TELL : +91 0 99 08 87 07 06

 Mcmed International ®: Largest Journal Publication in world

European journal of molecular biology and biochemistry

Volume 1, Issue 5, 2014
Mcmed International
European journal of molecular biology and biochemistry
Issn
2348 - 2192 (Print), 2348 - 2206 (Online)
Frequency
bi-annual
Email
editorejmbb@mcmed.us
Journal Home page
http://mcmed.us/journal/ejmbb
Recommend to
Purchase
Abstract
Title
AUTOMATED ANALYSIS OF ANGIOGRAPHY AND VENOGRAPHY REPORTS FOR DIAGNOSING THROMBOEMBOLIC DISEASES: A NATURAL LANGUAGE PROCESSING APPROACH
Author
Dr. Taheerkhan Tadakod
Email
keyword
Medical informatics, Phlebography, Kinematics, NLP, Phlebology, Affective disorders
Abstract
This study evaluates the application of natural language processing (NLP) techniques to analyze angiography and venography reports, focusing on diagnosing thromboembolic diseases and identifying key findings incidentally mentioned in the reports. A dataset comprising 573 identified radiology reports was manually annotated using NLP tools. Machine learning algorithms were then employed to identify clinically relevant findings, including deep vein thrombosis, pulmonary embolisms, and other incidental findings. Consideration was given to addressing data imbalances and report structure complexities during model development. The results demonstrated that the proposed approach facilitated the detection of pulmonary embolism, deep vein thrombosis, and incidental clinically relevant findings at a significantly lower cost, with faster and more efficient processing rates, achieving best F measures of 0.98, 1.00, and 0.80, respectively. The study highlights improvements in concept recognition, mode identification, and relationship analysis within radiology reports, paving the way for automated identification of medical terms, modalities, and relationships. Such advancements in NLP-based annotation and classification systems have the potential to enhance epidemiological surveillance, performance monitoring, and accreditation in the field of radiology
Back to Top >>>>