Jianguo Zhang PhD, Presenter: Nothing to Disclose

Weiling Zheng MS, Abstract Co-Author: Nothing to Disclose

Xiangjiao Chen MS, Abstract Co-Author: Nothing to Disclose

Jianyong Sun, Abstract Co-Author: Nothing to Disclose

Kai Zhang BS, Abstract Co-Author: Nothing to Disclose

Yuanyuan Yang MS, Abstract Co-Author: Nothing to Disclose

We designed an innovation visual indexing method to assign an anatomical 3D structure object to every patient visually to store indexes of the patients’ historical study information, and presented this prototype system in scientific presentation in 2011 RSNA Conference. In this presentation, we present a new approach to extract and process the semantic and characteristic information from RIS-integrated PACS to create the 3D content-based Visual Index (VI) to enable doctors to understand the content of reports and images before they retrieve and read them from PACS and RIS.

This approach includes following steps: (1) Building a medical characteristic semantic knowledge base based on SNOMED and ACR code standards to normalize the semantic description of reports and map them to related VI ; (2) Developing natural language processing (NLP) engine to perform semantic analysis and logical judgment on both free and sub-structured text reports; (3) Applying the knowledge base and NLP engine on reports to extract medical characteristics, and mapping extracted information to related organ/parts of 3D human model to create the VI. As the date/time information of studies is included in an electronic anatomic structure of 3D model, the dash board feature of examined body/parts are also achieved in this 3D Content-Based Visual Index.

We performed the testing procedures on 559 samples of radiological reports which include 853 focuses, and achieved 828 focuses’ information. The successful rate of focus extraction is 97.1%. Although more than 900 visual body parts in 3D model are segmented visually to represent the content of a VI object, there are still some features and content extracted being not mapped to the proper levels of the 3D anatomic model.

The developed method can efficiently convert the content of radiological reports to the VI of a patient to enable doctors to understand the content of reports and images before they retrieve and read them from PACS and RIS. The evaluation results showed that the VI concept has higher potential to be used in RIS/PACS quickly to search medical records and deeply to understand patients’ healthcare status of patients with a large number of examinations.

Visual Index can probe, mark out and display meaningful information with a 3D human model in different status for different body parts and medical conditions of a patient in PACS and RIS.

Zhang, J, Zheng, W, Chen, X, Sun, J, Zhang, K, Yang, Y, Semantic Extraction and Processing of Medical Records from RIS Integrated PACS for 3D Content-based Visual Index.  Radiological Society of North America 2012 Scientific Assembly and Annual Meeting, November 25 - November 30, 2012 ,Chicago IL. http://archive.rsna.org/2012/12029779.html