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

Silva-Rodríguez, Julio; Colomer, Adrián; Sales, María Á.; Molina, Rafael; Naranjo, Valery

doi:10.1016/j.cmpb.2020.105637

Cited by 95 publications

(87 citation statements)

References 35 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The gigapixel images were resampled to 10× resolution, and randomly clustered into three groups for training, validating and testing. Further, the external SICAP 2 database presented in [26] is used for testing. This dataset consists of 155 prostate WSIs with both global primary and secondary Gleason grades annotated by expert pathologists.…”

Section: Experiments and Resultsmentioning

confidence: 99%

“…The global Gleason score and Grade Group was predicted also using previously proposed methods based on the percentage of each cancerous grade in the tissue (GG%) using a kNN model as in [32] and a MLP used in [26]. It is noteworthy to mention that other learnable aggregation functions were tested to obtain the embedding of instance-level features instead of the proposed average pooling.…”

Section: Experiments and Resultsmentioning

confidence: 99%

“…Regarding previous state-of-the-art methods for patch-level Gleason grading based on supervised training on pixel-level annotations, our proposed teacher-student model using max-aggregation compared also favorable. In the supervised method in [26], which employs a CNN trained on annotated patches from SICAP database, a consistent drop in the κ metric was observed across the three datasets: SICAP test subset (κ = 0.77) to ARVANITI (κ = 0.64) and GERTYCH (κ = 0.51). In contrast, our weakly-supervised model obtained similar results in the three external datasets, with κ = 0.83 on SICAP, κ = 0.79 on ARVANITI and κ = 0.82 on GERTYCH.…”

Section: ) Grading Of Local Patternsmentioning

confidence: 85%

“…In contrast, our weakly-supervised model obtained similar results in the three external datasets, with κ = 0.83 on SICAP, κ = 0.79 on ARVANITI and κ = 0.82 on GERTYCH. These values demonstrate that the proposed weakly supervised pipeline outperforms the methodology presented in [26], showing a higher generalization ability and requiring a weaker supervision during training. This comparison extends to other prior works using supervised learning, which reach k values of 0.55/0.49 in [25] and 0.61 in [23] under different setups.…”

Section: ) Grading Of Local Patternsmentioning

confidence: 90%

See 3 more Smart Citations

Self-Learning for Weakly Supervised Gleason Grading of Local Patterns

Silva-Rodríguez

Colomer

Dolz

et al. 2021

IEEE J. Biomed. Health Inform.

Self Cite

View full text Add to dashboard Cite

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, computeraided-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 patchlevel 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.

show abstract

Section: Experiments and Resultsmentioning

confidence: 99%

Section: Experiments and Resultsmentioning

confidence: 99%

Section: ) Grading Of Local Patternsmentioning

confidence: 85%

Section: ) Grading Of Local Patternsmentioning

confidence: 90%

See 2 more Smart Citations

Self-Learning for Weakly Supervised Gleason Grading of Local Patterns

Silva-Rodríguez

Colomer

Dolz

et al. 2021

IEEE J. Biomed. Health Inform.

Self Cite

View full text Add to dashboard Cite

show abstract

“…Combined with Grad-CAM and UMAP embedding methods, we further provided an intuitive visualization of the local and global feature patterns of all EMB images learned by the VGG-19 model. Unlike other applications in cancer (24,(30)(31)(32), the implementation of this new model in myocardial injury reveals ill-defined histopathological patterns in local regions, providing a guideline and attention maps for welltrained pathologists. Therefore, integrating VGG-19 with Grad-CAM and UMAP embedding methods provides an interpretive DNN model for more accurate histopathological analyses.…”

Section: Discussionmentioning

confidence: 99%

Deep Neural Network-Aided Histopathological Analysis of Myocardial Injury

Jiao

Yuan

Sodimu

et al. 2022

Front. Cardiovasc. Med.

View full text Add to dashboard Cite

Deep neural networks have become the mainstream approach for analyzing and interpreting histology images. In this study, we established and validated an interpretable DNN model to assess endomyocardial biopsy (EMB) data of patients with myocardial injury. Deep learning models were used to extract features and classify EMB histopathological images of heart failure cases diagnosed with either ischemic cardiomyopathy or idiopathic dilated cardiomyopathy and non-failing cases (organ donors without a history of heart failure). We utilized the gradient-weighted class activation mapping (Grad-CAM) technique to emphasize injured regions, providing an entry point to assess the dominant morphology in the process of a comprehensive evaluation. To visualize clustered regions of interest (ROI), we utilized uniform manifold approximation and projection (UMAP) embedding for dimension reduction. We further implemented a multi-model ensemble mechanism to improve the quantitative metric (area under the receiver operating characteristic curve, AUC) to 0.985 and 0.992 on ROI-level and case-level, respectively, outperforming the achievement of 0.971 ± 0.017 and 0.981 ± 0.020 based on the sub-models. Collectively, this new methodology provides a robust and interpretive framework to explore local histopathological patterns, facilitating the automatic and high-throughput quantification of cardiac EMB analysis.

show abstract

Machine learning in computational histopathology: Challenges and opportunities

Cooper

Krishnan

2023

Genes Chromosomes & Cancer

View full text Add to dashboard Cite

Digital histopathological images, high-resolution images of stained tissue samples, are a vital tool for clinicians to diagnose and stage cancers. The visual analysis of patient state based on these images are an important part of oncology workflow. Although pathology workflows have historically been conducted in laboratories under a microscope, the increasing digitization of histopathological images has led to their analysis on computers in the clinic. The last decade has seen the emergence of machine learning, and deep learning in particular, a powerful set of tools for the analysis of histopathological images. Machine learning models trained on large datasets of digitized histopathology slides have resulted in automated models for prediction and stratification of patient risk. In this review, we provide context for the rise of such models in computational histopathology, highlight the clinical tasks they have found success in automating, discuss the various machine learning techniques that have been applied to this domain, and underscore open problems and opportunities.

show abstract

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

Cited by 95 publications

References 35 publications

Self-Learning for Weakly Supervised Gleason Grading of Local Patterns

Self-Learning for Weakly Supervised Gleason Grading of Local Patterns

Deep Neural Network-Aided Histopathological Analysis of Myocardial Injury

Machine learning in computational histopathology: Challenges and opportunities

Contact Info

Product

Resources

About