Advances In Liquid Biopsies Research Articles

Abstract A43: Plasma-derived circulating tumor DNA (ctDNA) as a surrogate biomarker for treatment response with the polo-like kinase 1 (PLK1) inhibitor, onvansertib, in combination with LDAC or decitabine in acute myeloid leukemia (AML)

Abstract Background: Cancer cells from solid tumors release ctDNA into the bloodstream, clinically enabling assessment, monitoring tumor mutational heterogeneity, treatment response, minimal residual disease (MRD), and disease progression. However, the clinical utility of ctDNA in AML has yet to be established. In AML, blast cells carrying a driver mutation originate in the bone marrow (BM) and migrate into circulation in peripheral blood (PB). Here, we aimed to determine the potential of ctDNA as a surrogate for response to treatment in the phase 1b/2 study (NCT03303339) testing the efficacy and safety of the PLK1 inhibitor, onvansertib, in combination with either decitabine or LDAC in relapsed or refractory AML (R/R AML). Methods: R/R AML patients were dosed for 5 days with onvansertib in combination with either decitabine (5 days) or LDAC (10 days) within 21 to 28-day cycle. BM aspirates and blood samples were collected pre-dose, at the end of cycles 1, 2 and every other cycle after. Genomic DNA (gDNA) was extracted from BM and PB mononuclear cells (BMMCs and PBMCs, respectively), while ctDNA was extracted from plasma. Clinically significant mutations were detected in gDNA from PBMCs and BMMCs using a next-generation sequencing panel targeting variants in 75 genes associated with AML. For each patient, a droplet digital PCR (ddPCR) assay was developed to assess mutant allele frequencies (MAF) of a single variant of interest across timepoints and three sample sources (BMMC, PBMC, and Plasma). MAF changes across all 3 sample sources were compared to the clinical response assessed by immunohistochemistry in bone marrow aspirate (BM-IHC). Results: As of October 16th, 16 patients were evaluated, totaling 33 timepoints with MAF measured in all 3 sources. MAF in plasma highly correlated with BMMCs MAF (R2 = 0.85), more so than PBMCs with BMMCs (R2 = 0.78) and over plasma with PBMCs (R2 = 0.68). Clinically, 4 patients had a complete response (with or without count recovery, CR+CRi) and showed a corresponding drop in ctDNA MAF at or before clinical diagnosis. Similarly, 3 of 4 evaluable patients with progressive disease showed an increase in ctDNA MAF before or at the timepoint when blast percentages rose in BM, with the fourth patient maintaining a stable MAF across timepoints. In 5 patients with stable disease, 2 showed a divergence between ctDNA MAF and BM-IHC, while 3 maintained congruency. Conclusions: Serial testing of plasma-derived ctDNA was representative of treatment response and relapse in R/R AML patients. The correlation between ctDNA MAF and BMMCs MAF supports the utility of ctDNA as a surrogate biomarker for disease status, and in some cases provides a more immediate diagnostic of response or nonresponse. More patients need to be evaluated to build an objective classification of clinical significance around ctDNA MAF, and, importantly, the biology of R/R AML subclonal heterogeneity and MRD likely necessitate interrogation and monitoring of multiple variants of interest. Citation Format: Errin Samuelsz, Maya Ridinger, Mark Erlander, Latifa Hassaine, Marion Luebbermann, Brittany Ross. Plasma-derived circulating tumor DNA (ctDNA) as a surrogate biomarker for treatment response with the polo-like kinase 1 (PLK1) inhibitor, onvansertib, in combination with LDAC or decitabine in acute myeloid leukemia (AML) [abstract]. In: Proceedings of the AACR Special Conference on Advances in Liquid Biopsies; Jan 13-16, 2020; Miami, FL. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(11_Suppl):Abstract nr A43.

Abstract A47: Circulating tumor cell-defined minimal residual disease in locally advanced rectal cancer treated with multimodality therapy

Abstract Background: Current treatment for locally advanced rectal cancer (LARC) includes multimodality therapy in the neoadjuvant and adjuvant settings. Response to neoadjuvant therapy (NT) is prognostic of long-term outcomes. While long-course chemoradiation (CRT) had been the traditional NT regimen, novel regimens adding systemic therapy and/or eliminating radiation have been introduced. We examined if circulating tumor cells provide prognostic information independent of the specific treatment regimen utilized. Specifically, we aimed to define the significance of CTC-defined minimal residual disease (MRD) in LARC. Methods: A prospective longitudinal protocol enrolled LARC patients (N=100; clinical stage II=4; stage III=96) undergoing NT. Peripheral blood was collected at baseline (t1, treatment-naive), after NT (t2, intraoperatively before tumor manipulation), after resection (t3, 2-8 weeks postoperatively), and after completion of adjuvant therapy (t4, 2-12 weeks after completion). CTC was enumerated by the CellSearch® platform within 72 hours of collection. Patients were followed for disease-free survival (DFS). Results: At t1, CTCs were detected in 28/90 (31.1%) patients with available samples with a median CTC count of 1.5 [IQR: 1.0, 2.75]. NT consisted of long-course pelvic CRT (65%), extended NT with systemic chemotherapy and long- or short-course pelvic CRT (22%), and a radiation-sparing regimen with systemic chemotherapy only (13%). At t2, CTCs were detected in 34/75 (45.3%) patients with a median CTC count of 1 [IQR: 1, 1]. Type of NAT did not correlate with CTC positivity (p=0.637). Ten (9.4%) patients declined surgery, while cPR occurred in 12/93 (12.9%) surgical patients. Postoperatively (t3), CTC was detected in 12/47 (25.5%) patients, with a median count of 1.5 (IQR: 1.0-3.0). After completing adjuvant chemotherapy (t4), only 3/29 (10.3%) patients had detectable CTC. With a median follow-up of 47.50 months from diagnosis for the entire cohort, CTC positivity at t4 significantly stratified DFS (p=0.035). Conclusions: One third of locally advanced RC patients harbored detectable CTCs at baseline. CTC detection after completion of curative-intent multimodality therapy (i.e., MRD) correlated with long-term DFS. Citation Format: Lucas Lee, Carol Hall, Antony Lucci, Brian Bednarski, Miguel Rodriguez-Bigas, George Chang, Y. Nancy You. Circulating tumor cell-defined minimal residual disease in locally advanced rectal cancer treated with multimodality therapy [abstract]. In: Proceedings of the AACR Special Conference on Advances in Liquid Biopsies; Jan 13-16, 2020; Miami, FL. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(11_Suppl):Abstract nr A47.

Abstract PR11: Multimodal analysis of circulating tumor cell RNA, circulating cell-free DNA, and genomic DNA from a single blood sample collected into a PAXgene Blood ccfDNA Tube*

Abstract Background: High demand for liquid biopsy tests in cancer research increases the need for combined approaches that allow multimodal testing from a limited blood volume. Although crosslinking-based fixatives allow for efficient preservation of the circulating cell-free DNA (ccfDNA) profile in whole blood, analysis of mRNA from circulating tumor cells (CTCs) is hampered. Currently, there are no blood collection tubes available that allow RNA analysis from CTCs after prolonged (>48 hours) storage. In this research study, we present a newly developed combined workflow for blood stabilization and subsequent analysis of CTC RNA, ccfDNA, and gDNA after prolonged (≥72 hours) storage. Methods: To establish the multimodal workflow, blood was collected from healthy donors into PAXgene Blood ccfDNA Tubes*, aliquoted and manually spiked with 20 individual LNCaP95 cells/5 mL blood as a CTC model or 20 µL PBS as a control. Blood samples were spiked and stored at 2–8°C for 3, 24, 48, and 72 hours before processing. CTCs were enriched and detected using the AdnaTest ProstateCancerPanel AR-V7† (QIAGEN). The plasma and cellular fraction of the CTC-depleted blood was then used to extract ccfDNA and gDNA, respectively. As a control for ccfDNA and gDNA yield measurements, plasma and the cellular fraction were also generated from whole samples (not CTC-depleted) spiked with 20 LNCaP95 cells/5 mL. The resulting multimodal workflow was tested: 1) in comparison to Streck Cell-Free DNA BCT; 2) to assess the impact of room temperature. Results: Tumor cells spiked into PAXgene Blood ccfDNA Tubes could be enriched and detected at experimental time points 3, 24, 48, and 72 hours after spiking with 100% test sensitivity within 24 hours storage and 91% test sensitivity after 72 hours storage at 2–8°C. In unspiked samples, the PAXgene stabilization solution did not cause false positive signals at any experimental time point. Yields of ccfDNA and gDNA were not statistically significantly affected by CTC capture (p>0.05). RNA from spiked tumor cells was detected in blood samples collected and stored in Streck Cell-Free DNA BCTs for 3 hours, but not at later time points. Storage of PAXgene stabilization solution-treated samples at room temperature is feasible, with 100% test sensitivity for CTC detection within 24 hours storage (similar to storage at 2–8°C) and 80% after 72 hours storage. Conclusions: Blood collected into PAXgene Blood ccfDNA Tubes and stored for up to 72 hours at 2–8°C or at room temperature can be used for multimodal analysis of CTC RNA, ccfDNA, and gDNA from a single blood sample. *For research use only. Not for use in diagnostic procedures. †For molecular biology applications. Not intended for the diagnosis, prevention, or treatment of a disease. This abstract is also being presented as Poster B18. Citation Format: Eric Provencher, Maike Remmel, Siegfried Hauch, Michael Otte, Andrea Ullius, Daniel Groelz, Anna Babayan. Multimodal analysis of circulating tumor cell RNA, circulating cell-free DNA, and genomic DNA from a single blood sample collected into a PAXgene Blood ccfDNA Tube* [abstract]. In: Proceedings of the AACR Special Conference on Advances in Liquid Biopsies; Jan 13-16, 2020; Miami, FL. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(11_Suppl):Abstract nr PR11.

Abstract A38: A novel clinical-grade liquid biopsy platform for multiple myeloma

Abstract Direct-to-patient (DTP) multiple myeloma (MM) research studies have been launched recently, including PCROWD (NCT02269592), PROMISE (NCT03689595), and the MMRF CureCloud Research Initiative (NCT03657251), aimed at enrolling thousands of individuals from whom comprehensive molecular and immune analyses will be generated from blood specimens and the resulting data aggregated with the correlating clinical information. To support the molecular characterization of liquid biopsies for such DTP efforts, a myeloma-specific hybrid selection panel was developed that captures 70 commonly altered genes. The assay detects somatic point mutations and indels present in a patient’s circulating-free DNA (cfDNA). For this MM 70-Gene cfDNA Assay, samples are received as blood in a Streck’s tube and DNA is extracted from buffy coat using magnetic bead-based chemistry. Coverage sequencing (80,000x depth) is performed and duplex BAM files are generated with UMI alignment and de-duplication. As will be presented, MM blood specimens present a unique challenge as circulating MM cells are often present at significant levels in the buffy coat blood fraction used as the source of normal genomic DNA. The performance of the 70-Gene cfDNA Assay was thoroughly validated in order to establish the sensitivity, specificity, and reproducibility of the technical approach. First, the reference genomic DNA from unrelated healthy individuals was sequenced in replicate at deep coverage. Next, two cohorts were used, one from Dana-Farber and one from the MMRF CureCloud pilot. For both cohorts, tumor DNA samples from bone marrow aspirates (BMAs) with matched normal DNA from blood were sequenced on an orthogonal platform and compared to results from the MM 70-Gene Assay on cfDNA extracted from the same individuals. The yield of extracted cfDNA ranged from 6 ng to 80 ng, and about two third of cases yielded enough material to attempt sequencing, with failures coming mostly from individuals in remission. As will be presented, there was a very strong correlation between BMA and cfDNA and additional events could actually be detected in the blood that were not seen in the BMAs. Because this MM 70-Gene cfDNA Assay may potentially be used by treating physicians for management of care, a clinical-grade (CLIA) pipeline was established. For that CLIA pipeline, the variants reported are a subset of all the events detected by the MM 70-Gene Assay. The events detected in the assay are reviewed by a Genomic Tumor Board within the appropriate subset of territory predefined for reporting. The territory limitations are defined by the Genomic Tumor Board knowledgebase of actionable genomic territory available. In summary, we have developed a robust and very sensitive clinical-grade next-gen liquid biopsy sequencing platform allowing for less invasive monitoring of MM disease genomics that can be used to complement other more classical approaches and to help support direct-to-patient Initiatives. Citation Format: Mark W. Bustoros, Carrie Cibulskis, Teni Dowdell, Svetlana Gavrilov, Cody Boehner, Jennifer Yesil, Steven E. Labkoff, Shaadi Mehr, Jihye Park, Romanos Sklavenitis Pistofidis, Salomon Manier, Annette S. Kim, Keith L. Ligon, Niall Lennon, Viktor Adalsteinsson, Jane Wilkinson, Irene M. Ghobrial, Daniel Auclair. A novel clinical-grade liquid biopsy platform for multiple myeloma [abstract]. In: Proceedings of the AACR Special Conference on Advances in Liquid Biopsies; Jan 13-16, 2020; Miami, FL. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(11_Suppl):Abstract nr A38.

Abstract A42: Development and validation of diagnostic biomarkers for B-cell lymphoma using EpiSwitchTM profiling of whole blood: From humans to canines

Abstract Backgound: Changes in chromosome conformation are strongly associated with different clinical phenotypes and outcomes for a variety of diseases. Here, we used the EpiSwitchTM platform technology to evaluate chromosome conformation signatures (CCS) as biomarkers for detection of canine diffuse large B-cell lymphoma (DLBCL). We examined whether established, systemic liquid biopsy biomarkers previously characterized in human DLBCL patients by EpiSwitchTM would translate to dogs with the homologous disease. Orthologous sequence conversion of CCS from humans to dogs was first verified and validated in control and lymphoma canine cohorts. Methods: Blood samples from dogs with DLBCL and from apparently healthy dogs were obtained as part of the comparative cancer genomics and veterinary clinical trials programs of the Animal Cancer Care and Research Program of the University of Minnesota. Dog owners provided signed informed consent, and all studies were done under protocols approved by the Institutional Animal Care Committee of the University of Minnesota. All of the dogs diagnosed with DLBCL were part of the LICKing Lymphoma trial. Blood samples were obtained from each dog prior to initiating treatment at day+5 after the experimental intervention, but prior to initiating doxorubicin chemotherapy. EpiSwitchTM technology (Oxford Biodynamics) was used to monitor systemic epigenetic biomarkers for CCS. This application was based on published methodology for validated predictive biomarkers for response to treatment, systemic blood-based monitoring of oncologic conditions, and proprietary programs in collaboration with Roche/Genentech for liquid biopsy development for DLBCL prognosis. Results: A 11-marker classifier was generated with whole blood from 28 dogs, 14 diagnosed with DLBCL and 14 controls with no apparent disease, from a pool of 75 EpiSwitchTM CCSs identified in human DLBCL. Validation of the developed diagnostic markers was performed on a second cohort of 10 dogs: 5 with DLBCL and 5 controls. The classifier delivered stratifications for DLBCL vs. non-DLBCL with 80% accuracy, 80% sensitivity, 80% specificity, 80% positive predictive value (PPV), and 80% negative predictive value (NPV) on the second cohort. Conclusions: The established EpiSwitchTM classifier contains strong systemic binary markers of epigenetic deregulation with features normally attributed to genetic markers: the binary status of these classifying markers is statistically significant for diagnosis, and it is translatable across species. The data suggest that certain features of the higher-order structural chromatin organization associated with DLBCL, and perhaps other homologous cancers, may be evolutionarily conserved. Altogether, these findings highlight the potential of the EpiSwitchTM approach for disease diagnosis and further applications using liquid biopsies in humans and dogs. Citation Format: Lauren J. Mills, Ewan Hunter, Aaron L. Sarver, Davis Seelig, Jennifer Back, Louis James, Navin Jandor, Polina Brayer, Thomas Lavin, Ryan Powell, Matthew Salter, Jayne Green, Aroul Ramadass, Alexandre Akoulitchev, Jaime F. Modiano. Development and validation of diagnostic biomarkers for B-cell lymphoma using EpiSwitchTM profiling of whole blood: From humans to canines [abstract]. In: Proceedings of the AACR Special Conference on Advances in Liquid Biopsies; Jan 13-16, 2020; Miami, FL. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(11_Suppl):Abstract nr A42.

Abstract IA13: Tracking and targeting colorectal cancer evolution

Abstract When metastatic cancers are challenged with targeted agents, almost invariably a subset of cells insensitive to the drug emerges. As a result, in most instances, targeted therapies are only transiently effective in patients. Strategies to prevent or overcome resistance are therefore essential to design the next generation of clinical trials. How can we overcome the near certainty of disease recurrence following treatment with targeted agents? Addressing this question means considering as a target not “only” individual oncogenes but also the evolving nature of human tumors. We used colorectal cancer (CRC) as a model system to test the hypothesis that by understanding tumors’ evolution, the emergence of drug resistance can be controlled. We find that clonal dynamics can be monitored in real time in the blood of patients, and liquid biopsies can be used to intercept the emergence of resistant clones before relapses are clinically manifest. We discovered that a multistep clonal evolution process driven by progressive increases in drug fitness underlies the development of resistance in cells and patient avatars. To have long-term efficacy, the use of targeted therapies must take into account the continuous evolution of cancer cells, that is to say, therapies must adapt to tumor evolution. One possibility is to anticipate the changes the tumors will make. For example, by knowing in advance how CRC cells overcome resistance to EGFR blockade, we devised further rounds of therapy. These findings became the bases of clinical trials aimed at targeting cancer evolution and exploiting liquid biopsies to monitor the emergence of drug resistance and to design further lines of therapies. Citation Format: Alberto Bardelli. Tracking and targeting colorectal cancer evolution [abstract]. In: Proceedings of the AACR Special Conference on Advances in Liquid Biopsies; Jan 13-16, 2020; Miami, FL. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(11_Suppl):Abstract nr IA13.

Abstract PR13: Comprehensive detection of ctDNA in localized head and neck cancer by genome- and methylome-based analysis

Abstract Head and neck squamous cell carcinoma (HNSCC) comprises 3% of all cancer cases worldwide. Despite intensive multimodal therapies, HNSCC patient outcomes remain heterogeneous with minimal improvements in survival. Utilization of fluid-based biomarkers for prognostication, risk stratification, and disease surveillance may improve patient outcomes by enabling more effective treatment decisions. Here, we describe the performance of comprehensive mutation and methylome analysis for highly sensitive detection of circulating tumor (ct)DNA in HNSCC plasma. To detect HNSCC-specific mutations and aberrant methylation in ctDNA, we conducted CAPP-Seq (CAncer Personalized Profiling by deep Sequencing) and cfMeDIP-seq (cell-free Methylated DNA ImmunoPrecipitation sequencing), respectively. For CAPP-Seq, we developed a HNSCC-specific hybrid capture panel and applied it to plasma DNA and peripheral blood leukocyte (PBL) genomic DNA from a cohort of HNSCC patients (n=32) and healthy controls (n=20). Single-nucleotide variants (SNVs) were identified by integrated digital error suppression (iDES). For cfMeDIP-seq, hypermethylated differentially methylated regions within regions of low PBL methylation were identified by DESeq2. A median of 3 mutations (range: 1-10) were detected by CAPP-Seq within plasma DNA of 20/32 (62.5%) HNSCC patients. Mean mutant allele frequency ranged from 0.1–5% (median: 0.92%) and correlated with tumor stage. Among these 20 patients, cfMeDIP-seq identified 860 DMRs that were enriched (hypermethylated) in HNSCC plasma DNA compared with healthy controls. These hypermethylated DMRs (hyper-DMRs) were over-represented by CpG islands and gene promoters and showed significant overlap with HNSCC-specific methylated regions in TCGA. When applied to all 32 HNSCC patients, hyper-DMR abundance was positively correlated with mutation-based ctDNA abundance (R=0.87; p=1e−16). Hyper-DMRs were capable of accurate discrimination of HNSCC patients and healthy controls (AUC=0.85). The median DNA fragment size within these hyperDMRs was lower in HNSCC patients compared to healthy controls—a characteristic of ctDNA described in previous studies—and correlated with both hyperDMR-based and mutation-based ctDNA abundance. We have conducted the first comparative analysis of genetic and epigenetic profiling approaches for ctDNA detection in HNSCC. Both CAPP-Seq and cfMeDIP-seq have the potential to detect ctDNA in patient plasma without prior knowledge of patient-specific tumor aberrations. With CAPP-seq, ctDNA was detectable at levels as low as 0.1%. Plasma methylome profiling using cfMeDIP-seq revealed hyper-DMRs with tumor-related features and that correlated with mutation-based ctDNA abundance. Future analysis will validate patient-specific mutations and methylation signals in tumor tissue and in longitudinally collected samples from this cohort. This abstract is also being presented as Poster B47. Citation Format: Justin M. Burgener, Jinfeng Zou, Zhen Zhao, Shu Y. Shen, Daniel D. De Carvalho, Scott V. Bratman. Comprehensive detection of ctDNA in localized head and neck cancer by genome- and methylome-based analysis [abstract]. In: Proceedings of the AACR Special Conference on Advances in Liquid Biopsies; Jan 13-16, 2020; Miami, FL. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(11_Suppl):Abstract nr PR13.

Abstract PR15: Charting extracellular transcriptomes in The Human Biofluid RNA Atlas

Abstract Liquid biopsies offer a minimally invasive alternative to tissue biopsies for both diagnosis and monitoring of treatment response. Extracellular RNAs present in biofluids have emerged as potential biomarkers in health and disease, including cancer. While current studies typically focus on plasma or serum, other biofluids may contain more informative RNA molecules for particular disease types. Here, we present an unprecedented atlas of messenger, circular, and small RNA transcriptomes of a comprehensive collection of 20 different human biofluids (amniotic fluid, aqueous humor, ascites, bile, bronchial lavage fluid, breast milk, cerebrospinal fluid, colostrum, gastric fluid, pancreatic cyst fluid, plasma, saliva, seminal fluid, serum, sputum, stool, synovial fluid, sweat, tear fluid, and urine). By means of 173 synthetic RNA spike-in controls, we compared RNA content across biofluids, revealing a more than 10,000-fold difference in RNA concentration. Of interest, the circular RNA fraction is significantly increased in nearly all biofluids compared to various solid tissues. Each biofluid transcriptome is enriched for RNA molecules derived from specific tissues and cell types. In addition, a subset of biofluids, including stool, sweat, saliva, and sputum, contain high levels of bacterial RNAs. These findings enable a more informed selection of the most relevant biofluid to monitor individual cancer types. To verify the biomarker potential in these biofluids, four validation cohorts representing a broad spectrum of diseases, including urine from bladder cancer patients and CSF from glioblastoma patients, were profiled, revealing differential RNAs between case and control subjects. Taken together, our results reveal novel insights in the RNA content of human biofluids and may serve as a valuable resource for future biomarker studies. This abstract is also being presented as Poster B01. Citation Format: Eva Hulstaert, Annelien Morlion, Francisco Avila Cobos, Kimberly Verniers, Justine Nuytens, Eveline Vanden Eynde, Nurten Yigit, Jasper Anckaert, Anja Geerts, Pieter Hindryckx, Peggy Jacques, Guy Brusselle, Ken Bracke, Tania Maes, Thomas Malfait, Thierry Derveaux, Virginie Ninclaus, Caroline Van Cauwenbergh, Kristien Roelens, Ellen Roets, Dimitri Hemelsoet, Kelly Tilleman, Lieve Brochez, Scott Kuersten, Lukas Simon, Sebastian Karg, Christa Nöhammer, Alexandra Kautzky-Willers, Michael Leutner, Ondrej Slaby, Gary Schroth, Jo Vandesompele, Pieter Mestdagh. Charting extracellular transcriptomes in The Human Biofluid RNA Atlas [abstract]. In: Proceedings of the AACR Special Conference on Advances in Liquid Biopsies; Jan 13-16, 2020; Miami, FL. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(11_Suppl):Abstract nr PR15.

Abstract A23: Ultrasensitive mutation detection in FFPE tissues and circulating tumor DNA using SiMSen-Seq

Abstract Novel diagnostic tools in oncology, prenatal diagnostics, and transplantation medicine depend on the detection of ultrarare variants. The analysis of cell-free DNA has the potential to revolutionize diagnosis, treatment monitoring, and recurrence and resistance of cancer. Yet these analysis remains challenging due to limited amounts of material and low allele frequencies of targeted variants. Simple, Multiplexed, PCR-based barcoding of DNA for Sensitive mutation detection using Sequencing (SiMSen-Seq) enables easy generation of libraries containing unique molecular identifiers (UMI) with minimal DNA input across several kilobases of DNA using two rounds of PCR and allows the detection of rare variants below 0.1% allele frequency. Clinical applications of ctDNA require highly optimized workflows to ensure high sensitivity and reproducibility. Here we present our state-of-the-art liquid biopsy workflow using SiMSen-Seq and simple quality control steps, applicable to almost any liquid biopsy protocol. In a series of case studies with melanoma patients we find that ctDNA mirrors the time-course of established protein markers such as S-100 and in some cases detects minimal residual months in advance of standard clinical diagnostics. SiMSen-Seq efficiently determined even rare mutations across multiple tumor forms in challenging sample types, such as FFPE tissue or cfDNA. We identified mutations in highly degraded FFPE tissue using a 60-plex SiMSen-Seq panel and then analyzed longitudinal cfDNA samples using patient-specific panels in a cohort of metastatic breast cancer patients. Our data indicate that ctDNA might be used as a biomarker for early detection of recurrence in these patients. We were able to detect treatment resistance variants as well as observe tumor recurrence well in advance of the onset of clinical symptoms. Thus, SiMSen-Seq offers an affordable, minimally invasive method for monitoring treatment efficiency and determination of drug resistance development in cancer. Citation Format: Stefan Filges, Daniel Andersson, Helena Kristiansson, Christoffer Vannas, Gustav Johansson, Junrui Li, Tony E. Godfrey, Max Levin, Barbro Linderholm, Anders Ståhlberg. Ultrasensitive mutation detection in FFPE tissues and circulating tumor DNA using SiMSen-Seq [abstract]. In: Proceedings of the AACR Special Conference on Advances in Liquid Biopsies; Jan 13-16, 2020; Miami, FL. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(11_Suppl):Abstract nr A23.

Abstract A49: Viable circulating ensembles of tumor-associated cells persist in pretreated patients with solid organ cancers showing no radiologically detectable disease

Abstract The success of treatment in solid organ cancers is ascertained by radiologic imaging as per standard-of-care protocol, with PET-CT being the modality of choice. However, even in cases where complete resolution of the disease is noted radiologically, recurrence or emergence of new metastases is not uncommon. To explore the underlying cause of such recurrence, we hypothesized that circulating metastatic disease (CMD) in the nature of viable tumor cells or clusters would be a persistent systemic feature of solid organ cancers, although there may be no overt evidence of disease. We identified 1,217 patients with known and previously treated cases of various solid organ cancers (metastatic and nonmetastatic), where a recent PET-CT scan showed no radiologic evidence of disease. 15 mL peripheral blood was collected from these patients. Peripheral blood mononuclear cells (PBMCs) were harvested by centrifugation. Circulating ensembles of tumor-associated cells (C-ETACs) were enriched from PBMCs using an epigenetically acting stabilization process, which is cytotoxic towards nonmalignant cells but confers survival privilege on apoptosis-resistant cells of tumorigenic origin as well as their heterotypic clusters (C-ETACs). C-ETACs were characterized by immunostaining for EpCAM, pan-CK, and CD45 as well as for organ-specific markers. Among the 1,217 patients with no radiologic evidence of disease, C-ETACs were detected in 1,118 (91.9%) and were confirmed by immunostaining. In a subset analysis, C-ETACs from 80 samples were immunostained for organ- or subtype-specific antigens including those from head and neck (P63), breast (GCDFP15), colorectum (CDX2), liver (Hep Par-1), ovary (CA125), prostate (PSMA), cervical (P63), and lung (TTF-1/Napsin-A). Circulating metastatic disease (CMD) in the form of viable C-ETACs is a potent threat in patients with solid organ cancers despite complete radiologic response to treatment. Citation Format: Sewanti A. Limaye, Timothy Crook, Dadasaheb Akolkar, Darshana Patil, Pradip Devhare, Revati Patil, Shoeb Patel, Shabista Khan, Raja Dhasarathan, Vineet Datta, Cynthe Sims, Ajay Srinivasan, Rajan Datar. Viable circulating ensembles of tumor-associated cells persist in pretreated patients with solid organ cancers showing no radiologically detectable disease [abstract]. In: Proceedings of the AACR Special Conference on Advances in Liquid Biopsies; Jan 13-16, 2020; Miami, FL. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(11_Suppl):Abstract nr A49.

Abstract B65: Precompetitive collaboration on liquid biopsy for early cancer assessment

Abstract Precompetitive collaboration is a term used when competitors share early stages of research and development that are beneficial to all stakeholders. This Public-Private Partnership Consortium supports research to advance technologies and use these technologies for early cancer detection, to distinguish benign disease from cancer, and to distinguish aggressive from nonaggressive cancer. Combining the complementary expertise of both academics and industry may increase the likelihood of developing a viable diagnostic assay. In the era of personalized medicine, it is desirable to have noninvasive or minimally invasive methods to determine and follow longitudinally the molecular composition and characterization of a patient’s tumor, which will help gain a broader understanding of the disease. Liquid biopsy (CTC, ctDNA, extracellular vesicles, etc.) can be used to molecularly characterize the tumor and monitor genetic changes over time using repeat sampling of biofluids. The current technologies have shown great potential in diagnosis, predicting response to therapy, and monitoring relapse; however, the biologic and analytical challenges have rendered their usefulness for early detection/screening more difficult to establish. To address these challenges, the NCI developed this cooperative agreement program to support the work of six teams. Each team consists of a group of experts that include oncologists, bioinformaticians, researchers, and at least one industry partner. The current teams are developing a variety of unique assays/technologies that range from capture enhancements, microscope integration, nanoplasmonics, to specialized machine learning algorithms and bioinformatic analyses. Ovarian, breast, lung, and brain cancers are the model systems being studied, and there are at least two teams focused on each cancer type. The Consortium conducts work with the assistance of two working subgroups, Study Design and Biorepository, in conjunction with NCI staff. To monitor progress, there are yearly face-to-face Steering Committee meetings and monthly conference calls/Webex meetings. Progress in the first year of funding has been noteworthy. One of the teams has already achieved Clinical Laboratory Improvement Amendments (CLIA) and College of American Pathologists Certification, while another team has an insurance reimbursable extracellular vesicles test for risk assessment for prostate cancer. Citation Format: Lynn Sorbara, David Wong, Carter Bob, Richard Cote, Hakho Lee, Nickolas Papadopoulos, Abhijit Patel, Sudhir Srivastava. Precompetitive collaboration on liquid biopsy for early cancer assessment [abstract]. In: Proceedings of the AACR Special Conference on Advances in Liquid Biopsies; Jan 13-16, 2020; Miami, FL. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(11_Suppl):Abstract nr B65.

Abstract B36: Identifying single-cell gene expression and EGFR mutation profile heterogeneity in NSCLC patients’ CTCs

Abstract The efficacy of targeted therapies was first shown in patients with non-small cell lung cancer (NSCLC) who possess activating epidermal growth factor receptor (EGFR) mutations that responded to tyrosine kinase inhibitors (TKIs). Current standard of care is to screen metastatic patients for the presence of EGFR mutations to determine if they qualify for TKI treatment. Due to the invasiveness of a tumor biopsy, this characterization is typically only done at the time of diagnosis. This single time-point determines the entire treatment plan. Despite strong initial response to TKI therapy, most patients develop resistance within months, most commonly by acquiring a secondary EGFR mutation. Identifying resistance through current monitoring techniques, such as measuring tumor volume, can lead to a delay in detection. Alternatively, circulating tumor cells (CTCs) are cancer cells present in the blood that provide access to tumor cells without the requirement of a biopsy. Shed from the tumor into the vasculature, CTCs circulate throughout the body before a fraction extravasate, leading to metastatic sites. CTCs have been shown to carry tumor-matched characteristics in both genotype and phenotype. Due to their ease of access, they can be used to serially track patients’ conditions to detect the early emergence of new tumor clones and, potentially, therapeutic resistance. One of the major limitations of CTCs’ clinical utility is their low abundance in the blood. Our group previously developed the Labyrinth, a high-throughput, label-free microfluidic technology, which enables rapid and efficient CTC enrichment. These CTCs are used for enumeration or single-cell transcriptomic analysis. Initial single-cell analysis was performed through highly multiplexed RT-qPCR that compared 96-gene expression profiles to unravel inter- and intrapatient CTC heterogeneity. We recently developed a complimentary single-cell analysis approach using digital PCR (dPCR) to detect the presence of sensitizing EGFR mutations, exon 19 deletion and L858R, and resistance mutation, T790M. dPCR is an ultrasensitive approach with single-molecule resolution, which enables detection and quantification at a single-cell level by partitioning the sample into individual PCR reaction droplets. This resolution can distinguish if a cell is homo- or heterozygous for a mutation and each allele’s relative expression level. We validated this system for single-cell analysis using lung cancer cell lines and demonstrated its agreement with qPCR results. We present initial findings of EGFR mutation screening in CTCs from a small cohort of NSCLC patients with known EGFR status. We highlight the single-cell heterogeneity of CTCs within a patient at both the gene expression and EGFR mutation level. By tracking these patients’ CTCs over time, we may be able to identify the emergence of resistant clones sooner than can be done in the clinic. This may allow for more rapid treatment modification and lead to improved patient outcomes. Citation Format: Sarah Owen, Ting-Wen Lo, Shamileh Fouladdel, Mina Zeinali, Evan Keller, Ebrahim Azizi, Nithya Ramnath, Sunitha Nagrath. Identifying single-cell gene expression and EGFR mutation profile heterogeneity in NSCLC patients’ CTCs [abstract]. In: Proceedings of the AACR Special Conference on Advances in Liquid Biopsies; Jan 13-16, 2020; Miami, FL. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(11_Suppl):Abstract nr B36.

Abstract B40: Analytical validation and initial clinical utility of multianalyte transcriptomic biomarker profiling of circulating tumor cells using automated exclusion-based sample preparation technology

Abstract The efficacy of androgen receptor (AR) signaling inhibitors (ARSIs) in prostate cancer is limited by both de novo and acquired resistance. Two well-described mechanisms of resistance to ARSIs are the development of constitutively activated AR splice variants (AR-Vs) or neuroendocrine prostate cancer (NEPC). Expression of NEPC markers is traditionally evaluated in solid tumor biopsies while AR-Vs can be detected in circulating tumor cells (CTCs) and have prognostic relevance in prostate cancer. Unlike solid tumor biopsy-based detection strategies, liquid biopsies can be performed periodically over the course of treatments, enabling early identification of resistance mechanisms. To evaluate for ARSI resistance, we developed a multiplex gene expression panel of AR-Vs and NEPC markers. We evaluated potential clinical utility of this assay in CTCs captured using an exclusion-based sample preparation technology. This test was evaluated with a cohort of 46 patients, showing 39% sensitivity and 94% specificity at detecting ARSI resistance. To develop a clinical-grade liquid biopsy assay with this multiplexed gene expression panel, we utilized a high-throughput automated exclusion-based sample preparation (ESP) technology for CTC capture, mRNA extraction, and gene expression analysis using RT-qPCR. Cell lines were used to evaluate the analytical performance of the methodology, and spike-in control materials were designed to monitor the efficiency of mRNA extraction for each individual patient sample. Serial dilutions of cell lines demonstrated the sensitivity to extract and quantify mRNA at the single-cell level with high precision between evaluations performed on three different days. Increasing numbers of cells resulted in the quantification of proportionally higher amounts of both AR-Vs and NEPC markers, demonstrating the analytical accuracy of the methodology (R2 values of 0.8 and above). To monitor for RNA degradation, a grape mRNA amplicon material was designed as an internal spike-in control for daily use in testing patient samples. Healthy donor PBMCs were used to confirm primers were specific for cells of cancer origin. ESP-mRNA extraction followed by RT-qPCR demonstrates single-cell resolution for the detection of AR-V/NEPC transcriptomic expression, with high precision in evaluating actual patient samples. The high analytical performance, and the promising preliminary correlation with ARSI resistance, warrant further efforts towards implementation into clinical laboratory testing environments and testing in prospective clinical trials. Citation Format: Zachery D. Schultz, Jennifer L. Schehr, Rory M. Bade, Molly M. Morgan, Madalyn S. Gill, Hannah M. Pezzi, Jaime M. Sperger, Charlotte N. Stahlfeld, Anupama Singh, Jay W. Warrick, David J. Beebe, Joshua M. Lang. Analytical validation and initial clinical utility of multianalyte transcriptomic biomarker profiling of circulating tumor cells using automated exclusion-based sample preparation technology [abstract]. In: Proceedings of the AACR Special Conference on Advances in Liquid Biopsies; Jan 13-16, 2020; Miami, FL. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(11_Suppl):Abstract nr B40.

Abstract B09: Identification of salivary exosomes derived miRNAs as potential early diagnostic markers in oral cancer patients: A liquid biopsy approach

Abstract Oral squamous cell carcinoma (OSCC) is one of the most prevalent forms of cancer globally. Late-stage diagnosis is one of the major confounders for poor prognosis, as there are no disease-specific biomarkers available for early risk prediction. Some of the major limitations of invasive tissue biopsies are inadequacy of tissue, difficulty in accessing deep sites, pain, and potential risk for the patient. Therefore, it is important to identify sensitive and specific circulatory biomarkers using noninvasive biopsy techniques. Some studies have investigated microRNA (miRNAs) expression in saliva as potential biomarkers; however, these studies suffer from extensive RNA degradation due to the presence of RNA-degrading enzymes in saliva. Since miRNAs associated with exosomes are known to be protected from enzymatic degradation, we evaluated whether salivary exosomes from OSCC patients were enriched with specific subsets of miRNAs that would be useful in the early diagnosis of OSCC. In this study, we performed miRNA and mRNA sequencing (Illumina) in leukoplakia and OSCC patient-derived salivary exosomes and matched tumors (n=20) compared to controls. The combination of miRNA and mRNA expression profiles from the same sample allowed us to significantly reduce the false positives in predicted miRNA targets. Potential target prediction, Gene Ontology (GO) annotation, and pathway enrichment analysis were conducted to explore the functional and biologic significance of the identified miRNA and gene networks. Subset of miRNAs and genes that were enriched in both tumor and exosomes were further evaluated using real-time quantitative PCR (qPCR) analysis. Our results demonstrated that exosomes from patients’ saliva have a distinct pattern of miRNA expression compared to healthy controls. miRNA profiles of salivary exosomes were similar to tumor profiles of both leukoplakia and oral cancer tumor samples. 26 miRNAs were expressed exclusively in leukoplakia (16 miRs) and OSCC patient population (10 miRs), but not in controls. We also identified miRNA regulated gene networks that were differentially expressed in various stages of oral cancer compared to controls with good discrimination power and were involved in signaling pathways such as PI3K/Akt, MAPK, and WNT. One of the noteworthy findings from this study was identification of 5 tumor-suppressor miRNAs that were exclusively expressed in healthy controls and were absent in both leukoplakia and oral cancer patients. Thus, this study shows that miRNAs exclusively enriched in exosomes of leukoplakia and OSCC patients could be exploited as potential noninvasive biomarkers for OSCC diagnosis. This study will also aid in gaining further insights on miRNA-mediated gene networks responsible for transforming epithelial phenotype to oral carcinoma. Citation Format: Shanaya Patel, Aditi Patel, Dhanvi Mathur, Vivek Tanavde. Identification of salivary exosomes derived miRNAs as potential early diagnostic markers in oral cancer patients: A liquid biopsy approach [abstract]. In: Proceedings of the AACR Special Conference on Advances in Liquid Biopsies; Jan 13-16, 2020; Miami, FL. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(11_Suppl):Abstract nr B09.

Abstract B31: Detection of viable tumor cells from cryopreserved buffy coat using the VTX-1 Liquid Biopsy Platform

Abstract Introduction: Several circulating tumor cell (CTC) technologies have been developed amidst growing interests in the “liquid biopsy” field. The fully automated VTX-1 platform is developed to gently isolate intact CTCs from whole blood with high recovery and viability, compatible with many applications (1-3). However, many CTC technologies require processing of fresh patient samples at off-site facilities. Here we present encouraging preliminary results showing viable cell isolation by VTX-1 from cryopreserved buffy coat, using a telomerase-selective replicating adenovirus (TelomeScan). This widens potential applications of viable CTC detection for retrospective clinical studies and drug development. Methods: Buffy coat preparation: Buffy coat was prepared by centrifugation of healthy donor blood collected in EDTA, heparin, or heparin CPT vacutainers. Cell spiking and processing: 1) Cancer cells were spiked into freshly prepared buffy coat and processed with the Vortex chip at different dilutions. 2) Cancer cells were labeled and spiked into healthy donor blood, and buffy coat was prepared and processed by VTX-1 either immediately or following cryopreservation up to 1 week. Imaging was evaluated for cell recovery. Viable, telomerase-activated tumor cell detection: VTX-1 enriched cells were directly collected in chamber slides and infected with the TelomeScan adenovirus (OBP-401) for 24 h (4). The percentage of GFP (+), viable, telomerase-expressing cells was evaluated alongside cell recovery. Results: MCF7 cells were spiked into freshly prepared buffy coat, resuspended in 8mL, and processed at 5X, 10X, and 20X dilutions. The best compromise between recovery, WBC count, and processing time was obtained for the 10X dilution with 46% recovery. The majority of spiked cells (MCF7 and H1299) collected from buffy coat were shown to be viable, as demonstrated by infection with the telomerase-selective replicating adenovirus (TelomeScan, OBP-401). Cryopreservation of the buffy coat samples showed a slight reduction following cryopreservation. Discussion: The optimal performance of the Vortex chip on buffy coat samples has previously been shown at 10X dilution. This capability of the VTX-1 was further demonstrated through infection with a telomerase-selective replicating adenovirus. This combined workflow has shown seamless integration from whole blood. Noteworthy recovery and sensitivity were also observed with cryopreserved buffy coat. Patient biobanked samples are currently being pursued. The overall advantages are: a) viable cell detection, b) direct collection of cells into culture media ready for in vitro testing, c) no requirements for fixation, labeling, backflushing, d) a very low WBC background eliminating false positives, and e) the potential to retrospectively analyze clinical biobanked samples at a centralized laboratory.

Abstract B16: Noninvasive liquid biopsies for guideline-compliant diagnostic assessment in ovarian cancers

Abstract Diagnosis of ovarian adenocarcinomas is based on histopathologic analysis of tumor tissue obtained by invasive biopsies. However, invasive biopsies are associated with procedural expenses, sequelae, and risk of complications. Though immunocytochemistry (ICC) profiling of circulating tumor cells (CTCs, defined as EpCAM+, panCK+, CD45-) has been previously considered for diagnosis in some cancers, the attempts have been unsuccessful. We used a novel approach for enrichment of sufficient viable CTCs that permit meaningful ICC profiling for diagnosis of ovarian cancers in suspected cases. We obtained 15 mL of venous blood draw from 248 patients with confirmed diagnosis of ovarian cancer. PMBCs were isolated by centrifugation. CTCs were enriched from PBMC using an epigenetically acting enrichment process that is lethal towards normal (nonmalignant) cells but confers survival privilege on apoptosis-resistant CTCs. Harvested CTCs were confirmed by immunostaining for EpCAM and pan-CK. Deep ICC profiling of these CTCs was performed with organ-specific markers CA125 and WT1 as well as P63 as a nonspecific marker. Viable CTCs could be obtained from 225 samples (90.7%) out of 248. Among the 100 samples characterized by deep ICC profiling, 70 samples (70%) were positive for CA125 and WT1. None of the samples was positive for P63. Our results show that ICC-based characterization of CTCs can provide necessary diagnostic information noninvasively to substitute conventional procedures dependent on tissue extraction. Citation Format: Dadasaheb B. Akolkar, Darshana Patil, Revati Patil, Pradip Devhare, Pradip Fulmali, Pooja Fulmali, Shoeb Patel, Navin Srivastava, Vineet Datta, Cynthe Sime, Ajay Srinivasan, Rajan Datar. Noninvasive liquid biopsies for guideline-compliant diagnostic assessment in ovarian cancers [abstract]. In: Proceedings of the AACR Special Conference on Advances in Liquid Biopsies; Jan 13-16, 2020; Miami, FL. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(11_Suppl):Abstract nr B16.

Abstract PR02: Epigenetic biomarkers in cell-free DNA for early detection of high-grade serous ovarian carcinoma

Abstract Background: High-grade serous ovarian carcinoma (HGSOC) is the most lethal gynecologic malignancy, with a 5-year survival rate of 40% or less when diagnosed with late-stage disease. There is a dramatic benefit to early diagnosis of HGSOC with a 5-year survival rate of nearly 90% for stage I disease. Unfortunately, the majority of HGSOC is diagnosed as late-stage disease and there are no reliable early screening or diagnostic tests. The objective of this study is to identify tumor-specific aberrant DNA methylation alterations detectable in circulation in order to develop a biomarker to detect early-stage HGSOC. Methods: We performed reduced representation bisulfite sequencing (RRBS) on 33 stage 1 HGSOC tissues and 10 normal fallopian tube tissue samples from contralateral ovaries of patients with ovarian cancer. We selected the top 82 regions from RRBS for validation by targeted bisulfite amplicon sequencing in plasma. We successfully designed assays for 35 out of 82 regions. Targeted bisulfite amplicon sequencing was performed on these top 35 regions in a separate cohort of circulating tumor DNA extracted from 27 stage 1 HGSOC and 11 benign ovarian masses. We constructed a novel classifier model that we used to differentiate between patients with HGSOC and those with benign ovarian lesions. Results: We identified 10,972 statistically significant differentially methylated regions between stage 1 HGSOC and normal fallopian tube tissues using RRBS. We then applied the classifier model on the top 35 regions determined by targeted bisulfite amplicon sequencing data on an independent cohort of plasma samples. This demonstrated that 13 of the 35 regions were able to discriminate stage 1 HGSOC from benign tumors with a high sensitivity and specificity (Area Under Curve = 0.909). Conclusions: These novel, noninvasive biomarkers show promise as an early detection test for high-grade serous ovarian cancer. This abstract is also being presented as Poster A08. Citation Format: Ben Yi Tew, Gerald Gooden, David N. Buckley, Heather Miller, Monique Spillman, Lynda Roman, Bodour Salhia. Epigenetic biomarkers in cell-free DNA for early detection of high-grade serous ovarian carcinoma [abstract]. In: Proceedings of the AACR Special Conference on Advances in Liquid Biopsies; Jan 13-16, 2020; Miami, FL. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(11_Suppl):Abstract nr PR02.

Abstract A03: Evaluation of the Oncomine Pan-Cancer Cell-Free Assay for liquid biopsy profiling

Abstract Genomic profiling of liquid biopsies is no longer an emerging area of clinical research, but quickly becoming translated as part of contemporary clinical trials for use in the diagnostic setting. Comparatively less invasive than traditional solid tissue biopsy approaches, liquid biopsy specimens such as blood or urine can be used not only for early detection, but also for the monitoring of therapeutic response and progression. While custom circulating tumor DNA (ctDNA) assays are being academically developed for the profiling of specific biomarkers in a specific disease or therapeutic context, there are commercially available pan-cancer circulating nucleic acid (cNA) panels available. These commercially available panels can be used to reveal the spectrum of genomic changes within liquid biopsies and determine the level of sensitivity that can be obtained. In this study, we profiled 39 patients (n= 78 samples) with matched plasma and solid tumor to evaluate the Oncomine Pan-Cancer Cell-Free Assay, a 52-gene total nucleic acid panel for the detection of hotspot mutation and copy-number changes in key cancer driving genes. The evaluation cohort comprises 30 invasive breast cancers, 5 lung cancers, and 4 cancers of indeterminate origin, ranging from early and localized cancers to those that were characterized as late or metastatic. The Oncomine Pan-Cancer Cell-Free Assay utilizes molecular tagging technology to enable variant detection as low as 0.1%, given an optimal input of 20 ng of ctDNA. To determine whether variants detected using the Oncomine Pan-Cancer Cell-Free Assay were represented in the solid tumor, the solid tumor tissues were assayed using the Oncomine Pan-Cancer Cell-Free Assay in addition to the Oncomine Comprehensive Assay v3 (OCAv3). The OCAv3 profiles 161 pan-cancer genes for both full exon coverage as well as hotspot mutation and copy-number detection. Using these commercially available panels, for which OCAv3 is currently being utilized in the NCI-MATCH Trial (NCT02465060), we were able to validate the detection of genomic changes in the ctDNAs to the matching solid tissues; among them, MET mutations detected in the plasma of invasive breast cancers were identified in the matching solid tumor, as were PIK3CA and TP53 mutations, in addition to amplification of ERBB2. Similarly, KRAS mutations detected in the plasma of lung cancer patients were identified in the matched solid tumor. This presentation will summarize those findings with respect to the limit of detection obtained and the implications of monitoring progression and therapeutic response. These data suggest that existing commercial panels such as the Oncomine Pan-Cancer Cell-Free Assay, the Ion Torrent Platform, and Ion Reporter have the potential to be used in a routine clinical setting. Citation Format: Jane Bayani, Megan Hopkins, Melanie Spears, John M. S. Bartlett. Evaluation of the Oncomine Pan-Cancer Cell-Free Assay for liquid biopsy profiling [abstract]. In: Proceedings of the AACR Special Conference on Advances in Liquid Biopsies; Jan 13-16, 2020; Miami, FL. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(11_Suppl):Abstract nr A03.

Abstract B20: Prolonged time to clearance of circulating-tumor DNA from patients with limited-stage small-cell lung cancer is associated with inferior progression-free and overall survival

Abstract Introduction: Despite recent incremental advances, patients with small-cell lung cancer (SCLC) continue to have a poor prognosis, with a median overall survival (OS) of 12-20 months in limited-stage disease (LS-SCLC) and approximately 12 months in extensive-stage disease (ES-SCLC). In an attempt to improve the detection and monitoring of SCLC, assays of circulating tumor DNA (ctDNA) via blood-based next-generation sequencing (NGS) have been validated. In this study, we used a blood-based 14 gene SCLC ctDNA NGS panel to evaluate the prognostic significance of the diagnostic maximum ctDNA variant allele frequency (VAF), diagnostic mean ctDNA VAF, and time to ctDNA clearance while on first-line therapy. Design: In our previous work, we developed a ctDNA assay to sequence 14 genes (TP53, RB1, BRAF, KIT, NOTCH1-4, PIK3CA, PTEN, FGFR1, MYC, MYCL1, and MYCN) that are commonly mutated in SCLC. A total of 104 plasma samples were analyzed from a cohort of 14 patients with LS-SCLC who completed definitive chemoradiation (n=13) or surgical resection (n=1) and had an end-of-treatment blood collection within nine weeks (mean 10.5 days, range 0-63 days) of completion of definitive initial therapy. We used a Cox Proportional Hazards model to estimate the hazard ratio (HR) for progression-free survival (PFS) or death based on the diagnostic maximum ctDNA VAF, diagnostic mean ctDNA VAF, and time to ctDNA clearance on first-line therapy. Results: In our 14-patient cohort, we did not observe any association between progression or death and maximum diagnostic ctDNA VAF (PFS HR: 1.01, CI 0.97-1.05; OS HR: 1, CI 0.95-1.04) or mean diagnostic ctDNA VAF (PFS HR: 1.01, CI 0.94-1.08; OS HR: 0.99, CI 0.91-1.07). Of the specific mutations representing the maximum diagnostic ctDNA VAF, TP53 represented 14.2% (n=2), while the remaining included RB1 associated mutations (n=5), PIK3CA (n=1), MYCL1 amplification (n=1), or no ctDNA at diagnosis (n=5). Among patients with clearance of ctDNA during first-line therapy (n=9), delayed time to ctDNA clearance was associated with inferior PFS (HR 1.1, CI 1.01-1.19) and OS (HR 1.07, CI 1.01-1.15), with median time to clearance of 63 days (range 29-92 days). Notably, 3 patients cleared ctDNA on their first on-treatment draw at approximately 30 days (29, days, 29 days, and 32 days). These patients have had no evidence of relapse and all remain alive since the start of first-line treatment (1467, 965, and 383 days of follow-up). Of the patients who had disease recurrence, the median time to clearance on first-line therapy was 65 days and median time to progression was 249 days, with all but 1 patient succumbing to their disease (median OS 437 days). Conclusion: In patients with LS-SCLC, prolonged time to ctDNA clearance during first-line therapy is associated with inferior PFS and OS. Larger patient cohorts are needed to validate this finding. Citation Format: Christopher Cann, Prasad Kopparapu, Yingjun Yan, Anel Muterspaugh, Heidi Chen, Zhiguo Zhao, Sally York, Leora Horn, Kristen Ancell, Kenneth Wyman, Caterina Bertucci, Tristan Shaffer, Lauren Hodson, Kavita Garg, Seyed Ali Hosseini, Lee Lim, Christine M. Lovly, Wade Iams. Prolonged time to clearance of circulating-tumor DNA from patients with limited-stage small-cell lung cancer is associated with inferior progression-free and overall survival [abstract]. In: Proceedings of the AACR Special Conference on Advances in Liquid Biopsies; Jan 13-16, 2020; Miami, FL. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(11_Suppl):Abstract nr B20.

Abstract A55: Circulating tumor DNA in newly diagnosed intermediate-risk rhabdomyosarcoma

Abstract Background: Patients with intermediate-risk rhabdomyosarcoma (IR RMS) have approximately 60-70% chance of survival with current therapies. However, biomarkers of outcome and response to therapy are lacking for these patients. Circulating tumor DNA (ctDNA) has been shown to be prognostic in a wide range of cancer types, but it is unknown whether patients with IR RMS have detectable levels of ctDNA. Objective: To study ctDNA prevalence and prognostic impact in newly diagnosed IR RMS, we utilized two next-generation sequencing (NGS) approaches that detect the presence of somatic copy-number changes and oncogenic translocations. Methods: Cell-free DNA was extracted from serum samples obtained from patients with newly diagnosed IR RMS who enrolled on the COG studies ARST0531 and D9803, including patients with embryonal (ERMS) and alveolar (ARMS) histology subtypes. While both subtypes are characterized by frequent somatic copy-number variants (CNVs), ARMS is also defined by recurrent translocations. To detect CNVs, we performed ultra-low passage whole-genome sequencing (ULP-WGS). Translocations were detected with a validated custom hybrid capture assay (TranSS-Seq). Results: Serum samples were analyzed from 132 patients with IR RMS, including 75 with ERMS and 57 with ARMS. ctDNA was detected by CNVs in 70% (92/132) of IR RMS patients with similar detection rates in each histology: 65% (49/75) in ERMS and 75% (43/57) in ARMS. Among the ARMS samples sequenced, only 37% (18/49) were positive by TranSS-Seq. Furthermore, copy-number events resulting in amplifications of the somatic translocation frequently resulted in miscalculation of ctDNA content by the TranSS-Seq method. Estimates of event-free and overall survival were lower in patients with detectable ctDNA, though the differences were not statistically significant. Conclusion: Patients with IR RMS frequently have detectable levels of ctDNA that can be measured by NGS assays designed to detect somatic structural events. This study relied on previously banked serum samples and a relatively small retrospective analysis. These findings provide justification for our current efforts to utilize a large prospective study, COG ARST1431, to collect pretreatment and serial blood samples using procedures optimized for ctDNA assays. Sequencing of matched tumor samples is ongoing to understand the differences in sensitivity between ULP-WGS and TranSS-Seq for ctDNA detection in patients with ARMS. Citation Format: Samuel Abbou, David Hall, Donald A. Barkauskas, Kelly Klega, Anwesha Nag, Aaron R. Thorner, Mark Krailo, Steven Dubois, Douglas S. Hawkins, Brian D. Crompton. Circulating tumor DNA in newly diagnosed intermediate-risk rhabdomyosarcoma [abstract]. In: Proceedings of the AACR Special Conference on Advances in Liquid Biopsies; Jan 13-16, 2020; Miami, FL. Philadelphia (PA): AACR; Clin Cancer Res 2020;26(11_Suppl):Abstract nr A55.