Our objectives were to develop a user-friendly graphic interface for a module that integrates traditional radiology reporting, natural language processing, and editing capabilities; to facilitate the structuring of radiology reports as part of routine clinical practice; to use a commercial speech recognition module for online transcription; and to implement the module in a hardware-independent environment.
After implementation, the module was tested with 150 chest radiology reports by two radiologists and assessed for ease of use and accuracy. Overall, accuracy was close to 90% and user satisfaction was high. When radiology reports are structured as a part of routine clinical practice, it is possible to accomplish intelligent indexing and retrieval to facilitate teaching and research.
Supported by National Institutes of Health grant NCI:2P01CA51198-06.
Sinha US, Dai BY, Johnson DB, Taira RK, Dionisio JDN, Tashima GH, Golamco M, Kangarloo H. “Interactive Software for Generation and Visualization of Structured Findings in Radiology Reports.” American Journal of Roentgenology 175:609-612, 2000.