ARA: accurate, reliable and active histopathological image classification framework with Bayesian deep learning

Abstract

Machine learning algorithms hold the promise to effectively automate the analysis of histopathological images that are routinely generated in clinical practice. Any machine learning method used in the clinical diagnostic process has to be extremely accurate and, ideally, provide a measure of uncertainty for its predictions. Such accurate and reliable classifiers need enough labelled data for training, which requires time-consuming and costly manual annotation by pathologists. Thus, it is critical to minimise the amount of data needed to reach the desired accuracy by maximising the efficiency of training. We propose an accurate, reliable and active (ARA) image classification framework and introduce a new Bayesian Convolutional Neural Network (ARA-CNN) for classifying histopathological images of colorectal cancer. The model achieves exceptional classification accuracy, outperforming other models trained on the same dataset. The network outputs an uncertainty measurement for each tested image. We show that uncertainty measures can be used to detect mislabelled training samples and can be employed in an efficient active learning workflow. Using a variational dropout-based entropy measure of uncertainty in the workflow speeds up the learning process by roughly 45%. Finally, we utilise our model to segment whole-slide images of colorectal tissue and compute segmentation-based spatial statistics.

Deep learning-based tumor microenvironment segmentation is predictive of tumor mutations and patient survival in non-small-cell lung cancer

“…This training effort can be minimised by utilising the active learning component of ARA, which shortens the number of iterations required to build an effective training dataset. For colorectal cancer, a pre-trained model is available from a previous study 37 .…”

Effective Melanoma Recognition Using Deep Convolutional Neural Network with Covariance Discriminant Loss

“…In addition, quantifying the uncertainty is another key factor for the computer-aided diagnosis methods. Based on the Bayesian theory, we can estimate the uncertainty of the predicted results [ 46 , 47 ]. The proposed approach is actually a data-driven method.…”

Semi-automated annotation of cell type-specific protein expression patterns in human testis based on immunohistochemistry

“…Previous studies conducted on histological images mainly focused on tissue type classi cation for disease detection 12 . Here, we present the rst approach in multi-label classi cation based on antibody-based proteomics, to recognize cell type-speci c protein expression patterns in eight different testicular cell types.…”

“…AI has been used for many different classi cation problems across various medical elds, especially in cancer and other diseases [6][7][8][9][10] . While an AI-driven approach holds much promise for e cient and accurate pattern recognition of histological images, few efforts were based on IHC images, and no previous study has used AI for distinguishing between cell type speci c protein expression patterns in human IHC samples [11][12] .…”