Digital pathology as a platform for primary diagnosis and augmentation via deep learning

Parwani, Anil V.

doi:10.1016/b978-0-323-67538-3.00006-3

Cited by 2 publications

(1 citation statement)

References 130 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Open source platforms such as HistoQC 13 , CellProfiler 3.0 17 and ImageJ 24,25 also leverage deep learning models for FQA. Moreover, there is advancement in artificial intelligence (AI) for medical diagnosis purposes in digital pathology [26][27][28][29][30][31][32][33][34] . Out-offocus regions in an image are a major contributor to systematic errors in these diagnoses 2,35,36 , highlighting the importance of reliable FQA methods to accompany these diagnosis tools.…”

Section: Introductionmentioning

confidence: 99%

Transferability of Deep Learning Models for Focus Quality Assessment in Digital Pathology

Miles

Hosseini

Tang

et al. 2021

Preprint

View full text Add to dashboard Cite

Out-of-focus sections of whole slide images are a significant source of false positives and other systematic errors in clinical diagnoses. As a result, focus quality assessment (FQA) methods must be able to quickly and accurately differentiate between focus levels in a scan. Recently, deep learning methods using convolutional neural networks (CNNs) have been adopted for FQA. However, the biggest obstacles impeding their wide usage in clinical workflows are their generalizability across different test conditions and their potentially high computational cost. In this study, we focus on the transferability and scalability of CNN-based FQA approaches. We carry out an investigation on ten architecturally diverse networks using five datasets with stain and tissue diversity. We evaluate the computational complexity of each network and scale this to realistic applications involving hundreds of whole slide images. We assess how well each full model transfers to a separate, unseen dataset without fine-tuning. We show that shallower networks transfer well when used on small input patch sizes, while deeper networks work more effectively on larger inputs. Furthermore, we introduce neural architecture search (NAS) to the field and learn an automatically designed low-complexity CNN architecture using differentiable architecture search which achieved competitive performance relative to established CNNs.

show abstract

Section: Introductionmentioning

confidence: 99%

Transferability of Deep Learning Models for Focus Quality Assessment in Digital Pathology

Miles

Hosseini

Tang

et al. 2021

Preprint

View full text Add to dashboard Cite

show abstract

Videoconference microscopy is a reliable alternative to conventional microscopy in the evaluation of Barrett’s esophagus: Zooming into a new era

Axley

Mitchell

Patel

et al. 2021

Diseases of the Esophagus

View full text Add to dashboard Cite

Telepathology, practicing pathology from a distance, allows experts to review cases without the need to transfer glass slides. Due to significant intra- and inter-observer variabilities in the histological evaluation of Barrett’s esophagus (BE), current guidelines recommend expert consultation in cases of dysplasia. We aimed to determine whether telepathology using microscope videoconferencing can be reliably used for evaluation of BE. Biopsies from 62 patients with endoscopic findings of salmon colored mucosa extending ≥1 cm proximal to the gastroesophageal junction were randomly selected to represent benign esophagus, non-dysplastic BE, low-grade dysplasia, high-grade dysplasia, and adenocarcinoma. Three gastrointestinal-trained pathologists reviewed the cases via videoconference microscopy followed by conventional microscopy. Intra-observer and pairwise inter-observer agreements between the conventional microscopy and videoconference methodologies were calculated for each of the three pathologists using Fleiss-Cohen weighted kappa (K) analysis. The intra-observer agreement for each pathologist’s assessment of videoconference microscopy and glass slide readings showed very good reliability (K = 0.94, 95% confidence interval = 0.89–0.99; 0.88, 95% confidence interval = 0.79–0.98; 0.93, 95% confidence interval = 0.90–0.97). Mean pairwise inter-observer agreement was 0.90 for videoconference and 0.91 for conventional microscopy. Diagnosis and grading of BE using videoconference microscopy show similar reliability as conventional microscopy. Based on our findings, we propose that videoconferencing pathology is a valid instrument for evaluating BE.

show abstract

Digital pathology as a platform for primary diagnosis and augmentation via deep learning

Cited by 2 publications

References 130 publications

Transferability of Deep Learning Models for Focus Quality Assessment in Digital Pathology

Transferability of Deep Learning Models for Focus Quality Assessment in Digital Pathology

Videoconference microscopy is a reliable alternative to conventional microscopy in the evaluation of Barrett’s esophagus: Zooming into a new era

Contact Info

Product

Resources

About