Search

e14051 Background: Adoption of next-generation sequencing (NGS) technology in routine clinical practice has enabled the detection of genetic aberrations such as single nucleotide variants, copy number alterations, and gene fusions. Pathway and network analyses (PNA) are key components for evaluation of NGS data in a clinical setting to explain findings involving thousands of altered genes and proteins with a smaller and more interpretable set of altered processes. Though PNA have been applied to identify driver genes and pathways in cohort-based analyses, its application in precision oncology remains unexplored. We investigate the potential utility of the Watson for Genomics (WfG) pathway analyses tool in interpreting complex and multiple genomic alterations in individual cancers. Methods: DNA and RNA isolated from 70 patient tumors across 30 different cancer types were processed with Illumina’s TST170 NGS platform. WfG’s feature of pathway analyses was used to identify gene variants, signaling pathways, networks, and the drugs targeting these alterations based on evidence in the clinical literature and FDA drug databases. Results: Analyses defined 5 different pathway/network models: 1) downstream therapeutic targets, 2) synthetic lethality, 3) combinatorial downstream targets + synthetic lethality, 4) two or more pathways converging to downstream targets, and 5) complex profile analyses. The five PNA models are illustrated by the following unique cases. 1) A thyroid cancer case with HRAS variant and activated RAF1 downstream pathway showed MAPK1/3 were suggestive of relevant targets. 2) An acute myeloid leukemia case with BRCA1, BRCA2 and PTEN variants, targeting a common synthetic lethal partner PARP1 was ideal for therapy. 3) A penile carcinoma case with BRAF, CDKN2A and TP53 variants, targeting the BRAF downstream pathway in combination with either CDKN2A or TP53 were the likely choice for therapy. 4) A glioma case with activated PI3K and MEK downstream pathway, targeting a common downstream marker would block both pathways. 5) A breast carcinoma case with a complex pathogenic variant profile provided relevant clinical information and levels of evidence for multiple drug targets. Conclusions: We discovered that the integrated WfG pathway analyses tool is ideal for visualization of the variants with levels of evidence from clinical literature and FDA drug databases that can help inform treatment options and provides a holistic understanding of a specific tumor profile allowing the treating clinician to select personalized targeted therapy.

e15672 Background: Rare tumors, such as cholangiocarcinoma (CC), have no established treatment protocols, forcing clinicians to select therapies based on educated guesses or small case series. Genomic tumor analyses fail to identify options for most patients. Here, we show the feasibility of using a high-throughput functional assay to test targeted drug libraries on individual patient-derived tumor organoids from CC to yield experimentally-validated therapeutic options. Methods: Three consecutive patients with CC who were undergoing a standard of care biopsy or resection were enrolled in an IRB-approved study. Patient characteristics, demographics, tumor pathology, including Next Generation Sequencing (NGS) data and organoid-derived drug recommendations are shown in the Table. Specimen (~ 200 mg of tumor) was shipped on ice overnight to SEngine Precision Medicine for tumor-derived organoid culture and sensitivity analysis. Organoids were evaluated using a multi-dose response to each drug in a library of 120 FDA-approved and investigational drugs and compared those responses to all prior patients. This established both functional sensitivity and the uniqueness of each patient’s response. Results: We tested operational logistics, tumor growth rates, and chemosensitivity reporting timelines. This has allowed us to: 1) operationalize the expansion process, 2) quantify turn-around time, and 3) identify any technical or logistic barriers. All specimens resulted in adequate growth for sensitivity characterization. None of the NGS studies resulted in recommendations for sensitivity to targeted chemotherapy. Conversely, tumor organoid assays rapidly identified selective, personalized drugs. Conclusions: We have shown the feasibility of this approach in CC. Our preliminary results will inform the design of a future prospective clinical trial to establish and validate this method to select personalized cancer treatments for rare malignancies.[Table: see text]

e18588Background: There were ~49,300 clinical research publications in oncology in the last 10 years. Oncologists do not have the time to read them all. AI and computationally assisted systems may ...

e22164 Background: Next generation sequencing (NGS) has enabled genomewide personalized oncology efforts at centers and companies with the specialty expertise and infrastructure required for action...

11097 Background: Advanced genomics technology can be used to characterize genomic alterations (GA) that potentially drive tumor growth. We have established a sequencing and array based pipeline to...

e13562 Background: Personalized oncology has advanced through genomic and proteomic platforms. BCR-abl; EGFr and ALK have provided drug-able targets and companion diagnostics in several diseases, yet many transforming events in humans are polygenic, complex and incompletely understood at a genomic level. Recognition that oncogenesis reflects changes in the cell and its micro environment has renewed interest in whole cell experimental models that capture native-state cell-cell, -stroma and -vascular signaling. Ex vivo analysis of programmed cell death (EVA/PCD) has been shown to correlate significantly with response, time to progression and survival (Nagourney, R. Curr. Treat Op Oncol, 2006).To explore EVA/PCD functional profiling in Brazil, hospital-based investigators, Centro de Genomas and Rational Therapeutics coordinated the transport of 67 surgical specimens for analyses. Methods: Dose-response curves using metabolic (ATP-content, mitochondrial) and morphologic endpoints, interpolated to LC50's and synergy by median-effect, were compared with databases to identify patient-specific profiles for cytotoxics, targeted agents and combinations. Reports provided day 7. Results: 62/67 (92%) provided adequate tumor for analysis; 39 male (58%); 28 (42%) female; 6 chemo-naive and 61 previously treated. Results provided for < 8 drugs in 14/62 (22%); 8-16 drugs in 34/62 (54%) and > 16 drugs in 14/62 (22%). Of 22 tumor types, breast (8); Melanoma (8); NSCLC (8); ovary (7); pancreas (7) and sarcoma (4) predominated. EVA/PCD was used to select the most active combinations. Agents selected, response rates and durations are being tabulated and will be reported. Conclusions: The transport, processing and reporting of human tumor primary culture analyses is feasible, providing results in 92% of specimens and median of 12 drugs evaluated (range 4-32). Preliminary outcomes in these drug-refractory patients reveal that novel, often unexpected drug combinations (Everolimus and Lapatinib in triple negative breast) were identified and provided objective tumor responses, supporting EVA/PCD in therapy selection (personalized therapy) and drug development.

e21032 Background: Personalized oncology offers the promise of selectively applying therapeutics to patients most likely to benefit, while sparing those unlikely to benefit from potentially toxic therapies. Oncotype DX is a 21-gene assay utilized to identify hormone-receptor positive (HR+), node negative, breast cancer (Br CA) patients who may be successfully treated with adjuvant hormonal therapy alone. We hypothesized that practice patterns with adjuvant therapy have changed since the commercial availability of Oncotype DX in 2004. Methods: The Public National Cancer Database was queried to identify patients age ≥ 50 with stage I or II Br CA diagnosed from 2000 to 2008. Patients were classified by adjuvant therapy including hormone, chemotherapy, hormone and chemotherapy, and no hormone or chemotherapy. Log-binomial regression was used to estimate prevalence ratios for the proportion of patients receiving adjuvant therapies from 2000-2003 compared to 2004-2008. Results: 833,018 patients age ≥ 50 with stage I or II Br CA were identified. The application of adjuvant therapies for the periods pre- and post-availability of Oncotype DX are detailed in the Table. Conclusions: There has been significant increase (13%) in the use of hormonal therapy alone as adjuvant therapy for patients age ≥ 50 with HR+ stage I-II Br CA since commercial availability of Oncotype DX. While this has been slightly offset by a decrease in the use of chemotherapy plus hormonal therapy, there has been a larger decrease in the use of “no adjuvant therapy”. Data regarding predictive biomarkers should be captured by registries in an effort to determine the true impact of these tests on treatment utilization. [Table: see text]