Enhancing Breast Cancer Detection with FxMammo: Results from a Multi-Experience AI Evaluation
Date:
28 August 2025Category:
Innovation and TechnologyType:
Research Project ReportsTags:
Innovation, Technology, FathomX, AI, breast cancer, IndonesiaPrint Article:
This study evaluates the impact of FxMammo, a deep-learning decision-support tool developed by FathomX, on breast cancer detection in Indonesia. The tool was assessed through a single-centre, multi-reader, multi-case study conducted at Universitas Gadjah Mada, involving 500 mammograms (250 cancer cases and 250 benign or normal cases). Three radiologists and three senior residents interpreted each case twice – once unaided and once with AI support – in a blinded reading trial. The use of AI improved average accuracy across both reader groups. Sensitivity increased or was maintained for all readers without compromising specificity, and no cancers identified unaided were missed when using AI. Inter-reader agreement improved, indicating reduced interpretive variability. The most significant improvements were observed in dense breast cases and amongst less-experienced readers. These findings suggest that integrating FxMammo as a second reader can enhance early breast cancer detection, standardise diagnostic interpretations, and help address workforce limitations – provided implementation is supported by adequate training and continuous performance monitoring.
