John Nicholas Morelli MD, Presenter: Nothing to Disclose

Valery Donfack, Abstract Co-Author: Nothing to Disclose

Regis M. Duvauferrier MD, Abstract Co-Author: Nothing to Disclose

Jeremy Lasbleiz MD, Abstract Co-Author: Nothing to Disclose

A number of different classification systems have been proposed for orthopedic tumors (Recist 1.1, WHO). The variety of systems in existence may result in confusion among radiologists who are faced with classifying and staging particular tumor. The goal of this work is to utilize two informatics tools—segmentation and ontology—to construct an interactive tool guiding bone tumor staging and followup.

The first part of this work consisted of contructing an ontology of bone tumor staging including the different information required to accurately classify a malignancy according to tumor type. Using Protégé, an ontology editor, a domain representation was created using concept, relation, and object properties. Two experts in the field validated the ontology. With owl API, an interactive form was constructed to allow users to interact with the ontology. Java has also was implemented to make the tool suitable for Internet use. The segmentation method was based on Image J, a public domain Java image processing program. The tool has been tested on ten actual patient studies.

The tool allows a user to select the appropriate staging form according to tumor type (osteosarcoma, chondrosarcoma, etc.). The corresponding form then must be populated by the user. The semiquantitative segmentation method also brings information (such as tumor size) that populates the form. In the end, a completed report is constructed with information relevant to the staging of the tumor. In the end an exploited document with an xml marks is obtained.

Bone malignancies are relatively uncommon, thus often rendering it difficult for radiologists to produce concise, accurate reports that aid oncologists in establishing tumor stage. This interactive tool decreases variability in radiologist reporting ensuring relevant reports are transmitted to the oncologist.

Utilization of ontology-driven informatics tools allow radiologists to construct clinically relevant reports in oncology patients.

Morelli, J, Donfack, V, Duvauferrier, R, Lasbleiz, J, An Ontology-based Informatics Tool for Bone Tumor Staging and Follow-up.  Radiological Society of North America 2012 Scientific Assembly and Annual Meeting, November 25 - November 30, 2012 ,Chicago IL. http://archive.rsna.org/2012/12035633.html