The hottest area in web search is image search and analysis. Needless to say, this will be a very important component of pathology informatics because of the evolving role of digital imaging in surgical pathology (see: "Eminence-Based" Surgical Pathology and the Digital Pathology Department; Image Analysis Using an iPhone Camera; Comparisons with Digital Pathology). Google scientists have achieved a breakthrough in image description using neural networks (see: Google’s Brain-Inspired Software Describes What It Sees in Complex Images). Below is an excerpt from the article describing the work:
Researchers at Google have created software that can use complete sentences to accurately describe scenes shown in photos....When shown a photo of a game of ultimate Frisbee, for example, the software responded with the description “A group of young people playing a game of frisbee.” The software can even count, giving answers such as “Two pizzas sitting on top of a stove top oven.”....Google’s researchers created the software through a kind of digital brain surgery, plugging together two neural networks developed separately for different tasks. One network had been trained to process images into a mathematical representation of their contents, in preparation for identifying objects. The other had been trained to generate full English sentences as part of automated translation software. When the networks are combined, the first can “look” at an image and then feed the mathematical description of what it “sees” into the second, which uses that information to generate a human-readable sentence. The combined network was trained to generate more accurate descriptions by showing it tens of thousands of images with descriptions written by humans....After that training process, the software was set loose on several large data sets of images from Flickr and other sources and asked to describe them. The accuracy of its descriptions was then judged with an automated test used to benchmark computer-vision software. Google’s software posted scores in the 60s on a 100-point scale. Humans doing the test typically score in 70s....That result suggests Google is far ahead of other researchers working to create scene-describing software.
When we get to the point of widely implemented digital pathology and image search, the surgical pathologist will never be redundant. Rather he or she will be empowered to operate more efficiently. We can draw an analogy between the current use of digital imaging in cytopathology and hematopathology. Here's a quote from an article about digital imaging in hematology (see: Digital imaging in hematology):
Digital imaging of Romanovsky-stained blood films can improve the work environment and alleviate some of the burdens associated with performing a WBC differential. With properly prepared and well-stained blood films, imaging technology can locate WBCs, capture images of the cells, perform a preliminary classification of cell type, and then display those images on a monitor. The advantages of this approach are immediately evident, since the technologist is freed from preparing slides and locating cells and can manipulate cells on the monitor to confirm or reclassify the identity of 100 (or more) WBCs. Metaphorically, the technologist becomes the editor, rather than having to author the book.
I think that we will see the following scenario evolving in the deployment of digital pathology in surgical pathology. Stained glass slides will undergo whole slide imaging and then a step in which the images are compared to those in an image archive seeking comparable cases. The surgical pathologist reading the case will then be directed to abnormal areas the slides with suggestions about what such cases have been diagnosed as previously. This is what my colleague Dr. Ulysses Balis has referred to as assisted, directed review. It may be the case that biopsies deemed totally normal by the search and analysis algorithms may be automatically signed out which is now the SOP for WBC differentials. However, any suggestion of abnormality in a slide will automatically require the attention of a pathologist.
Comments