Deep learning as a tool for increased accuracy and efficiency of histopathological diagnosis

Abstract

Pathologists face a substantial increase in workload and complexity of histopathologic cancer diagnosis due to the advent of personalized medicine. Therefore, diagnostic protocols have to focus equally on efficiency and accuracy. In this paper we introduce ‘deep learning’ as a technique to improve the objectivity and efficiency of histopathologic slide analysis. Through two examples, prostate cancer identification in biopsy specimens and breast cancer metastasis detection in sentinel lymph nodes, we show the potential of this new methodology to reduce the workload for pathologists, while at the same time increasing objectivity of diagnoses. We found that all slides containing prostate cancer and micro- and macro-metastases of breast cancer could be identified automatically while 30–40% of the slides containing benign and normal tissue could be excluded without the use of any additional immunohistochemical markers or human intervention. We conclude that ‘deep learning’ holds great promise to improve the efficacy of prostate cancer diagnosis and breast cancer staging.

A survey on deep learning in medical image analysis

“…Some papers combat this by adapting the loss function: Brosch et al (2016) defined it to be a weighted combination of the sensitivity and the specificity, with a larger weight for the specificity to make it less sensitive to the data imbalance. Others balance the data set by performing data augmentation on positive samples (Kamnitsas et al, 2017;Litjens et al, 2016;Pereira et al, 2016).…”

Leveraging uncertainty information from deep neural networks for disease detection

“…By stacking layers of linear convolutions with appropriate non-linearities 4 , abstract concepts can be learnt from high-dimensional input alleviating the challenging and time-consuming task of hand-crafting algorithms. Such DNNs are quickly entering the field of medical imaging and diagnosis [5][6][7][8][9][10][11][12][13][14][15] , outperforming state-of-the-art methods at disease detection or allowing one to tackle problems that had previously been out of reach. Applied at scale, such systems could considerably alleviate the workload of physicians by detecting patients at risk from a prescreening examination.…”

Avoiding Over-Detection: Towards Combined Object Detection and Counting

“…Deep neural networks generally require many thousands of labeled images to train effectively, but individual problems in biomedicine tend to avail neither thousands of images nor enough trained experts to label them all. Many proposed methods [1,11,14] circumvent this problem by using CNNs to perform pixel-wise binary classification. These networks take small image patches as input and output the probability of the central pixel in the patch being part of a target object.…”

“…FCNs are particularly useful when processing histological images due to their ability to naturally scale to images of arbitrary size, without needing to downsample large images to a fixed size. [11] train a standard CNN to classify the central pixel of image patches, then convert it to an FCN to perform pixel-wise classification over a whole image in one pass. This has performance benefits over processing patches one-by-one, since computations can be shared among overlapping image patches.…”