Lessons Learned from Designing an AI-Enabled Diagnosis Tool for Pathologists

Despite the promises of data-driven artificial intelligence (AI), little is known about how we can bridge the gulf between traditional physician-driven diagnosis and a plausible future of medicine automated by AI. Specifically, how can we involve AI usefully in physicians' diagnosis workflow given that most AI is still nascent and error-prone (e.g., in digital pathology)? To explore this question, we first propose a series of collaborative techniques to engage human pathologists with AI given AI's capabilities and limitations, based on which we prototype Impetus - a tool where an AI takes various degrees of initiatives to provide various forms of assistance to a pathologist in detecting tumors from histological slides. Finally, we summarize observations and lessons learned from a study with eight pathologists and discuss recommendations for future work on human-centered medical AI systems.

[1]  Zhi-Hua Zhou.  Multi-Instance Learning : A Survey , 2004 .

[2]  John Zimmerman,et al.  Investigating the Heart Pump Implant Decision Process: Opportunities for Decision Support Tools to Help , 2016, CHI.

[3]  Peter Bankhead,et al.  QuPath: Open source software for digital pathology image analysis , 2017, Scientific Reports.

[4]  M. S. Blois.  Clinical judgment and computers. , 1980, The New England journal of medicine.

[5]  Meyke Hermsen,et al.  1399 H&E-stained sentinel lymph node sections of breast cancer patients: the CAMELYON dataset , 2018, GigaScience.

[6]  Aniket Kittur,et al.  Apolo: making sense of large network data by combining rich user interaction and machine learning , 2011, CHI.

[7]  Samuel Leung,et al.  Diagnosis of Ovarian Carcinoma Cell Type is Highly Reproducible: A Transcanadian Study , 2010, The American journal of surgical pathology.

[8]  Anne E Carpenter,et al.  CellProfiler: image analysis software for identifying and quantifying cell phenotypes , 2006, Genome Biology.

[9]  J. C. R. Licklider,et al.  Man-Computer Symbiosis , 1960 .

[10]  Wei Liu,et al.  Scalable Histopathological Image Analysis via Active Learning , 2014, MICCAI.

[11]  Ullrich Köthe,et al.  Ilastik: Interactive learning and segmentation toolkit , 2011, 2011 IEEE International Symposium on Biomedical Imaging: From Nano to Macro.

[12]  Andreas Holzinger,et al.  Towards the Augmented Pathologist: Challenges of Explainable-AI in Digital Pathology , 2017, ArXiv.

[13]  Xiang 'Anthony' Chen,et al.  Forte: User-Driven Generative Design , 2018, CHI.

[14]  Trafton Drew,et al.  Informatics in radiology: what can you see in a single glance and how might this guide visual search in medical images? , 2013, Radiographics : a review publication of the Radiological Society of North America, Inc.

[15]  Sophia K Apple,et al.  Sentinel Lymph Node in Breast Cancer: Review Article from a Pathologist’s Point of View , 2016, Journal of pathology and translational medicine.

[16]  Geoffrey E. Hinton,et al.  Visualizing Data using t-SNE , 2008 .

[17]  Gilles Louppe,et al.  Collaborative analysis of multi-gigapixel imaging data using Cytomine , 2016, Bioinform..

[18]  David Maxwell Chickering,et al.  Machine Teaching: A New Paradigm for Building Machine Learning Systems , 2017, ArXiv.

[19]  Paul N. Bennett,et al.  Guidelines for Human-AI Interaction , 2019, CHI.

[20]  Xiang 'Anthony' Chen,et al.  CheXplain: Enabling Physicians to Explore and Understand Data-Driven, AI-Enabled Medical Imaging Analysis , 2020, CHI.

[21]  Martin Wattenberg,et al.  Human-Centered Tools for Coping with Imperfect Algorithms During Medical Decision-Making , 2019, CHI.

[22]  Yi Yang,et al.  Diagnose like a Radiologist: Attention Guided Convolutional Neural Network for Thorax Disease Classification , 2018, ArXiv.

[23]  Sanghoon Lee,et al.  Interactive phenotyping of large-scale histology imaging data with HistomicsML , 2017, bioRxiv.

[24]  Tony J Collins,et al.  ImageJ for microscopy. , 2007, BioTechniques.

[25]  Kevin W. Eliceiri,et al.  ImageJ2: ImageJ for the next generation of scientific image data , 2017, BMC Bioinformatics.

[26]  Sanghoon Lee,et al.  The Digital Slide Archive: A Software Platform for Management, Integration, and Analysis of Histology for Cancer Research. , 2017, Cancer research.

[27]  Yu Zhou,et al.  An Image Analysis Resource for Cancer Research: PIIP-Pathology Image Informatics Platform for Visualization, Analysis, and Management. , 2017, Cancer research.

[28]  Eric Horvitz,et al.  Principles of mixed-initiative user interfaces , 1999, CHI '99.

[29]  Kevin W Eliceiri,et al.  NIH Image to ImageJ: 25 years of image analysis , 2012, Nature Methods.

[30]  Sergey Ioffe,et al.  Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift , 2015, ICML.

[31]  R A Miller,et al.  The Demise of the “Greek Oracle” Model for Medical Diagnostic Systems , 1990, Methods of Information in Medicine.

[32]  Yifan Yu,et al.  CheXpert: A Large Chest Radiograph Dataset with Uncertainty Labels and Expert Comparison , 2019, AAAI.

[33]  Zhuowen Tu,et al.  Multiple clustered instance learning for histopathology cancer image classification, segmentation and clustering , 2012, 2012 IEEE Conference on Computer Vision and Pattern Recognition.

[34]  Yi Gao,et al.  A Containerized Software System for Generation, Management, and Exploration of Features from Whole Slide Tissue Images. , 2017, Cancer research.

[35]  John Zimmerman,et al.  Unremarkable AI: Fitting Intelligent Decision Support into Critical, Clinical Decision-Making Processes , 2019, CHI.

[36]  Qian Yang,et al.  Designing Theory-Driven User-Centric Explainable AI , 2019, CHI.

[37]  Ruiming Cao,et al.  Joint Prostate Cancer Detection and Gleason Score Prediction in mp-MRI via FocalNet , 2019, IEEE Transactions on Medical Imaging.

[38]  Xiaojin Zhu,et al.  The Label Complexity of Mixed-Initiative Classifier Training , 2016, ICML.

[39]  Xiang 'Anthony' Chen,et al.  Outlining the Design Space of Explainable Intelligent Systems for Medical Diagnosis , 2019, IUI Workshops.

[40]  Matthew Lease,et al.  Believe it or not: Designing a Human-AI Partnership for Mixed-Initiative Fact-Checking , 2018, UIST.

[41]  Lin Yang,et al.  MDNet: A Semantically and Visually Interpretable Medical Image Diagnosis Network , 2017, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[42]  Johannes E. Schindelin,et al.  Fiji: an open-source platform for biological-image analysis , 2012, Nature Methods.

[43]  Burr Settles,et al.  Active Learning Literature Survey , 2009 .

[44]  Guokun Lai,et al.  Explicit factor models for explainable recommendation based on phrase-level sentiment analysis , 2014, SIGIR.

[45]  Max Welling,et al.  Attention-based Deep Multiple Instance Learning , 2018, ICML.

[46]  Desney S. Tan,et al.  Overview based example selection in end user interactive concept learning , 2009, UIST '09.

[47]  Enrico Bertini,et al.  The Exploratory Labeling Assistant: Mixed-Initiative Label Curation with Large Document Collections , 2018, UIST.

[48]  Johannes Gehrke,et al.  Intelligible Models for HealthCare: Predicting Pneumonia Risk and Hospital 30-day Readmission , 2015, KDD.

[49]  Zhi-Hua Zhou,et al.  Isolation Forest , 2008, 2008 Eighth IEEE International Conference on Data Mining.

[50]  Thomas J. Fuchs,et al.  Clinical-grade computational pathology using weakly supervised deep learning on whole slide images , 2019, Nature Medicine.

[51]  Hans-Peter Kriegel,et al.  A Density-Based Algorithm for Discovering Clusters in Large Spatial Databases with Noise , 1996, KDD.