Background: The increased complexity of the decisional process in breast cancer surgery is well documented. With this study we aimed to create a software tool able to assist patients and surgeons in taking proper decisions.
Methodology: We hypothesized that the endpoints of breast cancer surgery could be addressed combining a set of decisional drivers. We created a decision support system software tool (DSS) and an interactive decision tree. A formal analysis estimated the information gain derived from each feature in the process. We tested the DSS on 52 patients and we analyzed the concordance of decisions obtained by different users and between the DSS suggestions and the actual surgery. We also tested the ability of the system to prevent post breast conservation deformities.
Results: The information gain revealed that patients preferences are the root of our decision tree. An observed concordance respectively of 0.98 and 0.88 was reported when the DSS was used twice by an expert operator or by a newly trained operator vs. an expert one. The observed concordance between the DSS suggestion and the actual decision was 0.69. A significantly higher incidence of post breast conservation defects was reported among patients who did not follow the DSS decision (Type III of Fitoussi, N = 4; 33.3%, p = 0.004).
Conclusion: The DSS decisions can be reproduced by operators with different experience. The concordance between suggestions and actual decision is quite low, however the DSS is able to prevent post- breast conservation deformities.
Keywords: Breast reconstructive surgery; Cosmetic outcome; Decision making in breast surgery; Oncoplastic surgery; Prevention of defects.
Copyright © 2016 Elsevier Ltd. All rights reserved.