Background. The applications of artificial intelligence (AI) processes have grown significantly in all medical disciplines during the last decades. Two main types of AI have been applied in Medicine: symbolic AI (e.g. knowledge base and ontologies) and non-symbolic AI (e.g. machine learning and artificial neural networks). Consequently, AI has also been applied across most Obstetrics and Gynecology domains (general obstetrics, gynecology surgery, fetal ultrasound, Assisted Reproductive Medicine…).
Objective. The objective of this study was to provide a systematic review to establish the actual contributions of AI reported Obstetrics and Gynecology discipline journals.
Methods. The PubMed database was searched for citations indexed with “artificial intelligence” and at least one of the following MeSH terms: “obstetrics”, “gynecology”, “reproductive techniques, assisted” or “pregnancy”, between 01/01/2000 and 04/30/2020. All publications in Obstetrics and Gynecology core disciplines journals were considered. The selection of journals was based on disciplines defined in Web of Science. The publications were excluded if no artificial intelligence process was used in the study. Review, editorial and commentary articles were also excluded. The study analysis comprises 1) the classification of publications into Obstetrics and Gynecology domains, 2) the description of AI methods, 3) the description of AI algorithms, 4) the description of datasets, 5) the description of AI contributions and 6) the description of the validation of the AI process.
Results. The PubMed search retrieved 579 citations and 66 publications met the selection criteria. All OB/GYN sub-domains were covered: Obstetrics (41%), Gynecology (3%), Assisted Reproductive Medicine (33%), Early pregnancy (2%) and Fetal Medicine (21%). Both Machine Learning methods (n=39/66) and Knowledge Base methods (n=25/66) were represented. Machine Learning used imaging, numerical and clinical datasets. Knowledge Base methods used mostly omics datasets. The actual contributions of AI were method/algorithm development (53%), hypothesis generation (42%) or software development (3%). Validation was performed on 1 dataset (87%) and no external validation was reported. We observed a general rising trend in publications related to AI in OB/GYN over the last two decades. Most of these publications (82%) remain out of the scope of the usual OB/GYN journals.
Conclusions. In Obstetrics and Gynecology discipline journals, mostly preliminary work (e.g. proof of concept algorithm or method) in AI applied to this discipline is reported and clinical validation remains an unmet prerequisite. Improvement driven by new AI research guidelines is expected. However, these guidelines are covering only a part of AI approaches (non-symbolic) reported in this review, hence updates need to be considered.


Author