ºÝºÝߣshows by User: StevenHart2 / http://www.slideshare.net/images/logo.gif ºÝºÝߣshows by User: StevenHart2 / Thu, 11 Oct 2018 12:53:56 GMT ºÝºÝߣShare feed for ºÝºÝߣshows by User: StevenHart2 Spitz vs conventional /slideshow/spitz-vs-conventional/119114542 spitzvsconventional-181011125356
Examination of hematoxylin and eosin staining (H&E) stained slides by light microscopy has been the cornerstone of histopathology for over a century.  During microscopic examination, a pathologist uses salient clinical information, pattern matching and feature recognition (shape, color, structure, etc.) to render a diagnosis.  Recently, whole-slide image (WSI) scanners have made it possible to fully digitize pathology slides.  In addition to enabling long term slide preservation and facilitating slide sharing for collaboration or second opinions, digitization of pathology slides allows for the development and utilization Artificial Intelligence (AI)-driven diagnostic tools. We conducted a pilot study to test the ability an AI convolutional neural network (CNN) to distinguish between two types of melanocytic lesions, Conventional and Spitz nevi.  We sought to determine the added value of pathologist-assisted training by comparing training effectiveness of complete slide analysis versus training on pathologist selected image patches. Images were classified by a deep CNN using Google’s TensorFlow framework.  We found significant improvement in classification accuracy when the model was trained from the pathologist-curated image set. These data provide strong evidence for the continued development of AI-driven diagnostic tools in digital pathology, and highlights the added value of domain experts when building AI workflows. Future directions of this work include expanding the number melanocytic lesions recognized by this tool, and enhancing its clinical performance through incorporation of molecular, demographic, and outcomes data. ]]>
Examination of hematoxylin and eosin staining (H&E) stained slides by light microscopy has been the cornerstone of histopathology for over a century.  During microscopic examination, a pathologist uses salient clinical information, pattern matching and feature recognition (shape, color, structure, etc.) to render a diagnosis.  Recently, whole-slide image (WSI) scanners have made it possible to fully digitize pathology slides.  In addition to enabling long term slide preservation and facilitating slide sharing for collaboration or second opinions, digitization of pathology slides allows for the development and utilization Artificial Intelligence (AI)-driven diagnostic tools. We conducted a pilot study to test the ability an AI convolutional neural network (CNN) to distinguish between two types of melanocytic lesions, Conventional and Spitz nevi.  We sought to determine the added value of pathologist-assisted training by comparing training effectiveness of complete slide analysis versus training on pathologist selected image patches. Images were classified by a deep CNN using Google’s TensorFlow framework.  We found significant improvement in classification accuracy when the model was trained from the pathologist-curated image set. These data provide strong evidence for the continued development of AI-driven diagnostic tools in digital pathology, and highlights the added value of domain experts when building AI workflows. Future directions of this work include expanding the number melanocytic lesions recognized by this tool, and enhancing its clinical performance through incorporation of molecular, demographic, and outcomes data. ]]> Thu, 11 Oct 2018 12:53:56 GMT /slideshow/spitz-vs-conventional/119114542 StevenHart2@slideshare.net(StevenHart2) Spitz vs conventional StevenHart2 Examination of hematoxylin and eosin staining (H&E) stained slides by light microscopy has been the cornerstone of histopathology for over a century.  During microscopic examination, a pathologist uses salient clinical information, pattern matching and feature recognition (shape, color, structure, etc.) to render a diagnosis.  Recently, whole-slide image (WSI) scanners have made it possible to fully digitize pathology slides.  In addition to enabling long term slide preservation and facilitating slide sharing for collaboration or second opinions, digitization of pathology slides allows for the development and utilization Artificial Intelligence (AI)-driven diagnostic tools. We conducted a pilot study to test the ability an AI convolutional neural network (CNN) to distinguish between two types of melanocytic lesions, Conventional and Spitz nevi.  We sought to determine the added value of pathologist-assisted training by comparing training effectiveness of complete slide analysis versus training on pathologist selected image patches. Images were classified by a deep CNN using Google’s TensorFlow framework.  We found significant improvement in classification accuracy when the model was trained from the pathologist-curated image set. These data provide strong evidence for the continued development of AI-driven diagnostic tools in digital pathology, and highlights the added value of domain experts when building AI workflows. Future directions of this work include expanding the number melanocytic lesions recognized by this tool, and enhancing its clinical performance through incorporation of molecular, demographic, and outcomes data. <img style="border:1px solid #C3E6D8;float:right;" alt="" src="https://cdn.slidesharecdn.com/ss_thumbnails/spitzvsconventional-181011125356-thumbnail.jpg?width=120&amp;height=120&amp;fit=bounds" /><br> Examination of hematoxylin and eosin staining (H&amp;E) stained slides by light microscopy has been the cornerstone of histopathology for over a century.  During microscopic examination, a pathologist uses salient clinical information, pattern matching and feature recognition (shape, color, structure, etc.) to render a diagnosis.  Recently, whole-slide image (WSI) scanners have made it possible to fully digitize pathology slides.  In addition to enabling long term slide preservation and facilitating slide sharing for collaboration or second opinions, digitization of pathology slides allows for the development and utilization Artificial Intelligence (AI)-driven diagnostic tools. We conducted a pilot study to test the ability an AI convolutional neural network (CNN) to distinguish between two types of melanocytic lesions, Conventional and Spitz nevi.  We sought to determine the added value of pathologist-assisted training by comparing training effectiveness of complete slide analysis versus training on pathologist selected image patches. Images were classified by a deep CNN using Google’s TensorFlow framework.  We found significant improvement in classification accuracy when the model was trained from the pathologist-curated image set. These data provide strong evidence for the continued development of AI-driven diagnostic tools in digital pathology, and highlights the added value of domain experts when building AI workflows. Future directions of this work include expanding the number melanocytic lesions recognized by this tool, and enhancing its clinical performance through incorporation of molecular, demographic, and outcomes data.
Spitz vs conventional from Steven Hart
]]> 86 1 https://cdn.slidesharecdn.com/ss_thumbnails/spitzvsconventional-181011125356-thumbnail.jpg?width=120&height=120&fit=bounds document Black http://activitystrea.ms/schema/1.0/post http://activitystrea.ms/schema/1.0/posted 0 https://cdn.slidesharecdn.com/profile-photo-StevenHart2-48x48.jpg?cb=1539262436