Accurate medical image segmentation is crucial for clinical diagnosis and treatment planning. However, class imbalance and vagueness of boundary in medical images make it challenging to achieve accurate and precise results. In particular, 3D multi-organ segmentation is a complex process. These challenges are further exacerbated in semi-supervised learning settings with limited labeled data. Existing methods rarely effectively incorporate boundary information to alleviate class imbalance, leading to biased predictions and suboptimal segmentation accuracy. To address these limitations, we propose DBANet, a dual-model framework integrating three key modules. The Confidence-Guided Pseudo-Label Fusion (CPF) module enhances pseudo-label reliability by selecting high-confidence logits. This improves training stability in limited annotation settings. The Boundary Distribution Awareness (BDA) module dynamically adjusts class weights based on boundary distributions, alleviating class imbalance and enhancing segmentation performance. Additionally, the Boundary Vagueness Awareness (BVA) module further refines boundary delineation by prioritizing regions with blurred boundaries. Experiments on two benchmark datasets validate the effectiveness of DBANet. On the Synapse dataset, DBANet achieves average Dice score improvements of 3.56%, 2.17%, and 5.12% under 10%, 20%, and 40% labeled data settings, respectively. Similarly, on the WORD dataset, DBANet achieves average Dice score improvements of 1.72%, 0.97%, and 0.65% under 2%, 5%, and 10% labeled data settings, respectively. These results highlight the potential of boundary-aware adaptive weighting for advancing semi-supervised medical image segmentation.

Recently, virtual staining has emerged as a promising alternative to revolutionize histological staining by digitally generating stains. However, most existing methods suffer from the curse of staining unreality and unreliability. In this paper, we propose the Orthogonal Decoupling Alignment Generative Adversarial Network (ODA-GAN) for unpaired virtual immunohistochemistry (IHC) staining. Our approach is based on the assumption that an image consists of IHC staining-related features, which influence staining distribution and intensity, and staining-unrelated features, such as tissue morphology. Leveraging a pathology foundation model, we first develop a weakly-supervised segmentation pipeline as an alternative to expert annotations. We introduce an Orthogonal MLP (O-MLP) module to project image features into an orthogonal space, decoupling them into staining-related and unrelated components. Additionally, we propose a Dual-stream PatchNCE (DPNCE) loss to resolve contrastive learning contradictions in the staining-related space, thereby enhancing staining accuracy. To further improve realism, we introduce a Multi-layer Domain Alignment (MDA) module to bridge the domain gap between generated and real IHC images. Evaluations on three benchmark datasets show that our ODA-GAN reaches state-of-the-art (SOTA) performance. Our source code is available at https://github.com/ittong/ODA-GAN.

Classifying the status of NSCLC PD-L1 on chest CT is a cost-effective and non-invasive method. The existing multiple instance learning (MIL) methods are not effective for this task, due to the lack of an efficient feature encoder for 3D instances and ignoring the importance of representative instance selection. Thus, they cannot capture weak visual cues related to PD-L1 status on CT images. To address this, we propose a 2.5D top-K ranked multiple instance learning method. We design a 2.5D instance feature encoder, which takes advantage of knowledge from a 2D pre-trained model and has trainable parameters to learn information for 3D instances. In addition, we design a top-K ranked multiple instance learning strategy, which fully exploits the bag-level labels to select representative instances to eliminate the effect of atypical instances and guide the network to learn effective information. We demonstrate that our method can not only outperform state-of-the-art MIL methods on the PD-L1 status classification but also generalize well on a COVID-19 classification task.