Search

e16519 Background: The pan-FGFR inhibitor erdafitinib (erda) is FDA-approved for pretreated mUC pts harboring FGFR2/3 alterations. We explored concordance of FGFR3 alt profiles between primary tumor and cfDNA using the next generation sequencing assay MSK-ACCESS. We also correlated changes in FGFR3 cfDNA mutant allele fraction (MAF) with response to erda. Methods: After consent on an approved biomarker protocol, plasma samples were collected from mUC pts on erda at baseline, on treatment (tx), and at progression along with patient clinical characteristics. Baseline tumors were sequenced with MSK-IMPACT and plasma samples sequenced with MSK-ACCESS, a cfDNA platform that sequences select exons and introns of 129 genes and uses unique molecular indexes to detect somatic mutations down to 0.1% MAF. Results: Between 8/2019-12/2020, 18 pts started erda 8mg daily and had plasma drawn for MSK-ACCESS. Three pts increased to 9mg daily, 7 required dose reductions and 10 had dose interruptions. Treatment was discontinued in 13 pts for disease progression and 3 for toxicity. Median PFS was 3.7 months. FGFR3 S249C was the most frequent alt detected (11/18, 61%), then Y373C (3/18, 16%), and R248C (2/18, 11%) (Table). FGFR3 alt were detected in 15/18 (83%) baseline plasma samples, all of which were of the same alt as tumor tissues. In 3 samples, additional FGFR3 alt were detected, including 1 pt with an FGFR3-TACC3 fusion and hotspot mutations found only in cfDNA. FGFR3 MAF decreased in 7 of 9 pts on erda, 2 of whom declined to undetectable levels. Conclusions: A high degree of concordance of FGFR3 alt was observed between primary tumors and cfDNA. Most erda responders displayed reduction of FGFR3 cfDNA MAF. FGFR3 alts exclusive to cfDNA were found in a small subset of pts. Further pt accrual and follow-up are ongoing to assess for correlations between erda response/progression and changes in FGFR3 cfDNA MAF, and to assess whether cfDNA can identify resistance mechanisms.[Table: see text]

9009 Background: The effectiveness of circulating tumor DNA (ctDNA) at matching patients to life prolonging therapy has been studied mostly in small cohorts with limited follow up. The prognostic value of ctDNA alterations, particularly those absent on tissue, is also unclear. To address these questions, we studied survival outcomes in a prospective cohort of patients (N = 1002) with non-small cell lung cancer (NSCLC). Methods: Adults with metastatic or recurrent NSCLC were eligible if they had no known driver mutation or a known driver with progression following targeted therapy. Patients were enrolled at Memorial Sloan Kettering Cancer Center (New York, NY) starting October 21, 2016; analysis here is from a snapshot November 1, 2020. All patients had ctDNA sequenced via the Resolution ctDx Lung platform. To reduce inclusion of incidental germline mutations, we excluded non-functionally significant mutations with an allele frequency 35-65% that were present in gnomAD. Patients could also receive, at their provider’s discretion, tissue sequencing with MSK-IMPACT, which filters germline and clonal hematopoietic (CH) mutations with matched white blood cell sequencing. We performed survival analyses using Cox proportional hazards models from time of diagnosis of advanced disease to death, left truncating at time of study entry. Results: Of 1002 patients, 348 (35%) were treated with targeted therapy; in 181 of these (52%) the targetable alteration was detected in ctDNA. Patients treated with targeted therapy had prolonged survival whether matched by tissue-based methods (HR 0.39, 95%CI 0.30-0.51) or ctDNA (HR 0.47, 95%CI 0.37-0.61). These benefits persisted across multiple subgroups. ctDNA alterations themselves were associated with worse survival (HR 2.2, 95%CI 1.8-2.8), in a manner that scaled with allele fraction and burden. Of 401 patients with time-matched tissue sampling, 62 (15%) had ctDNA alterations that were absent on IMPACT (“unique” ctDNA alterations). Three such patients had unique ctDNA EGFR T790M mutations leading to changes in therapy. However, unique ctDNA alterations were generally associated with worse survival than no ctDNA alterations (HR 2.5, 95%CI 1.7-3.7) and even tissue-matched ctDNA alterations (HR 1.7, 95%CI 1.1-2.4). Of 98 unique ctDNA mutations, 48 (49%) were detectable in tissue at subthreshold levels, 12 (12%) were filtered by IMPACT as CH or germline, and 38 mutations (39%) were absent even at subthreshold levels. ctDNA alteration burden correlated with radiographic disease extent. In multivariate models with radiographic disease extent and other clinical variables, ctDNA alterations were the strongest independent predictor of worse survival. Conclusions: Our results show that ctDNA may match patients to life-prolonging targeted therapy and have prognostic importance. ctDNA may provide data about a patient’s cancer missed by spatially restricted tissue sequencing. Clinical trial information: NCT01775072.

3529 Background: Circulating cell-free DNA (cfDNA) next-generation sequencing (NGS) is a promising strategy for non-invasive molecular profiling of cancers. MSK-ACCESS (Analysis of Circulating cfDNA to Evaluate Somatic Status) is a hybridization-capture targeted NGS assay that detects somatic variants in select exons of 129 genes with matched white blood cell sequencing. We present the initial clinical experience with MSK-ACCESS among patients with advanced non-small cell lung cancer (NSCLC). Methods: Patients with stage IV NSCLC underwent prospective MSK-ACCESS testing at initial diagnosis or progression of disease on targeted therapy between June 2019 and January 2020. A subset of patients had matched tissue-based NGS testing with the MSK-IMPACT 468 gene assay. We assessed oncogenic driver detection, turnaround time, plasma-tissue concordance, and matching to therapy. National Comprehensive-Cancer Network designated driver alterations were included in evaluation of tissue-plasma concordance (EGFR, ALK, KRAS, MET, RET, BRAF, HER2, ROS1, NTRK). Turnaround time was compared by a two-sided Wilcoxon signed-rank test. Results: A total of 201 patients with NSCLC had MSK-ACCESS testing at initial diagnosis (n = 79) or following progression of disease (n = 122). The median turn-around-time from plasma collection to MSK-ACCESS report was 16 days (range: 9 – 36 days) compared to 19 days from lab receipt of tissue to report (range: 12 – 57) for MSK-IMPACT (p < 0.001). Among patients with a driver detected on MSK-ACCESS, 100% (92/92) had an identical driver detected on MSK-IMPACT. Among patients with a driver detected on MSK-IMPACT, 75% (92/123) had an identical driver detected on MSK-ACCESS. This rate was similar among patients who were treatment-naive (74%; 64/86) and had disease progression (76%, 28/37) at the time of MSK-ACCESS. MSK-ACCESS identified driver alterations that directly guided first-line targeted therapy (n = 18) with response in all patients with available radiographic follow-up (n = 10), including a patient without confirmatory tissue testing. MSK-ACCESS identified resistance alterations among patients with disease progression including EGFR T790M, EGFR C797S, ROS1 G2032R, as well as a BRAF fusion. Conclusions: MSK-ACCESS successfully identified driver alterations with high concordance to tissue-based testing, directly guided patients to therapy with clinical responses, and detected known and novel resistance mechanisms. This assay warrants further clinical development to guide and facilitate precision oncology.

e17034 Background: The pan-FGFR inhibitor erda was recently FDA-approved for pretreated mUC pts harboring FGFR2/3 alterations. We explored concordance of FGFR3 mutation profiles between the primary tumor and cfDNA using the MSK-ACCESS platform. We also correlated changes in FGFR3 cfDNA mutant allele fraction (MAF) with response to drug. Methods: Plasma samples were collected from mUC pts started on erda at baseline, on treatment (tx), and at progression. Demographic and clinical characteristics were obtained. Baseline tumors were sequenced with MSK-IMPACT (ref) and plasma samples were sequenced using MSK-ACCESS, a cfDNA platform that sequences 129 genes using unique molecular indexes to generate > 15,000x coverage and detection of somatic mutations down to 0.1% MAF. Results: Between 08/2019-1/2020, 11 pts received erda 8mg daily. Of these, 3 pts increased dose to 9mg daily based on phosphorus level, 4 required dose reductions and 6 dose interruptions. In 5 pts, erda was discontinued for disease progression. FGFR3 S249C was the most frequent alteration detected (64%) followed by Y373C (18%), R248C and S371C (both 9%). Pre-treatment plasma FGFR3 profiles were concordant with tissue in 91% (10/11) of pts and additional FGFR3 mutations were detected in 3 cases (27%, Table), including 1 pt with an FGFR3-TACC3 fusion and hotspot mutations only in cfDNA. In 2 responding pts, the mutant allele was undetectable on erda. Conclusions: A high degree of concordance between primary tumor and cfDNA FGFR3 mutation detection was observed. FGFR3 mutations exclusive to cfDNA were found in a subset of pts. Further pt accrual and follow-up are ongoing to assess for correlations between erda response and tx-related changes in cfDNA MAF, and to assess whether cfDNA can identify resistance mechanisms. [Table: see text]

9006 Background: MET exon 14 alterations comprise a novel class of lung cancer drivers. MET tyrosine kinase inhibitors (TKIs) are active in patients with these cancers, but objective response rates (ORRs) are modest (~30%-40%). A subset of these cancers may harbor intrinsic resistance. Moreover, patients with initial benefit invariably develop acquired resistance. We set out to identify potential resistance mechanisms. Methods: We studied patients with stage IV MET exon 14-altered lung cancers who received a MET TKI. When feasible, tumor and/or plasma samples were collected, prioritizing paired pre- and post-TKI collection. Tumor samples underwent targeted mass spectrometry analysis (Nantomics) and DNA- (including MSK-IMPACT)/RNA-based (MSK-Fusion) next-generation sequencing (NGS). Plasma cfDNA underwent targeted NGS. ORR and progression-free survival (PFS) were assessed (RECIST v1.1). Results: 74 patients received a MET TKI (1 TKI n = 55; ≥2 TKIs n = 19). 91% received crizotinib as their 1st TKI. Pre-TKI MET levels in tumor tissue (range 0-2120 amol/µg) were associated with outcomes: ORR 63% (n = 7/11) and median PFS 6.9 mos with detectable MET vs ORR 0% (n = 0/5) and median PFS 4.6 mos with undetectable MET (HR for PFS 0.3). Pre-TKI RAS pathway activation was associated with response: ORR 0% (n = 0/6) with KRAS/NF1/RASA1 mutation vs ORR 29% (n = 25/87) in others. Similar outcomes were observed with pre-TKI KRAS expression (n = 16, all with detectable KRAS levels): ORR 0% (n = 0/2) in KRAS ≥700 amol/µg vs ORR 50% (n = 7/14) < 700 amol/µg. Acquired resistance (Jackman criteria) was seen in 29 patients, 9 with paired pre-/post-treatment samples. On-target acquired resistance was found in 2/9 patients (22%): MET D1228N (n = 1), HGF amplification (n = 1). Potential off-target acquired resistance mechanisms were found in 5/9 pts (44%): KRAS G13V (n = 1), RASA1 S742* (n = 1), MDM2 amplification (n = 2), EGFR amplification (n = 1). Conclusions: Lack of MET expression or RAS pathway activation is associated with poor MET TKI outcomes in MET exon14-altered lung cancers. On-target acquired resistance is found in < 25% of patients; HGF amplification is a novel mechanism. Off-target intrinsic/acquired resistance may be mediated by RAS/MDM2/EGFR pathway activation.

640 Background: Unclassified renal cell carcinoma (uRCC) constitutes a large portion of aggressive non-clear cell RCC with limited response to standard therapy. Clinicopathologic parameters or biomarkers to stratify this heterogeneous group of tumors are currently lacking. In a recently reported analysis of 62 high-grade primary uRCC [“discovery cohort (DC)”], we identified distinct molecular subsets. We aimed to validate this molecular schema in an independent clinical cohort and further delineate the clinicopathologic and molecular features that may refine prognostication and management. Methods: All cases was reviewed by experienced GU pathologists based on the current WHO criteria. Primary (n = 54) or metastatic (n = 21) tumor samples from 75 uRCC patients [“validation cohort (VC)”] were analyzed by a CLIA-approved targeted NGS platform for somatic alterations. 37 had germline testing results available. We performed integrative analysis of both VC and DC. Results: Somatic mutations found in VC were NF2 (24%), SETD2 (13%), SMARCB1 (9%), TP53 (9%), TSC1 (9%), FH (8%), TSC2 (5%), MTOR (5%), EP300 (5%), BAP1 (5%), PBRM1 (5%) and PIK3CA (5%), highly consistent with findings in DC. Germline alterations [ FH (11), SDHB (4), and SMARCB1 (1)] were detected in previously unsuspected patients. Integrative analysis supported the presence of NF2-loss (NF2), hyperactive mTOR-driven (MTOR), FH/SDH-deficient (FH/SDH), and chromatin/DNA damage response (Chrom/DDR) molecular subsets. Univariate analysis of combined DC and VC (n = 137, median f/u 26 mos, death 74%) showed a significantly higher risk associated with NF2 subset than the MTOR group (Table). Clonality analysis confirmed NF2 inactivating mutation as a main driver in the NF2 subset. Rare cases with alterations indicating sensitivity or resistance to immunotherapy were also identified. Conclusions: Molecular features of high-grade uRCC improve risk stratification and provide rationale for distinct management strategies. [Table: see text]

4573 Background: Rapalogs, inhibitors of mTOR, are approved for treatment of advanced RCC. Recent reports of clonal heterogeneity challenge the concept of targeted monotherapy and the development of genomic biomarkers. Still, a subset of patients (pts) derives extended benefit from single agent rapalogs. This study analyzed such outliers so as to explore the genomic background of rapalog sensitivity in the setting of clonal heterogeneity. Methods: Cases were chosen based on time to treatment failure > 20 mos and tissue availability. DNA was extracted from spatially separate areas in primary tumors, metastases and the germline. Custom target capture and ultra-deep sequencing identified small indels, single base pair substitutions and copy number changes across all exons of 230 target genes. Results: 4 pts contributed 13 specimens (11 primary tumor samples, 2 metastases); mean exon coverage was 443X. Genomic alterations with activating effect on PI3K pathway signaling were seen in 11 of 13 specimens (Table). Clonal heterogeneity was present in all pts. For 2 pts, different mechanisms activated the pathway in separate sites, in 1 pt through separate genes. Conclusions: Pathway-activating genomic events across all sites explain rapalog benefit in 3 of 4 pts. Different disease sites in the same pt can harbor separate mechanisms activating the targeted pathway. This suggests that clonal convergence within the PI3K pathway can create an oncogenomic landscape sensitive to rapalogs despite branching clonal evolution. It supports the notion of sampling >1 disease site during future biomarker development. [Table: see text]

4522 Background: The objective of this study is to create a molecular tool that can be applied widely to clinical specimens using existing transcript signatures for use in clinical risk prediction of clear cell Renal Cell Carcinoma (ccRCC) to improve personalized disease management. Methods: We developed a 34-gene subtype predictor to classify clear cell tumors according to two subtypes, clear cell A (ccA) or B (ccB). The training set consisted of 72 ccRCC microarray-analyzed tumor samples that had previously been classified by unsupervised clustering and logical analysis of data (LAD). The predictor was developed from a panel of genes significantly expressed in ccA and ccB tumors and associated with prognosis. The prognostic value of the algorithm was corroborated in RNA-sequencing data from 379 ccRCC samples from The Cancer Genome Atlas (TCGA) and further validated using the NanoString platform with a cohort of 163 archival fixed samples collected at the University of North Carolina. Results: Risk associated molecular subtypes, ccA and ccB, were classified in TCGA and NanoString cohorts. Subtype classification showed significant prognostic outcomes for overall survival (p

412 Background: Clear cell renal cell carcinoma (ccRCC) displays molecular and histological heterogeneity. Previously described subsets of this disease, ccA and ccB, were defined based on multi-gene expression profiles, but it is unclear if these subgroupings reflect the full spectrum of disease or how these molecular subtypes relate to histological descriptions or gender. We sought to determine whether additional subtypes of ccRCC exist, and whether these subtypes are related to VHL inactivation, HIF1/HIF2 expression, tumor histology, or gender. Methods: Six large publically available ccRCC gene expression databases were identified that cumulatively provided data for 480 tumors for meta-analysis via meta-array compilation. Unsupervised consensus clustering was performed on the meta-arrays. Tumors were examined for the relationship of multigene-defined consensus subtypes and expression signatures of VHL mutation and HIF status, tumor histology, and gender. Results: Two dominant subsets of ccRCC were observed. However, a minor third cluster was revealed which correlated strongly with a wild type VHL expression profile and indications of variant histologies. When variant histologies were removed, ccA tumors naturally divided by gender. This technique is limited by potential for persistent batch effect, tumor sampling bias, and restrictions of annotated information. Conclusions: ccA and ccB subsets of ccRCC are robust in meta-analysis among histologically conventional ccRCC tumors. A third group of tumors was identified, which may represent a new variant of ccRCC. Within definitively clear cell tumors, gender may delineate tumors in such a way that could have implications regarding current treatments and future drug development.