Kim Submits NIH Grant Application on Cancer Patient Exercise

Kim Submits NIH Grant Application on Cancer Patient Exercise

hyeoneui kimKudos to Hyeoneui Kim, associate professor, for the submission of her NIH U01 application entitled "Global Harmonization of Data from Multiple Sources to Support Evidence-based Research on Exercise Interventions for Cancer Patients." This proposal requests funds for a 3 year period with a start date of July 1, 2020.

Many cancer patients and cancer survivors endure a suboptimal quality of life due to the adverse symptoms associated with the disease process and its treatments. As of January 2019, there were an estimated 16.9 million cancer survivors in the United States and who will all require lifetime monitoring and management of the condition. This growing trend highlights the importance of cancer research and clinical practice that are geared toward promoting the successful return to and maintenance of a healthy life for this population. There is overwhelming scientific evidence that physical activity has significantly positive effects on adverse effects, treatment effects, psychological well-being, and general quality of health. However, adherence to exercise recommendations for these patients remains low, currently less than 30%. Among many factors that hamper translating research evidence to practice is the lack of a systematic means of identifying scientific evidence best suited for a specific patient case. Determining what constitutes the right level and type of physical activity for an individual patient is not always straightforward and can be influenced by the cancer type/subtype, comorbidities, biological determinants, demographics, and various environmental or behavioral context.

The main purpose of the proposed study therefore is to develop methods and resources to extract the evidence details regarding exercise effects on cancer from scientific papers and clinical trial reports so that this evidence can be systematically searched, reviewed, compared, and integrated with other data sources for further analyses. The proposed study addresses major issues related to heterogeneous, fragmented representations that an information system will encounter in attempting to operationalize knowledge reported in the primary literature for end-user research and clinical applications. The technical contributions of the study are within the areas of: 1) Natural language processing based text-mining; 2) Data standardization to facilitate interoperability of the extracted evidence, and 3) Adoption of semantic web technologies and deep-learning based graph embedding to facilitate and enhance query processing. The system requirements and design, and prototype development and testing will consider both translational research and clinical research.

The three aims that this study intends to achieve are: Aim 1) developing the core technologies for the Literature to Knowledgebase (L2KB) text-mining pipeline and building an evidence knowledgebase (eKB) where key information is encoded with standardized biomedical terminologies and ontologies, Aim 2) operationalizing the eKB developed in Aim 1 with a Resource Description Framework (RDF) transformation (i.e., eKB-RDF) with graph embeddings, as well as a GUI-based query application, and Aim 3) evaluating the performance, usability, and utility of the methods and resources developed in Aim 1 and 2 using query scenarios and research use cases in exercise related cancer research.

Scroll back to top automatically