多光谱图像
边距(机器学习)
手术切缘
乳腺肿瘤
近红外光谱
校准
保乳手术
核医学
医学
乳腺癌
计算机科学
光学
地质学
遥感
外科
物理
癌症
乳房切除术
切除术
机器学习
内科学
量子力学
作者
Hao Wang,Shuyu Chen,Sisi Guo,Jing Gao,Xudong Zhou,Qian Chen,Houpu Yang,Fei Xie,Shu Wang,Kun Gao,Lihui Liu,Ruoyu Zhang,Jianfeng Wang
标识
DOI:10.1002/jbio.202400370
摘要
ABSTRACT A visible–near infrared (VIS–NIR) projected multispectral imaging (Proj‐MSI) system consisting of an MSI subsystem and a compact projector for intraoperative breast tumor margin assessment was developed. MSI with an average spectral resolution of 24 nm was realized through sequential illumination of 26 sets of VIS–NIR light‐emitting diodes and synchronized single NIR‐sensitive camera image capture. Rapid (~1 min) tumor margin delineation revealed significantly ( p < 0.01, Student's t ‐test) reduced reflectivity associated with breast tumor. Speeded‐up robust features algorithm‐based field of view calibration allowed the MSI identified tumor margins to be projected directly onto the breast‐conserving surgery (BCS) surgical cavity with a projection error of < 1 mm. Besides, the projected tumor margin boundaries were outlined using Indian ink to simulate breast tumor removal, demonstrating the potential of the Proj‐MSI system developed for intraoperative tumor margin assessment during BCS.
科研通智能强力驱动
Strongly Powered by AbleSci AI