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

Abstract

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.

Machine learning in oncology—Perspectives in patient-reported outcome research

“…One of the main applications for machine learning methods is for the diagnosis and screening of cancer. Comprising both radiology and pathology, the majority of machine learning research in oncology has been conducted in this area [8,23]. Machine learning can be used to detect malignant tissue via imaging techniques and by training models to recognize cancer in images.…”

Machine Learning in der Onkologie – Perspektiven in der Patient-Reported-Outcome-Forschung

“…Die Anwendung von Machine Learning zur Diagnosestellung und zum onkologischen Screening wird klassischerweise vor allem in der Radiologie und Pathologie genutzt. Der Großteil der bisherigen Machine-Learning-Forschung in der Onkologie findet in diesem Bereich statt [8,21]. Machine Learning wird hier vor allem dazu genutzt, malignes Gewebe über bildgebende Verfahren zu erkennen.…”

The Impact of Digital Histopathology Batch Effect on Deep Learning Model Accuracy and Bias

“…The Cancer Genome Atlas (TCGA) has been critical for development of deep learning histology models, containing over 20,000 digital slide images from 24 tumor types, along with associated clinical, genomic, and radiomic data 18 . Due to the propensity of machine learning algorithms to overfit, performance is typically reported in a reserved testing set or evaluated with cross-validation, to avoid biased estimates of accuracy 19 . However, the propensity for overfitting of digital histology models to site level characteristics has been incompletely characterized and is infrequently accounted for in internal validation of deep learning models.…”

“…This is followed by a single hidden layer with width 500 and a softmax layer for prediction. Further details regarding implementation have been previously published 16,19 .…”