Exact graph-based analysis of scientific articles on clinical trials

This article describes Amorpha, a software package based on new concept of exact graph-based linguistic analysis. Analytical capabilities of Amorpha are demonstrated using analysis of scientific abstracts on clinical trials from PubMed. Current trends in therapy of breast cancer and psoriatic arthritis were analyzed using 400 abstracts on breast cancer and 131 abstracts on psoriatic arthritis. The spectrum of diseases that currently treated with paclitaxel was extracted from 400 most recent abstracts on paclitaxel. In addition to text representation, analytical results are presented as graph images showing essential concepts of a text. Amorpha is not designed specifically to analyze clinical trials and will be also useful for analysis of biological scientific articles and regulatory documents. Amorpha does not require any preliminary knowledge base, ensures full coverage of target text and 100% accuracy of obtained results.