Deep learning methods have shown great potential in processing multi-modal Magnetic Resonance Imaging (MRI) data, enabling improved accuracy in brain tumor segmentation. However, the performance of these methods can suffer when dealing with incomplete modalities, which is a common issue in clinical practice. Existing solutions, such as missing modality synthesis, knowledge distillation, and architecture-based methods, suffer from drawbacks such as long training times, high model complexity, and poor scalability. This paper proposes IMSTrans, a novel lightweight scalable Swin Transformer network by utilizing a single encoder to extract latent feature maps from all available modalities. This unified feature extraction process enables efficient information sharing and fusion among the modalities, resulting in efficiency without compromising segmentation performance even in the presence of missing modalities. Two datasets, BraTS 2018 and BraTS 2020, containing incomplete modalities for brain tumor segmentation are evaluated against popular benchmarks. On the BraTS 2018 dataset, our model achieved higher average Dice similarity coefficient (DSC) scores for the whole tumor, tumor core, and enhancing tumor regions (86.57, 75.67, and 58.28, respectively), in comparison with a state-of-the-art model, i.e. mmFormer (86.45, 75.51, and 57.79, respectively). Similarly, on the BraTS 2020 dataset, our model scored higher DSC scores in these three brain tumor regions (87.33, 79.09, and 62.11, respectively) compared to mmFormer (86.17, 78.34, and 60.36, respectively). We also conducted a Wilcoxon test on the experimental results, and the generated -value confirmed that our model’s performance was statistically significant. Moreover, our model exhibits significantly reduced complexity with only 4.47 M parameters, 121.89G FLOPs, and a model size of 77.13 MB, whereas mmFormer comprises 34.96 M parameters, 265.79 G FLOPs, and a model size of 559.74 MB. These indicate our model, being light-weighted with significantly reduced parameters, is still able to achieve better performance than a state-of-the-art model. By leveraging a single encoder for processing the available modalities, IMSTrans offers notable scalability advantages over methods that rely on multiple encoders. This streamlined approach eliminates the need for maintaining separate encoders for each modality, resulting in a lightweight and scalable network architecture. The source code of IMSTrans and the associated weights are both publicly available at .

中文翻译：

---

可扩展的 Swin Transformer 网络用于从不完整的 MRI 模式中分割脑肿瘤

深度学习方法在处理多模态磁共振成像（MRI）数据方面显示出巨大潜力，可以提高脑肿瘤分割的准确性。然而，在处理不完整的模式时，这些方法的性能可能会受到影响，这是临床实践中的常见问题。现有的解决方案，例如缺少模态合成、知识蒸馏和基于架构的方法，都存在训练时间长、模型复杂性高和可扩展性差等缺点。本文提出了 IMSTrans，这是一种新颖的轻量级可扩展 Swin Transformer 网络，它利用单个编码器从所有可用模态中提取潜在特征图。这种统一的特征提取过程可以实现模态之间的有效信息共享和融合，即使在存在模态缺失的情况下，也能在不影响分割性能的情况下提高效率。根据流行的基准对包含不完整脑肿瘤分割模式的两个数据集 BraTS 2018 和 BraTS 2020 进行了评估。在 BraTS 2018 数据集上，与现有模型相比，我们的模型在整个肿瘤、肿瘤核心和增强肿瘤区域获得了更高的平均 Dice 相似系数 (DSC) 得分（分别为 86.57、75.67 和 58.28）。 -艺术模型，即mmFormer（分别为86.45、75.51和57.79）。同样，在 BraTS 2020 数据集上，与 mmFormer（分别为 86.17、78.34 和 60.36）相比，我们的模型在这三个脑肿瘤区域（分别为 87.33、79.09 和 62.11）获得了更高的 DSC 分数。我们还对实验结果进行了 Wilcoxon 测试，生成的 值证实了我们的模型的性能具有统计显着性。此外，我们的模型的复杂性显着降低，只有 4.47 M 参数、121.89G FLOP 和 77.13 MB 的模型大小，而 mmFormer 包含 34.96 M 参数、265.79 G FLOP 和 559.74 MB 的模型大小。这些表明我们的模型是轻量级的，参数显着减少，仍然能够实现比最先进的模型更好的性能。通过利用单个编码器来处理可用模态，IMSTrans 比依赖多个编码器的方法具有显着的可扩展性优势。这种简化的方法无需为每种模态维护单独的编码器，从而形成轻量级且可扩展的网络架构。IMSTrans 的源代码和相关权重均可在 上公开获取。