首页 /研究 /Digital Pathology
SURGICAL

Digital Pathology

Alessandro Caputo, Antonio D’Antonio

发表年份
2021
引用次数
19

摘要

Adopting digital pathology has some very interesting implications, the most exciting of which will be briefly discussed in this article. FOR ROUTINE DIAGNOSIS The most obvious improvements are in routine diagnostic activity. Measurements of all kinds can be obtained on digital slides: some, like simple length measurements and mitotic count, are only faster and more accurate than with conventional microscopy; others, on the other hand, are so tedious to obtain with conventional microscopy that one could say that virtual microscopy enables the pathologist to obtain them. These include, for example, nonlinear lengths, surface area measurements, cell counts, average cell diameter, or even nucleolar size. The percentage of cells positive for a marker (e.g. Ki67) can be obtained with unprecedented speed and accuracy. If combined with artificial intelligence (AI), this system can even count only cells of interest, such as epithelial cells excluding stromal cells.[1,2] Consultations will not require shipment of glass slides or paraffin-embedded tissue blocks, but simply sending a link to the chosen colleague. Frozen sections can be handled in a conceptually similar way, but with a different system in keeping with their time-sensitive nature: robotic, remote-controlled microscopes exist, with a remote-streaming camera instead of oculars, so that the slide placed beneath the objective can be examined instantly by a pathologist even from a distance. The pathologist can move the stage, adjust the focus, and change objective from another hospital or from his home. After all, “it is often easier to move an image around than it is to move a patient or a pathologist.”[3] ?Numerous studies have assessed the non-inferiority of digital pathology versus conventional pathology in many aspects of routine practice, including for primary diagnosis of a wide variety of surgical pathology cases.[4,5] In the vast majority of cases, the same pathologist renders the same diagnosis using virtual microscopy and conventional microscopy.[4] Reading time is also similar, with a clear learning effect: pathologists that are initially slower with virtual microscopy tend to reach similar proficiency and comparable reading times with experience.[4] FOR LABORATORY TECHNICIANS In a properly designed laboratory, the racks of glass slides that come out of the stainer/cover slipper can be inserted directly in the slide scanner.[6] The digital slides will be delivered to the pathologists automatically and instantly once they have been scanned, and the rack coming out of the scanner can be archived right away. For the same reason, retrieval of an old case from the archive will not involve any friction, because each slide will be instantly accessible from the workstation. FOR TEACHING As with glass slides, interesting cases can be gathered, annotated, and used to assemble teaching files. The difference with digital teaching files is that they can benefit all interested people simultaneously, from all around the world. A notable example is the Juan Rosai Collection of Surgical Pathology Seminars, which includes over 18,000 cases.[7] Along the same lines, annotated digital slides can be used for quality assurance to ensure homogeneous reporting worldwide.[8] Teaching files for residents can contain verbose descriptions and annotations. FOR RESEARCH AI in pathology, another frontier, is also advancing quickly. One big drawback of AI is its requirement for large amounts of data to train algorithms. Hundreds if not thousands of human-annotated cases must be fed into an algorithm to train it before it can be of any use. This has been slowing down the adoption of AI because of the effort required to build such datasets. With digital slides, on the other hand, building a digital slide archive will be much easier and faster. Despite all its benefits, there are some issues that must be addressed to ease the adoption of digital pathology. Adopting a digital pathology workflow require

关键词

Digital pathologyVirtual microscopyComputer scienceFocus (optics)TelepathologyMicroscopyArtificial intelligenceComputer visionPathologyBiomedical engineering

相关论文

查看 SURGICAL 分类全部论文