Inferring Negative Molecular Biomarker Data at Scale.

PURPOSE: Patients who represent the negative biomarker population, those tested for a biomarker but found to be negative, are a critical component of the growing molecular data repository. Many next-generation sequencing (NGS)-based tumor sequencing panels test hundreds of genes, but most laboratories do not provide explicit negative results on test reports nor in their structured data. However, the need for a complete picture of the testing landscape is significant. Syapse has created an internal ingestion and data transformation pipeline that uses the power of natural language processing (NLP), terminology management, and internal rulesets to semantically align data and infer negative results not explicitly stated. PATIENTS AND METHODS: Patients within the learning health network with a cancer diagnosis and at least one NGS-based molecular report were included. To obtain this critical negative result data, laboratory gene panel information was extracted and transformed using NLP techniques into a semistructured format for analysis. A normalization ontology was created in tandem. With this approach, we were able to successfully leverage positive biomarker data to derive negative data and create a comprehensive data set for molecular testing paradigms. RESULTS: The application of this process resulted in a drastic improvement in data completeness and clarity, especially when compared with other similar data sets. CONCLUSION: The ability to accurately determine positivity and testing rates among patient populations is imperative. With only positive results, it is impossible to draw conclusions about the entire tested population or the characteristics of the subgroup who are negative for the biomarker in question. We leverage these values to perform quality checks on ingested data, and end users can easily monitor their adherence to testing recommendations.

Customizable Natural Language Processing Biomarker Extraction Tool.

PURPOSE: Natural language processing (NLP) in pathology reports to extract biomarker information is an ongoing area of research. MetaMap is a natural language processing tool developed and funded by the National Library of Medicine to map biomedical text to the Unified Medical Language System Metathesaurus by applying specific tags to clinically relevant terms. Although results are useful without additional postprocessing, these tags lack important contextual information. METHODS: Our novel method takes terminology-driven semantic tags and incorporates those into a semantic frame that is task-specific to add necessary context to MetaMap. We use important contextual information to capture biomarker results to support Community Health System's use of Precision Medicine treatments for patients with cancer. For each biomarker, the name, type, numeric quantifiers, non-numeric qualifiers, and the time frame are extracted. These fields then associate biomarkers with their context in the pathology report such as test type, probe intensity, copy-number changes, and even failed results. A selection of 6,713 relevant reports contained the following standard-of-care biomarkers for metastatic breast cancer: breast cancer gene 1 and 2, estrogen receptor, progesterone receptor, human epidermal growth factor receptor 2, and programmed death-ligand 1. RESULTS: The method was tested on pathology reports from the internal pathology laboratory at Henry Ford Health System. A certified tumor registrar reviewed 400 tests, which showed > 95% accuracy for all extracted biomarker types. CONCLUSION: Using this new method, it is possible to extract high-quality, contextual biomarker information, and this represents a significant advance in biomarker extraction.