Artificial selection

Contributions of artificial intelligence reported in obstetrics and gynecology journals: a systematic review

Background: Applications of artificial intelligence (AI) processes have grown significantly in all medical disciplines over the past decades. Two main types of AI have been applied in medicine: symbolic AI (eg, knowledge base and ontologies) and non-symbolic AI (eg, machine learning and artificial neural networks). Therefore, AI has also been applied in most areas of obstetrics and gynecology (OB/GYN), including general obstetrics, gynecological surgery, fetal ultrasound, and reproductive medicine. assisted, among others.

Goal: The purpose of this study was to provide a systematic review to establish the actual contributions of AI reported in journals of the OB/GYN discipline.

Methods : The PubMed database was searched for citations indexed with “artificial intelligence” and at least one of the following medical subject title (MeSH) terms between January 1, 2000 and April 30, 2020: “obstetrics”; “gynecology”; “assisted reproductive techniques”; or “pregnancy”. All publications in journals of the main OB/GYN disciplines were taken into account. The selection of journals was based on the disciplines defined in Web of Science. Publications were excluded if no AI process was used in the study. Review articles, editorials and commentaries were also excluded. The study analysis included (1) classification of publications in OB/GYN fields, (2) description of AI methods, (3) description of AI algorithms, (4) description of datasets, (5) description of AI contributions, and (6) description of AI process validation.

Results: The PubMed search retrieved 579 citations and 66 publications met the selection criteria. All OB/GYN subfields were covered: obstetrics (41%, 27/66), gynecology (3%, 2/66), assisted reproductive medicine (33%, 22/66), early pregnancy (2% , 1/66), and fetal medicine (21%, 14/66). Machine learning methods (39/66) and knowledge base methods (25/66) were represented. Machine learning used imaging, numerical and clinical datasets. The knowledge base methods mainly used omics datasets. Actual AI contributions were method/algorithm development (53%, 35/66), hypothesis generation (42%, 28/66) or software development (3%, 2/66). Validation was performed on one dataset (86%, 57/66) and no external validation was reported. We have observed a general upward trend in AI-related publications in OB/GYN over the past two decades. Most of these publications (82%, 54/66) remain outside the scope of the usual OB/GYN reviews.

Conclusion : In journals of the OB/GYN discipline, most of the preliminary work (e.g. algorithm or proof-of-concept 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 only cover part of the (non-symbolic) AI approaches reported in this review; therefore, updates should be considered.