Joint categorical and ordinal learning for cancer grading in pathology images

Vuong, Trinh Thi Le; Kim, Kyungeun; Song, Bong Keun; Kwak, Jin Tae

doi:10.1016/j.media.2021.102206

Cited by 22 publications

(5 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The feature extractor is followed by a classifier that predicts the legion grade of the slides. Since the four classes are ordinal instead of categorical, 41 as described in Equation (), squared error instead of cross entropy is used as the loss function, where

σ (x)

is the sigmoid function of

x

{boldh}_{i}

is the output vector for sample i , and

{\overset{y}{true}}_{i}

is the true grade of sample i . It is worth noting that

h \in {normalℝ}^{m \times 3}

, and that

y_{i} \in [0, 1, 2, 3], i = 1, 2, \dots, m

.…”

Section: Methodsmentioning

confidence: 99%

Registration‐enhanced multiple instance learning for cervical cancer whole slide image classification

He,

Wang,

Zeng

et al. 2023

Int J Imaging Syst Tech

View full text Add to dashboard Cite

Histopathological diagnosis is the golden standard for cancer diagnosis and grading. The past decade has witnessed great successes of vision‐based deep learning in computer‐aided cancer diagnosis. However, fully‐supervised methods require collecting large amounts of annotated data, which is a major challenge. In this research, we propose RMIL, a novel registration‐enhanced multiple instance learning pipeline, for cervical cancer biopsy grading of whole slide images as well as patch‐level regions. It requires only slide‐level annotations and a small amount of patch‐level annotations for training. To enhance it, an efficient registration method is implemented to generate more labeled patches with discriminative local features as data augmentations. In addition, attention mechanisms and self‐supervised learning are introduced to enhance feature extraction. Experiments showed that its classification accuracy is increased by 11.11%, compared to the vanilla multiple instance learning methods. The method has great potential for transferring to other scenarios.

show abstract

σ (x)

is the sigmoid function of

x

{boldh}_{i}

is the output vector for sample i , and

{\overset{y}{true}}_{i}

is the true grade of sample i . It is worth noting that

h \in {normalℝ}^{m \times 3}

, and that

y_{i} \in [0, 1, 2, 3], i = 1, 2, \dots, m

.…”

Section: Methodsmentioning

confidence: 99%

Registration‐enhanced multiple instance learning for cervical cancer whole slide image classification

He,

Wang,

Zeng

et al. 2023

Int J Imaging Syst Tech

View full text Add to dashboard Cite

show abstract

“…We employ two colorectal cancer datasets that are publicly available. 8 The first dataset includes 9,857 image patches of size 1024 × 1024 pixels (∼ 258 µm × 258 µm), digitized using an Aperio digital slide scanner (Leica Biosystems). These image patches are grouped into four categories, including 1,600 benign (BN), 2,322 welldifferentiated (WD) cancer, 4,105 moderately-differentiated (MD) cancer, and 1,830 poorly-differentiated (PD) cancer, and are divided into a training set of 7,027 images (T rain), a validation set of 1,242 images (V alidation), and a test set of 1,588 images (T estI).…”

Section: Datasetmentioning

confidence: 99%

Patch stitching data augmentation for cancer classification in pathology images

Wang,

Kim,

Kwak

2024

Medical Imaging 2024: Digital and Computational Pathology

View full text Add to dashboard Cite

Computational pathology, integrating computational methods and digital imaging, has shown to be effective in advancing disease diagnosis and prognosis. In recent years, the development of machine learning and deep learning has greatly bolstered the power of computational pathology. However, there still remains the issue of data scarcity and data imbalance, which can have an adversarial effect on any computational method. In this paper, we introduce an efficient and effective data augmentation strategy to generate new pathology images from the existing pathology images and thus enrich datasets without additional data collection or annotation costs. To evaluate the proposed method, we employed two sets of colorectal cancer datasets and obtained improved classification results, suggesting that the proposed simple approach holds the potential for alleviating the data scarcity and imbalance in computational pathology.

show abstract

“…To accurately distinguish the similar visual appearance between tumors with similar GS and consider the ordinal characteristics of prostate cancer aggressiveness, the multiple losses of prediction model is required. Therefore, to effectively learn the characteristics of prostate tumors according to their aggressiveness, we propose multiple losses consisting of triplet loss [14], mean squared error (MSE) loss, and cross-entropy ordinal (CEO) loss [15].…”

Section: Multiple Losses Of Prediction Modelmentioning

confidence: 99%

Deep regression model with ordinal and triplet loss for the prediction of prostate cancer aggressiveness in multi-parametric MR images

Kim¹,

Lee²,

Hong³

et al. 2023

Medical Imaging 2023: Computer-Aided Diagnosis

View full text Add to dashboard Cite

Assessment of prostate cancer aggressiveness is important because the effectiveness of treatment vary depending on the aggressiveness. The use of multi-parametric MR imaging prior to biopsy is recommended for accurate prostate cancer aggressiveness assessment but suffers from similar visual appearance of tumors between adjacent grades. To improve the predictive performance of prostate cancer aggressiveness, this study proposes a deep regression model involving sizenormalized patch generation and multiple losses. First, we generate two types of input patches such as tumor-centered patch and size-normalized patch to effectively learn the characteristics of small tumors. Second, we propose a multiple loss functions consisting of triplet loss, mean squared error, and cross-entropy ordinal loss to increase the ability to discriminate between tumors with similar visual appearance and different aggressiveness. As a result, the proposed model trained with the size-normalized ADC map showed the highest performance with an accuracy of 78.85%, specificity of 89.66%, and AUC of 0.77. The ensemble model of tumor-centered T2w image and size-normalized ADC map improved sensitivity by 8.69% and showed the best performance with accuracy of 78.85%.

show abstract

Joint categorical and ordinal learning for cancer grading in pathology images

Cited by 22 publications

References 31 publications

Registration‐enhanced multiple instance learning for cervical cancer whole slide image classification

Registration‐enhanced multiple instance learning for cervical cancer whole slide image classification

Patch stitching data augmentation for cancer classification in pathology images

Deep regression model with ordinal and triplet loss for the prediction of prostate cancer aggressiveness in multi-parametric MR images

Contact Info

Product

Resources

About