BCR-ABL1 positive leukemias account for a substantial portion of adult leukemia. Tyrosine kinase inhibitors (TKIs) have dramatically changed survival outlook. However, current protocols recommend patients receive TKI chemotherapy agents indefinitely, causing long-term toxicity and substantive quality of life deficits, leading to decreased TKI compliance. Moreover, >50% of patients who stop TKIs ultimately relapse and are not as responsive to post-relapse treatment; additionally, mutation resistance is becoming an increasing issue with TKIs as patients live longer. It is hypothesized patient heterogeneity within BCR-ABL1 leukemias are a major driving factor on outcome and strongly influences optimal TKI selection and cessation. However, small cohort size and disease rarity has impacted large, pragmatic clinical trials, necessitating a big data approach. The overall goal is to quilt together individual studies to produce a comprehensive view of BCR-ABL1 leukemias that includes epidemiology, etiology, assessment, and therapy, as well their inter-relationships. With a comprehensive view, personalized, predictive medicine becomes possible. This project utilizes literature mining and ?big data? techniques to analyze four major categories: epidemiology (who gets BCR-ABL1 leukemias, how response correlates to patient characteristics, etc.); etiology (what factors trigger mutation, mechanisms to improve TKI specificity, preclinical model metrics, prognostic indicators of recurrence/relapse, etc.); assessment (identifying new diagnostic/prognostic metrics, improving polymerase chain reaction (PCR) protocols, objective staging criteria); and treatment (aggregate effect sizes among different therapies, short and long-term side effect profiles, TKI selection protocols, adjunctive and combination therapies, criteria for TKI cessation, etc.).
The specific aims of the project include: 1) prototype a data path and construct infrastructure for BCR-ABL1 data curation from literature and/or clinical sources; 2) construct literature ontological field map to quantify topic depth, aggregate data, and identify relationships within and between categories; 3) perform exploratory analysis to assess aggregate statistical power and prototype predictive models for TKI optimization. In summary, the present project delivers a 21st century, big data approach for BCR-ABL1 leukemia to optimize clinical management and expedite basic preclinical research.
