In this work, a novel brain tumor classification model—DMA-Net—is proposed to address the challenges of limited annotations and class imbalance in medical MRI analysis. DMA-Net integrates dynamic meta-learning with anatomy-aware channel attention to improve generalization under few-shot conditions. The framework introduces three core innovations: (1) a task-aware channel reconfiguration module that aligns meta-learned representations with anatomical priors, (2) a dual-temperature focal loss that adaptively re-weights intra- and inter-class features, and (3) an augmentation engine incorporating morphology-driven transformations. Experiments on the BraTS 2018 dataset show that DMA-Net achieves 85.7% accuracy and an AUC of 0.893 using only 120 training cases, outperforming several baselines including MAML-CNN. The model also demonstrates strong interpretability (Dice = 0.892), low latency (47 ms/case), and preliminary capacity to predict IDH mutation status. These results indicate that DMA-Net provides a scalable, explainable, and deployment-ready solution for intelligent brain tumor diagnosis. Future directions include multimodal PET–MRI modeling, federated learning for cross-institutional training, and clinical validation through prospective trials.

Bohan Wang. DMA-Net: A Dynamic Meta-Attentive Network for Few-Shot Brain Tumor Classification with Anatomical Priors and Clinical Interpretability. International Journal of Frontiers in Engineering Technology (2025), Vol. 7, Issue 3: 58-70. https://doi.org/10.25236/IJFET.2025.070309.

[1] Louis, D.N., Perry, A., Wesseling, P., et al. (2021) The 2021 WHO Classification of Tumors of the Central Nervous System: a summary. Neuro-Oncology, 23(8): 1231–1251.

[2] Brat, D.J., Aldape, K.D., Colman, H., et al. (2018) cIMPACT-NOW update 3: recommended diagnostic criteria for "Diffuse astrocytic glioma, IDH-wildtype, with molecular features of glioblastoma". Acta Neuropathologica, 136(5): 805–810.

[3] Menze, B.H., Jakab, A., Bauer, S., et al. (2015) The Multimodal Brain Tumor Image Segmentation Benchmark (BRATS). IEEE Trans. Med. Imaging, 34(10): 1993–2024.

[4] Zhao, Y., Li, J., Zhang, Q., Wang, C., Liu, S., Chen, W. (2022) A hybrid deep CNN model for brain tumor classification based on radiomic features. Medical Image Analysis, 75: 102303.

[5] Liu, X., Zhang, Y., Chen, Z., et al. (2023) Advancing brain tumor classification through fine-tuned Vision Transformers. Frontiers in Neuroscience, 17: 1053533.

[6] He, K., Chen, X., Shen, Y., et al. (2022) A data-efficient training framework for Transformer models in medical imaging. In: Medical Image Computing and Computer Assisted Intervention. Singapore. pp. 112–125.

[7] Litjens, G., Kooi, T., Bejnordi, B.E., et al. (2017) A survey on deep learning in medical image analysis. Medical Image Analysis, 42: 60–88.

[8] Myronenko, A. (2018) 3D MRI brain tumor segmentation using autoencoder regularization. In: MICCAI BraTS Challenge. Granada, Spain. pp. 89–101.

[9] Işın, A., Direkoğlu, C., Şah, M. (2021) Review of medical image classification using deep learning. Computerized Medical Imaging and Graphics, 88: 101820.

[10] Gu, Z., Cheng, J., Fu, H., et al. (2022) CE-Net: Context encoder network for 2D medical image segmentation. IEEE Trans. Med. Imaging, 41(1): 24–34.

[11] Chen, X., Wang, Y., Zhang, Y., et al. (2021) Meta-learning with radiological priors for few-shot chest X-ray classification. Nature Machine Intelligence, 3(7): 555–563.

[12] Wang, L., Zhao, X., Li, J., et al. (2022) Anatomic-aware meta-learning for lung nodule classification. Medical Image Analysis, 77: 102351.

[13] Zhang, J., Wang, Y., Liu, S., et al. (2021) SCA-Net: Spatial-channel attention network for AD classification. In: Information Processing in Medical Imaging. Lisbon, Portugal. pp. 234–246.

[14] Liu, C., Wang, Z., Zhang, H., et al. (2022) Anatomy-aware deformable attention for liver tumor segmentation. In: IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops. New Orleans, LA. pp. 1024–1035.

[15] Xu, L., Yu, H., Wang, T., et al. (2023) Homomorphic attention transformation for stable few-shot medical classification. In: Advances in Neural Information Processing Systems. New Orleans, LA. pp. 27890–27903.

[16] Sudre, C.H., Li, W., Vercauteren, T., et al. (2017) Generalised Dice overlap as a deep learning loss function for highly unbalanced segmentations. Medical Image Analysis, 18: 285–294.

[17] Yan, K., Zhang, Y., Li, W., et al. (2021) Compound Focal Loss: Tackling Class Imbalance in Medical Imaging. IEEE Trans. Med. Imaging, 40(12): 3641–3651.

[18] Hu, J., Shen, L., Sun, G. (2018) Squeeze-and-Excitation Networks. In: IEEE Conference on Computer Vision and Pattern Recognition. Salt Lake City, UT. pp. 7132–7141.

[19] Lin, T.-Y., Goyal, P., Girshick, R., et al. (2017) Focal Loss for Dense Object Detection. In: IEEE International Conference on Computer Vision. Venice, Italy. pp. 2980–2988.

[20] Finn, C., Abbeel, P., Levine, S. (2017) Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks. In: International Conference on Machine Learning. Sydney, Australia. pp. 1126–1135.

[21] Shorten, C., Khoshgoftaar, T.M. (2019) A survey on Image Data Augmentation for Deep Learning. Journal of Big Data, 6(1): 1–48.

[22] Tustison, N.J., Avants, B.B., Cook, P.A., et al. (2010) N4ITK: Improved N3 Bias Correction. IEEE Trans. Med. Imaging, 29(6): 1310–1320.

[23] Simard, P.Y., Steinkraus, D., Platt, J.C. (2003) Best Practices for Convolutional Neural Networks Applied to Visual Document Analysis. In: International Conference on Document Analysis and Recognition. Edinburgh, UK. pp. 958–962.

[24] Shen, D., Wu, G., Suk, H.-I. (2017) Deep Learning in Medical Image Analysis. Annual Review of Biomedical Engineering, 19: 221–248.

[25] Akkus, Z., Galimzianova, A., Hoogi, A., et al. (2017) Deep learning for brain MRI segmentation: state of the art and future directions. Journal of Digital Imaging, 30(4): 449–459.