P05.82 Clinical decision making. The route from an ontology to a prediction model for stereotactic radiotherapy: PRE.M.I.S.E. (PREdiction Models in Stereotactic External Radiotherapy). (19th September 2018)

• Record Type: Journal Article
• Authors: Bianciardi, F; Chiesa, S; Tolu, B; Longo, S; Minniti, G; Nardiello, B; Cellini, F; Rea, F; Gentile, P; Valentini, V
• Is Part Of: Neuro-oncology. Volume 20(2018)Supplement 3
• Page Start: iii322
• Page End: iii322
• Publication Date: 2018-09-19
• Subjects: Brain Neoplasms -- Periodicals; Brain -- Tumors -- Periodicals; Brain -- Cancer -- Periodicals; Nervous system -- Cancer -- Periodicals; 616.99481
• DOI: 10.1093/neuonc/noy139.408
• Languages: English
• ISSNs: 1522-8517

Abstract

Abstract: Background: It becomes more and more clear that people with the same cancer can have different forms of the disease, so responses to treatment and their toxicity can vary. Beside the widely accepted and old fashioned clinical guidelines, personalized medicine needs the DSS (Decision Support System) to be implemented in order to offer a valuable assistance in clinical decision making and in daily practice. DSS are represented by softwares, apps, nomograms all based on the creation of predictive models (PM) that require standardization of data collection. Ontology represents a fundamental tool to establish a common language without knowing, beforehand, which one of the described features could be the most relevant for a specific hypothesis. Through the creation of a ontology for a specific technique [Stereotactic Radiotherapy (SRT)], this multicentric study aims at collecting standardized data in order to build up a predictive model. We started focusing on brain metastases SRT. Material and Methods: The first two centres involved selected personnel in order to create a multidisciplinary team made of physicians, physicist, nurses and therapists. They meet twice a month. The first step of the project was to identify variables, validate them and built up the ontology. The second step was to implement a system that defines variables characteristics and the relationships among them. The semantic web technology was put into practice and the team is now collecting data on a dedicated software called BOA (Beyond Ontology Awareness) with its own WEB platform. The future development is to involve other RT centres to combine multiple datasets. Results: The drafted ontology is made up of more than 130 variables related with brain SRT, including both general "brain" nomenclature and SRT specific features. All the features were collected and organized in three levels ('registry', 'procedural' and 'research' level) that classify all the information in order to easily diversify queries. The BOA Web platform was shared between the first two centers involved and data collection process is ongoing by using Case Report Form (CRF). Conclusion: We drafted a brain SRT ontology in order to create a common language with the aim of enlarging our databases by involving different centres in collecting clinical data. The creation of a centralized multicentre large database is therefore essential for the development of ad hoc prediction tools. The aim is to implement a system that analyse large heterogeneous datasets to develop and validate multifactorial models that are able to assist clinical decision making based on specific pre-treatment characteristics in each individual patient.