Semi-supervised training of deep convolutional neural networks with heterogeneous data and few local annotations: An experiment on prostate histopathology image classification

Marini, Niccolò; Otálora, Sebastian; Müller, Henning; Atzori, Manfredo

doi:10.1016/j.media.2021.102165

Cited by 40 publications

(21 citation statements)

References 25 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…We used the preprocessed MRI data as an input to 3D-CNN for MSFI embedding to improve the robustness of MSFI embedding. We expect that we can further improve the pCR prediction performance by applying semi-supervised training techniques for heterogeneous medical image data [51].…”

Section: Discussionmentioning

confidence: 99%

Pathologic Complete Response Prediction after Neoadjuvant Chemoradiation Therapy for Rectal Cancer Using Radiomics and Deep Embedding Network of MRI

et al. 2021

View full text Add to dashboard Cite

Assessment of magnetic resonance imaging (MRI) after neoadjuvant chemoradiation therapy (nCRT) is essential in rectal cancer staging and treatment planning. However, when predicting the pathologic complete response (pCR) after nCRT for rectal cancer, existing works either rely on simple quantitative evaluation based on radiomics features or partially analyze multi-parametric MRI. We propose an effective pCR prediction method based on novel multi-parametric MRI embedding. We first seek to extract volumetric features of tumors that can be found only by analyzing multiple MRI sequences jointly. Specifically, we encapsulate multiple MRI sequences into multi-sequence fusion images (MSFI) and generate MSFI embedding. We merge radiomics features, which capture important characteristics of tumors, with MSFI embedding to generate multi-parametric MRI embedding and then use it to predict pCR using a random forest classifier. Our extensive experiments demonstrate that using all given MRI sequences is the most effective regardless of the dimension reduction method. The proposed method outperformed any variants with different combinations of feature vectors and dimension reduction methods or different classification models. Comparative experiments demonstrate that it outperformed four competing baselines in terms of the AUC and F1-score. We use MRI sequences from 912 patients with rectal cancer, a much larger sample than in any existing work.

show abstract

Section: Discussionmentioning

confidence: 99%

Pathologic Complete Response Prediction after Neoadjuvant Chemoradiation Therapy for Rectal Cancer Using Radiomics and Deep Embedding Network of MRI

et al. 2021

View full text Add to dashboard Cite

show abstract

“…Although much progress has been achieved, the abovementioned patch-level prediction methods still have several inherent drawbacks. First, the patch-based methods required labor-sensitive patch-level annotation, which would increase the workload for the pathologist [ 118 ]. Second, most of the existing patch-based methods usually assumed that the diagnosis or survival information with each randomly selected patch was the same as its corresponding WSI, which neglected the fact that WSI usually had large heterogenous patterns and thus the patch-level label would not always match the WSI-level label [ 119 ].…”

Section: Cancer Diagnosis and Prognosismentioning

confidence: 99%

Recent Advances of Deep Learning for Computational Histopathology: Principles and Applications

Cheng

Huang

et al. 2022

Cancers

View full text Add to dashboard Cite

With the remarkable success of digital histopathology, we have witnessed a rapid expansion of the use of computational methods for the analysis of digital pathology and biopsy image patches. However, the unprecedented scale and heterogeneous patterns of histopathological images have presented critical computational bottlenecks requiring new computational histopathology tools. Recently, deep learning technology has been extremely successful in the field of computer vision, which has also boosted considerable interest in digital pathology applications. Deep learning and its extensions have opened several avenues to tackle many challenging histopathological image analysis problems including color normalization, image segmentation, and the diagnosis/prognosis of human cancers. In this paper, we provide a comprehensive up-to-date review of the deep learning methods for digital H&E-stained pathology image analysis. Specifically, we first describe recent literature that uses deep learning for color normalization, which is one essential research direction for H&E-stained histopathological image analysis. Followed by the discussion of color normalization, we review applications of the deep learning method for various H&E-stained image analysis tasks such as nuclei and tissue segmentation. We also summarize several key clinical studies that use deep learning for the diagnosis and prognosis of human cancers from H&E-stained histopathological images. Finally, online resources and open research problems on pathological image analysis are also provided in this review for the convenience of researchers who are interested in this exciting field.

show abstract

“…Methods have become more reliable, achieving in some cases a performance that is comparable to pathologists for specific segmentation and classification tasks [1, 5, 2, 15, 9, 12]. Despite the remarkable good performance of some methods, there are still technical barriers that prevent the translation of these advances into clinical applications [17].…”

Section: Motivation and Significancementioning

confidence: 99%

stainlib: a python library for augmentation and normalization of histopathology H&E images

Otálora

Marini

Podareanu

et al. 2022

Preprint

Self Cite

View full text Add to dashboard Cite

Computational pathology is a domain of increasing scientific and social interest. The automatic analysis of histopathology images stained with Hematoxylin and Eosin (H&E) can help clinicians diagnose and quantify diseases. Computer vision methods based on deep learning can perform on par or better than pathologists in specific tasks. Nevertheless, the visual heterogeneity in histopathology images due to batch effects, differences in preparation in different pathology laboratories, and the scanner can produce tissue appearance changes in the digitized whole-slide images. Such changes impede the application of the trained models in clinical scenarios where there is high variability in the images. We introduce stainlib, an easy-to-use and expandable python3 library that collects and unifies state-of-the-art methods for color augmentation and normalization of histopathology H&E images. stainlib also contains recent deep learning-based approaches that perform a robust stain-invariant training of CNN models. stainlib can help researchers build models robust to color domain shift by augmenting and harmonizing the training data, allowing the deployment of better models in the digital pathology practice.

show abstract

Semi-supervised training of deep convolutional neural networks with heterogeneous data and few local annotations: An experiment on prostate histopathology image classification

Cited by 40 publications

References 25 publications

Pathologic Complete Response Prediction after Neoadjuvant Chemoradiation Therapy for Rectal Cancer Using Radiomics and Deep Embedding Network of MRI

Pathologic Complete Response Prediction after Neoadjuvant Chemoradiation Therapy for Rectal Cancer Using Radiomics and Deep Embedding Network of MRI

Recent Advances of Deep Learning for Computational Histopathology: Principles and Applications

stainlib: a python library for augmentation and normalization of histopathology H&E images

Contact Info

Product

Resources

About