Systematic drug repositioning through mining adverse event data in ClinicalTrials.gov

Abstract

Drug repositioning (i.e., drug repurposing) is the process of discovering new uses for marketed drugs. Historically, such discoveries were serendipitous. However, the rapid growth in electronic clinical data and text mining tools makes it feasible to systematically identify drugs with the potential to be repurposed. Described here is a novel method of drug repositioning by mining ClinicalTrials.gov. The text mining tools I2E (Linguamatics) and PolyAnalyst (Megaputer) were utilized. An I2E query extracts "Serious Adverse Events" (SAE) data from randomized trials in ClinicalTrials.gov. Through a statistical algorithm, a PolyAnalyst workflow ranks the drugs where the treatment arm has fewer predefined SAEs than the control arm, indicating that potentially the drug is reducing the level of SAE. Hypotheses could then be generated for the new use of these drugs based on the predefined SAE that is indicative of disease (for example, cancer).

Keywords: ClinicalTrials.gov; Drug repositioning; Drug repurposing; Indication discovery; Text mining.

Figure 1. The I2E query.

See Supplemental Information 1 to reproduce the query by copying and pasting the YAML script into the I2E Pro interface. The query was run on the I2E index that covers the data posted in ClinicalTrials.gov up to August 14, 2016.

Figure 2. An example of the data extracted from ClinicalTrials.gov (A) into Excel (B) by the I2E query described above.

The top two rows in (B) show the data extracted from the table in (A). The precision of the I2E query described above is 100%, and the recall is estimated as 99% assuming 1% of the cancer terms that the trial sponsors used are not among the cancer synonyms collected by MeSH or NCI.