Leakage-Resistant Colorectal Histopathology Classification Using ConvNeXt-Tiny and the Midnight-12k Pathology Foundation Model

Fadhil, Ahmed F.

doi:10.26706/ijceae.7.4.20260808

Abstract

The classification of colorectal histopathology images is crucial for computer-assisted pathological diagnosis, however the similarity of morphological structures between the categories of tissue makes multiclass classification challenging. This paper explores whether a Patho generative foundation model can better differentiate colorectal tissues in a leakage-safe evaluation protocol. The four-class dataset consisted of 6,160 images of histopathology tissue types — Normal, Serrated, Adenocarcinoma, and Adenoma — that were tested with ConvNeXt-Tiny, the first proposed framework and Midnight-12k, the last.

After applying the official test benchmark, the overall accuracy and other metrics for Midnight-12k were: 81.39 % accuracy, 81.50 % weighted F1-score, 93.89 % macro specificity, 94.90 % macro-ROC-AUC, 83.17 % macro PR-AUC, and MCC 0.7366. The precision (95.69%), recall (97.86%), F1-score (96.76%) and ROC-AUC (99.60%) were well achieved for Adenocarcinoma. F1 score of 82.18% with normal tissue and precision of 92.20%, specificity of 99.41% with Adenoma achieved. Overall macro precision, recall, and F1-score were 79.73%, 73.47%, and 74.43%, respectively. Midnight-12k boosted the accuracy from 74.58% to 77.24% and macro F1-score from 70.38% to 73.18% compared to ConvNeXt-Tiny under strict five-fold evaluation. The results illustrate the promise of pathologists' foundation models in colorectal histopathology classification and underscore the difficulty of distinguishing between Serrated and Adenoma.

Keywords

Deep Learning histopathology images ConvNeXt-Tiny Midnight-12k Pathology foundation model

References

World Health Organization, “Colorectal cancer,” WHO, Feb. 13, 2026. [Online]. Available: https://www.who.int/news-room/fact-sheets/detail/colorectal-cancer.
D. Komura, M. Ochi, and S. Ishikawa, “Machine learning methods for histopathological image analysis: Updates in 2024,” Comput. Struct. Biotechnol. J., vol. 27, pp. 383–400, 2025, doi: 10.1016/j.csbj.2024.12.033.
C. Zhu, W. Chen, T. Peng, Y. Wang, and M. Jin, “Hard sample aware noise robust learning for histopathology image classification,” IEEE Trans. Med. Imaging, vol. 41, no. 4, pp. 881–894, 2022, doi: 10.1109/TMI.2021.3125459.
Z. Liu, H. Mao, C.-Y. Wu, C. Feichtenhofer, T. Darrell, and S. Xie, “A ConvNet for the 2020s,” in Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit. (CVPR), New Orleans, LA, USA, 2022, pp. 11966–11976, doi: 10.1109/CVPR52688.2022.01167.
R. J. Chen, T. Ding, M. Y. Lu, et al., “Towards a general-purpose foundation model for computational pathology,” Nat. Med., vol. 30, pp. 850–862, 2024, doi: 10.1038/s41591-024-02857-3.
A. Filiot, P. Jacob, A. Mac Kain, and C. Saillard, “Phikon-v2, a large and public feature extractor for biomarker prediction,” arXiv preprint arXiv:2409.09173, Sep. 2024, doi: 10.48550/arXiv.2409.09173.
M. Karasikov, J. van Doorn, N. Känzig, M. E. Cesur, H. M. Horlings, R. Berke, F. Tang, and S. Otálora, “Training state-of-the-art pathology foundation models with orders of magnitude less data,” in Medical Image Computing and Computer Assisted Intervention – MICCAI 2025, Lecture Notes in Computer Science, vol. 15967. Cham, Switzerland: Springer Nature Switzerland, 2025, pp. 573–583.
P. C. Neto, D. Montezuma, S. P. Oliveira, et al., “An interpretable machine learning system for colorectal cancer diagnosis from pathology slides,” npj Precis. Oncol., vol. 8, Art. no. 56, 2024, doi: 10.1038/s41698-024-00539-4.
S. B. Yengec-Tasdemir, Z. Aydin, E. Akay, S. Dogan, and B. Yilmaz, “An effective colorectal polyp classification for histopathological images based on supervised contrastive learning,” Comput. Biol. Med., vol. 172, Art. no. 108267, Apr. 2024, doi: 10.1016/j.compbiomed.2024.108267.
Z. Fu, Q. Chen, M. Wang, and C. Huang, “Transformer based on multi-scale local feature for colon cancer histopathological image classification,” Biomed. Signal Process. Control, vol. 100, pt. A, Art. no. 106970, Feb. 2025, doi: 10.1016/j.bspc.2024.106970.
Q. Ke, Y. C. Hum, W.-S. Yap, T. S. Tan, H. Nisar, H. Mokayed, A. Li, R. Gao, and Y. Gan, “Histopathological classification of colorectal cancer based on domain-specific transfer learning and multi-model feature fusion,” Sci. Rep., vol. 15, Art. no. 35155, Oct. 2025, doi: 10.1038/s41598-025-19134-z.
G. Campanella, S. Chen, M. Singh, et al., “A clinical benchmark of public self-supervised pathology foundation models,” Nat. Commun., vol. 16, Art. no. 3640, 2025, doi: 10.1038/s41467-025-58796-1.
P. Neidlinger, O. S. M. El Nahhas, H. S. Muti, et al., “Benchmarking foundation models as feature extractors for weakly supervised computational pathology,” Nat. Biomed. Eng., vol. 10, pp. 1113–1123, 2026, doi: 10.1038/s41551-025-01516-3.
R. Bareja, F. Carrillo-Perez, Y. Zheng, et al., “A benchmark study of vision and pathology foundation models for computational pathology,” Nat. Commun., 2026, doi: 10.1038/s41467-026-76004-6.
R. Albasrawi, M. Ilyas, and Q. Althebyan, “Modified YOLOv8x model for coronary stenosis detection and troponin risk stratification,” Discov. Comput., vol. 29, Art. no. 34, 2026, doi: 10.1007/s10791-026-09913-1.
R. Albasrawi, M. Ilyas, and O. Bayat, “A modified YOLO-based deep learning model for accurate ST-elevation myocardial infarction detection in ECG images,” Neural Comput. Appl., vol. 38, Art. no. 617, 2026, doi: 10.1007/s00521-026-12347-x.
A. B. Salih, R. B. Mohammed, A. A. Othman, S. A. Mostafa, and M. Kamalrudin, “YOLOv8-based real-time pedestrian detection model for urban traffic surveillance systems,” Int. J. Inform. Vis., vol. 10, no. 2, pp. 556–564, Mar. 2026.
S. I. Hamad, “Utilizing convolutional neural networks for the identification of lung cancer,” International Journal of Computational and Electronics Aspects in Engineering, vol. 6, no. 1, pp. 35–41, Mar. 2025, doi: 10.26706/ijceae.6.1.20250206.
N. Waisi, “Comparative evaluation of ResNet-50 and EfficientNet-B1 for pneumonia detection in chest X-ray images using transfer learning,” International Journal of Computational and Electronics Aspects in Engineering, vol. 6, no. 2, pp. 89–97, Jun. 2025, doi: 10.26706/ijceae.6.2.20250405.
M. K. Yousif, “Enhanced diabetic retinopathy classification using hybrid EfficientNetB0 with CBAM,” International Journal of Computational and Electronics Aspects in Engineering, vol. 7, no. 2, pp. 176–188, May 2026, doi: 10.26706/ijceae.7.2.20260307.
N. Waisi, “Deep feature fusion method for images classification,” International Journal of Computational and Electronics Aspects in Engineering, vol. 5, no. 4, pp. 148–153, Dec. 2024, doi: 10.26706/ijceae.5.4.20241103.
I. S. Seger, “Evaluating and selecting optimal CNN architectures for accurate pneumonia detection in chest X-rays,” Int. J. Comput. Electron. Aspects Eng., vol. 5, no. 4, pp. 183–193, Dec. 2024, doi: 10.26706/ijceae.5.4.20241106.

[1] World Health Organization, “Colorectal cancer,” WHO, Feb. 13, 2026. [Online]. Available: https://www.who.int/news-room/fact-sheets/detail/colorectal-cancer.

[2] D. Komura, M. Ochi, and S. Ishikawa, “Machine learning methods for histopathological image analysis: Updates in 2024,” Comput. Struct. Biotechnol. J., vol. 27, pp. 383–400, 2025, doi: 10.1016/j.csbj.2024.12.033.

[3] C. Zhu, W. Chen, T. Peng, Y. Wang, and M. Jin, “Hard sample aware noise robust learning for histopathology image classification,” IEEE Trans. Med. Imaging, vol. 41, no. 4, pp. 881–894, 2022, doi: 10.1109/TMI.2021.3125459.

[4] Z. Liu, H. Mao, C.-Y. Wu, C. Feichtenhofer, T. Darrell, and S. Xie, “A ConvNet for the 2020s,” in Proc. IEEE/CVF Conf. Comput. Vis. Pattern Recognit. (CVPR), New Orleans, LA, USA, 2022, pp. 11966–11976, doi: 10.1109/CVPR52688.2022.01167.

[5] R. J. Chen, T. Ding, M. Y. Lu, et al., “Towards a general-purpose foundation model for computational pathology,” Nat. Med., vol. 30, pp. 850–862, 2024, doi: 10.1038/s41591-024-02857-3.

[6] A. Filiot, P. Jacob, A. Mac Kain, and C. Saillard, “Phikon-v2, a large and public feature extractor for biomarker prediction,” arXiv preprint arXiv:2409.09173, Sep. 2024, doi: 10.48550/arXiv.2409.09173.

[7] M. Karasikov, J. van Doorn, N. Känzig, M. E. Cesur, H. M. Horlings, R. Berke, F. Tang, and S. Otálora, “Training state-of-the-art pathology foundation models with orders of magnitude less data,” in Medical Image Computing and Computer Assisted Intervention – MICCAI 2025, Lecture Notes in Computer Science, vol. 15967. Cham, Switzerland: Springer Nature Switzerland, 2025, pp. 573–583.

[8] P. C. Neto, D. Montezuma, S. P. Oliveira, et al., “An interpretable machine learning system for colorectal cancer diagnosis from pathology slides,” npj Precis. Oncol., vol. 8, Art. no. 56, 2024, doi: 10.1038/s41698-024-00539-4.

[9] S. B. Yengec-Tasdemir, Z. Aydin, E. Akay, S. Dogan, and B. Yilmaz, “An effective colorectal polyp classification for histopathological images based on supervised contrastive learning,” Comput. Biol. Med., vol. 172, Art. no. 108267, Apr. 2024, doi: 10.1016/j.compbiomed.2024.108267.

[10] Z. Fu, Q. Chen, M. Wang, and C. Huang, “Transformer based on multi-scale local feature for colon cancer histopathological image classification,” Biomed. Signal Process. Control, vol. 100, pt. A, Art. no. 106970, Feb. 2025, doi: 10.1016/j.bspc.2024.106970.

[11] Q. Ke, Y. C. Hum, W.-S. Yap, T. S. Tan, H. Nisar, H. Mokayed, A. Li, R. Gao, and Y. Gan, “Histopathological classification of colorectal cancer based on domain-specific transfer learning and multi-model feature fusion,” Sci. Rep., vol. 15, Art. no. 35155, Oct. 2025, doi: 10.1038/s41598-025-19134-z.

[12] G. Campanella, S. Chen, M. Singh, et al., “A clinical benchmark of public self-supervised pathology foundation models,” Nat. Commun., vol. 16, Art. no. 3640, 2025, doi: 10.1038/s41467-025-58796-1.

[13] P. Neidlinger, O. S. M. El Nahhas, H. S. Muti, et al., “Benchmarking foundation models as feature extractors for weakly supervised computational pathology,” Nat. Biomed. Eng., vol. 10, pp. 1113–1123, 2026, doi: 10.1038/s41551-025-01516-3.

[14] R. Bareja, F. Carrillo-Perez, Y. Zheng, et al., “A benchmark study of vision and pathology foundation models for computational pathology,” Nat. Commun., 2026, doi: 10.1038/s41467-026-76004-6.

[15] R. Albasrawi, M. Ilyas, and Q. Althebyan, “Modified YOLOv8x model for coronary stenosis detection and troponin risk stratification,” Discov. Comput., vol. 29, Art. no. 34, 2026, doi: 10.1007/s10791-026-09913-1.

[16] R. Albasrawi, M. Ilyas, and O. Bayat, “A modified YOLO-based deep learning model for accurate ST-elevation myocardial infarction detection in ECG images,” Neural Comput. Appl., vol. 38, Art. no. 617, 2026, doi: 10.1007/s00521-026-12347-x.

[17] A. B. Salih, R. B. Mohammed, A. A. Othman, S. A. Mostafa, and M. Kamalrudin, “YOLOv8-based real-time pedestrian detection model for urban traffic surveillance systems,” Int. J. Inform. Vis., vol. 10, no. 2, pp. 556–564, Mar. 2026.

[18] S. I. Hamad, “Utilizing convolutional neural networks for the identification of lung cancer,” International Journal of Computational and Electronics Aspects in Engineering, vol. 6, no. 1, pp. 35–41, Mar. 2025, doi: 10.26706/ijceae.6.1.20250206.

[19] N. Waisi, “Comparative evaluation of ResNet-50 and EfficientNet-B1 for pneumonia detection in chest X-ray images using transfer learning,” International Journal of Computational and Electronics Aspects in Engineering, vol. 6, no. 2, pp. 89–97, Jun. 2025, doi: 10.26706/ijceae.6.2.20250405.

[20] M. K. Yousif, “Enhanced diabetic retinopathy classification using hybrid EfficientNetB0 with CBAM,” International Journal of Computational and Electronics Aspects in Engineering, vol. 7, no. 2, pp. 176–188, May 2026, doi: 10.26706/ijceae.7.2.20260307.

[21] N. Waisi, “Deep feature fusion method for images classification,” International Journal of Computational and Electronics Aspects in Engineering, vol. 5, no. 4, pp. 148–153, Dec. 2024, doi: 10.26706/ijceae.5.4.20241103.

[22] I. S. Seger, “Evaluating and selecting optimal CNN architectures for accurate pneumonia detection in chest X-rays,” Int. J. Comput. Electron. Aspects Eng., vol. 5, no. 4, pp. 183–193, Dec. 2024, doi: 10.26706/ijceae.5.4.20241106.