Linguamatics Text Mining Summit 2018

The Linguamatics Text Mining Summit 2018 will take place on October 15 - 17, 2018 at the beautiful Wentworth by the Sea in New Castle, New Hampshire, USA. Join the conversation on Twitter using #TMS18.

Speakers include Eli Lilly, Mercy, Bristol-Myers Squibb, Atrius, Novo Nordisk, Sanofi, Regeneron Pharmaceuticals, Secure Exchange Solutions, GSK. More speakers will be announced soon! Interested in speaking? Please email speakers@linguamatics.com

Why attend?

This event is the ideal opportunity for experienced, new and potential I2E users to:

• Network and exchange ideas with peers and text mining experts

• Understand the challenges other pharmaceutical and healthcare professionals are facing, and explore solutions to these challenges

• Gain hands-on experience of NLP text-mining and where it can fit into your organization, through workshops and training

Want to find out more?

Learn more about the Linguamatics Text Mining Summit 2018.

You can also read and download the final TMS agenda.

Presentations and Abstracts

Eric Su, Principal Research Scientist at Eli Lilly

Presentation: Mining ClinicalTrials.gov to Support Clinical Trial Design

Abstract: ClinicalTrials.gov (CT.gov) has grown to include over 280,000 trials by 2018, from over 200 countries globally.  Much knowledge can be learned from CT.gov, to assist in trial planning, protocol design, and more. However, manual review of all relevant trials is not feasible with such a large database.  I2E has the capability of extracting data and text with high precision and recall.  I will present two examples of data extraction from CT.gov to support clinical trial design at Lilly.

Yun Yun Yang, Senior Patent Analyst at Bristol-Myers Squibb

Presentation: A Multiple Step Approach for Finding Clinical Trials For GPCR Target Class

Abstract: G-protein coupled receptors (GPCRs) represent a large class of important therapeutic targets.  The first objective of this work was to provide therapeutic area trend analysis, based on clinical trial information, on GPCR target classes.  The second objective was to identify clinical trial information associated with specific GPCR targets.  We used the value-added information that Cortellis provides around drugs and target-based actions, and applied i2e disease ontologies to achieve our objectives.  This presentation provides a framework on how to find clinical trials associated with a target class or specific targets. These targets can then be categorized into therapeutic areas of interest and visualized for trend analysis.

Thierry Breyette, Associate Director, Information Analytics at Novo Nordisk

Presentation: Using NLP at Novo Nordisk to Generate Actionable Insights from Real World Data

Abstract: The density and variability of the information landscape is making it increasingly difficult to identify meaningful trends in data.  Traditional data sources such as clinical trial data and publication data are one piece of an increasingly complex information puzzle.  As data capture and publishing platforms explode, newer and highly varied data sources are available for analysis, including internally generated data, social data, patient data, clinician data, market data, hospital data, etc. Much of these data are in unstructured, textual format, making it difficult to extract and analyse using traditional search methods. Building a forward-looking analytics framework to tackle these new data challenges requires both extensible and flexible tools, and creative thinking.

At Novo Nordisk we are using advanced tools and technologies such as natural language processing (NLP) to gain real value from sources including call centre feeds, information from medical liaisons and health care providers. These enable us to identify macro and micro healthcare market trends in the US, detect patterns in clinical trial protocol deviations, and discern patterns in patient sentiment, compliance, routines, behaviors, and overall treatment satisfaction and outcomes. The talk will focus on our approach to these projects, the outcome and impact.

Cheng Zhu, Senior Scientist, Translational Bioinformatics and Research Informatics at Sanofi

Presentation: Integrated text mining approaches for drug target new indication search

Abstract: There is strong interest within biopharma to expand the therapeutic base of drugs on the pipelines. To address this, one way is to identify novel drug target to disease associations that can be considered as potential new indications. The traditional approach on this can be described as an ad hoc process that relies on expert knowledge and experiment observations, which are usually limited and time consuming. Text mining provides a powerful approach that enables us to quickly and systematically discover the hidden linkages between target and new indications. In this presentation, we will introduce several text mining strategies and use cases on using I2E to interconnect drug target and diseases.  Our approaches attempt to capture and integrate all the available information of drug target from literature, including genetic causal or risk mutations to diseases, and the underlying causal pathways, cell types and clinical phenotypes that shared by target and diseases.  The approaches allow us to quickly identify new indication opportunities from target-disease pairs that have various commonalities based on broad scientific, medical and strategic values. We have applied the approaches in several disease areas, such as respiratory and rare diseases as proof of principle.

Peng Zhang, Sr Staff Scientist, Target Information Group, Regeneron Pharmaceuticals, Inc.

Presentation: Knowledge Management and Data Integration – Building A Resource for GPCR Drug Discovery

Abstract: The Target Information Group (TIG) at Regeneron uses a wide variety of public and proprietary information sources to help Regeneron researchers address their scientific questions in target and disease biology. G-protein coupled receptors (GPCRs) family has traditionally been the biggest class of targets for drug discovery in pharmaceutical industry, with the efforts mostly focused on small molecule class of drugs. As a biologics-focused company, Regeneron has decided to reassess the landscape of this drug discovery space and to identify potential opportunities for developing novel or better drugs by leveraging our unique VelociSuiteÒ technologies. A web-based dashboard was built to provide a focused information portal for all GPCRs from human and mouse. The information integrated together includes GPCR hierarchical family classifications, known natural ligand and signal transduction mechanism. This dashboard also contains all drugs that are known to target GPCRs and provides ways to quickly drill down for more information according to their molecular type, development status, mechanism of action, etc. Linguamatics I2E was used to mine large scale clinical trials information related to these known drugs and the result provided unique insights on the potential reasons behind development status change. This resource has been widely used within the company and provided a starting point for various drug discovery projects.

Craig Monsen, Chief Medical Information Officer at Atrius Health

Presentation: Operationalizing NLP to support value-based care at Atrius Health

Abstract: Healthcare providers are facing an urgent need to streamline operations while improving quality of care and patient satisfaction. With a wealth of technology hype around AI, Natural Language Processing (NLP) and big data, how are providers to know what investments to make and how to bring these technologies into production use? During this talk, we bring together leading healthcare organization, Atrius Health, and NLP experts, Linguamatics, to explore the practical uses of NLP in healthcare and give real-life examples of how Atrius Health has implemented processes to improve clinical documentation, identify at-risk patients and streamline their ACO reporting.

Tom McGraw, Senior Vice President of Product Development for Secure Exchange Solutions (SES)

Presentation: SPOT for Clinical Review

Abstract: Secure Exchange Solutions (SES) launched its SPOT for Clinical Review product in June of this year.  SPOT for Clinical Review leverages the natural language capabilities in Linguamatics I2E product to read medical records and other freeform text and normalized data to ascertain if payer’s prior authorization, claims adjudication, or other requirements have been met.  SPOT leverages SES’ position as a leader in digital communication of clinical healthcare information and is designed to speed up medical review, lower its costs, and improve consistency.  Designed for the healthcare payer market, the product has also received interest from the healthcare provider, workers compensation carrier, and life insurance markets.  An overview of SPOT for Clinical Review and application of payers’ rules in several areas will be presented.

Chaya Duraiswami, Associate Fellow, Compliance Solutions Manager, GSK

Presentation: Use of Text Analytics to Enable Data-Driven Risk Management

Abstract: Biopharm Product Development and Supply (BPDS) within GSK utilizes a data-driven approach to risk management, by consolidating internal data feeds from Deviations, Corrective and preventative actions (CAPAs), Risks, and Response to questions (RTQs) to form a Data Lake. The Data Lake also receives external data feeds from FDA Warning Letters (483s), Biological License Application (BLA) Review Reports, White Papers and industry benchmark repositories that add to the broader context of relevancy. This establishes a broad knowledge-base for analysing risk rating and frequency, understanding risk relation to industry practices and applying thresholds for risk management (accept, transfer, mitigate, accept). The Data Lake of unstructured or semi-structured data is structured using Linguamatics which then enables the extraction of intelligence, (concept and sentiments) embedded in the data. The value proposition is further maximized with simple to understand visualizations (that are easy to drill down), sustainable (up-dates contemporaneously) and scalable reporting of risks, its analysis, and recommendations to act. This presentation will demonstrate how Linguamatics, which uses Natural Language Processing and Machine Learning algorithms, can be used for addressing emerging concerns.

Kerry Bommarito, Director Data Science at Mercy

Presentation: Mercy’s Experience: Natural Language Processing.

Stuart Murray, Research Fellow/Director of Informatics, Agios Pharmaceuticals, Agios

Using I2E to Accelerate Tool Compound Discovery for Chemical Genetic Screens

Genetic changes such as mutations in cancers create potential druggable Achilles Heels. At Agios we use genomics, proteomics and metabolomics date to build an understanding of how these genetic changes perturb metabolic biology in a tumour. This insight is used to design screening strategies to identify potential new drug targets. One such strategy is a Chemical Genetics Screen. At Agios a Chemical Genetics Screen explores the function of metabolic proteins and pathways in cancer cells by the using of chemical libraries of small molecules (tool compounds). In this presentation, we will highlight results from using I2E to build libraries of tool compounds that have known activities and defined characteristics. We then used these tool compounds in medium throughput chemical genetic screen across a broad cancer cell line panel. The outcome of the screen revealed important genes or pathways involved in cell growth and proliferation. We will highlight the benefits of using this kind of approach to speed the identification of potential new genes of interest and the development of tools to validate new anti-cancer targets.