Implementation of an ontological reasoning to support the guideline-based management of primary breast cancer patients in the DESIREE project

Abstract

The DESIREE project has developed a platform offering several complementary therapeutic decision support modules to improve the quality of care for breast cancer patients. All modules are operating consistently with a common breast cancer knowledge model (BCKM) following the generic entity-attribute-value model. The BCKM is formalized as an ontology including both the data model to represent clinical patient information and the termino-ontological model to represent the application domain concepts. This ontological model is used to describe data semantics and to allow for reasoning at different levels of abstraction. We present the guidelinebased decision support module (GL-DSS). Three breast cancer clinical practice guidelines have been formalized as decision rules including evidence levels, conformance levels, and two types of dependency, "refinement" and "complement", used to build complete care plans from the reconciliation of atomic recommendations. The system has been assessed on 138 decisions previously made without the system and re-played with the system after a washout period on simulated tumor boards (TBs) in three pilot sites. When TB clinicians changed their decision after using the GL-DSS, it was for a better decision than the decision made without the system in 75 % of the cases.(consolidated figures in 2018), breast cancer mortality is declining in France and breast cancer is one of the best prognosis cancers with among the best five and ten-year survival rates (87 % and 76 %, respectively). However, it still remains a therapeutic challenge especially for triple negative breast cancers and HER2+ breast cancers, for which improvements are both possible and necessary [1].Clinical practice guidelines (CPGs) are free-text documents developed by National agencies or academic associations to provide the best recommendations for the management of a set of selected patient profiles. These recommendations are built from published clinical research results and represent the state of the art following evidencebased medicine principles [2,3]. Although studies have shown that implementing oncology CPGs does improve clinical outcomes in both overall and recurrence-free survivals of cancer patients [4][5][6][7][8][9][10][11], there

Reusing Decisions Made with One Decision Support System to Assess a Second Decision Support System: Introducing the Notion of Complex Cases

“…The BCKM allows for rule-based and subsumption-based reasoning to provide best patientcentered guideline-based recommendations. The BCKM relies on a data model based on the generic Entity-Attribute-Value (EAV) model [2] (see Figure 1), the main entities being the patient, the breast side, and the lesion. CPGs used in the project, including the APHP CPGs considered as the update of CancerEst CPGs, are formalized as sets of Natural Rule Language (NRL) rules [6] matching the data model described in the BCKM.…”

“…DESIREE 2 is a recent European-funded project which aimed at developing a webbased platform to improve the management of primary breast cancer patients. Among other services, DESIREE includes a guideline-based decision support system (GL-DSS) that some authors of this article have developed [2]. As part of the DESIREE project, we had to evaluate the GL-DSS.…”

National study on use of artificial intelligence in breast disease and cancer

“…Breast cancer is the most common malignancy in women and the leading cause of cancer death. Now, with the help of AI, there is more optimism for a universal resolution to this important public health problem (4,5).…”

Artificial intelligence (AI) in breast cancer care - Leveraging multidisciplinary skills to improve care

“…On the methodological side, contributions on decision support systems come from the papers by Ref. [ 14 , 15 ]. Bouaud et al.…”

“…On the methodological side, contributions on decision support systems come from the papers by Ref. [14,15]. Bouaud Big data analysis refers to the analysis of huge quantities of data which are now collected and stored in any human activity, including healthcare.…”