Papers
Topics
Authors
Recent
Gemini 2.5 Flash
Gemini 2.5 Flash
156 tokens/sec
GPT-4o
7 tokens/sec
Gemini 2.5 Pro Pro
45 tokens/sec
o3 Pro
4 tokens/sec
GPT-4.1 Pro
38 tokens/sec
DeepSeek R1 via Azure Pro
28 tokens/sec
2000 character limit reached

Mitosis Detection, Fast and Slow: Robust and Efficient Detection of Mitotic Figures (2208.12587v2)

Published 26 Aug 2022 in cs.CV

Abstract: Counting of mitotic figures is a fundamental step in grading and prognostication of several cancers. However, manual mitosis counting is tedious and time-consuming. In addition, variation in the appearance of mitotic figures causes a high degree of discordance among pathologists. With advances in deep learning models, several automatic mitosis detection algorithms have been proposed but they are sensitive to {\em domain shift} often seen in histology images. We propose a robust and efficient two-stage mitosis detection framework, which comprises mitosis candidate segmentation ({\em Detecting Fast}) and candidate refinement ({\em Detecting Slow}) stages. The proposed candidate segmentation model, termed \textit{EUNet}, is fast and accurate due to its architectural design. EUNet can precisely segment candidates at a lower resolution to considerably speed up candidate detection. Candidates are then refined using a deeper classifier network, EfficientNet-B7, in the second stage. We make sure both stages are robust against domain shift by incorporating domain generalization methods. We demonstrate state-of-the-art performance and generalizability of the proposed model on the three largest publicly available mitosis datasets, winning the two mitosis domain generalization challenge contests (MIDOG21 and MIDOG22). Finally, we showcase the utility of the proposed algorithm by processing the TCGA breast cancer cohort (1,125 whole-slide images) to generate and release a repository of more than 620K mitotic figures.

Definition Search Book Streamline Icon: https://streamlinehq.com
References (82)
  1. Leveraging unlabeled whole-slide-images for mitosis detection, in: Computational Pathology and Ophthalmic Medical Image Analysis. Springer, pp. 69–77.
  2. Nuclei detection using mixture density networks, in: International workshop on machine learning in medical imaging, Springer. pp. 241–248.
  3. Nuclear instance segmentation using a proposal-free spatially aware deep learning framework, in: International conference on medical image computing and computer-assisted intervention, Springer. pp. 622–630.
  4. Nuclick: a deep learning framework for interactive segmentation of microscopic images. Medical Image Analysis 65, 101771.
  5. Self-path: Self-supervision for classification of pathology images with limited annotations. IEEE Transactions on Medical Imaging 40, 2845–2856.
  6. Interobserver variability of mitotic index and utility of phh3 for risk stratification in gastrointestinal stromal tumors. American journal of clinical pathology 143, 385–392.
  7. Towards launching ai algorithms for cellular pathology into clinical & pharmaceutical orbits. arXiv preprint arXiv:2112.09496 .
  8. Mitosis domain generalization challenge 2022. URL: https://doi.org/10.5281/zenodo.6362337, doi:10.5281/zenodo.6362337.
  9. Mitosis domain generalization challenge. 24th International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI 2021) URL: https://doi.org/10.5281/zenodo.4573978, doi:10.5281/zenodo.4573978.
  10. Mitosis domain generalization in histopathology images–the midog challenge. arXiv preprint arXiv:2204.03742 .
  11. Diagnostic assessment of deep learning algorithms for detection of lymph node metastases in women with breast cancer. Jama 318, 2199–2210.
  12. A large-scale dataset for mitotic figure assessment on whole slide images of canine cutaneous mast cell tumor. Scientific data 6, 1–9.
  13. Are pathologist-defined labels reproducible? comparison of the tupac16 mitotic figure dataset with an alternative set of labels, in: Interpretable and Annotation-Efficient Learning for Medical Image Computing. Springer, pp. 204–213.
  14. Evaluating reproducibility of ai algorithms in digital pathology with dapper. PLoS computational biology 15, e1006269.
  15. Molecular architecture of the kinetochore–microtubule interface. Nature reviews Molecular cell biology 9, 33–46.
  16. A simple framework for contrastive learning of visual representations, in: International conference on machine learning, PMLR. pp. 1597–1607.
  17. Domain-robust mitotic figure detection with style transfer, in: Biomedical Image Registration, Domain Generalisation and Out-of-Distribution Analysis, MICCAI 2021 Challenges L2R, MIDOG and MOOD. Springer, pp. 23–31.
  18. Mitosis detection in breast cancer histology images with deep neural networks, in: International conference on medical image computing and computer-assisted intervention, Springer. pp. 411–418.
  19. Counting mitoses: Si (ze) matters! Modern Pathology 34, 1651–1657.
  20. Accurate and reproducible invasive breast cancer detection in whole-slide images: A deep learning approach for quantifying tumor extent. Scientific reports 7, 46450.
  21. Deep learning methods for mitosis detection in breast cancer histopathological images: A comprehensive review. Artificial Intelligence and Machine Learning for Digital Pathology , 279–306.
  22. Precision histology: how deep learning is poised to revitalize histomorphology for personalized cancer care. NPJ precision oncology 1, 22.
  23. Pathology reporting of breast disease. NHS publication .
  24. Pathological prognostic factors in breast cancer. i. the value of histological grade in breast cancer: experience from a large study with long-term follow-up. Histopathology 19, 403–410.
  25. Domain-specific cycle-gan augmentation improves domain generalizability for mitosis detection, in: Biomedical Image Registration, Domain Generalisation and Out-of-Distribution Analysis, MICCAI 2021 Challenges L2R, MIDOG and MOOD. Springer, pp. 40–47.
  26. Pannuke dataset extension, insights and baselines. arXiv preprint arXiv:2003.10778 .
  27. Domain-adversarial training of neural networks. The journal of machine learning research 17, 2096–2030.
  28. Lizard: a large-scale dataset for colonic nuclear instance segmentation and classification, in: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 684–693.
  29. Hover-net: Simultaneous segmentation and classification of nuclei in multi-tissue histology images. Medical Image Analysis 58, 101563. URL: https://www.sciencedirect.com/science/article/pii/S1361841519301045, doi:https://doi.org/10.1016/j.media.2019.101563.
  30. Mask r-cnn, in: Proceedings of the IEEE international conference on computer vision, pp. 2961–2969.
  31. Deep residual learning for image recognition, in: Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR).
  32. Squeeze-and-excitation networks, in: Proceedings of the IEEE conference on computer vision and pattern recognition, pp. 7132–7141.
  33. Assessment of mitotic activity in breast cancer: revisited in the digital pathology era. Journal of Clinical Pathology 75, 365–372.
  34. Improving mitotic cell counting accuracy and efficiency using phosphohistone-h3 (phh3) antibody counterstained with haematoxylin and eosin as part of breast cancer grading. Histopathology 82, 393–406.
  35. Nuclick: From clicks in the nuclei to nuclear boundaries, in: MICCAI 2019 Workshop COMPAY.
  36. Segmentation of skin lesions and their attributes using multi-scale convolutional neural networks and domain specific augmentations. arXiv preprint arXiv:1809.10243 .
  37. Self-supervised visual feature learning with deep neural networks: A survey. IEEE transactions on pattern analysis and machine intelligence 43, 4037–4058.
  38. Thinking, fast and slow. Macmillan.
  39. Smallmitosis: small size mitotic cells detection in breast histopathology images. IEEE Access 9, 905–922.
  40. Supervised contrastive learning. Advances in Neural Information Processing Systems 33, 18661–18673.
  41. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980 .
  42. Multi-source domain adaptation using gradient reversal layer for mitotic cell detection. URL: https://arxiv.org/abs/2109.01503, doi:10.48550/ARXIV.2109.01503.
  43. A Deep Learning based Ensemble Model for Generalized Mitosis detection in H&E stained Whole Slide Images. URL: https://doi.org/10.5281/zenodo.7030876, doi:10.5281/zenodo.7030876.
  44. Imagenet classification with deep convolutional neural networks. Advances in neural information processing systems 25.
  45. Learning domain-invariant representations of histological images. Frontiers in medicine 6, 162.
  46. Visual assessment of mitotic figures in breast cancer: a comparative study between light microscopy and whole slide images. Histopathology 79, 913–925.
  47. Weakly supervised mitosis detection in breast histopathology images using concentric loss. Medical image analysis 53, 165–178.
  48. A survey of convolutional neural networks: analysis, applications, and prospects. IEEE transactions on neural networks and learning systems .
  49. Detecting mitosis against domain shift using a fused detector and deep ensemble classification model for midog challenge, in: Biomedical Image Registration, Domain Generalisation and Out-of-Distribution Analysis, MICCAI 2021 Challenges L2R, MIDOG and MOOD. Springer, pp. 68–72.
  50. Focal loss for dense object detection, in: Proceedings of the IEEE international conference on computer vision, pp. 2980–2988.
  51. Mitosis detection in breast cancer histological images an icpr 2012 contest. Journal of pathology informatics 4, 8.
  52. Artificial intelligence-based mitosis detection in breast cancer histopathology images using faster r-cnn and deep cnns. Journal of clinical medicine 9, 749.
  53. Computational methods for automated mitosis detection in histopathology images: A review. Biocybernetics and Biomedical Engineering 41, 64–82.
  54. Thinking fast and slow: Efficient text-to-visual retrieval with transformers, in: Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, pp. 9826–9836.
  55. Observer reliability in assessment of mitotic activity and mib-1—determined proliferation rate in pediatric sarcomas. Annals of diagnostic pathology 4, 228–235.
  56. Two-step domain adaptation for mitotic cell detection in histopathology images, in: Biomedical Image Registration, Domain Generalisation and Out-of-Distribution Analysis, MICCAI 2021 Challenges L2R, MIDOG and MOOD. Springer, pp. 32–39.
  57. Mitosis detection for invasive breast cancer grading in histopathological images. IEEE transactions on image processing 24, 4041–4054.
  58. TIAToolbox as an end-to-end library for advanced tissue image analytics. Communications Medicine 2, 120. URL: https://www.nature.com/articles/s43856-022-00186-5, doi:10.1038/s43856-022-00186-5.
  59. Prognostic significance of nottingham histologic grade in invasive breast carcinoma. Journal of clinical oncology 26, 3153–3158.
  60. Swish: a self-gated activation function. arXiv preprint arXiv:1710.05941 7, 5.
  61. Cascade r-cnn for midog challenge, in: Biomedical Image Registration, Domain Generalisation and Out-of-Distribution Analysis, MICCAI 2021 Challenges L2R, MIDOG and MOOD. Springer, pp. 81–85.
  62. U-net: Convolutional networks for biomedical image segmentation, in: International Conference on Medical image computing and computer-assisted intervention, Springer. pp. 234–241.
  63. A study about color normalization methods for histopathology images. Micron 114, 42–61.
  64. Attention based Transformers for Mitosis Domain Generalization in H&E stained WSIs. URL: https://doi.org/10.5281/zenodo.7029973, doi:10.5281/zenodo.7029973.
  65. Global and mitosis-specific interobserver variation in mitotic count scoring and implications for malignant melanoma staging. Histopathology 76, 803–813.
  66. Mobilenetv2: Inverted residuals and linear bottlenecks, in: Proceedings of the IEEE conference on computer vision and pattern recognition, pp. 4510–4520.
  67. Maskmitosis: a deep learning framework for fully supervised, weakly supervised, and unsupervised mitosis detection in histopathology images. Medical & Biological Engineering & Computing 58, 1603–1623.
  68. Simultaneous nuclear instance and layer segmentation in oral epithelial dysplasia, in: Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 552–561.
  69. Measuring domain shift for deep learning in histopathology. IEEE journal of biomedical and health informatics 25, 325–336.
  70. Efficientnet: Rethinking model scaling for convolutional neural networks, in: International Conference on Machine Learning, PMLR. pp. 6105–6114.
  71. Efficientdet: Scalable and efficient object detection, in: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp. 10781–10790.
  72. Whole-slide mitosis detection in h&e breast histology using phh3 as a reference to train distilled stain-invariant convolutional networks. IEEE transactions on medical imaging 37, 2126–2136.
  73. Quantifying the effects of data augmentation and stain color normalization in convolutional neural networks for computational pathology. Medical image analysis 58, 101544.
  74. Structure-preserving color normalization and sparse stain separation for histological images. IEEE transactions on medical imaging 35, 1962–1971.
  75. Assessment of algorithms for mitosis detection in breast cancer histopathology images. Medical Image Analysis 20, 237–248.
  76. Predicting breast tumor proliferation from whole-slide images: the tupac16 challenge. Medical image analysis 54, 111–121.
  77. Mitosis counting in breast cancer: Object-level interobserver agreement and comparison to an automatic method. PloS one 11, e0161286.
  78. Nuclear segmentation and classification: On color & compression generalization, in: International Workshop on Machine Learning in Medical Imaging, Springer. pp. 10–19.
  79. Impash: A novel domain-shift resistant representation for colorectal cancer tissue classification, in: Medical Computer Vision Workshop (MCV).
  80. Influence of inter-annotator variability on automatic mitotic figure assessment, in: Bildverarbeitung für die Medizin 2021: Proceedings, German Workshop on Medical Image Computing, Regensburg, March 7-9, 2021, Springer. pp. 241–246.
  81. Domain adversarial retinanet as a reference algorithm for the mitosis domain generalization (midog) challenge, in: Biomedical Image Registration, Domain Generalisation and Out-of-Distribution Analysis, MICCAI 2021 Challenges L2R, MIDOG and MOOD. Springer, pp. 5–13.
  82. Sk-unet model with fourier domain for mitosis detection, in: Biomedical Image Registration, Domain Generalisation and Out-of-Distribution Analysis, MICCAI 2021 Challenges L2R, MIDOG and MOOD. Springer, pp. 86–90.
Citations (8)
View on Semantic Scholar

Summary

We haven't generated a summary for this paper yet.

Ai Sparkles Streamline Icon: https://streamlinehq.com Summarize Now