Beyond attention: deriving biologically interpretable insights from
weakly-supervised multiple-instance learning models
- URL: http://arxiv.org/abs/2309.03925v1
- Date: Thu, 7 Sep 2023 09:44:35 GMT
- Title: Beyond attention: deriving biologically interpretable insights from
weakly-supervised multiple-instance learning models
- Authors: Willem Bonnaff\'e, CRUK ICGC Prostate Group, Freddie Hamdy, Yang Hu,
Ian Mills, Jens Rittscher, Clare Verrill, Dan J. Woodcock
- Abstract summary: We introduce prediction-attention-weighted (PAW) maps by combining tile-level attention and prediction scores produced by a refined encoder.
We also introduce a biological feature instantiation technique by integrating PAW maps with nuclei segmentation masks.
Our approach reveals that regions that are predictive of adverse prognosis do not tend to co-locate with the tumour regions.
- Score: 2.639541396835675
- License: http://creativecommons.org/licenses/by/4.0/
- Abstract: Recent advances in attention-based multiple instance learning (MIL) have
improved our insights into the tissue regions that models rely on to make
predictions in digital pathology. However, the interpretability of these
approaches is still limited. In particular, they do not report whether
high-attention regions are positively or negatively associated with the class
labels or how well these regions correspond to previously established clinical
and biological knowledge. We address this by introducing a post-training
methodology to analyse MIL models. Firstly, we introduce
prediction-attention-weighted (PAW) maps by combining tile-level attention and
prediction scores produced by a refined encoder, allowing us to quantify the
predictive contribution of high-attention regions. Secondly, we introduce a
biological feature instantiation technique by integrating PAW maps with nuclei
segmentation masks. This further improves interpretability by providing
biologically meaningful features related to the cellular organisation of the
tissue and facilitates comparisons with known clinical features. We illustrate
the utility of our approach by comparing PAW maps obtained for prostate cancer
diagnosis (i.e. samples containing malignant tissue, 381/516 tissue samples)
and prognosis (i.e. samples from patients with biochemical recurrence following
surgery, 98/663 tissue samples) in a cohort of patients from the international
cancer genome consortium (ICGC UK Prostate Group). Our approach reveals that
regions that are predictive of adverse prognosis do not tend to co-locate with
the tumour regions, indicating that non-cancer cells should also be studied
when evaluating prognosis.
Related papers
- Histopathologic Cancer Detection [0.0]
This work uses the PatchCamelyon benchmark datasets and trains them in a multi-layer perceptron and convolution model to observe the model's performance in terms of precision Recall, F1 Score, Accuracy, and AUC Score.
Also, this paper introduced ResNet50 and InceptionNet models with data augmentation, where ResNet50 is able to beat the state-of-the-art model.
arXiv Detail & Related papers (2023-11-13T19:51:46Z) - Tertiary Lymphoid Structures Generation through Graph-based Diffusion [54.37503714313661]
In this work, we leverage state-of-the-art graph-based diffusion models to generate biologically meaningful cell-graphs.
We show that the adopted graph diffusion model is able to accurately learn the distribution of cells in terms of their tertiary lymphoid structures (TLS) content.
arXiv Detail & Related papers (2023-10-10T14:37:17Z) - Active Learning Enhances Classification of Histopathology Whole Slide
Images with Attention-based Multiple Instance Learning [48.02011627390706]
We train an attention-based MIL and calculate a confidence metric for every image in the dataset to select the most uncertain WSIs for expert annotation.
With a novel attention guiding loss, this leads to an accuracy boost of the trained models with few regions annotated for each class.
It may in the future serve as an important contribution to train MIL models in the clinically relevant context of cancer classification in histopathology.
arXiv Detail & Related papers (2023-03-02T15:18:58Z) - Detecting Histologic & Clinical Glioblastoma Patterns of Prognostic
Relevance [6.281092892485014]
Glioblastoma is the most common and aggressive malignant adult tumor of the central nervous system.
Since adopting the current standard-of-care treatment 18 years ago, no substantial prognostic improvement has been noticed.
Here, we focus on identifying prognostically relevant characteristics from H&E stained WSI & clinical data relating to OS.
arXiv Detail & Related papers (2023-02-01T18:56:09Z) - Mixed Supervision of Histopathology Improves Prostate Cancer
Classification from MRI [0.9395521049323435]
Non-invasive prostate cancer detection from MRI has the potential to revolutionize patient care.
We present an MRI-based deep learning method for predicting clinically significant prostate cancer applicable to a patient population.
arXiv Detail & Related papers (2022-12-13T02:34:57Z) - A Pathologist-Informed Workflow for Classification of Prostate Glands in
Histopathology [62.997667081978825]
Pathologists diagnose and grade prostate cancer by examining tissue from needle biopsies on glass slides.
Cancer's severity and risk of metastasis are determined by the Gleason grade, a score based on the organization and morphology of prostate cancer glands.
This paper proposes an automated workflow that follows pathologists' textitmodus operandi, isolating and classifying multi-scale patches of individual glands.
arXiv Detail & Related papers (2022-09-27T14:08:19Z) - Weakly-supervised learning for image-based classification of primary
melanomas into genomic immune subgroups [1.4585861543119112]
We develop deep learning models to classify gigapixel H&E stained pathology slides into immune subgroups.
We leverage a multiple-instance learning approach, which only requires slide-level labels and uses an attention mechanism to highlight regions of high importance to the classification.
arXiv Detail & Related papers (2022-02-23T13:57:35Z) - Lung Cancer Lesion Detection in Histopathology Images Using Graph-Based
Sparse PCA Network [93.22587316229954]
We propose a graph-based sparse principal component analysis (GS-PCA) network, for automated detection of cancerous lesions on histological lung slides stained by hematoxylin and eosin (H&E)
We evaluate the performance of the proposed algorithm on H&E slides obtained from an SVM K-rasG12D lung cancer mouse model using precision/recall rates, F-score, Tanimoto coefficient, and area under the curve (AUC) of the receiver operator characteristic (ROC)
arXiv Detail & Related papers (2021-10-27T19:28:36Z) - Divide-and-Rule: Self-Supervised Learning for Survival Analysis in
Colorectal Cancer [9.431791041887957]
We propose a self-supervised learning method that learns a representation of tissue regions as well as a metric of the clustering to obtain their underlying patterns.
We show that the proposed approach can benefit from linear predictors to avoid overfitting in patient outcomes predictions.
arXiv Detail & Related papers (2020-07-07T09:15:36Z) - Weakly supervised multiple instance learning histopathological tumor
segmentation [51.085268272912415]
We propose a weakly supervised framework for whole slide imaging segmentation.
We exploit a multiple instance learning scheme for training models.
The proposed framework has been evaluated on multi-locations and multi-centric public data from The Cancer Genome Atlas and the PatchCamelyon dataset.
arXiv Detail & Related papers (2020-04-10T13:12:47Z) - Stan: Small tumor-aware network for breast ultrasound image segmentation [68.8204255655161]
We propose a novel deep learning architecture called Small Tumor-Aware Network (STAN) to improve the performance of segmenting tumors with different size.
The proposed approach outperformed the state-of-the-art approaches in segmenting small breast tumors.
arXiv Detail & Related papers (2020-02-03T22:25:01Z)
This list is automatically generated from the titles and abstracts of the papers in this site.
This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.