Background: Deep learning diagnostic algorithms are proving comparable results with human experts in a wide variety of tasks, and they still require a huge amount of well-annotated data for training, which is often non affordable. Metric learning techniques have allowed a reduction in the required annotated data allowing few-shot learning over deep learning architectures. Aims and Objectives: In this work, we analyze the state-of-the-art loss functions such as triplet loss, contrastive loss, and multi-class N-pair loss for the visual embedding extraction of hematoxylin and eosin (H&E) microscopy images and we propose a novel constellation loss function that takes advantage of the visual distances of the embeddings of the negative samples and thus, performing a regularization that increases the quality of the extracted embeddings. Materials and Methods: To this end, we employed the public H&E imaging dataset from the University Medical Center Mannheim (Germany) that contains tissue samples from low-grade and high-grade primary tumors of digitalized colorectal cancer tissue slides. These samples are divided into eight different textures (1. tumour epithelium, 2. simple stroma, 3. complex stroma, 4. immune cells, 5. debris and mucus, 6. mucosal glands, 7. adipose tissue and 8. background,). The dataset was divided randomly into train and test splits and the training split was used to train a classifier to distinguish among the different textures with just 20 training images. The process was repeated 10 times for each loss function. Performance was compared both for cluster compactness and for classification accuracy on separating the aforementioned textures. Results: Our results show that the proposed loss function outperforms the other methods by obtaining more compact clusters (Davis-Boulding: 1.41 ± 0.08, Silhouette: 0.37 ± 0.02) and better classification capabilities (accuracy: 85.0 ± 0.6) over H and E microscopy images. We demonstrate that the proposed constellation loss can be successfully used in the medical domain in situations of data scarcity.

Background: Prescreening of biopsies has the potential to improve pathologists' workflow. Tools that identify features and display results in a visually thoughtful manner can enhance efficiency, accuracy, and reproducibility. Machine learning for detection of glomeruli ensures comprehensive assessment and registration of four different stains allows for simultaneous navigation and viewing. Methods: Medical renal core biopsies (4 stains each) were digitized using a Leica SCN400 at ×40 and loaded into the Corista Quantum research platform. Glomeruli were manually annotated by pathologists. The tissue on the 4 stains was registered using a combination of keypoint- and intensity-based algorithms, and a 4-panel simultaneous viewing display was created. Using a training cohort, machine learning convolutional neural net (CNN) models were created to identify glomeruli in all stains, and merged into composite fields of views (FOVs). The sensitivity and specificity of glomerulus detection, and FOV area for each detection were calculated. Results: Forty-one biopsies were used for training (28) and same-batch evaluation (6). Seven additional biopsies from a temporally different batch were also evaluated. A variant of AlexNet CNN, used for object recognition, showed the best result for the detection of glomeruli with same-batch and different-batch evaluation: Same-batch sensitivity 92%, “modified” specificity 89%, average FOV size represented 0.8% of the total slide area; different-batch sensitivity 90%, “modified” specificity 98% and average FOV size 1.6% of the total slide area. Conclusions: Glomerulus detection in the best CNN model shows that machine learning algorithms may be accurate for this task. The added benefit of biopsy registration with simultaneous display and navigation allows reviewers to move from one machine-generated FOV to the next in all 4 stains. Together these features could increase both efficiency and accuracy in the review process.

A whole-slide imaging (WSI) system is a digital color imaging system used in digital pathology with the potential to substitute the conventional light microscope. A WSI system digitalizes a glass slide by converting the optical image to digital data with a scanner and then converting the digital data back to the optical image with a display. During the digital-to-optical or optical-to-digital conversion, a color space is required to define the mapping between the digital domain and the optical domain so that the numerical data of each color pixel can be interpreted meaningfully. Unfortunately, many current WSI products do not specify the designated color space clearly, which leaves the user using the universally default color space, sRGB. sRGB is a legacy color space that has a limited color gamut, which is known to be unable to reproduce all color shades present in histology slides. In this work, experiments were conducted to quantitatively investigate the limitation of the sRGB color space used in WSI systems. Eight hematoxylin and eosin (H and E)-stained tissue samples, including human bladder, brain, breast, colon, kidney, liver, lung, and uterus, were measured with a multispectral imaging system to obtain the true colors at the pixel level. The measured color truth of each pixel was converted into the standard CIELAB color space to test whether it was within the color gamut of the sRGB color space. Experiment results show that all the eight images have a portion of pixels outside the sRGB color gamut. In the worst-case scenario, the bladder sample, about 35% of the image exceeded the sRGB color gamut. The results suggest that the sRGB color space is inadequate for WSI scanners to encode H and E-stained whole-slide images, and an sRGB display may have insufficient color gamut for displaying H and E-stained histology images.

Background: Clinicopathological scores are used to predict the likelihood of recurrence-free survival for patients with clear cell renal cell carcinoma (ccRCC) after surgery. These are fallible, particularly in the middle range. This inevitably means that a significant proportion of ccRCC patients who will not develop recurrent disease enroll into clinical trials. As an exemplar of using digital pathology, we sought to improve the predictive power of “recurrence free” designation in localized ccRCC patients, by precise measurement of ccRCC nuclear morphological features using computational image analysis, thereby replacing manual nuclear grade assessment. Materials and Methods: TNM 8 UICC pathological stage pT1-pT3 ccRCC cases were recruited in Scotland and in Singapore. A Leibovich score (LS) was calculated. Definiens Tissue studio® (Definiens GmbH, Munich) image analysis platform was used to measure tumor nuclear morphological features in digitized hematoxylin and eosin (H&E) images. Results: Replacing human-defined nuclear grade with computer-defined mean perimeter generated a modified Leibovich algorithm, improved overall specificity 0.86 from 0.76 in the training cohort. The greatest increase in specificity was seen in LS 5 and 6, which went from 0 to 0.57 and 0.40, respectively. The modified Leibovich algorithm increased the specificity from 0.84 to 0.94 in the validation cohort. Conclusions: CcRCC nuclear mean perimeter, measured by computational image analysis, together with tumor stage and size, node status and necrosis improved the accuracy of predicting recurrence-free in the localized ccRCC patients. This finding was validated in an ethnically different Singaporean cohort, despite the different H and E staining protocol and scanner used. This may be a useful patient selection tool for recruitment to multicenter studies, preventing some patients from receiving unnecessary additional treatment while reducing the number of patients required to achieve adequate power within neoadjuvant and adjuvant clinical studies.