HistoViT：用于准确且可扩展的组织病理学癌症诊断的视觉Transformer
HistoViT: Vision Transformer for Accurate and Scalable Histopathological Cancer Diagnosis
AI总结：
本研究提出基于Vision Transformer的深度学习框架HistoViT，通过简化的预处理管道处理组织病理学图像，在四个基准数据集上实现高准确率分类，验证了Transformer架构在数字病理学中的临床潜力。
摘要：
准确且可扩展的癌症诊断仍然是现代病理学中的一个关键挑战，特别是对于表现出复杂组织学变异性的乳腺、前列腺、骨骼和宫颈等恶性肿瘤。在本研究中，我们提出了一个基于Transformer的深度学习框架，用于组织病理学图像中的多类肿瘤分类。利用微调的Vision Transformer (ViT)架构，我们的方法解决了传统卷积神经网络的关键局限性，提供了改进的性能、减少的预处理需求和跨组织类型的增强可扩展性。为了使模型适应组织病理学癌症图像，我们实现了一个简化的预处理管道，将分块的全切片图像转换为PyTorch张量，并通过数据归一化对其进行标准化。这确保了与ViT架构的兼容性，并增强了收敛稳定性和整体分类性能。我们在四个基准数据集上评估了我们的模型：ICIAR2018（乳腺）、SICAPv2（前列腺）、UT-Osteosarcoma（骨骼）和SipakMed（宫颈）数据集——展示了对现有深度学习方法的一致性超越。我们的方法在乳腺、前列腺、骨骼和宫颈癌上分别实现了99.32%、96.92%、95.28%和96.94%的分类准确率，所有数据集的ROC曲线下面积（AUC）得分均超过99%。这些结果证实了基于Transformer的架构在数字病理学中的鲁棒性、泛化性和临床潜力。我们的工作代表了在可靠、自动化和可解释的癌症诊断系统方面的重要进展，这些系统可以减轻诊断负担并改善医疗保健结果。
英文摘要：
Accurate and scalable cancer diagnosis remains a critical challenge in modern pathology, particularly for malignancies such as breast, prostate, bone, and cervical, which exhibit complex histological variability. In this study, we propose a transformer-based deep learning framework for multi-class tumor classification in histopathological images. Leveraging a fine-tuned Vision Transformer (ViT) architecture, our method addresses key limitations of conventional convolutional neural networks, offering improved performance, reduced preprocessing requirements, and enhanced scalability across tissue types. To adapt the model for histopathological cancer images, we implement a streamlined preprocessing pipeline that converts tiled whole-slide images into PyTorch tensors and standardizes them through data normalization. This ensures compatibility with the ViT architecture and enhances both convergence stability and overall classification performance. We evaluate our model on four benchmark datasets: ICIAR2018 (breast), SICAPv2 (prostate), UT-Osteosarcoma (bone), and SipakMed (cervical) dataset -- demonstrating consistent outperformance over existing deep learning methods. Our approach achieves classification accuracies of 99.32%, 96.92%, 95.28%, and 96.94% for breast, prostate, bone, and cervical cancers respectively, with area under the ROC curve (AUC) scores exceeding 99% across all datasets. These results confirm the robustness, generalizability, and clinical potential of transformer-based architectures in digital pathology. Our work represents a significant advancement toward reliable, automated, and interpretable cancer diagnosis systems that can alleviate diagnostic burdens and improve healthcare outcomes.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
