Dense Steerable Filter CNNs for Exploiting Rotational Symmetry in Histology Images

Integrating SAM priors with U-Net for enhanced multiclass cell detection in digital pathology

In digital pathology, the accurate detection, segmentation, and classification of cells are pivotal for precise pathological diagnosis. Traditionally, pathologists manually segment cells from pathological images to facilitate diagnosis based on these results and other critical indicators. However, this manual approach is not only time-consuming but also prone to subjective biases, which significantly hampers its efficiency and consistency in large-scale applications. While classic segmentation networks like U-Net have gained widespread adoption in medical imaging, their integration with external prior features remains limited, thereby restricting the potential enhancement of segmentation accuracy. Although the large model SAM, renowned for its capability to "segment anything", has shown promise, its application in the specialized field of medical image processing presents considerable challenges. Direct application of SAM to medical scenarios often fails to yield optimal results. To overcome these limitations, this paper proposes a novel prior-guided joint attention mechanism. This method effectively integrates the prior features generated by SAM with the foundational features extracted by U-Net, achieving superior cell segmentation and classification. Extensive experiments on public datasets reveal that the proposed method significantly surpasses both standalone U-Net and approaches that merely augment inputs by overlaying prior features onto color channels. This advancement not only enhances the utility of large models in medical applications but also lays the groundwork for the evolution of intelligent pathological diagnostic technologies.

SE(3) group convolutional neural networks and a study on group convolutions and equivariance for DWI segmentation

We present an SE(3) Group Convolutional Neural Network along with a series of networks with different group actions for segmentation of Diffusion Weighted Imaging data. These networks gradually incorporate group actions that are natural for this type of data, in the form of convolutions that provide equivariant transformations of the data. This knowledge provides a potentially important inductive bias and may alleviate the need for data augmentation strategies. We study the effects of these actions on the performances of the networks by training and validating them using the diffusion data from the Human Connectome project. Unlike previous works that use Fourier-based convolutions, we implement direct convolutions, which are more lightweight. We show how incorporating more actions - using the SE(3) group actions - generally improves the performances of our segmentation while limiting the number of parameters that must be learned.

Computational pathology: A survey review and the way forward

Computational Pathology (CPath) is an interdisciplinary science that augments developments of computational approaches to analyze and model medical histopathology images. The main objective for CPath is to develop infrastructure and workflows of digital diagnostics as an assistive CAD system for clinical pathology, facilitating transformational changes in the diagnosis and treatment of cancer that are mainly address by CPath tools. With evergrowing developments in deep learning and computer vision algorithms, and the ease of the data flow from digital pathology, currently CPath is witnessing a paradigm shift. Despite the sheer volume of engineering and scientific works being introduced for cancer image analysis, there is still a considerable gap of adopting and integrating these algorithms in clinical practice. This raises a significant question regarding the direction and trends that are undertaken in CPath. In this article we provide a comprehensive review of more than 800 papers to address the challenges faced in problem design all-the-way to the application and implementation viewpoints. We have catalogued each paper into a model-card by examining the key works and challenges faced to layout the current landscape in CPath. We hope this helps the community to locate relevant works and facilitate understanding of the field's future directions. In a nutshell, we oversee the CPath developments in cycle of stages which are required to be cohesively linked together to address the challenges associated with such multidisciplinary science. We overview this cycle from different perspectives of data-centric, model-centric, and application-centric problems. We finally sketch remaining challenges and provide directions for future technical developments and clinical integration of CPath. For updated information on this survey review paper and accessing to the original model cards repository, please refer to GitHub. Updated version of this draft can also be found from arXiv.

Nuclei Detection and Segmentation of Histopathological Images Using a Feature Pyramidal Network Variant of a Mask R-CNN

Cell nuclei interpretation is crucial in pathological diagnostics, especially in tumor specimens. A critical step in computational pathology is to detect and analyze individual nuclear properties using segmentation algorithms. Conventionally, a semantic segmentation network is used, where individual nuclear properties are derived after post-processing a segmentation mask. In this study, we focus on showing that an object-detection-based instance segmentation network, the Mask R-CNN, after integrating it with a Feature Pyramidal Network (FPN), gives mature and reliable results for nuclei detection without the need for additional post-processing. The results were analyzed using the Kumar dataset, a public dataset with over 20,000 nuclei annotations from various organs. The dice score of the baseline Mask R-CNN improved from 76% to 83% after integration with an FPN. This was comparable with the 82.6% dice score achieved by modern semantic-segmentation-based networks. Thus, evidence is provided that an end-to-end trainable detection-based instance segmentation algorithm with minimal post-processing steps can reliably be used for the detection and analysis of individual nuclear properties. This represents a relevant task for research and diagnosis in digital pathology, which can improve the automated analysis of histopathological images.

Contrastive and uncertainty-aware nuclei segmentation and classification

Nuclei segmentation and classification play a crucial role in pathology diagnosis, enabling pathologists to analyze cellular characteristics accurately. Overlapping cluster nuclei, misdetection of small-scale nuclei, and pleomorphic nuclei-induced misclassification have always been major challenges in the nuclei segmentation and classification tasks. To this end, we introduce an auxiliary task of nuclei boundary-guided contrastive learning to enhance the representativeness and discriminative power of visual features, particularly for addressing the challenge posed by the unclear contours of adherent nuclei and small nuclei. In addition, misclassifications resulting from pleomorphic nuclei often exhibit low classification confidence, indicating a high level of uncertainty. To mitigate misclassification, we capitalize on the characteristic clustering of similar cells to propose a locality-aware class embedding module, offering a regional perspective to capture category information. Moreover, we address uncertain classification in densely aggregated nuclei by designing a top-k uncertainty attention module that leverages deep features to enhance shallow features, thereby improving the learning of contextual semantic information. We demonstrate that the proposed network outperforms the off-the-shelf methods in both nuclei segmentation and classification experiments, achieving the state-of-the-art performance.

ViT-PSO-SVM: Cervical Cancer Predication Based on Integrating Vision Transformer with Particle Swarm Optimization and Support Vector Machine

Cervical cancer (CCa) is the fourth most prevalent and common cancer affecting women worldwide, with increasing incidence and mortality rates. Hence, early detection of CCa plays a crucial role in improving outcomes. Non-invasive imaging procedures with good diagnostic performance are desirable and have the potential to lessen the degree of intervention associated with the gold standard, biopsy. Recently, artificial intelligence-based diagnostic models such as Vision Transformers (ViT) have shown promising performance in image classification tasks, rivaling or surpassing traditional convolutional neural networks (CNNs). This paper studies the effect of applying a ViT to predict CCa using different image benchmark datasets. A newly developed approach (ViT-PSO-SVM) was presented for boosting the results of the ViT based on integrating the ViT with particle swarm optimization (PSO), and support vector machine (SVM). First, the proposed framework extracts features from the Vision Transformer. Then, PSO is used to reduce the complexity of extracted features and optimize feature representation. Finally, a softmax classification layer is replaced with an SVM classification model to precisely predict CCa. The models are evaluated using two benchmark cervical cell image datasets, namely SipakMed and Herlev, with different classification scenarios: two, three, and five classes. The proposed approach achieved 99.112% accuracy and 99.113% F1-score for SipakMed with two classes and achieved 97.778% accuracy and 97.805% F1-score for Herlev with two classes outperforming other Vision Transformers, CNN models, and pre-trained models. Finally, GradCAM is used as an explainable artificial intelligence (XAI) tool to visualize and understand the regions of a given image that are important for a model's prediction. The obtained experimental results demonstrate the feasibility and efficacy of the developed ViT-PSO-SVM approach and hold the promise of providing a robust, reliable, accurate, and non-invasive diagnostic tool that will lead to improved healthcare outcomes worldwide.

Latent Diffusion Models with Image-Derived Annotations for Enhanced AI-Assisted Cancer Diagnosis in Histopathology

Artificial Intelligence (AI)-based image analysis has immense potential to support diagnostic histopathology, including cancer diagnostics. However, developing supervised AI methods requires large-scale annotated datasets. A potentially powerful solution is to augment training data with synthetic data. Latent diffusion models, which can generate high-quality, diverse synthetic images, are promising. However, the most common implementations rely on detailed textual descriptions, which are not generally available in this domain. This work proposes a method that constructs structured textual prompts from automatically extracted image features. We experiment with the PCam dataset, composed of tissue patches only loosely annotated as healthy or cancerous. We show that including image-derived features in the prompt, as opposed to only healthy and cancerous labels, improves the Fréchet Inception Distance (FID) by 88.6. We also show that pathologists find it challenging to detect synthetic images, with a median sensitivity/specificity of 0.55/0.55. Finally, we show that synthetic data effectively train AI models.

Built to Last? Reproducibility and Reusability of Deep Learning Algorithms in Computational Pathology

Recent progress in computational pathology has been driven by deep learning. While code and data availability are essential to reproduce findings from preceding publications, ensuring a deep learning model's reusability is more challenging. For that, the codebase should be well-documented and easy to integrate into existing workflows and models should be robust toward noise and generalizable toward data from different sources. Strikingly, only a few computational pathology algorithms have been reused by other researchers so far, let alone employed in a clinical setting. To assess the current state of reproducibility and reusability of computational pathology algorithms, we evaluated peer-reviewed articles available in PubMed, published between January 2019 and March 2021, in 5 use cases: stain normalization; tissue type segmentation; evaluation of cell-level features; genetic alteration prediction; and inference of grading, staging, and prognostic information. We compiled criteria for data and code availability and statistical result analysis and assessed them in 160 publications. We found that only one-quarter (41 of 160 publications) made code publicly available. Among these 41 studies, three-quarters (30 of 41) analyzed their results statistically, half of them (20 of 41) released their trained model weights, and approximately a third (16 of 41) used an independent cohort for evaluation. Our review is intended for both pathologists interested in deep learning and researchers applying algorithms to computational pathology challenges. We provide a detailed overview of publications with published code in the field, list reusable data handling tools, and provide criteria for reproducibility and reusability.

Dense Dilated Multi-Scale Supervised Attention-Guided Network for histopathology image segmentation

Over the last couple of decades, the introduction and proliferation of whole-slide scanners led to increasing interest in the research of digital pathology. Although manual analysis of histopathological images is still the gold standard, the process is often tedious and time consuming. Furthermore, manual analysis also suffers from intra- and interobserver variability. Separating structures or grading morphological changes can be difficult due to architectural variability of these images. Deep learning techniques have shown great potential in histopathology image segmentation that drastically reduces the time needed for downstream tasks of analysis and providing accurate diagnosis. However, few algorithms have clinical implementations. In this paper, we propose a new deep learning model Dense Dilated Multiscale Supervised Attention-Guided (D2MSA) Network for histopathology image segmentation that makes use of deep supervision coupled with a hierarchical system of novel attention mechanisms. The proposed model surpasses state-of-the-art performance while using similar computational resources. The performance of the model has been evaluated for the tasks of gland segmentation and nuclei instance segmentation, both of which are clinically relevant tasks to assess the state and progress of malignancy. Here, we have used histopathology image datasets for three different types of cancer. We have also performed extensive ablation tests and hyperparameter tuning to ensure the validity and reproducibility of the model performance. The proposed model is available at www.github.com/shirshabose/D2MSA-Net.

Unleashing the potential of AI for pathology: challenges and recommendations

Computational pathology is currently witnessing a surge in the development of AI techniques, offering promise for achieving breakthroughs and significantly impacting the practices of pathology and oncology. These AI methods bring with them the potential to revolutionize diagnostic pipelines as well as treatment planning and overall patient care. Numerous peer-reviewed studies reporting remarkable performance across diverse tasks serve as a testimony to the potential of AI in the field. However, widespread adoption of these methods in clinical and pre-clinical settings still remains a challenge. In this review article, we present a detailed analysis of the major obstacles encountered during the development of effective models and their deployment in practice. We aim to provide readers with an overview of the latest developments, assist them with insights into identifying some specific challenges that may require resolution, and suggest recommendations and potential future research directions. © 2023 The Authors. The Journal of Pathology published by John Wiley & Sons Ltd on behalf of The Pathological Society of Great Britain and Ireland.

Topology-Guided Multi-Class Cell Context Generation for Digital Pathology

In digital pathology, the spatial context of cells is important for cell classification, cancer diagnosis and prognosis. To model such complex cell context, however, is challenging. Cells form different mixtures, lineages, clusters and holes. To model such structural patterns in a learnable fashion, we introduce several mathematical tools from spatial statistics and topological data analysis. We incorporate such structural descriptors into a deep generative model as both conditional inputs and a differentiable loss. This way, we are able to generate high quality multi-class cell layouts for the first time. We show that the topology-rich cell layouts can be used for data augmentation and improve the performance of downstream tasks such as cell classification.

From modern CNNs to vision transformers: Assessing the performance, robustness, and classification strategies of deep learning models in histopathology

While machine learning is currently transforming the field of histopathology, the domain lacks a comprehensive evaluation of state-of-the-art models based on essential but complementary quality requirements beyond a mere classification accuracy. In order to fill this gap, we developed a new methodology to extensively evaluate a wide range of classification models, including recent vision transformers, and convolutional neural networks such as: ConvNeXt, ResNet (BiT), Inception, ViT and Swin transformer, with and without supervised or self-supervised pretraining. We thoroughly tested the models on five widely used histopathology datasets containing whole slide images of breast, gastric, and colorectal cancer and developed a novel approach using an image-to-image translation model to assess the robustness of a cancer classification model against stain variations. Further, we extended existing interpretability methods to previously unstudied models and systematically reveal insights of the models' classification strategies that allow for plausibility checks and systematic comparisons. The study resulted in specific model recommendations for practitioners as well as putting forward a general methodology to quantify a model's quality according to complementary requirements that can be transferred to future model architectures.

Handcrafted Histological Transformer (H2T): Unsupervised representation of whole slide images

Diagnostic, prognostic and therapeutic decision-making of cancer in pathology clinics can now be carried out based on analysis of multi-gigapixel tissue images, also known as whole-slide images (WSIs). Recently, deep convolutional neural networks (CNNs) have been proposed to derive unsupervised WSI representations; these are attractive as they rely less on expert annotation which is cumbersome. However, a major trade-off is that higher predictive power generally comes at the cost of interpretability, posing a challenge to their clinical use where transparency in decision-making is generally expected. To address this challenge, we present a handcrafted framework based on deep CNN for constructing holistic WSI-level representations. Building on recent findings about the internal working of the Transformer in the domain of natural language processing, we break down its processes and handcraft them into a more transparent framework that we term as the Handcrafted Histological Transformer or H2T. Based on our experiments involving various datasets consisting of a total of 10,042 WSIs, the results demonstrate that H2T based holistic WSI-level representations offer competitive performance compared to recent state-of-the-art methods and can be readily utilized for various downstream analysis tasks. Finally, our results demonstrate that the H2T framework can be up to 14 times faster than the Transformer models.

An Atrous Convolved Hybrid Seg-Net Model with residual and attention mechanism for gland detection and segmentation in histopathological images

PURPOSE

A clinically compatible computerized segmentation model is presented here that aspires to supply clinical gland informative details by seizing every small and intricate variation in medical images, integrate second opinions, and reduce human errors.

APPROACH

It comprises of enhanced learning capability that extracts denser multi-scale gland-specific features, recover semantic gap during concatenation, and effectively handle resolution-degradation and vanishing gradient problems. It is having three proposed modules namely Atrous Convolved Residual Learning Module in the encoder as well as decoder, Residual Attention Module in the skip connection paths, and Atrous Convolved Transitional Module as the transitional and output layer. Also, pre-processing techniques like patch-sampling, stain-normalization, augmentation, etc. are employed to develop its generalization capability. To verify its robustness and invigorate network invariance against digital variability, extensive experiments are carried out employing three different public datasets i.e., GlaS (Gland Segmentation Challenge), CRAG (Colorectal Adenocarcinoma Gland) and LC-25000 (Lung Colon-25000) dataset and a private HosC (Hospital Colon) dataset.

RESULTS

The presented model accomplished combative gland detection outcomes having F1-score (GlaS(Test A(0.957), Test B(0.926)), CRAG(0.935), LC 25000(0.922), HosC(0.963)); and gland segmentation results having Object-Dice Index (GlaS(Test A(0.961), Test B(0.933)), CRAG(0.961), LC-25000(0.940), HosC(0.929)), and Object-Hausdorff Distance (GlaS(Test A(21.77) and Test B(69.74)), CRAG(87.63), LC-25000(95.85), HosC(83.29)). In addition, validation score (GlaS (Test A(0.945), Test B(0.937)), CRAG(0.934), LC-25000(0.911), HosC(0.928)) supplied by the proficient pathologists is integrated for the end segmentation results to corroborate the applicability and appropriateness for assistance at the clinical level applications.

CONCLUSION

The proposed system will assist pathologists in devising precise diagnoses by offering a referential perspective during morphology assessment of colon histopathology images.

Symmetry-based regularization in deep breast cancer screening

Breast cancer is the most common and lethal form of cancer in women. Recent efforts have focused on developing accurate neural network-based computer-aided diagnosis systems for screening to help anticipate this disease. The ultimate goal is to reduce mortality and improve quality of life after treatment. Due to the difficulty in collecting and annotating data in this domain, data scarcity is - and will continue to be - a limiting factor. In this work, we present a unified view of different regularization methods that incorporate domain-known symmetries in the model. Three general strategies were followed: (i) data augmentation, (ii) invariance promotion in the loss function, and (iii) the use of equivariant architectures. Each of these strategies encodes different priors on the functions learned by the model and can be readily introduced in most settings. Empirically we show that the proposed symmetry-based regularization procedures improve generalization to unseen examples. This advantage is verified in different scenarios, datasets and model architectures. We hope that both the principle of symmetry-based regularization and the concrete methods presented can guide development towards more data-efficient methods for breast cancer screening as well as other medical imaging domains.

Finding a Suitable Class Distribution for Building Histological Images Datasets Used in Deep Model Training-The Case of Cancer Detection

The class distribution of a training dataset is an important factor which influences the performance of a deep learning-based system. Understanding the optimal class distribution is therefore crucial when building a new training set which may be costly to annotate. This is the case for histological images used in cancer diagnosis where image annotation requires domain experts. In this paper, we tackle the problem of finding the optimal class distribution of a training set to be able to train an optimal model that detects cancer in histological images. We formulate several hypotheses which are then tested in scores of experiments with hundreds of trials. The experiments have been designed to account for both segmentation and classification frameworks with various class distributions in the training set, such as natural, balanced, over-represented cancer, and over-represented non-cancer. In the case of cancer detection, the experiments show several important results: (a) the natural class distribution produces more accurate results than the artificially generated balanced distribution; (b) the over-representation of non-cancer/negative classes (healthy tissue and/or background classes) compared to cancer/positive classes reduces the number of samples which are falsely predicted as cancer (false positive); (c) the least expensive to annotate non-ROI (non-region-of-interest) data can be useful in compensating for the performance loss in the system due to a shortage of expensive to annotate ROI data; (d) the multi-label examples are more useful than the single-label ones to train a segmentation model; and (e) when the classification model is tuned with a balanced validation set, it is less affected than the segmentation model by the class distribution of the training set.

A semi-supervised learning framework for micropapillary adenocarcinoma detection

PURPOSE

Micropapillary adenocarcinoma is a distinctive histological subtype of lung adenocarcinoma with poor prognosis. Computer-aided diagnosis method has the potential to provide help for its early diagnosis. But the implementation of the existing methods largely relies on massive manually labeled data and consumes a lot of time and energy. To tackle these problems, we propose a framework that applies semi-supervised learning method to detect micropapillary adenocarcinoma, which aims to utilize labeled and unlabeled data better.

METHODS

The framework consists of a teacher model and a student model. The teacher model is first obtained by using the labeled data. Then, it makes predictions on unlabeled data as pseudo-labels for students. Finally, high-quality pseudo-labels are selected and associated with the labeled data to train the student model. During the learning process of the student model, augmentation is added so that the student model generalizes better than the teacher model.

RESULTS

Experiments are conducted on our own whole slide micropapillary lung adenocarcinoma histopathology image dataset and we selected 3527 patches for the experiment. In the supervised learning, our detector achieves a precision of 0.762 and recall of 0.884. In the semi-supervised learning, our method achieves a precision of 0.775 and recall of 0.896; it is superior to other methods.

CONCLUSION

We proposed a semi-supervised learning framework for micropapillary adenocarcinoma detection, which has better performance in utilizing both labeled and unlabeled data. In addition, the detector we designed improves the detection accuracy and speed and achieves promising results in detecting micropapillary adenocarcinoma.

SONNET: A self-guided ordinal regression neural network for segmentation and classification of nuclei in large-scale multi-tissue histology images

Automated nuclei segmentation and classification are the keys to analyze and understand the cellular characteristics and functionality, supporting computer-aided digital pathology in disease diagnosis. However, the task still remains challenging due to the intrinsic variations in size, intensity, and morphology of different types of nuclei. Herein, we propose a self-guided ordinal regression neural network for simultaneous nuclear segmentation and classification that can exploit the intrinsic characteristics of nuclei and focus on highly uncertain areas during training. The proposed network formulates nuclei segmentation as an ordinal regression learning by introducing a distance decreasing discretization strategy, which stratifies nuclei in a way that inner regions forming a regular shape of nuclei are separated from outer regions forming an irregular shape. It also adopts a self-guided training strategy to adaptively adjust the weights associated with nuclear pixels, depending on the difficulty of the pixels that is assessed by the network itself. To evaluate the performance of the proposed network, we employ large-scale multi-tissue datasets with 276349 exhaustively annotated nuclei. We show that the proposed network achieves the state-of-the-art performance in both nuclei segmentation and classification in comparison to several methods that are recently developed for segmentation and/or classification.

Toward Scalable and Unified Example-Based Explanation and Outlier Detection

When neural networks are employed for high-stakes decision-making, it is desirable that they provide explanations for their prediction in order for us to understand the features that have contributed to the decision. At the same time, it is important to flag potential outliers for in-depth verification by domain experts. In this work we propose to unify two differing aspects of explainability with outlier detection. We argue for a broader adoption of prototype-based student networks capable of providing an example-based explanation for their prediction and at the same time identify regions of similarity between the predicted sample and the examples. The examples are real prototypical cases sampled from the training set via a novel iterative prototype replacement algorithm. Furthermore, we propose to use the prototype similarity scores for identifying outliers. We compare performance in terms of the classification, explanation quality and outlier detection of our proposed network with baselines. We show that our prototype-based networks extending beyond similarity kernels deliver meaningful explanations and promising outlier detection results without compromising classification accuracy.

Self-Path: Self-Supervision for Classification of Pathology Images With Limited Annotations

While high-resolution pathology images lend themselves well to 'data hungry' deep learning algorithms, obtaining exhaustive annotations on these images for learning is a major challenge. In this article, we propose a self-supervised convolutional neural network (CNN) framework to leverage unlabeled data for learning generalizable and domain invariant representations in pathology images. Our proposed framework, termed as Self-Path, employs multi-task learning where the main task is tissue classification and pretext tasks are a variety of self-supervised tasks with labels inherent to the input images. We introduce novel pathology-specific self-supervision tasks that leverage contextual, multi-resolution and semantic features in pathology images for semi-supervised learning and domain adaptation. We investigate the effectiveness of Self-Path on 3 different pathology datasets. Our results show that Self-Path with the pathology-specific pretext tasks achieves state-of-the-art performance for semi-supervised learning when small amounts of labeled data are available. Further, we show that Self-Path improves domain adaptation for histopathology image classification when there is no labeled data available for the target domain. This approach can potentially be employed for other applications in computational pathology, where annotation budget is often limited or large amount of unlabeled image data is available.

Multi-Class Cell Detection Using Spatial Context Representation

In digital pathology, both detection and classification of cells are important for automatic diagnostic and prognostic tasks. Classifying cells into subtypes, such as tumor cells, lymphocytes or stromal cells is particularly challenging. Existing methods focus on morphological appearance of individual cells, whereas in practice pathologists often infer cell classes through their spatial context. In this paper, we propose a novel method for both detection and classification that explicitly incorporates spatial contextual information. We use the spatial statistical function to describe local density in both a multi-class and a multi-scale manner. Through representation learning and deep clustering techniques, we learn advanced cell representation with both appearance and spatial context. On various benchmarks, our method achieves better performance than state-of-the-arts, especially on the classification task. We also create a new dataset for multi-class cell detection and classification in breast cancer and we make both our code and data publicly available.

Resolution-based distillation for efficient histology image classification

Developing deep learning models to analyze histology images has been computationally challenging, as the massive size of the images causes excessive strain on all parts of the computing pipeline. This paper proposes a novel deep learning-based methodology for improving the computational efficiency of histology image classification. The proposed approach is robust when used with images that have reduced input resolution, and it can be trained effectively with limited labeled data. Moreover, our approach operates at either the tissue- or slide-level, removing the need for laborious patch-level labeling. Our method uses knowledge distillation to transfer knowledge from a teacher model pre-trained at high resolution to a student model trained on the same images at a considerably lower resolution. Also, to address the lack of large-scale labeled histology image datasets, we perform the knowledge distillation in a self-supervised fashion. We evaluate our approach on three distinct histology image datasets associated with celiac disease, lung adenocarcinoma, and renal cell carcinoma. Our results on these datasets demonstrate that a combination of knowledge distillation and self-supervision allows the student model to approach and, in some cases, surpass the teacher model's classification accuracy while being much more computationally efficient. Additionally, we observe an increase in student classification performance as the size of the unlabeled dataset increases, indicating that there is potential for this method to scale further with additional unlabeled data. Our model outperforms the high-resolution teacher model for celiac disease in accuracy, F1-score, precision, and recall while requiring 4 times fewer computations. For lung adenocarcinoma, our results at 1.25× magnification are within 1.5% of the results for the teacher model at 10× magnification, with a reduction in computational cost by a factor of 64. Our model on renal cell carcinoma at 1.25× magnification performs within 1% of the teacher model at 5× magnification while requiring 16 times fewer computations. Furthermore, our celiac disease outcomes benefit from additional performance scaling with the use of more unlabeled data. In the case of 0.625× magnification, using unlabeled data improves accuracy by 4% over the tissue-level baseline. Therefore, our approach can improve the feasibility of deep learning solutions for digital pathology on standard computational hardware and infrastructures.

Evaluation of Deep Learning Architectures for Complex Immunofluorescence Nuclear Image Segmentation

Separating and labeling each nuclear instance (instance-aware segmentation) is the key challenge in nuclear image segmentation. Deep Convolutional Neural Networks have been demonstrated to solve nuclear image segmentation tasks across different imaging modalities, but a systematic comparison on complex immunofluorescence images has not been performed. Deep learning based segmentation requires annotated datasets for training, but annotated fluorescence nuclear image datasets are rare and of limited size and complexity. In this work, we evaluate and compare the segmentation effectiveness of multiple deep learning architectures (U-Net, U-Net ResNet, Cellpose, Mask R-CNN, KG instance segmentation) and two conventional algorithms (Iterative h-min based watershed, Attributed relational graphs) on complex fluorescence nuclear images of various types. We propose and evaluate a novel strategy to create artificial images to extend the training set. Results show that instance-aware segmentation architectures and Cellpose outperform the U-Net architectures and conventional methods on complex images in terms of F1 scores, while the U-Net architectures achieve overall higher mean Dice scores. Training with additional artificially generated images improves recall and F1 scores for complex images, thereby leading to top F1 scores for three out of five sample preparation types. Mask R-CNN trained on artificial images achieves the overall highest F1 score on complex images of similar conditions to the training set images while Cellpose achieves the overall highest F1 score on complex images of new imaging conditions. We provide quantitative results demonstrating that images annotated by under-graduates are sufficient for training instance-aware segmentation architectures to efficiently segment complex fluorescence nuclear images.