (all 1%). KRAS was mutated in 27%; 6% had G12C alterations. Of the 23 assessable high TMB cases, 8 (35%) had a tobacco and 5 (22%) a UV light or mismatch repair mutational signature. 20% of cases harboured a putative cancer-associated germline DNA mutation. Conclusions 32% of CUP pts would have been potentially eligible for MGT in CUPISCO. Future studies including additional biomarkers, such as PD-L1–positivity and gLOH, may identify a greater proportion of CUP pts potentially benefitting from individualised treatment. Editorial acknowledgement Support for third-party editing assistance for this abstract, furnished by Daniel Clyde, PhD, of Health Interactions, was provided by F. Hoffmann-La Roche Ltd, Basel, Switzerland. Legal entity responsible for the study F. Hoffmann-La Roche Ltd. Funding F. Hoffmann-La Roche Ltd. Disclosure J.S. Ross: Leadership role, Shareholder / Stockholder / Stock options, Full / Part-time employment: Foundation Medicine, Inc.; Non-remunerated activity/ies, Support for third-party editing assistance: F. Hoffmann-La Roche Ltd. E.S. Sokol: Shareholder / Stockholder / Stock options, Full / Part-time employment: Foundation Medicine, Inc; Non-remunerated activity/ies, Support for third-party editing assistance: F. Hoffmann-La Roche Ltd. H. Moch: Advisory / Consultancy, Research grant / Funding (institution), Travel / Accommodation / Expenses: F. Hoffmann-La Roche Ltd ; Advisory / Consultancy, Travel / Accommodation / Expenses: Definiens AG; Non-remunerated activity/ies, Support for third-party editing assistance: F. Hoffmann-La Roche Ltd. L. Mileshkin: Travel / Accommodation / Expenses: F. Hoffmann-La Roche Ltd; Travel / Accommodation / Expenses: Beigene; Non-remunerated activity/ies, Support for third-party editing assistance: F. Hoffmann-La Roche Ltd. G. Baciarello: Honoraria (self), Travel / Accommodation / Expenses: F. Hoffmann-La Roche Ltd; Honoraria (self), Travel / Accommodation / Expenses: Astellas; Honoraria (self), Advisory / Consultancy, Travel / Accommodation / Expenses: Janssen; Honoraria (self), Travel / Accommodation / Expenses: Ipsen; Honoraria (self): Amgen; Travel / Accommodation / Expenses: AstraZeneca; Non-remunerated activity/ies, Support for third-party editing assistance: F. Hoffmann-La Roche Ltd; Advisory / Consultancy, Travel / Accommodation / Expenses: Sanofi. F. Losa: Advisory / Consultancy, Speaker Bureau / Expert testimony, Research grant / Funding (institution), Travel / Accommodation / Expenses: F. Hoffmann-La Roche Ltd; Advisory / Consultancy, Speaker Bureau / Expert testimony: Sanofi; Advisory / Consultancy, Research grant / Funding (institution): Amgen; Advisory / Consultancy, Research grant / Funding (institution), Travel / Accommodation / Expenses: Merck; Advisory / Consultancy: Servier; Non-remunerated activity/ies, Support for third-party editing assistance: F. Hoffmann-La Roche Ltd. A. Beringer: Full / Part-time employment, Non-remunerated activity/ies, Support for third-party editing assistance: F. Hoffmann-La Roche Ltd. M. Thomas: Leadership role, Shareholder / Stockholder / Stock options, Full / Part-time employment, Non-remunerated activity/ies, Support for third-party editing assistance: F. Hoffmann-La Roche Ltd. S. Foser: Leadership role, Shareholder / Stockholder / Stock options, Full / Part-time employment, Non-remunerated activity/ies, Support for third-party editing assistance: F. Hoffmann-La Roche Ltd. J. Elvin: Shareholder / Stockholder / Stock options, Full / Part-time employment: Foundation Medicine, Inc; Non-remunerated activity/ies, Support for third-party editing assistance: F. Hoffmann-La Roche Ltd. N. Danziger: Shareholder / Stockholder / Stock options, Full / Part-time employment: Foundation Medicine, Inc; Non-remunerated activity/ies, Support for third-party editing assistance: F. Hoffmann-La Roche Ltd. N. Ngo: Shareholder / Stockholder / Stock options, Full / Part-time employment: Foundation Medicine, Inc; Non-remunerated activity/ies, Support for third-party editing assistance: F. Hoffmann-La Roche Ltd. J.Y. Tse: Shareholder / Stockholder / Stock options, Full / Part-time employment: Foundation Medicine, Inc; Non-remunerated activity/ies, Support for third-party editing assistance: F. Hoffmann-La Roche Ltd. K. Killian: Shareholder / Stockholder / Stock options, Full / Part-time employment: Foundation Medicine, Inc; Full / Part-time employment, Special Volunteer: National Cancer Institute; Shareholder / Stockholder / Stock options, Non-remunerated activity/ies, Support for third-party editing assistance: F. Hoffmann-La Roche Ltd. D.X. Jin: Shareholder / Stockholder / Stock options, Full / Part-time employment: Foundation Medicine, Inc; Shareholder / Stockholder / Stock options, Non-remunerated activity/ies, Support for third-party editing assistance: F. Hoffmann-La Roche Ltd. L.M. Gay: Full / Part-time employment, Current: Ellem Consulting; Non-remunerated activity/ies, Support for third-party editing assistance: F. Hoffmann-La Roche Ltd; Full / Part-time employment, Previous: Foundation Medicine, Inc. A. Kramer: Honoraria (self), Honoraria (institution), Advisory / Consultancy, Travel / Accommodation / Expenses, Non-remunerated activity/ies, Support for third-party editing assistance: F. Hoffmann-La Roche Ltd; Honoraria (self), Advisory / Consultancy, Travel / Accommodation / Expenses: Daiichi Sankyo; Honoraria (self), Advisory / Consultancy: Abbvie; Honoraria (institution), Research grant / Funding (institution): Bayer; Advisory / Consultancy: BMS; Travel / Accommodation / Expenses: Celgene; Research grant / Funding (institution): Merck.

Background GBAC is an aggressive neoplasm with an abysmal 5-year survival rate and no approved targeted therapies. Incidence of GBAC varies greatly across global geographical regions, with highest reported rates in South and East Asia. We evaluated whether GBAC patients have differential genomic profiles as correlated to ancestry. Methods CGP was performed by an adaptor-ligation/hybrid capture-based assay of coding DNA of 315 cancer-related genes and select introns of 28 genes that are frequently involved in genomic rearrangements to detect base substitutions, insertions and deletions, copy number alterations, and rearrangements. Tumor mutational burden (TMB) was determined on 1.1 Mbp of sequenced DNA and microsatellite instability (MSI) was assessed across 114 homopolymeric loci. Patient ancestry was determined by single nucleotide polymorphism microarray data, ancestry informative markers, and principal component analysis. Results The genomic profiles of tumour samples from 881 patients with GBAC were reviewed. Of these samples, 22 (2.50%) were derived from individuals of South Asian ancestry (SAS), 61 (6.92%) from East Asian (EAS) ancestry and the remainder were distributed among African, American, and European ancestry (Others). Age and gender were similar across all three groups, except for an enrichment in SAS male patients (SAS male 68.2%, EAS male 32.7%, Others male 31.2%; p = 0.001). MSI was uncommon in all three groups (0.007% overall) and TMB was low (median 2.6 mut/mb). The most commonly altered genes across all three groups were TP53, CDKN2A/B and SMAD4 (all: 64%, 53%, 17%) while alterations in ERBB2 (23% v 30% v 14%, p = 0.03) and CTNNB1 (14% v 11% v 3%, p = 0.001) were enriched in SAS and EAS populations. Amongst ERBB2 altered cases, amplifications were seen at similar frequencies across all three groups (80% v 72% v 72%), while less than half of all ERBB2 point mutations occurred at the S310 codon (18/67, 27%). Conclusions GBAC patients of SAS and EAS ancestry were enriched for ERBB2 and CTNBB1 relative to other ancestries. Follow-up studies are needed to understand the genomic context of this finding, as well as the influence of cancer predisposition genes in the context of ancestry specific oncogenesis. Legal entity responsible for the study Foundation Medicine Inc. Funding Foundation Medicine Inc. Disclosure M. Javle: Research grant / Funding (self): QED Therapeutics; Honoraria (self): Merck & Co; Advisory / Consultancy: EDO Pharma; Advisory / Consultancy: More Health; Honoraria (self): Taiho Pharmaceutical; Honoraria (self): Seattle Genetics; Research grant / Funding (self): Meclun; Research grant / Funding (institution): Arqule; Research grant / Funding (institution): Lilly; Advisory / Consultancy: Origimed; Research grant / Funding (self): Novartis. S. Akumalla: Full / Part-time employment: Foundation Medicine; Shareholder / Stockholder / Stock options: Roche. R. Madison: Full / Part-time employment: Foundation Medicine; Shareholder / Stockholder / Stock options: Roche. J. Newburg: Shareholder / Stockholder / Stock options: Foundation Medicine Inc; Shareholder / Stockholder / Stock options: Roche. B.M. Alexander: Leadership role, Full / Part-time employment, Officer / Board of Directors: Foundation Medicine Inc; Shareholder / Stockholder / Stock options: Roche; Advisory / Consultancy, Personal Fees: AbbVie; Advisory / Consultancy, Personal Fees: Schlesinger Associates; Advisory / Consultancy, Personal Fees: Bristol-Myers Squibb; Advisory / Consultancy, Personal Fees: Precision Health Economics; Research grant / Funding (self), grants outside submitted work: Puma Biotechnology; Research grant / Funding (self), grants outside submitted work: Celgene; Research grant / Funding (self), grants outside submitted work: Eli Lily. J. Chung: Full / Part-time employment: Foundation Medicine Inc; Shareholder / Stockholder / Stock options: Roche. V.A. Miller: Full / Part-time employment: Foundation Medicine Inc; Shareholder / Stockholder / Stock options: Roche; Advisory / Consultancy, Scientific Advisory Board: Revolution Medicines. J. Lee: Full / Part-time employment: Foundation Medicine Inc; Shareholder / Stockholder / Stock options: Roche. J.S. Ross: Leadership role, Full / Part-time employment: Foundation Medicine Inc; Shareholder / Stockholder / Stock options: Roche; Officer / Board of Directors: Celcius Therapeutics. A.B. Schrock: Full / Part-time employment: Foundation Medicine Inc; Shareholder / Stockholder / Stock options: Roche. G.M. Frampton: Full / Part-time employment: Foundation Medicine Inc; Shareholder / Stockholder / Stock options: Roche. S.M. Ali: Full / Part-time employment: Foundation Medicine Inc; Shareholder / Stockholder / Stock options: Roche; Advisory / Consultancy, Scientific Advisory Board: Incysus Therapeutics, Inc.; Advisory / Consultancy: Revolution Medicines. All other authors have declared no conflicts of interest.

Background Succinate dehydrogenase (SDH) is a highly conserved component of energy production and metabolism. Genetic or epigenetic inactivation of SDH (dSDH) creates a state of intracellular pseudohypoxia that is uniquely oncogenic for subtypes of GIST, renal cell carcinoma (RCC), paraganglioma/pheochromocytoma (PGL/Pheo), and pituitary adenoma/carcinoma (PAC). dSHD stems from homozygous loss of one of four subunit encoding genes SDHA, SDHB, SDHC, or SDHD (aka, SDHx). dSDH is not targetable, and preclinical models are poorly viable. In the current study we analyze comprehensive genomic profiles (CGP) from a dSDH pan-tumor cohort for potential therapy-enabling non-SDHx genomic alterations and biomarkers (GA). Methods 231,706 clinical grade CGP were searched for dSDH GIST, RCC, PGL/Pheo, and PAC. GA were determined on 1.1 Mb genome sequence encompassing up to 324 cancer genes. dSDH was further established by histopathology review and analysis of SDHx homozygous loss. Germline and zygosity status of SDHx variants were performed by SGZ algorithm. Non-SDHx GA were assessed for therapeutic actionability. Results 82 SDH-deficient neoplasms were analyzed (Table). 0 of 38 GIST had highly actionable GA. In one case a subclonal KIT variant was identified in a recurrence of an SDHA-mutant GIST treated with Imatinib for several years. No clonal canonical driver kinase mutations were identified in KIT, PDGFRA, NF1, BRAF, FGFR1 or NTRK1-3. 0 of 36 dSDH PGL/Pheo and 0 of 1 PAC had potential therapeutic options. 2 of 7 dSDH RCC reported potential treatment options for oncogenic variants in EGFR and CDK4. No dSDH tumors had high TMB, microsatellite instability, or LOH score indicative of homologous recombination defect (HRD). Overall, 2 of 82 tumors (2.4%) harbored highly actionable GA. Table . 2039P 82 SDH-deficient tumors GIST (n = 38) PGL/Pheo (n = 36) RCC (n = 7) PAC (n = 1) Age (mean) 33 40 49 44 Sex ratio (M/F) 16/21 24/12 4/3 0/1 SDHA 11 12 3 1 SDHB 7 21 3 0 SDHC 4 1 1 0 SDHD 2 2 0 0 SDHx-WT 14 0 0 0 %SDHx germline 69 64 100 50 GA/tumor (pathogenic non-SDHx) 0.58 0.89 2.29 0 MSI 0 0 0 0 TMB (mut/Mb, mean) 1.86 2.4 3 0.9 HRD (LOHscore >16%) 0 0 0 0 Conclusions This study identifies SDH deficiency as a lone driver of malignancy and without associated actionable GA in > 97% of cases, and underscores the need for novel therapeutic approaches targeting SDH deficiency itself. Legal entity responsible for the study Foundation Medicine. Funding Foundation Medicine. Disclosure J.K. Killian: Full / Part-time employment: Foundation Medicine. D.C. Pavlick: Full / Part-time employment: Foundation Medicine. E.S. Sokol: Full / Part-time employment: Foundation Medicine. M. Montesion: Full / Part-time employment: Foundation Medicine. D.X. Jin: Full / Part-time employment: Foundation Medicine. B. Kaplan: Full / Part-time employment: Foundation Medicine. D. Lin: Full / Part-time employment: Foundation Medicine. J. Vergilio: Full / Part-time employment: Foundation Medicine. J.A. Elvin: Full / Part-time employment: Foundation Medicine. N. Ngo: Full / Part-time employment: Foundation Medicine. E. Severson: Full / Part-time employment: Foundation Medicine. S. Ramkissoon: Full / Part-time employment: Foundation Medicine. D. Duncan: Full / Part-time employment: Foundation Medicine. C. Edgerly: Full / Part-time employment: Foundation Medicine. A. Hemmerich: Full / Part-time employment: Foundation Medicine. G.M. Frampton: Full / Part-time employment: Foundation Medicine. V.A. Miller: Full / Part-time employment: Foundation Medicine. S.M. Ali: Full / Part-time employment: Foundation Medicine. J.S. Ross: Full / Part-time employment: Foundation Medicine. All other authors have declared no conflicts of interest.

Background KRAS GA are common oncogenic drivers. Effective systemic treatment of KRAS altered cancers has been largely elusive; however covalent inhibitors of G12C (AMG510, MRTX849) and SHP2 inhibitors (TNO155, RMC-4630) have recently entered the clinic in multiple tumor types. Methods Hybrid capture-based comprehensive genomic profiling (CGP) was performed on tumor samples from 213,312 unique patients with solid or hematological malignancies. Tumor mutational burden (TMB) was determined on 0.8-1.1 Mbp of sequenced DNA. PD-L1 expression was determined by IHC (22C3 or SP142 antibodies) for 17% of cases. Results KRAS GA were detected in 22% (n = 46,182) of cases: 88% mutations (m; >99% substitutions and A), and G12V (primarily G>T) and G12C (exclusively G>T) from transversions. G12C was most frequent in LUAD (14%, 3,613/25,968) and PSC (11%, 40/353) but uncommon in pancreatic (1.7%, 161/9,723) and GI (2.7%, 941/35,019) where G12D/V were most common. Tobacco signature (TS; 20% vs 4.5% vs 3.5%), elevated TMB (median 7.0 vs 3.5 vs 3.5 mut/Mb) and PD-L1 positivity (56% vs 29% vs 31%) were all significantly associated with G12C vs non-G12C KRASm and non-KRASm cancers (all p Conclusions Diverse KRAS GA are frequent across cancers and subtypes are associated with different solid malignancies. G12C is most common in LUAD and associated with smoking, elevated TMB and PD-L1 positivity. As multiple trials of novel therapeutic agents are currently enrolling for KRAS GA, CGP to identify these alterations is needed. Legal entity responsible for the study The authors. Funding Foundation Medicine. Disclosure S.I. Ou: Honoraria (self), Advisory/Consultancy, Research grant/Funding (institution): Novartis; Honoraria (self), Advisory/Consultancy, Research grant/Funding (institution): Pfizer; Honoraria (self), Advisory/Consultancy, Research grant/Funding (institution): Roche; Honoraria (self), Advisory/Consultancy, Speaker Bureau/Expert testimony: Genentech; Honoraria (self), Advisory/Consultancy, Speaker Bureau/Expert testimony, Research grant/Funding (institution): Ariad; Research grant/Funding (institution): Ignyta; Honoraria (self), Advisory/Consultancy, Speaker Bureau/Expert testimony: Takeda; Honoraria (self), Advisory/Consultancy, Research grant/Funding (institution): AstraZeneca; Research grant/Funding (institution): MedImmune; Research grant/Funding (institution): Clovis; Research grant/Funding (institution): Peregrine; Research grant/Funding (institution): GlaxoSmithKline; Research grant/Funding (institution): Astellas; Research grant/Funding (institution): TP Therapeutics; Research grant/Funding (institution): Blueprint Medicines. E.S. Sokol: Shareholder/Stockholder/Stock options, Full/Part-time employment: Foundation Medicine; Shareholder/Stockholder/Stock options: Roche. R. Madison: Shareholder/Stockholder/Stock options, Full/Part-time employment: Foundation Medicine; Shareholder/Stockholder/Stock options: Roche. J. Chung: Shareholder/Stockholder/Stock options: Roche; Shareholder/Stockholder/Stock options, Full/Part-time employment: Foundation Medicine. J.S. Ross: Leadership role, Shareholder/Stockholder/Stock options, Full/Part-time employment: Foundation Medicine; Shareholder/Stockholder/Stock options: Roche. V.A. Miller: Leadership role, Shareholder/Stockholder/Stock options, Full/Part-time employment: Foundation Medicine; Shareholder/Stockholder/Stock options: Roche; Advisory/Consultancy: Revolution Medicines. B.M. Alexander: Leadership role, Full/Part-time employment: Foundation Medicine; Shareholder/Stockholder/Stock options: Roche. S.M. Ali: Shareholder/Stockholder/Stock options, Full/Part-time employment: Foundation Medicine; Shareholder/Stockholder/Stock options: Roche; Advisory/Consultancy: Revolution Medicines. A.B. Schrock: Full/Part-time employment: Foundation Medicine; Shareholder/Stockholder/Stock options: Roche. S.S. Ramalingam: Advisory/Consultancy, Research grant/Funding (institution): AstraZeneca; Advisory/Consultancy, Research grant/Funding (institution): Amgen; Advisory/Consultancy, Research grant/Funding (institution): Brystol Myers Squib; Advisory/Consultancy: Genentech/Roche; Advisory/Consultancy, Research grant/Funding (institution): Merck; Advisory/Consultancy, Research grant/Funding (institution): Tesaro; Advisory/Consultancy: Takeda; Advisory/Consultancy: Nektar.

Background Interim results from APACHE (NCT03081923) have been presented, with the monotherapy arm stopped for futility (Necchi, Eur Urol 2018). Updated results with the expanded combination therapy cohort are presented. Methods Apache is an open-label, randomized, 3-stage, phase 2 study. Pts who have failed ≥2 chemotherapy (CT) regimens received Durva, 1.5 gr q4w, x 13 cycles (arm A) or Durva plus with Treme, 75 mg q4w, x 4 cycles, followed by Durva alone x total 13 cycles (arm B). The primary endpoint is the modified objective response-rate (mORR=RECIST 1.1 complete or partial response [PR] or stable disease [SD]+STM reduction >10%). H0: mORR rate ≤10%, H1: mORR ≥25% (α and β = 10%). The total sample size of 120 pts is split into 3 stages: in stage 1, each arm is terminated whenever no response is observed in 11 pts/arm. In stage 2, 29 additional pts/arm are enrolled. Biomarker analyses include: PD-L1 expression on immune cells (Ventana SP142) and comprehensive genomic profiling (CGP) with FoundationOne assay. Results From 02/17-11/18, 29 pts were enrolled (11 arm A and 18 arm B). 22 had gonadal and 7 extragonadal GCT, 19 had received >/=3 prior CT regimens. Median tumour mutational burden (TMB) was 4 mutations (mut)/mb, 1 pt showed microsatellite instability in arm A. 10 pts (34.5%) experienced any-Grade AEs, without differences between arms. In the expanded arm B, 2 responses (11.1%, 1 RECIST-PR in seminoma and 1 SD with STM reduction in nonseminoma) and 2 SD were observed. PD occurred regardless of PD-L1 expression and TMB. After median follow-up of 16.3 months, 12-month OS in arm A was 30% (95%CI: 7.2-57.7) and in arm B was 29.6% (95%CI: 10.2-52.2). Further CT post-PD was administered in 8 pts (27.6%), but was mostly ineffective (7/8 SD/PD). Conclusions Despite objective responses were rare with IO, long-term OS results are encouraging in very highly pre-treated GCT pts, suggesting a role for maintenance approach in earlier setting (2nd-line CT), and privileging Durva+Treme due to more responses. Conventional IO biomarkers did not allow for patient selection, and further research is warranted. Clinical trial identification NCT03081923. Legal entity responsible for the study AstraZeneca. Funding AstraZeneca. Disclosure S.M. Ali: Full / Part-time employment: Foundation Medicine. J. Chung: Full / Part-time employment: Foundation Medicine. J.S. Ross: Full / Part-time employment: Foundation Medicine. A. Necchi: Honoraria (institution), Advisory / Consultancy: Merck; Honoraria (institution), Advisory / Consultancy: AstraZeneca; Honoraria (institution), Advisory / Consultancy, Spouse / Financial dependant: Bayer; Advisory / Consultancy: BMS; Advisory / Consultancy: Clovis; Honoraria (institution), Advisory / Consultancy: Janssen; Honoraria (institution), Advisory / Consultancy: Roche. All other authors have declared no conflicts of interest.

Background Using comprehensive genomic profiling (CGP) and immunohistochemistry (IHC) we compared the frequency of IO biomarkers in ACB, UBC and SCCB. We also report the interim results of neoadjuvant pembrolizumab in patients (pts) with muscle-invasive bladder cancer (MIBC) and variant histologies from PURE01 (NCT02736266). Methods Within FMI database, 143 cases of ACB, 2,142 cases of UCB and 83 cases of SCCB were subjected to CGP using a hybrid-capture based assay. Tumor mutational burden (TMB) was determined on 1.1 Mbp of sequenced DNA and microsatellite instability (MSI) was determined on 114 loci. PD-L1 expression was determined by IHC using the Ventana SP-142 assay with >1% tumor cell or immunocyte (TILs) scoring positive. Among 96 pts of PURE01, 15 had predominant variant SCCB histology. Results ACB patients were younger and more often female than UBC and SCCB (P /= 20 mut/Mb (P Table . 930P Adenocarcinoma (ACB) Urothelial Carcinoma (UBC) Squamous Cell Carcinoma (SCCB) Cases 143 2,142 83 Median Age (Range) In years 58 (24-83) 67 (19-88) 62 (31-88) Males/Females 57/86 1597/545 45/38 GA/tumor 5.4 7.7 8.2 MSI High 2/106 (2%) 11/1661 (1%) 1/69 (1%) CD274 (PD-L1) gene amplification 0 (0%) 17 (1%) 4 (5%) Mean TMB 2.4 mut/Mb 9.9 mut/Mb 10.4 mut/Mb TMB >/= 10 mut/Mb 14 (10%) 697 (32%) 26 (31%) TMB >/= 20 mut/Mb 4 (3%) 243 (11%) 13 (16%) PD-L1 IHC Positive Tumor Cells 2/11 (18%) 76/244 (31%) 3/10 (30%) PD-L1 IHC Positive TILs 0/11 (0%) 74/244 (29%) 3/10 (30%) Conclusions Deep sequencing reveals significant differences in IO biomarkers among the 3 major types of bladder carcinomas. UBC and SCCB have higher frequencies of high TMB and PD-L1 expression than ACB and SCCB has the highest frequency of CD274amplification. Neoadjuvant pembrolizumab showed preliminary activity in SCCB. Further study of IO biomarkers in coordination with therapy response and inclusion of SCCB appear warranted in perioperative IO trials. Legal entity responsible for the study Foundation Medicine. Funding Foundation Medicine. Disclosure J.S. Ross: Full / Part-time employment: Foundation Medicine. S.M. Ali: Full / Part-time employment: Foundation Medicine. J. Chung: Full / Part-time employment: Foundation Medicine. A.B. Schrock: Full / Part-time employment: Foundation Medicine. R. Madison: Full / Part-time employment: Foundation Medicine. B.M. Alexander: Full / Part-time employment: Foundation Medicine. A. Necchi: Advisory / Consultancy: Merck; Honoraria (self), Honoraria (institution), Advisory / Consultancy: Roche; Honoraria (institution), Advisory / Consultancy: AstraZeneca; Honoraria (institution), Advisory / Consultancy: BMS; Honoraria (institution), Advisory / Consultancy: Janssen; Honoraria (institution), Advisory / Consultancy: Clovis; Honoraria (institution), Advisory / Consultancy: Bayer. All other authors have declared no conflicts of interest.

Objective Targeted next generation sequencing (NGS) was evaluated for its ability to identify unanticipated targetable genomic alterations (GA) for patients with relapsed ovarian epithelial carcinoma (OC). Methods DNA sequencing was performed for 3320 exons of 182 cancer-related genes and 37 introns of 14 genes frequently rearranged in cancer on indexed, adaptor ligated, hybridization-captured libraries using DNA isolated from FFPE sections from 48 histologically verified relapsed OC specimens. The original primary tumor was sequenced in 26 (54%) of the cases and recurrent/metastatic tumor site biopsies were sequenced in 22 (46%) of the cases. Actionability was defined as: GA that predict sensitivity or resistance to approved or standard therapies or are inclusion or exclusion criteria for specific experimental therapies in NCI registered clinical trials. Results There were 38 (80%) serous, 5 (10%) endometrioid, 3 (6%) clear cell, 1 mucinous (2%) and 1 (2%) undifferentiated carcinomas. 141 GA were identified with an average of 2.9 GA (range 0–8) per tumor, of which 67 were actionable for an average of 1.4 actionable GA per patient (range 0–5). 33/48 (69%) of OC patient samples harbored at least one actionable GA. Most common GA were TP53 (79%); MYC (25%); BRCA1 / 2 (23%); KRAS (16.6%) and NF1 (14.5%). One tumor featured an ERBB2 point mutation. One of 3 (33%) of clear cell tumors featured cMET amplification validated by both FISH and IHC. Conclusions NGS assessment of therapy resistant OC identifies an unexpectedly high frequency of GA that could influence targeted therapy selection for the disease.

Introduction L’association d’inhibiteurs de BRAF et de MEK a recemment obtenu l’AMM dans le traitement du melanome mute V600, avec des taux de reponse avoisinant les 70 %. Des reponses aux therapies ciblees ont aussi ete rapportees pour des mutations autres que BRAF V600. Ainsi, une detection sensible, fiable et non exhaustive des alterations de BRAF est indispensable pour pouvoir offrir aux patients le plus de chances d’avoir acces a une therapie ciblee correspondant au profil genetique de leur melanome. Materiel et methodes Etude multicentrique retrospective de 385 cas consecutifs de melanomes pour lesquels une mutation ou un rearrangement de BRAF a ete identifie par technique NGS (sequencage de nouvelle generation) a partir de capture par hybridation. Resultats Les resultats d’analyse de BRAF, faite avant NGS a partir de capture par hybridation, etaient disponibles pour 79 cas (21 %), par PCR (n = 30), Sanger (n = 13), immuno-histochimie (n = 10), NGS sans capture par hybridation (n = 9), ou autre methode non precisee (n = 17). Parmi les cas mutes BRAF V600, 11/57 (19 %) avec donnees disponibles etaient negatifs par la technique anterieure, dont 2/11 (18 %) avec confirmation qu’il s’agissait du meme prelevement analyse. Parmi les cas avec mutations de BRAF V600 finalement trouvee, il n’y avait pas de difference significative en termes de frequence allelique (mediane 35 % vs. 40 %, p = 0,25) ou de pourcentage de cellules tumorales (mediane 50 % dans les deux groupes, p = 0,97) entre les prelevements ayant un test auparavant positif vs auparavant negatif pour BRAF. Des resultats auparavant negatifs pour BRAF etaient aussi trouves dans 16/20 (80 %) des cas avec mutation non-V600, parmi lesquels deux presentaient de multiples alterations de BRAF [K601E (4), D594A/G/N (4), S467L (2), L584F (2), G464 V, G466 V, G469 V, E586 K, N581I, L597Q, A589_T599insT]. 2/2 (100 %) des cas avec fusion activatrice de BRAF identifiee par capture par hybridation avaient un test auparavant negatif. Les reponses cliniques de certains patients seront presentees. Discussion Bien que la recherche de mutation du gene BRAF soit desormais pratique courante dans les centres suivant des patients atteints de melanome metastatique, il existe une grande diversite de methodes approuvees. Dans cette etude, le NGS a partir de capture par hybridation a permis d’identifier des mutations et fusions activatrices variees, dont la mutation BRAF V600E, dans une proportion de cas pour lesquels la methode de reference locale avait rendu un resultat negatif pour BRAF. En raison du benefice clinique demontre des therapies ciblees en cas de melanome avec alteration de BRAF, le NGS a partir de capture par hybridation devrait pouvoir etre propose a tout patient atteint de melanome metastatique, en cas de test initial negatif.