Uninformed Teacher-Student for hard-samples distillation in weakly supervised mitosis localization.

BACKGROUND AND OBJECTIVE: Mitotic activity is a crucial biomarker for diagnosing and predicting outcomes for different types of cancers, particularly breast cancer. However, manual mitosis counting is challenging and time-consuming for pathologists, with moderate reproducibility due to biopsy slide size, low mitotic cell density, and pattern heterogeneity. In recent years, deep learning methods based on convolutional neural networks (CNNs) have been proposed to address these limitations. Nonetheless, these methods have been hampered by the available data labels, which usually consist only of the centroids of mitosis, and by the incoming noise from annotated hard negatives. As a result, complex algorithms with multiple stages are often required to refine the labels at the pixel level and reduce the number of false positives. METHODS: This article presents a novel weakly supervised approach for mitosis detection that utilizes only image-level labels on histological hematoxylin and eosin (H&E) images, avoiding the need for complex labeling scenarios. Also, an Uninformed Teacher-Student (UTS) pipeline is introduced to detect and distill hard samples by comparing weakly supervised localizations and the annotated centroids, using strong augmentations to enhance uncertainty. Additionally, an automatic proliferation score is proposed that mimicks the pathologist-annotated mitotic activity index (MAI). The proposed approach is evaluated on three publicly available datasets for mitosis detection on breast histology samples, and two datasets for mitotic activity counting in whole-slide images. RESULTS: The proposed framework achieves competitive performance with relevant prior literature in all the datasets used for evaluation without explicitly using the mitosis location information during training. This approach challenges previous methods that rely on strong mitosis location information and multiple stages to refine false positives. Furthermore, the proposed pipeline for hard-sample distillation demonstrates promising dataset-specific improvements. Concretely, when the annotation has not been thoroughly refined by multiple pathologists, the UTS model offers improvements of up to ∼4% in mitosis localization, thanks to the detection and distillation of uncertain cases. Concerning the mitosis counting task, the proposed automatic proliferation score shows a moderate positive correlation with the MAI annotated by pathologists at the biopsy level on two external datasets. CONCLUSIONS: The proposed Uninformed Teacher-Student pipeline leverages strong augmentations to distill uncertain samples and measure dissimilarities between predicted and annotated mitosis. Results demonstrate the feasibility of the weakly supervised approach and highlight its potential as an objective evaluation tool for tumor proliferation.

Labeling confidence for uncertainty-aware histology image classification.

Deep learning-based models applied to digital pathology require large, curated datasets with high-quality (HQ) annotations to perform correctly. In many cases, recruiting expert pathologists to annotate large databases is not feasible, and it is necessary to collect additional labeled data with varying label qualities, e.g., pathologists-in-training (henceforth, non-expert annotators). Learning from datasets with noisy labels is more challenging in medical applications since medical imaging datasets tend to have instance-dependent noise and suffer from high inter/intra-observer variability. In this paper, we design an uncertainty-driven labeling strategy with which we generate soft labels from 10 non-expert annotators for multi-class skin cancer classification. Based on this soft annotation, we propose an uncertainty estimation-based framework to handle these noisy labels. This framework is based on a novel formulation using a dual-branch min-max entropy calibration to penalize inexact labels during the training. Comprehensive experiments demonstrate the promising performance of our labeling strategy. Results show a consistent improvement by using soft labels with standard cross-entropy loss during training (∼4.0% F1-score) and increases when calibrating the model with the proposed min-max entropy calibration (∼6.6% F1-score). These improvements are produced at negligible cost, both in terms of annotation and calculation.

Constrained unsupervised anomaly segmentation.

Current unsupervised anomaly localization approaches rely on generative models to learn the distribution of normal images, which is later used to identify potential anomalous regions derived from errors on the reconstructed images. To address the limitations of residual-based anomaly localization, very recent literature has focused on attention maps, by integrating supervision on them in the form of homogenization constraints. In this work, we propose a novel formulation that addresses the problem in a more principled manner, leveraging well-known knowledge in constrained optimization. In particular, the equality constraint on the attention maps in prior work is replaced by an inequality constraint, which allows more flexibility. In addition, to address the limitations of penalty-based functions we employ an extension of the popular log-barrier methods to handle the constraint. Last, we propose an alternative regularization term that maximizes the Shannon entropy of the attention maps, reducing the amount of hyperparameters of the proposed model. Comprehensive experiments on two publicly available datasets on brain lesion segmentation demonstrate that the proposed approach substantially outperforms relevant literature, establishing new state-of-the-art results for unsupervised lesion segmentation.

Proportion constrained weakly supervised histopathology image classification.

Multiple instance learning (MIL) deals with data grouped into bags of instances, of which only the global information is known. In recent years, this weakly supervised learning paradigm has become very popular in histological image analysis because it alleviates the burden of labeling all cancerous regions of large Whole Slide Images (WSIs) in detail. However, these methods require large datasets to perform properly, and many approaches only focus on simple binary classification. This often does not match the real-world problems where multi-label settings are frequent and possible constraints must be taken into account. In this work, we propose a novel multi-label MIL formulation based on inequality constraints that is able to incorporate prior knowledge about instance proportions. Our method has a theoretical foundation in optimization with log-barrier extensions, applied to bag-level class proportions. This encourages the model to respect the proportion ordering during training. Extensive experiments on a new public dataset of prostate cancer WSIs analysis, SICAP-MIL, demonstrate that using the prior proportion information we can achieve instance-level results similar to supervised methods on datasets of similar size. In comparison with prior MIL settings, our method allows for ∼13% improvements in instance-level accuracy, and ∼3% in the multi-label mean area under the ROC curve at the bag-level.

WeGleNet: A weakly-supervised convolutional neural network for the semantic segmentation of Gleason grades in prostate histology images.

BACKGROUND AND OBJECTIVE: Prostate cancer is one of the main diseases affecting men worldwide. The Gleason scoring system is the primary diagnostic tool for prostate cancer. This is obtained via the visual analysis of cancerous patterns in prostate biopsies performed by expert pathologists, and the aggregation of the main Gleason grades in a combined score. Computer-aided diagnosis systems allow to reduce the workload of pathologists and increase the objectivity. Nevertheless, those require a large number of labeled samples, with pixel-level annotations performed by expert pathologists, to be developed. Recently, efforts have been made in the literature to develop algorithms aiming the direct estimation of the global Gleason score at biopsy/core level with global labels. However, these algorithms do not cover the accurate localization of the Gleason patterns into the tissue. These location maps are the basis to provide a reliable computer-aided diagnosis system to the experts to be used in clinical practice by pathologists. In this work, we propose a deep-learning-based system able to detect local cancerous patterns in the prostate tissue using only the global-level Gleason score obtained from clinical records during training. METHODS: The methodological core of this work is the proposed weakly-supervised-trained convolutional neural network, WeGleNet, based on a multi-class segmentation layer after the feature extraction module, a global-aggregation, and the slicing of the background class for the model loss estimation during training. RESULTS: Using a public dataset of prostate tissue-micro arrays, we obtained a Cohen's quadratic kappa (κ) of 0.67 for the pixel-level prediction of cancerous patterns in the validation cohort. We compared the model performance for semantic segmentation of Gleason grades with supervised state-of-the-art architectures in the test cohort. We obtained a pixel-level κ of 0.61 and a macro-averaged f1-score of 0.58, at the same level as fully-supervised methods. Regarding the estimation of the core-level Gleason score, we obtained a κ of 0.76 and 0.67 between the model and two different pathologists. CONCLUSIONS: WeGleNet is capable of performing the semantic segmentation of Gleason grades similarly to fully-supervised methods without requiring pixel-level annotations. Moreover, the model reached a performance at the same level as inter-pathologist agreement for the global Gleason scoring of the cores.

Self-Learning for Weakly Supervised Gleason Grading of Local Patterns.

Prostate cancer is one of the main diseases affecting men worldwide. The gold standard for diagnosis and prognosis is the Gleason grading system. In this process, pathologists manually analyze prostate histology slides under microscope, in a high time-consuming and subjective task. In the last years, computer-aided-diagnosis (CAD) systems have emerged as a promising tool that could support pathologists in the daily clinical practice. Nevertheless, these systems are usually trained using tedious and prone-to-error pixel-level annotations of Gleason grades in the tissue. To alleviate the need of manual pixel-wise labeling, just a handful of works have been presented in the literature. Furthermore, despite the promising results achieved on global scoring the location of cancerous patterns in the tissue is only qualitatively addressed. These heatmaps of tumor regions, however, are crucial to the reliability of CAD systems as they provide explainability to the system's output and give confidence to pathologists that the model is focusing on medical relevant features. Motivated by this, we propose a novel weakly-supervised deep-learning model, based on self-learning CNNs, that leverages only the global Gleason score of gigapixel whole slide images during training to accurately perform both, grading of patch-level patterns and biopsy-level scoring. To evaluate the performance of the proposed method, we perform extensive experiments on three different external datasets for the patch-level Gleason grading, and on two different test sets for global Grade Group prediction. We empirically demonstrate that our approach outperforms its supervised counterpart on patch-level Gleason grading by a large margin, as well as state-of-the-art methods on global biopsy-level scoring. Particularly, the proposed model brings an average improvement on the Cohen's quadratic kappa ( κ) score of nearly 18% compared to full-supervision for the patch-level Gleason grading task. This suggests that the absence of the annotator's bias in our approach and the capability of using large weakly labeled datasets during training leads to higher performing and more robust models. Furthermore, raw features obtained from the patch-level classifier showed to generalize better than previous approaches in the literature to the subjective global biopsy-level scoring.

Going deeper through the Gleason scoring scale: An automatic end-to-end system for histology prostate grading and cribriform pattern detection.

BACKGROUND AND OBJECTIVE: Prostate cancer is one of the most common diseases affecting men worldwide. The Gleason scoring system is the primary diagnostic and prognostic tool for prostate cancer. Furthermore, recent reports indicate that the presence of patterns of the Gleason scale such as the cribriform pattern may also correlate with a worse prognosis compared to other patterns belonging to the Gleason grade 4. Current clinical guidelines have indicated the convenience of highlight its presence during the analysis of biopsies. All these requirements suppose a great workload for the pathologist during the analysis of each sample, which is based on the pathologist's visual analysis of the morphology and organisation of the glands in the tissue, a time-consuming and subjective task. In recent years, with the development of digitisation devices, the use of computer vision techniques for the analysis of biopsies has increased. However, to the best of the authors' knowledge, the development of algorithms to automatically detect individual cribriform patterns belonging to Gleason grade 4 has not yet been studied in the literature. The objective of the work presented in this paper is to develop a deep-learning-based system able to support pathologists in the daily analysis of prostate biopsies. This analysis must include the Gleason grading of local structures, the detection of cribriform patterns, and the Gleason scoring of the whole biopsy. METHODS: The methodological core of this work is a patch-wise predictive model based on convolutional neural networks able to determine the presence of cancerous patterns based on the Gleason grading system. In particular, we train from scratch a simple self-design architecture with three filters and a top model with global-max pooling. The cribriform pattern is detected by retraining the set of filters of the last convolutional layer in the network. Subsequently, a biopsy-level prediction map is reconstructed by bi-linear interpolation of the patch-level prediction of the Gleason grades. In addition, from the reconstructed prediction map, we compute the percentage of each Gleason grade in the tissue to feed a multi-layer perceptron which provides a biopsy-level score. RESULTS: In our SICAPv2 database, composed of 182 annotated whole slide images, we obtained a Cohen's quadratic kappa of 0.77 in the test set for the patch-level Gleason grading with the proposed architecture trained from scratch. Our results outperform previous ones reported in the literature. Furthermore, this model reaches the level of fine-tuned state-of-the-art architectures in a patient-based four groups cross validation. In the cribriform pattern detection task, we obtained an area under ROC curve of 0.82. Regarding the biopsy Gleason scoring, we achieved a quadratic Cohen's Kappa of 0.81 in the test subset. Shallow CNN architectures trained from scratch outperform current state-of-the-art methods for Gleason grades classification. Our proposed model is capable of characterising the different Gleason grades in prostate tissue by extracting low-level features through three basic blocks (i.e. convolutional layer + max pooling). The use of global-max pooling to reduce each activation map has shown to be a key factor for reducing complexity in the model and avoiding overfitting. Regarding the Gleason scoring of biopsies, a multi-layer perceptron has shown to better model the decision-making of pathologists than previous simpler models used in the literature.