Cargando…

PathFlowAI: A High-Throughput Workflow for Preprocessing, Deep Learning and Interpretation in Digital Pathology

The diagnosis of disease often requires analysis of a biopsy. Many diagnoses depend not only on the presence of certain features but on their location within the tissue. Recently, a number of deep learning diagnostic aids have been developed to classify digitized biopsy slides. Clinical workflows of...

Descripción completa

Detalles Bibliográficos
Autores principales: Levy, Joshua J., Salas, Lucas A., Christensen, Brock C., Sriharan, Aravindhan, Vaickus, Louis J.
Formato: Online Artículo Texto
Lenguaje:English
Publicado: 2020
Materias:
Acceso en línea:https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6919317/
https://www.ncbi.nlm.nih.gov/pubmed/31797614
_version_ 1783480742304546816
author Levy, Joshua J.
Salas, Lucas A.
Christensen, Brock C.
Sriharan, Aravindhan
Vaickus, Louis J.
author_facet Levy, Joshua J.
Salas, Lucas A.
Christensen, Brock C.
Sriharan, Aravindhan
Vaickus, Louis J.
author_sort Levy, Joshua J.
collection PubMed
description The diagnosis of disease often requires analysis of a biopsy. Many diagnoses depend not only on the presence of certain features but on their location within the tissue. Recently, a number of deep learning diagnostic aids have been developed to classify digitized biopsy slides. Clinical workflows often involve processing of more than 500 slides per day. But, clinical use of deep learning diagnostic aids would require a preprocessing workflow that is cost-effective, flexible, scalable, rapid, interpretable, and transparent. Here, we present such a workflow, optimized using Dask and mixed precision training via APEX, capable of handling any patch-level or slide level classification and prediction problem. The workflow uses a flexible and fast preprocessing and deep learning analytics pipeline, incorporates model interpretation and has a highly storage-efficient audit trail. We demonstrate the utility of this package on the analysis of a prototypical anatomic pathology specimen, liver biopsies for evaluation of hepatitis from a prospective cohort. The preliminary data indicate that PathFlowAI may become a cost-effective and time-efficient tool for clinical use of Artificial Intelligence (AI) algorithms.
format Online
Article
Text
id pubmed-6919317
institution National Center for Biotechnology Information
language English
publishDate 2020
record_format MEDLINE/PubMed
spelling pubmed-69193172020-01-01 PathFlowAI: A High-Throughput Workflow for Preprocessing, Deep Learning and Interpretation in Digital Pathology Levy, Joshua J. Salas, Lucas A. Christensen, Brock C. Sriharan, Aravindhan Vaickus, Louis J. Pac Symp Biocomput Article The diagnosis of disease often requires analysis of a biopsy. Many diagnoses depend not only on the presence of certain features but on their location within the tissue. Recently, a number of deep learning diagnostic aids have been developed to classify digitized biopsy slides. Clinical workflows often involve processing of more than 500 slides per day. But, clinical use of deep learning diagnostic aids would require a preprocessing workflow that is cost-effective, flexible, scalable, rapid, interpretable, and transparent. Here, we present such a workflow, optimized using Dask and mixed precision training via APEX, capable of handling any patch-level or slide level classification and prediction problem. The workflow uses a flexible and fast preprocessing and deep learning analytics pipeline, incorporates model interpretation and has a highly storage-efficient audit trail. We demonstrate the utility of this package on the analysis of a prototypical anatomic pathology specimen, liver biopsies for evaluation of hepatitis from a prospective cohort. The preliminary data indicate that PathFlowAI may become a cost-effective and time-efficient tool for clinical use of Artificial Intelligence (AI) algorithms. 2020 /pmc/articles/PMC6919317/ /pubmed/31797614 Text en http://creativecommons.org/licenses/by-nc/4.0/ Open Access chapter published by World Scientific Publishing Company and distributed under the terms of the Creative Commons Attribution Non-Commercial (CC BY-NC) 4.0 License.
spellingShingle Article
Levy, Joshua J.
Salas, Lucas A.
Christensen, Brock C.
Sriharan, Aravindhan
Vaickus, Louis J.
PathFlowAI: A High-Throughput Workflow for Preprocessing, Deep Learning and Interpretation in Digital Pathology
title PathFlowAI: A High-Throughput Workflow for Preprocessing, Deep Learning and Interpretation in Digital Pathology
title_full PathFlowAI: A High-Throughput Workflow for Preprocessing, Deep Learning and Interpretation in Digital Pathology
title_fullStr PathFlowAI: A High-Throughput Workflow for Preprocessing, Deep Learning and Interpretation in Digital Pathology
title_full_unstemmed PathFlowAI: A High-Throughput Workflow for Preprocessing, Deep Learning and Interpretation in Digital Pathology
title_short PathFlowAI: A High-Throughput Workflow for Preprocessing, Deep Learning and Interpretation in Digital Pathology
title_sort pathflowai: a high-throughput workflow for preprocessing, deep learning and interpretation in digital pathology
topic Article
url https://www.ncbi.nlm.nih.gov/pmc/articles/PMC6919317/
https://www.ncbi.nlm.nih.gov/pubmed/31797614
work_keys_str_mv AT levyjoshuaj pathflowaiahighthroughputworkflowforpreprocessingdeeplearningandinterpretationindigitalpathology
AT salaslucasa pathflowaiahighthroughputworkflowforpreprocessingdeeplearningandinterpretationindigitalpathology
AT christensenbrockc pathflowaiahighthroughputworkflowforpreprocessingdeeplearningandinterpretationindigitalpathology
AT sriharanaravindhan pathflowaiahighthroughputworkflowforpreprocessingdeeplearningandinterpretationindigitalpathology
AT vaickuslouisj pathflowaiahighthroughputworkflowforpreprocessingdeeplearningandinterpretationindigitalpathology