Unleashing the potential of digital pathology data by training computer-aided diagnosis models without human annotations

Abstract

The digitalization of clinical workflows and the increasing performance of deep learning algorithms are paving the way towards new methods for tackling cancer diagnosis. However, the availability of medical specialists to annotate digitized images and free-text diagnostic reports does not scale with the need for large datasets required to train robust computer-aided diagnosis methods that can target the high variability of clinical cases and data produced. This work proposes and evaluates an approach to eliminate the need for manual annotations to train computer-aided diagnosis tools in digital pathology. The approach includes two components, to automatically extract semantically meaningful concepts from diagnostic reports and use them as weak labels to train convolutional neural networks (CNNs) for histopathology diagnosis. The approach is trained (through 10-fold cross-validation) on 3’769 clinical images and reports, provided by two hospitals and tested on over 11’000 images from private and publicly available datasets. The CNN, trained with automatically generated labels, is compared with the same architecture trained with manual labels. Results show that combining text analysis and end-to-end deep neural networks allows building computer-aided diagnosis tools that reach solid performance (micro-accuracy = 0.908 at image-level) based only on existing clinical data without the need for manual annotations.

Empowering digital pathology applications through explainable knowledge extraction tools

“…Between the 2 tasks, text classification has a prominent role as it provides weak annotations that can be used to reduce the high costs of training cancer assisted diagnosis tools – which prevent unleashing the full potential of digital pathology applications. 34 …”

“…The use of SKET for training deep image classifiers without human intervention paves the way to ML models in the clinical practice. 34 …”

“…Moreover, SKET labels can also be used to supervise cancer-assisted diagnosis tools. 34

Fig. 8 The SKET ecosystem.

…”

“…The labels produced by SKET are used to reduce supervised-training limitations for colon cancer-assisted diagnosis tools 34 – limitations that prevent the full exploitation of digital pathology applications. In other words, SKET labels serve as weak labels to train a deep image classifier.…”

“…Beyond explainability, we also report different digital pathology applications where SKET has been successfully integrated as a core system. 33 , 34 In particular, we deepen the use of SKET in such applications and the advantages it entails.…”