| Abstrakt: |
To compile evidence regarding the use of Machine Learning models to diagnose intraosseous lesions in gnathic bones, and to understand the impact and usefulness of these. The present SR was conducted following the guidelines of the PRISMA 2022 and registered in the PROSPERO database (CRD42022379298). The acronym PICOS was used to develop the focused review question "Should Computer vision be used to diagnose intraosseous lesions in radiographic images?" and the eligibility criteria. Electronic database search encompassed PubMed, Embase, Scopus, Cochrane Library, Web of Science, Lilacs, and IEEE Xplore, and Gray Literature (Google Scholar and ProQuest). Risk of Bias (RoB) was assessed through PROBAST, and the results were synthesized based on the task and according to the dataset sampling strategy. A total of 12 studies were included (21,146 radiographic images). Ameloblastomas, odontogenic keratocysts, dentigerous cysts, and periapical cysts were the most commonly investigated lesions. According to the TRIPOD, the majority of studies were classified as type 2 (randomly splited). Artificial intelligence application, safety, and generalization must be tested using more universal methodologies, since the diagnostic performance may differ according to the algorithms used. [ABSTRACT FROM AUTHOR] |
Full Text from ScienceDirect