Abstract P3-17-04: Presentation of a Bayesian decision model for the treatment of ductal carcinoma in situ (DCIS) of the breast

• Published in: Cancer research (Chicago, Ill.) Vol. 77; no. 4_Supplement; pp. P3 - P3-17-04
• Main Authors: Belda, T, Giménez, J, García, JM, Tortajada, S, Vila, J, Estevan, R
• Format: Journal Article
• Language: English
• Published: 15.02.2017
• ISSN: 0008-5472; 1538-7445
• DOI: 10.1158/1538-7445.SABCS16-P3-17-04

Abstract INTRODUCTION: Breast carcinoma (BC) is the most common cancer in women worldwide with an increasing incidence by 2% annually, representing a major health problem in developed countries. The National Health Systems are focused on early diagnosis to minimize the fatal consequences of the disease, conducting screening mammography in women aged 40 to 70 years. However, early diagnosis has dramatically increased the diagnosis of ductal carcinoma in situ (DCIS). Medical treatment decisions are based on Clinical Practice Guidelines (CPG), which provide physicians general recommendations to help obtain the best optimal patient care. Many studies have reported improper use in routine clinical practice as they do not take into account the uncertainty of the decision or the risks and benefits involved. Thus, the increased frequency of patients diagnosed with DCIS secondary to breast screening programs and the deficits in the current making decision process based on GPC, are the main reasons we have taken into account to build a System Decision Support (SDS) for the treatment of patients with DCIS. MATERIAL AND METHODS: The research centers involved in this study were the Valencian Institute of Oncology (IVO) and the Polytechnic University of Valencia and within the Institute for the Application of Information Technology and Advanced Communications (ITACA). Researchers from both centers have collaborated for the creation and validation the model. RESULTS: A decision tree was developed based on the accepted treatment alternatives from the NCCN guidelines 2015 and the possibility of uncertain events resulting from treatment decisions. The usefulness was measured in recurrence-free survival. The optimal treatment decision was to administer Radiation therapy regardless of the status of estrogen receptors. For the decision to give hormone therapy, the optimal decision is to administer HT if the percentage of cells staining positive for estrogen receptor was high and not administer if the percentage was low. We found that the decisions obtained with the model are consistent with current recommendations and with accepted scientific evidence. We also made the computer adaptation of the model for easy use in routine clinical practice. We collected data from 266 patients treated at the IVO in the last 17 ​​years, with a mean follow-up of 75 months. The patients had an overall survival of 99.25 % with a recurrence -free survival of 86.8 % after a median follow up of 75 months. CONCLUSIONS: In this study we developed a decision treatment model for patients with DCIS based on Bayesian Decision Theory. The results were consistent with scientific evidence and clinical practice. Finally, we performed the computer implementation of the model in order to be used in routine clinical practice. DISCUSION: The model developed in the study is the first decision model that takes into account the uncertainty calculation based on the expected utility of each decision on recurrence-free survival for each patient. Citation Format: Belda T, Giménez J, García JM, Tortajada S, Vila J, Estevan R. Presentation of a Bayesian decision model for the treatment of ductal carcinoma in situ (DCIS) of the breast [abstract]. In: Proceedings of the 2016 San Antonio Breast Cancer Symposium; 2016 Dec 6-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2017;77(4 Suppl):Abstract nr P3-17-04.