Publicly available datasets of breast histopathology H&E whole-slide images: A scoping review

Advancements in digital pathology and computing resources have made a significant impact in the field of computational pathology for breast cancer diagnosis and treatment. However, access to high-quality labeled histopathological images of breast cancer is a big challenge that limits the development of accurate and robust deep learning models. In this scoping review, we identified the publicly available datasets of breast H&E-stained whole-slide images (WSIs) that can be used to develop deep learning algorithms. We systematically searched 9 scientific literature databases and 9 research data repositories and found 17 publicly available datasets containing 10 385 H&E WSIs of breast cancer. Moreover, we reported image metadata and characteristics for each dataset to assist researchers in selecting proper datasets for specific tasks in breast cancer computational pathology. In addition, we compiled 2 lists of breast H&E patches and private datasets as supplementary resources for researchers. Notably, only 28% of the included articles utilized multiple datasets, and only 14% used an external validation set, suggesting that the performance of other developed models may be susceptible to overestimation. The TCGA-BRCA was used in 52% of the selected studies. This dataset has a considerable selection bias that can impact the robustness and generalizability of the trained algorithms. There is also a lack of consistent metadata reporting of breast WSI datasets that can be an issue in developing accurate deep learning models, indicating the necessity of establishing explicit guidelines for documenting breast WSI dataset characteristics and metadata.

A Pragmatic Machine Learning Approach to Quantify Tumor-Infiltrating Lymphocytes in Whole Slide Images

Simple Summary

Tumor tissues sampled from patients contain prognostic and predictive information beyond what is currently being used in clinical practice. Large-scale digitization enables new ways of exploiting this information. The most promising analysis pipelines include deep learning/artificial intelligence (AI). However, to ensure success, AI often requires a time-consuming curation of data. In our approach, we repurposed AI pipelines and training data for cell segmentation and classification to identify tissue-infiltrating lymphocytes (TILs) in lung cancer tissue. We showed that our approach is able to identify TILs and provide prognostic information in an unseen dataset from lung cancer patients. Our methods can be adapted in myriad ways and may help pave the way for the large-scale deployment of digital pathology.

Abstract

Increased levels of tumor-infiltrating lymphocytes (TILs) indicate favorable outcomes in many types of cancer. The manual quantification of immune cells is inaccurate and time-consuming for pathologists. Our aim is to leverage a computational solution to automatically quantify TILs in standard diagnostic hematoxylin and eosin-stained sections (H&E slides) from lung cancer patients. Our approach is to transfer an open-source machine learning method for the segmentation and classification of nuclei in H&E slides trained on public data to TIL quantification without manual labeling of the data. Our results show that the resulting TIL quantification correlates to the patient prognosis and compares favorably to the current state-of-the-art method for immune cell detection in non-small cell lung cancer (current standard CD8 cells in DAB-stained TMAs HR 0.34, 95% CI 0.17–0.68 vs. TILs in HE WSIs: HoVer-Net PanNuke Aug Model HR 0.30, 95% CI 0.15–0.60 and HoVer-Net MoNuSAC Aug model HR 0.27, 95% CI 0.14–0.53). Our approach bridges the gap between machine learning research, translational clinical research and clinical implementation. However, further validation is warranted before implementation in a clinical setting.

Integrative, multi-omics, analysis of blood samples improves model predictions: applications to cancer

Background

Cancer genomic studies often include data collected from several omics platforms. Each omics data source contributes to the understanding of the underlying biological process via source specific (“individual”) patterns of variability. At the same time, statistical associations and potential interactions among the different data sources can reveal signals from common biological processes that might not be identified by single source analyses. These common patterns of variability are referred to as “shared” or “joint”. In this work, we show how the use of joint and individual components can lead to better predictive models, and to a deeper understanding of the biological process at hand. We identify joint and individual contributions of DNA methylation, miRNA and mRNA expression collected from blood samples in a lung cancer case–control study nested within the Norwegian Women and Cancer (NOWAC) cohort study, and we use such components to build prediction models for case–control and metastatic status. To assess the quality of predictions, we compare models based on simultaneous, integrative analysis of multi-source omics data to a standard non-integrative analysis of each single omics dataset, and to penalized regression models. Additionally, we apply the proposed approach to a breast cancer dataset from The Cancer Genome Atlas.

Results

Our results show how an integrative analysis that preserves both components of variation is more appropriate than standard multi-omics analyses that are not based on such a distinction. Both joint and individual components are shown to contribute to a better quality of model predictions, and facilitate the interpretation of the underlying biological processes in lung cancer development.

Conclusions

In the presence of multiple omics data sources, we recommend the use of data integration techniques that preserve the joint and individual components across the omics sources. We show how the inclusion of such components increases the quality of model predictions of clinical outcomes.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12859-021-04296-0.

Surgeon's experience and clinical outcome after retropubic tension-free vaginal tape-A case series

INTRODUCTION

The retropubic tension-free vaginal tape procedure has been the preferred method for primary surgical treatment of stress and stress-dominant mixed urinary incontinence in women for more than 20 years. In this study, we assessed associations between surgeon's experience with the primary tension-free vaginal tape procedure and both perioperative complications and recurrence rates.

MATERIAL AND METHODS

Using a consecutive case-series design, we assessed 596 patients treated with primary retropubic tension-free vaginal tape surgery performed by 18 surgeons from 1998 through 2012, with follow up through 2015 (maximum follow-up time: 10 years per patient). Data on perioperative complications and recurrence of stress urinary incontinence from medical records was transferred to a case report form. Surgeon's experience with the tension-free vaginal tape procedure was defined as number of such procedures performed as lead surgeon (1-19 ["beginners"], 20-49 and ≥50 procedures). All analyses were done with a 5% level of statistical significance. We applied the Chi-square test in the assessment of perioperative complications. The regression analyses of recurrence rate by number of tension-free vaginal tape procedures performed were restricted to the three surgeons who performed ≥50 procedures.

RESULTS

We found a significantly higher rate of bladder perforations (P = .03) and a higher rate of urinary retentions among patients whose tension-free vaginal tape procedures were performed by "beginners" (P = .06). We observed a significant reduction in recurrence rates with increasing number of tension-free vaginal tape procedures for one surgeon (P = .03).

CONCLUSIONS

Surgeon's experience with the tension-free vaginal tape procedure is associated with the risk of bladder perforation and urinary retention, and may be associated with the long-term effectiveness of the procedure.

Reproduction study using public data of: Development and validation of a deep learning algorithm for detection of diabetic retinopathy in retinal fundus photographs

We have attempted to reproduce the results in Development and validation of a deep learning algorithm for detection of diabetic retinopathy in retinal fundus photographs, published in JAMA 2016; 316(22), using publicly available data sets. We re-implemented the main method in the original study since the source code is not available. The original study used non-public fundus images from EyePACS and three hospitals in India for training. We used a different EyePACS data set from Kaggle. The original study used the benchmark data set Messidor-2 to evaluate the algorithm's performance. We used another distribution of the Messidor-2 data set, since the original data set is no longer available. In the original study, ophthalmologists re-graded all images for diabetic retinopathy, macular edema, and image gradability. We have one diabetic retinopathy grade per image for our data sets, and we assessed image gradability ourselves. We were not able to reproduce the original study's results with publicly available data. Our algorithm's area under the receiver operating characteristic curve (AUC) of 0.951 (95% CI, 0.947-0.956) on the Kaggle EyePACS test set and 0.853 (95% CI, 0.835-0.871) on Messidor-2 did not come close to the reported AUC of 0.99 on both test sets in the original study. This may be caused by the use of a single grade per image, or different data. This study shows the challenges of reproducing deep learning method results, and the need for more replication and reproduction studies to validate deep learning methods, especially for medical image analysis. Our source code and instructions are available at: https://github.com/mikevoets/jama16-retina-replication.

Comparison of computer systems and ranking criteria for automatic melanoma detection in dermoscopic images

Melanoma is the deadliest form of skin cancer, and early detection is crucial for patient survival. Computer systems can assist in melanoma detection, but are not widespread in clinical practice. In 2016, an open challenge in classification of dermoscopic images of skin lesions was announced. A training set of 900 images with corresponding class labels and semi-automatic/manual segmentation masks was released for the challenge. An independent test set of 379 images, of which 75 were of melanomas, was used to rank the participants. This article demonstrates the impact of ranking criteria, segmentation method and classifier, and highlights the clinical perspective. We compare five different measures for diagnostic accuracy by analysing the resulting ranking of the computer systems in the challenge. Choice of performance measure had great impact on the ranking. Systems that were ranked among the top three for one measure, dropped to the bottom half when changing performance measure. Nevus Doctor, a computer system previously developed by the authors, was used to participate in the challenge, and investigate the impact of segmentation and classifier. The diagnostic accuracy when using an automatic versus the semi-automatic/manual segmentation is investigated. The unexpected small impact of segmentation method suggests that improvements of the automatic segmentation method w.r.t. resemblance to semi-automatic/manual segmentation will not improve diagnostic accuracy substantially. A small set of similar classification algorithms are used to investigate the impact of classifier on the diagnostic accuracy. The variability in diagnostic accuracy for different classifier algorithms was larger than the variability for segmentation methods, and suggests a focus for future investigations. From a clinical perspective, the misclassification of a melanoma as benign has far greater cost than the misclassification of a benign lesion. For computer systems to have clinical impact, their performance should be ranked by a high-sensitivity measure.

Performance of a dermoscopy-based computer vision system for the diagnosis of pigmented skin lesions compared with visual evaluation by experienced dermatologists

BACKGROUND

It is often difficult to differentiate early melanomas from benign melanocytic nevi even by expert dermatologists, and the task is even more challenging for primary care physicians untrained in dermatology and dermoscopy. A computer system can provide an objective and quantitative evaluation of skin lesions, reducing subjectivity in the diagnosis.

OBJECTIVE

Our objective is to make a low-cost computer aided diagnostic tool applicable in primary care based on a consumer grade camera with attached dermatoscope, and compare its performance to that of experienced dermatologists.

METHODS AND MATERIALS

We propose several new image-derived features computed from automatically segmented dermoscopic pictures. These are related to the asymmetry, color, border, geometry, and texture of skin lesions. The diagnostic accuracy of the system is compared with that of three dermatologists.

RESULTS

With a data set of 206 skin lesions, 169 benign and 37 melanomas, the classifier was able to provide competitive sensitivity (86%) and specificity (52%) scores compared with the sensitivity (85%) and specificity (48%) of the most accurate dermatologist using only dermoscopic images.

CONCLUSION

We show that simple statistical classifiers can be trained to provide a recommendation on whether a pigmented skin lesion requires biopsy to exclude skin cancer with a performance that is comparable to and exceeds that of experienced dermatologists.

Unsupervised segmentation for digital dermoscopic images

BACKGROUND

Skin cancer is among the most common types of cancer. Melanoma is the most fatal of all skin cancer types. The only effective treatment is early excision. Recognising melanoma is challenging both for general physicians and for expert dermatologists. A computer-aided diagnostic system improving diagnostic accuracy would be of great importance. Segmenting the lesion from the skin is the first step in this process.

METHODS

The present segmentation algorithm uses a multiscale approach for density analysis. Only the skin mode is found by density analysis and then the location of the lesion mode is estimated. The density estimates are attained by Gaussian kernel smoothing with several bandwidths. A new algorithm for hair recognition based on morphological operations on binary images is incorporated into the segmentation algorithm.

RESULTS

The algorithm provides correct segmentation for both unimodal and multimodal densities. The segmentation is totally unsupervised, with a digital image as the only input. The algorithm has been tested on an independent set of images collected in dermatological practice, and the segmentation is verified by three dermatologists.

CONCLUSION

The present segmentation algorithm is fast and intuitive. It gives correct segmentation for most types of skin lesions, but fails when the lesion is brighter than the surrounding skin.