光学相干层析成像
模态(人机交互)
胶囊内镜
医学
放射科
内窥镜检查
癌症检测
生物医学工程
计算机科学
癌症
人工智能
内科学
作者
Qian Li,Zakiullah Ali,Christian Zakian,Massimiliano di Pietro,Judith Honing,Maria O’Donovan,Krzysztof Flisikowski,Vassilis Sarantos,Guillaume Saint Pierre,Jerome Gloriod,Wolfgang Drexler,Vasilis Ntziachristos
标识
DOI:10.1038/s41551-025-01462-0
摘要
Abstract Endoscopic detection of oesophageal cancer (EC) often occurs late in disease development, leading to high mortality rates. Improved technologies are urgently needed for earlier EC detection. Here we research an endoscopic ultra-broadband acoustic detection scheme and introduce a 360-degree hybrid optoacoustic and optical coherence endoscopy to enable interrogation of surface and subsurface precancerous and cancerous features at a three-dimensional micrometre scale. In the following pilot tissue investigation, the dual-modal imaging features are assessed for classifying different mucosal types in Barrett’s oesophagus (BE)—a precursor of EC. We find that human lesions of different grades, such as metaplastic, dysplastic and cancerous mucosa, exhibit distinctly different imaging features that are unique to the hybrid modality. Based on these features, a classification system is developed and evaluated for identifying BE neoplasia. The results show accurate BE neoplasia detection due to the complementarity of the two imaging modalities. Therefore, this study highlights the ability of the new dual-modality feature set to improve the detection performance of any of the two modalities operating in stand-alone mode and enhance diagnostic accuracy.
科研通智能强力驱动
Strongly Powered by AbleSci AI