Zhaoshuo Diao

Accurate tumor segmentation in Positron Emission Tomography/Computed Tomography (PET/CT) is essential for clinical diagnosis, staging, and treatment planning. However, the reliance on time-consuming pixel-level annotations limits the practical deployment of deep learning-based segmentation methods. Weakly supervised segmentation using image-level classification labels provides a promising alternative, yet generating accurate supervision remains highly challenging. Specifically, PET provides critical metabolic information but is susceptible to physiological hypermetabolism and background noise, while CT offers anatomical structure but suffers from low soft-tissue contrast and boundary uncertainty. Consequently, static or single-modality pseudo-labels derived from Class Activation Maps (CAMs) often suffer from boundary ambiguity and false-positive activations, failing to reliably guide segmentation. To address the core problem of dynamically evaluating and correcting supervision reliability, we propose a Progressive Reliability-Guided Refinement (PRGR) framework. Specifically, pixel-level supervision reliability is modeled by integrating PET metabolic consistency and CT structural gradients, leveraging complementary cross-modal information without introducing additional trainable parameters. Based on the estimated reliability, the image domain is adaptively partitioned into regions with different supervision credibility levels, enabling progressive optimization with differentiated learning constraints. Furthermore, pseudo-labels are iteratively updated using reliability estimates and segmentation predictions to progressively correct erroneous supervision. Experiments on two public PET/CT datasets demonstrate that our method effectively suppresses false positives and improves boundary delineation. Quantitatively, the proposed PRGR achieves Dice scores of 0.673 and 0.596 on the STS and HeadNeck datasets, respectively, yielding absolute improvements of 1.8% and 8.4% over the second-best baseline. Furthermore, it achieves superior Precision scores of 0.776 and 0.684, demonstrating its effectiveness in narrowing the performance gap between weakly and fully supervised segmentation while reducing the dependence on pixel-level annotations.

Background and Objective: When hypertensive cerebral hemorrhage causes secondary intraventricular hemorrhage, it usually significantly increases the complexity of the patient’s condition and the risk of poor prognosis. Deep learning methods can automatically and quickly segment intraparenchymal hemorrhage and intraventricular hemorrhage, and quantitatively analyze hematoma-related properties to provide auxiliary information for subsequent diagnosis and treatment. Methods: We retrospectively enrolled 351 patients with hypertensive cerebral hemorrhage, 219 of whom had secondary intraventricular hemorrhage. All patients underwent computed tomography within 1 week after diagnosis. Based on 3D U-Net, we developed a deep learning network with a multi-scale deformable convolution module and a softened anatomical consistency loss. The multi-scale deformable convolution module can enhance the learning ability of multi-deformation features and increase the receptive field of the network. The anatomical consistency loss, built upon softened labels, can alleviate the impact of label noise. Results: We evaluated our model at pixel, volume, and morphology levels. It achieved Dice of 0.8990 ± 0.1169 for intraparenchymal hemorrhage and 0.7124 ± 0.1227 for intraventricular hemorrhage, both higher than that of the comparison model. Compared with the Coniglobus method, our model has a narrower consistency limit and more concentrated predicted values. Additionally, in segmenting irregular and different-sized hematomas, it generates the smallest centroid, volume, position, and morphology deviations. Conclusions: The proposed model can automatically and accurately segment two types of hematomas and quantify multiple attributes, is robust in multi-deformation and label noise scenarios, and has the potential to assist in clinical diagnosis and treatment; its actual impact on clinical decision-making and patient outcomes requires prospective validation.

Objective. Automatic segmentation of head and neck tumours from PET/CT images is important for radiotherapy planning and related image-guided workflows, but remains challenging because PET and CT provide complementary yet heterogeneous information. Existing fusion strategies often do not explicitly model metabolic hotspots in PET and anatomical boundary information in CT, which can impair the delineation of small lesions and lesions with indistinct boundaries. Approach. We propose CED-Diff, a diffusion-based tumour segmentation framework that integrates differentiated feature extraction (DFE) and task-oriented auxiliary supervision (TAS). DFE explicitly models metabolic activity in PET and anatomical boundary information in CT to derive more discriminative features for segmentation, while TAS provides auxiliary constraints to improve feature learning. Main results. On the HeadNeck dataset, CED-Diff achieved the highest Dice score among the evaluated methods, with a Dice score of 66.28%, along with a Sensitivity of 78.70% and a Precision of 69.08%. Significance. These results suggest that explicit modelling of complementary PET and CT information can improve the accuracy and robustness of automatic head and neck tumour segmentation.

Lung cancer is the leading cause of cancer-related mortality, with small cell lung cancer and non-small cell lung cancer being the primary subtypes that exhibit distinct treatment approaches and prognostic outcomes. Accurate identification of these lung cancer classes holds significant importance in clinical practice. This study introduces an experience-driven interpretable multi-task network to concurrently perform segmentation and classification of small cell and non-small cell lung cancer. The core architecture of this multi-task model is based on StarNet, featuring a shared feature extraction branch and task-specific decoding branches for tumor segmentation and classification. Leveraging clinical knowledge of small cell lung cancer characteristics, such as indistinct edges, tissue invasion, and limited large cavity areas, two auxiliary branches are proposed: edge uncertainty estimation and tumor core area reconstruction. The values from edge uncertainty estimation and reconstruction integrity estimation are utilized in the classification branch to facilitate small cell lung cancer classification. Furthermore, for enhanced interpretability, bottleneck layer features are extracted for comparative learning, and a three-level contrastive loss is proposed to improve the differentiation of disease features. Lastly, an interpretable strategy based on trained feature query matching is presented, providing radiologists with clinical insights and reference images while the model outputs recognition predictions. Experimental results on the public dataset demonstrate that the proposed multi-task model not only outperforms single-task models but also offers a certain level of interpretability, thus enhancing radiologists' clinical decision-making processes.

Background: Cancer metastasis usually means that cancer cells spread to other tissues or organs, and the condition worsens. Identifying whether cancer has metastasized can help doctors infer the progression of a patient's condition and is an essential prerequisite for devising treatment plans. Fluorine 18 fluorodeoxyglucose positron emission tomography/computed tomography ( ¹⁸F -FDG PET/CT) is an advanced cancer diagnostic imaging technique that provides both metabolic and structural information. Method: In cancer metastasis recognition tasks, effectively integrating metabolic and structural information stands as a key technology to enhance feature representation and recognition performance. This paper proposes a cancer metastasis identification network based on dynamic coordinated metabolic attention and structural attention to address these challenges. Specifically, metabolic and structural features are extracted by incorporating a dynamic coordinated attention module (DCAM) into two branches of ResNet networks, thereby amalgamating high metabolic spatial information from PET images with texture structure information from CT images, and dynamically adjusting this process through iterations. Discussion: Next, to improve the efficacy of feature expression, a multi-receptive field feature fusion module (MRFM) is included in order to execute multi-receptive field fusion of semantic features. Result: To validate the effectiveness of our proposed model, experiments were conducted on both a private lung lymph nodes dataset and a public soft tissue sarcomas dataset. Conclusion: The accuracy of our method reached 76.0% and 75.1% for the two datasets, respectively, demonstrating an improvement of 6.8% and 5.6% compared to ResNet, thus affirming the efficacy of our method.

Accurate segmentation of endoscopic instruments is essential in robot-assisted surgery, supporting precise navigation, enhancing safety, and advancing surgical automation. However, this task is challenging due to factors like complex environments, instrument-tissue similarity, and lighting variations. Instruments, due to their material properties, have distinct depth distributions compared to surrounding tissues. This aspect is often overlooked in monocular video segmentation methods.To address this issue, we propose EISegNet, a multi-task framework that prioritizes instrument segmentation with an auxiliary disparity estimation task. The framework integrates an asymmetric cross-attention mechanism to enhance segmentation performance by fusing features from both tasks. Moreover, by leveraging the geometric properties of motion, EISegNet adapts the stereo disparity estimation strategy for dual-view depth estimation, broadening its applicability to various endoscopic surgeries beyond laparoscopic procedures. Furthermore, EISegNet incorporates a Gaussian-weighted loss function to emphasize edge features, which are particularly challenging for disparity estimation. This function reduces overall loss and improves segmentation accuracy. Extensive cross-dataset experiments demonstrate the superior accuracy and generalization of our method, achieving a 5.97% increase in IoU (Intersection over Union). Qualitative evaluations on clinical datasets further demonstrate the promising performance in real-world scenarios. The code and private dataset will be available at the following link: https://github.com/ILSR-SIA/EISegNet

Cervical cancer is a malignant tumor that endangers women's life and health. While deep learning has enhanced the accuracy of cervical cell classification, there remain obstacles impeding further performance enhancement, including the similarities between different categories, variability between single cells and cell clusters, as well as the accuracy of annotations. To address these issues, a novel multi-task collaborative framework for cervical cell classification is proposed. Specifically, to solve the similarity between different categories, we propose a grouping cell contrast auxiliary branch, which divides cervical cells into different groups and utilizes supervised contrastive learning to learn representative feature between different categories. And we introduce a multi-level cell classification auxiliary branch that simultaneously performs 5-class, 3-class, and 2-class classification tasks, and explicitly constrains the inter-class relationship learning of cervical cells. Furthermore, to solve the variations within the same category of single cells and cell clusters, we propose an image reconstruction auxiliary branch, which encourages the model to learn more contextual features. Finally, to solve subjectivity and accuracy of annotations, we introduce a soft label distillation auxiliary branch, which constrains the consistency of probability distributions between the encoder and the momentum encoder. It is worth noting that these auxiliary branches only work during training and will not add additional computational consumption during inference. We validate on the HSJCC, DSCC and SIPaKMeD datasets. Compared to existing methods, our approach has achieved outstanding performance and effectively mitigates the issues raised, demonstrating its effectiveness in automated cervical cell classification.