Artificial intelligence in digital pathology — new tools for diagnosis and precision oncology

Nature Reviews Clinical Oncology, Journal Year: 2019, Volume and Issue: 16(11), P. 703 - 715

Published: Aug. 9, 2019

Language: Английский

---

Deep learning can predict microsatellite instability directly from histology in gastrointestinal cancer

Nature Medicine, Journal Year: 2019, Volume and Issue: 25(7), P. 1054 - 1056

Published: June 3, 2019

Language: Английский

---

Targeted drug delivery strategies for precision medicines

Nature Reviews Materials, Journal Year: 2021, Volume and Issue: 6(4), P. 351 - 370

Published: Feb. 2, 2021

Language: Английский

---

Deep learning in histopathology: the path to the clinic

Nature Medicine, Journal Year: 2021, Volume and Issue: 27(5), P. 775 - 784

Published: May 1, 2021

Language: Английский

---

Deep neural network models for computational histopathology: A survey

Medical Image Analysis, Journal Year: 2020, Volume and Issue: 67, P. 101813 - 101813

Published: Sept. 25, 2020

Language: Английский

---

Transfer learning for medical image classification: a literature review

BMC Medical Imaging, Journal Year: 2022, Volume and Issue: 22(1)

Published: April 13, 2022

Abstract Background Transfer learning (TL) with convolutional neural networks aims to improve performances on a new task by leveraging the knowledge of similar tasks learned in advance. It has made major contribution medical image analysis as it overcomes data scarcity problem well saves time and hardware resources. However, transfer been arbitrarily configured majority studies. This review paper attempts provide guidance for selecting model TL approaches classification task. Methods 425 peer-reviewed articles were retrieved from two databases, PubMed Web Science, published English, up until December 31, 2020. Articles assessed independent reviewers, aid third reviewer case discrepancies. We followed PRISMA guidelines selection 121 studies regarded eligible scope this review. investigated focused backbone models including feature extractor, extractor hybrid, fine-tuning scratch. Results The (n = 57) empirically evaluated multiple deep 33) shallow 24) models. Inception, one models, was most employed literature 26). With respect TL, 46) benchmarked identify optimal configuration. rest applied only single approach which 38) scratch 27) favored approaches. Only few hybrid 7) 3) pretrained Conclusion demonstrated efficacy despite scarcity. encourage scientists practitioners use (e.g. ResNet or Inception) extractors, can save computational costs without degrading predictive power.

Language: Английский

---

Deep learning in cancer pathology: a new generation of clinical biomarkers

British Journal of Cancer, Journal Year: 2020, Volume and Issue: 124(4), P. 686 - 696

Published: Nov. 17, 2020

Abstract Clinical workflows in oncology rely on predictive and prognostic molecular biomarkers. However, the growing number of these complex biomarkers tends to increase cost time for decision-making routine daily practice; furthermore, often require tumour tissue top diagnostic material. Nevertheless, routinely available contains an abundance clinically relevant information that is currently not fully exploited. Advances deep learning (DL), artificial intelligence (AI) technology, have enabled extraction previously hidden directly from histology images cancer, providing potentially useful information. Here, we outline emerging concepts how DL can extract summarise studies basic advanced image analysis cancer histology. Basic tasks include detection, grading subtyping images; they are aimed at automating pathology consequently do immediately translate into clinical decisions. Exceeding such approaches, has also been used tasks, which potential affecting processes. These approaches inference features, prediction survival end-to-end therapy response. Predictions made by systems could simplify enrich decision-making, but rigorous external validation settings.

Language: Английский

---

Molecular and histological correlations in liver cancer

Journal of Hepatology, Journal Year: 2019, Volume and Issue: 71(3), P. 616 - 630

Published: June 10, 2019

Language: Английский

---

A deep learning model to predict RNA-Seq expression of tumours from whole slide images

Nature Communications, Journal Year: 2020, Volume and Issue: 11(1)

Published: Aug. 3, 2020

Deep learning methods for digital pathology analysis are an effective way to address multiple clinical questions, from diagnosis prediction of treatment outcomes. These have also been used predict gene mutations images, but no comprehensive evaluation their potential extracting molecular features histology slides has yet performed. We show that HE2RNA, a model based on the integration data modes, can be trained systematically RNA-Seq profiles whole-slide images alone, without expert annotation. Through its interpretable design, HE2RNA provides virtual spatialization expression, as validated by CD3- and CD20-staining independent dataset. The transcriptomic representation learned transferred other datasets, even small size, increase performance specific phenotypes. illustrate use this approach in purposes such identification tumors with microsatellite instability.

Language: Английский

---

Deep Learning Models for Histopathological Classification of Gastric and Colonic Epithelial Tumours

Scientific Reports, Journal Year: 2020, Volume and Issue: 10(1)

Published: Jan. 30, 2020

Abstract Histopathological classification of gastric and colonic epithelial tumours is one the routine pathological diagnosis tasks for pathologists. Computational pathology techniques based on Artificial intelligence (AI) would be high benefit in easing ever increasing workloads pathologists, especially regions that have shortages access to services. In this study, we trained convolutional neural networks (CNNs) recurrent (RNNs) biopsy histopathology whole-slide images (WSIs) stomach colon. The models were classify WSI into adenocarcinoma, adenoma, non-neoplastic. We evaluated our three independent test sets each, achieving area under curves (AUCs) up 0.97 0.99 adenocarcinoma respectively, 0.96 adenoma respectively. results demonstrate generalisation ability promising potential deployment a practical histopathological diagnostic workflow system.

Language: Английский

---