Weakly Supervised Deep Learning for Whole Slide Lung Cancer Image Analysis

Wang, Xi; Chen, Hao; Gan, Caixia; Lin, Huangjing; Dou, Qi; Tsougenis, Efstratios; Huang, Qiuxia; Cai, Muyan; Heng, Pheng‐Ann

doi:10.1109/tcyb.2019.2935141

Cited by 319 publications

(187 citation statements)

References 38 publications

Supporting

Mentioning

186

Contrasting

Order By: Relevance

“…Automating histopathology workflows by DL For many years, digital pathology publications have described and iteratively refined basic image analysis tasks such as tumour detection, 28 tumour subtyping, 29 quantification of cell numbers 30 and classification of cell types. 31 What these approaches have in common is that the ground-truth method and the DL system use the same image data as input for their prediction.…”

Section: Basic Applications Of Dl: Tumour Detection Grading and Subtmentioning

confidence: 99%

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

et al. 2020

View full text Add to dashboard Cite

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

show abstract

Section: Basic Applications Of Dl: Tumour Detection Grading and Subtmentioning

confidence: 99%

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

et al. 2020

View full text Add to dashboard Cite

show abstract

“…In one such study, 928 WSIs from 330 patients were used in the testing phase and the developed algorithm accurately distinguished between benign breast proliferations and invasive breast cancer (AUC = 0.96) 24 . Most recently, Wang et al 25 2019 developed a weakly supervised deep learning system for the detection of lung cancer from WSIs. Their system demonstrated 97.3% accuracy 25 …”

Section: Discussionmentioning

confidence: 99%

The use of artificial intelligence, machine learning and deep learning in oncologic histopathology

Sultan

Elgharib

Tavares

et al. 2020

J Oral Pathology Medicine

View full text Add to dashboard Cite

The integration of artificial intelligence (AI), machine learning, deep learning, robotics, and Big Data into our physical daily lives has defined the Fourth Industrial Age. Clinicians are now provided with vast new opportunities to improve health care and optimize their approach to patient care. AI can be conceptualized into two main fields, namely artificial general intelligence (AGI) and artificial narrow intelligence (ANI). AGI is generally concerned with the design of machines that can mimic human intelligence with the goal of creating machines that can perform functions similar to humans or even functions that surpass what a human can do (superintelligence).

show abstract

“…noisy ones [6]- [10]. However, these methods can only take into account patterns present within individual patches, neglecting the potential relationships among them.…”

Section: The What and The Where Problemsmentioning

confidence: 99%