Lifelong Whole Slide Image Analysis: Online Vision-Language Adaptation and Past-to-Present Gradient Distillation

• URL: http://arxiv.org/abs/2505.01984v1
• Date: Sun, 04 May 2025 04:46:08 GMT
• Title: Lifelong Whole Slide Image Analysis: Online Vision-Language Adaptation and Past-to-Present Gradient Distillation
• Authors: Doanh C. Bui, Hoai Luan Pham, Vu Trung Duong Le, Tuan Hai Vu, Van Duy Tran, Khang Nguyen, Yasuhiko Nakashima,
• Abstract summary: Whole Slide Images (WSIs) play a crucial role in accurate cancer diagnosis and prognosis.<n>Given that WSIs are gigapixels in size, they present difficulties in terms of storage, processing, and model training.<n>We introduce ADaFGrad, a method designed to enhance lifelong learning for whole-slide image (WSI) analysis.
• Score: 1.1497371646067622
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Whole Slide Images (WSIs) play a crucial role in accurate cancer diagnosis and prognosis, as they provide tissue details at the cellular level. However, the rapid growth of computational tasks involving WSIs poses significant challenges. Given that WSIs are gigapixels in size, they present difficulties in terms of storage, processing, and model training. Therefore, it is essential to develop lifelong learning approaches for WSI analysis. In scenarios where slides are distributed across multiple institutes, we aim to leverage them to develop a unified online model as a computational tool for cancer diagnosis in clinical and hospital settings. In this study, we introduce ADaFGrad, a method designed to enhance lifelong learning for whole-slide image (WSI) analysis. First, we leverage pathology vision-language foundation models to develop a framework that enables interaction between a slide's regional tissue features and a predefined text-based prototype buffer. Additionally, we propose a gradient-distillation mechanism that mimics the gradient of a logit with respect to the classification-head parameters across past and current iterations in a continual-learning setting. We construct a sequence of six TCGA datasets for training and evaluation. Experimental results show that ADaFGrad outperforms both state-of-the-art WSI-specific and conventional continual-learning methods after only a few training epochs, exceeding them by up to +5.068% in the class-incremental learning scenario while exhibiting the least forgetting (i.e., retaining the most knowledge from previous tasks). Moreover, ADaFGrad surpasses its baseline by as much as +40.084% in accuracy, further demonstrating the effectiveness of the proposed modules.

Related papers

• ZeroSlide: Is Zero-Shot Classification Adequate for Lifelong Learning in Whole-Slide Image Analysis in the Era of Pathology Vision-Language Foundation Models? [0.5057850174013127]
  Lifelong learning for whole slide images (WSIs) poses the challenge of training a unified model to perform multiple WSI-related tasks.<n>This is the first study to compare conventional continual-learning approaches with vision-language zero-shot classification for WSIs.
  arXiv  Detail & Related papers  (2025-04-22T06:38:37Z)
• From Pixels to Histopathology: A Graph-Based Framework for Interpretable Whole Slide Image Analysis [81.19923502845441]
  We develop a graph-based framework that constructs WSI graph representations.<n>We build tissue representations (nodes) that follow biological boundaries rather than arbitrary patches.<n>In our method's final step, we solve the diagnostic task through a graph attention network.
  arXiv  Detail & Related papers  (2025-03-14T20:15:04Z)
• Fake It Till You Make It: Using Synthetic Data and Domain Knowledge for Improved Text-Based Learning for LGE Detection [11.532639713283226]
  We use strategies rooted in domain knowledge to train a model for LGE detection using text from clinical reports.<n>We standardize the orientation of the images in an anatomy-informed way to enable better alignment of spatial and text features.<n> ablation studies are carried out to elucidate the contributions of each design component to the overall performance of the model.
  arXiv  Detail & Related papers  (2025-02-18T15:30:48Z)
• Benchmarking Embedding Aggregation Methods in Computational Pathology: A Clinical Data Perspective [32.93871326428446]
  Recent advances in artificial intelligence (AI) are revolutionizing medical imaging and computational pathology.<n>A constant challenge in the analysis of digital Whole Slide Images (WSIs) is the problem of aggregating tens of thousands of tile-level image embeddings to a slide-level representation.<n>This study conducts a benchmarking analysis of ten slide-level aggregation techniques across nine clinically relevant tasks.
  arXiv  Detail & Related papers  (2024-07-10T17:00:57Z)
• A self-supervised framework for learning whole slide representations [52.774822784847565]
  We present Slide Pre-trained Transformers (SPT) for gigapixel-scale self-supervision of whole slide images. We benchmark SPT visual representations on five diagnostic tasks across three biomedical microscopy datasets.
  arXiv  Detail & Related papers  (2024-02-09T05:05:28Z)
• Context-Aware Self-Supervised Learning of Whole Slide Images [0.0]
  A novel two-stage learning technique is presented in this work. A graph representation capturing all dependencies among regions in the WSI is very intuitive. The entire slide is presented as a graph, where the nodes correspond to the patches from the WSI. The proposed framework is then tested using WSIs from prostate and kidney cancers.
  arXiv  Detail & Related papers  (2023-06-07T20:23:05Z)
• Vision-Language Modelling For Radiological Imaging and Reports In The Low Data Regime [70.04389979779195]
  This paper explores training medical vision-language models (VLMs) where the visual and language inputs are embedded into a common space. We explore several candidate methods to improve low-data performance, including adapting generic pre-trained models to novel image and text domains. Using text-to-image retrieval as a benchmark, we evaluate the performance of these methods with variable sized training datasets of paired chest X-rays and radiological reports.
  arXiv  Detail & Related papers  (2023-03-30T18:20:00Z)
• Learning to Exploit Temporal Structure for Biomedical Vision-Language Processing [53.89917396428747]
  Self-supervised learning in vision-language processing exploits semantic alignment between imaging and text modalities. We explicitly account for prior images and reports when available during both training and fine-tuning. Our approach, named BioViL-T, uses a CNN-Transformer hybrid multi-image encoder trained jointly with a text model.
  arXiv  Detail & Related papers  (2023-01-11T16:35:33Z)
• LifeLonger: A Benchmark for Continual Disease Classification [59.13735398630546]
  We introduce LifeLonger, a benchmark for continual disease classification on the MedMNIST collection. Task and class incremental learning of diseases address the issue of classifying new samples without re-training the models from scratch. Cross-domain incremental learning addresses the issue of dealing with datasets originating from different institutions while retaining the previously obtained knowledge.
  arXiv  Detail & Related papers  (2022-04-12T12:25:05Z)
• Deep Low-Shot Learning for Biological Image Classification and Visualization from Limited Training Samples [52.549928980694695]
  In situ hybridization (ISH) gene expression pattern images from the same developmental stage are compared. labeling training data with precise stages is very time-consuming even for biologists. We propose a deep two-step low-shot learning framework to accurately classify ISH images using limited training images.
  arXiv  Detail & Related papers  (2020-10-20T06:06:06Z)

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.