Weakly supervised learning for multi-class medical image segmentation via feature decomposition.

Weakly supervised learning with image-level labels, releasing deep learning from highly labor-intensive pixel-wise annotation, has gained great attention for medical image segmentation. However, existing weakly supervised methods are mainly designed for single-class segmentation while leaving multi-class medical image segmentation rarely-explored. Different from natural images, label symbiosis, together with location adjacency, are much more common in medical images, making it more challenging for multi-class segmentation. In this paper, we propose a novel weakly supervised learning method for multi-class medical image segmentation with image-level labels. In terms of the multi-class classification backbone, a multi-level classification network encoding multi-scale features is proposed to produce binary predictions, together with the corresponding CAMs, of each class separately. To address the above issues (i.e., label symbiosis and location adjacency), a feature decomposition module based on semantic affinity is first proposed to learn both class-independent and class-dependent features by maximizing the inter-class feature distance. Through a cross-guidance loss to jointly utilize the above features, label symbiosis is largely alleviated. In terms of location adjacency, a mutually exclusive loss is constructed to minimize the overlap among regions corresponding to different classes. Experimental results on three datasets demonstrate the superior performance of the proposed weakly-supervised framework for both single-class and multi-class medical image segmentation. We believe the analysis in this paper would shed new light on future work for multi-class medical image segmentation. The source code of this paper is publicly available at https://github.com/HustAlexander/MCWSS.

HCN1 pathogenic variants associated with childhood epilepsy in a cohort of Chinese patients.

OBJECTIVE: HCN ion channel family has a widespread expression in neurons, and recently, increasing studies have demonstrated their roles in epilepsies. METHODS: Clinical data of the patients were gathered in a retrospective study. Exon sequencing was used for the patients with unexplained recurrent seizures and varying levels of developmental delay. RESULTS: In this study, eight de novo variants of HCN1 genes were uncovered in eight patients, including six missense variants, one nonsense variant and one frameshift insertion variant; five of them were reported for the first time. The onset age for eight patients ranges from one month to one year. Their main clinical manifestations are epilepsy and varying degrees of developmental delay, and the main type of seizure is focal secondary generalized tonic-clonic seizure. Importantly, in our study, one case presented with a form of migrating focal seizure that has not been reported in the literature. Seizures from five of the eight children were effectively controlled with antiepileptic drugs including valproic acid, levetiracetam and oxcarbazepine. One child developed normally and four children developed mild delay. One child was treated with topiramate, and the convulsion was partially controlled and showed moderate to severe developmental delay. The antiepileptic treatment failed for the other two children, and the two children were treated with sodium valproate, oxcarbazepine, lamotrigine, chlorbazan, levetiracetam and nitrodiazepam successively, but their convulsions were not controlled and showed moderate to severe developmental delay. SIGNIFICANCE: Our research reported eight variants in HCN1 gene causing epilepsy; among these variants, five variants were never reported before. HCN1-related epilepsy usually starts infantile period, and focal secondary generalized tonic-clonic seizure is the most common seizure type. Importantly, we reported the case with migrating focal seizure was rarely reported. Our study expanded both genotype and phenotype for HCN1-related epilepsy.

NEDD4-2 and the CLC-2 channel regulate neuronal excitability in the pathogenesis of mesial temporal lobe epilepsy.

An increasing number of studies have focused on the role of NEDD4-2 in regulating neuronal excitability and the mechanism of epilepsy. However, the exact mechanism has not yet been elucidated. Here, we explored the roles of NEDD4-2 and the CLC-2 channel in regulating neuronal excitability and mesial temporal lobe epilepsy (MTLE) pathogenesis. First, chronic MTLE models were induced by lithium-pilocarpine in developmental rats. Coimmunoprecipitation analysis revealed that the interaction between CLC-2 and NEDD4-2. Western blot analyses indicated that NEDD4-2 expression was downregulated, while phosphorylated (P-) NEDD4-2 and CLC-2 expression was upregulated in adult MTLE rats. Then, the primary hippocampal neuronal cells were isolated and cultured, and the NEDD4-2 was knocked down by shRNA vector, resulting in decreased protein levels of CLC-2. While CLC-2 absence caused increased NEDD4-2 in cells. Next, in an epileptic cell model induced by a Mg2+-free culture, whole-cell current-clamp recording demonstrated that NEDD4-2 deficiency inhibited the spontaneous action potentials of cells, and CLC-2 absence caused more significant decrease in the spontaneous action potentials of cells. In conclusion, we herein revealed that NEDD4-2 regulates the expression of CLC-2, which is involved in neuronal excitability, and participates in the pathogenesis of MTLE.

Clinical characteristics, immunological alteration and distinction of MOG-IgG-associated disorders and GFAP-IgG-associated disorders.

The classification of autoimmune encephalitis (AE) is based on the presence of different types of antibodies. Currently, the clinical manifestations and treatment regimens of patients with all types of AE exhibit similarities. However, the presence of immunological distinctions among different types of AE remains uncertain. In this study, we prospectively collected clinical data, as well as blood and cerebrospinal fluid (CSF) samples from patients diagnosed with MOG antibody-associated disease (MOGAD) or GFAP astrocytopathy (GFAP-A), in order to assess changes in inflammatory biomarkers such as immunoglobulin oligoclonal bands, cytokines in serum and CSF, as well as peripheral blood lymphocyte subtypes within different subsets. To further distinguish the immune response in patients with MOGAD and GFAP-A from that of healthy individuals, we prospectively recruited 20 hospitalized patients diagnosed with AE. Among them, 15 (75%) tested positive for MOG antibodies, 4 (20%) tested positive for GFAP antibodies, and 1 (5%) tested positive for both MOG and GFAP antibodies. These patients were then followed up for a period of 18 months. Compared to healthy controls (HC), AE patients exhibited elevated levels of MIP-1beta, SDF-1alpha, IL-12p70, IL-5, IL-1RA, IL-8 and decreased levels of IL-23, IL-31, IFN-alpha, IL-7, TNF-beta and TNF-alpha in serum. The CSF of AE patients showed increased levels of IL-1RA, IL-6 and IL-2 while decreased levels of RANTES, IL-18,IL-7,TNF-beta,TNF-alpha,RANTES,Eotaxin,and IL-9. The level of MCP-1 in the CSF of GFAP-A patients was found to be lower compared to that of MOGAD patients, while RANTES levels were higher. And the levels of IL-17A, Eotaxin, GRO-alpha, IL-8, IL-1beta, MIP-1beta were higher in the CSF of patients with epilepsy. The presence of intrathecal immune responses is also observed in patients with spinal muscular atrophy (SMA). However, no biomarker was found to be associated with disease severity in patients with AE. Among the 17 patients, recovery was observed, while 2 patients experienced persistent symptoms after an 18-month follow-up period. Additionally, within one year of onset, 8 patients had a single recurrence. Therefore, the immunological profiles of MOGAD and GFAP-A patients differ from those of normal individuals, and the alterations in cytokine levels may also exhibit a causal association with the clinical presentations, such as seizure.

Cluster-Re-Supervision: Bridging the Gap Between Image-Level and Pixel-Wise Labels for Weakly Supervised Medical Image Segmentation.

Weakly supervised learning, releasing deep learning from highly labor-intensive pixel-wise annotations, has gained great attention, especially for medical image segmentation. With only image-level labels, pixel-wise segmentation/localization usually is achieved based on class activation maps (CAMs) containing the most discriminative regions. One common consequence of CAM-based approaches is incomplete foreground segmentation, i.e. under-segmentation/false negatives. Meanwhile, suffering from relatively limited medical imaging data, class-irrelevant tissues can hardly be suppressed during classification, resulting in incorrect background identification, i.e. over-segmentation/false positives. The above two issues are determined by the loose-constraint nature of image-level labels penalizing on the entire image space, and thus how to develop pixel-wise constraints based on image-level labels is the key for performance improvement which is under-explored. In this paper, based on unsupervised clustering, we propose a new paradigm called cluster-re-supervision to evaluate the contribution of each pixel in CAMs to final classification and thus generate pixel-wise supervision (i.e., clustering maps) for CAMs refinement on both over- and under-segmentation reduction. Furthermore, based on self-supervised learning, an inter-modality image reconstruction module, together with random masking, is designed to complement local information in feature learning which helps stabilize clustering. Experimental results on two popular public datasets demonstrate the superior performance of the proposed weakly-supervised framework for medical image segmentation. More importantly, cluster-re-supervision is independent of specific tasks and highly extendable to other applications.

Fracture R-CNN: An anchor-efficient anti-interference framework for skull fracture detection in CT images.

BACKGROUND: Skull fracture, as a common traumatic brain injury, can lead to multiple complications including bleeding, leaking of cerebrospinal fluid, infection, and seizures. Automatic skull fracture detection (SFD) is of great importance, especially in emergency medicine. PURPOSE: Existing algorithms for SFD, developed based on hand-crafted features, suffer from low detection accuracy due to poor generalizability to unseen samples. Deploying deep detectors designed for natural images like Faster Region-based Convolutional Neural Network (R-CNN) for SFD can be helpful but are of high redundancy and with nonnegligible false detections due to the cranial suture and skull base interference. Therefore, we, for the first time, propose an anchor-efficient anti-interference deep learning framework named Fracture R-CNN for accurate SFD with low computational cost. METHODS: The proposed Fracture R-CNN is developed by incorporating the prior knowledge utilized in clinical diagnosis into the original Faster R-CNN. Specifically, based on the distributions of skull fractures, we first propose an adaptive anchoring region proposal network (AA-RPN) to generate proposals for diverse-scale fractures with low computational complexity. Then, based on the prior knowledge that cranial sutures exist in the junctions of bones and usually contain sclerotic margins, we design an anti-interference head (A-Head) network to eliminate the cranial suture interference for better SFD detection. In addition, to further enhance the anti-interference ability of the proposed A-Head, a difficulty-balanced weighted loss function is proposed to emphasize more on distinguishing the interference areas from the skull base and the cranial sutures during training. RESULTS: Experimental results demonstrate that the proposed Fracture R-CNN outperforms the current state-of-the-art (SOTA) deep detectors for SFD with a higher recall and fewer false detections. Compared to Faster R-CNN, the proposed Fracture R-CNN improves the average precision (AP) by 11.74% and the free-response receiver operating characteristic (FROC) score by 11.08%. Through validating on various backbones, we further demonstrate the architecture independence of Fracture R-CNN, making it extendable to other detection applications. CONCLUSIONS: As the customized deep learning-based framework for SFD, Fracture R-CNN can effectively overcome the unique challenges in SFD with less computational cost, leading to a better detection performance compared to the SOTA deep detectors. Moreover, we believe the prior knowledge explored for Fracture R-CNN would shed new light on future deep learning approaches for SFD.

Uncertainty-Aware Deep Learning With Cross-Task Supervision for PHE Segmentation on CT Images.

Perihematomal edema (PHE) volume, surrounding spontaneous intracerebral hemorrhage (SICH), is an important biomarker for the presence of SICH-associated diseases. However, due to irregular shapes and extremely low contrast of PHE on CT images, manually annotating PHE in pixel-wise is time-consuming and labour intensive even for experienced experts, which makes it almost infeasible to deploy current supervised deep learning approaches for automated PHE segmentation. How to develop annotation-efficient deep learning to achieve accurate PHE segmentation is an open problem. In this paper, we, for the first time, propose a cross-task supervised framework by introducing slice-level PHE labels and pixel-wise SICH annotations, which are more accessible in clinical scenarios compared to pixel-wise PHE annotations. Specifically, we first train a multi-level classifier based on slice-level PHE labels to produce high-quality class activation maps (CAMs) as pseudo PHE annotations. Then, we train a deep learning model to produce accurate PHE segmentation by iteratively refining the pseudo annotations via an uncertainty-aware corrective training strategy for noise removal and a distance-aware loss for background compression. Experimental results demonstrate that, the proposed framework achieves a comparative performance with the fully supervised methods on PHE segmentation, and largely improves the baseline performance where only pseudo PHE labels are used for training. We believe the findings from this study of using cross-task supervision for annotation-efficient deep learning can be applied to other medical imaging applications.

Symmetry-Aware Deep Learning for Cerebral Ventricle Segmentation With Intra-Ventricular Hemorrhage.

Cerebral ventricles are one of the prominent structures in the brain, segmenting which can provide rich information for brain-related disease diagnosis. Unfortunately, cerebral ventricle segmentation in complex clinical cases, such as in the coexistence with other lesions/hemorrhages, remains unexplored. In this paper, we, for the first time, focus on cerebral ventricle segmentation with the presence of intra-ventricular hemorrhages (IVH). To overcome the occlusions formed by IVH, we propose a symmetry-aware deep learning approach inspired by contrastive self-supervised learning. Specifically, for each slice, we jointly employ the raw slice and the horizontally flipped slice as inputs and penalize the consistency loss between the corresponding segmentation maps in addition to their segmentation losses. In this way, the symmetry of cerebral ventricles is enforced to eliminate the occlusions brought by IVH. Extensive experimental results show that the proposed symmetry-aware deep learning approach achieves consistent performance improvements for ventricle segmentation in both normal (i.e. without IVH) and challenging cases (i.e. with IVH). Through evaluation of multiple backbone networks, we demonstrate the architecture-independence of the proposed approach for performance improvements. Moreover, we re-design an end-to-end version of symmetry-aware deep learning, making it more extendable to other approaches for brain-related analysis.

Ψ-Net: Focusing on the border areas of intracerebral hemorrhage on CT images.

BACKGROUND AND OBJECTIVE: The volume of the intracerebral hemorrhage (ICH) obtained from CT scans is essential for quantification and treatment planning. However,a fast and accurate volume acquisition brings great challenges. On the one hand, it is both time consuming and operator dependent for manual segmentation, which is the gold standard for volume estimation. On the other hand, low contrast to normal tissues, irregular shapes and distributions of the hemorrhage make the existing automatic segmentation methods hard to achieve satisfactory performance. METHOD: To solve above problems, a CNN-based architecture is proposed in this work, consisting of a novel model, which is named as Ψ-Net and a multi-level training strategy. In the structure of Ψ-Net, a self-attention block and a contextual-attention block is designed to suppresses the irrelevant information and segment border areas of the hemorrhage more finely. Further, an multi-level training strategy is put forward to facilitate the training process. By adding the slice-level learning and a weighted loss, the multi-level training strategy effectively alleviates the problems of vanishing gradient and the class imbalance. The proposed training strategy could be applied to most of the segmentation networks, especially for complex models and on small datasets. RESULTS: The proposed architecture is evaluated on a spontaneous ICH dataset and a traumatic ICH dataset. Compared to the previous works on the ICH sementation, the proposed architecture obtains the state-of-the-art performance(Dice of 0.950) on the spontaneous ICH, and comparable results(Dice of 0.895) with the best method on the traumatic ICH. On the other hand, the time consumption of the proposed architecture is much less than the previous methods on both training and inference. Morever, experiment results on various of models prove the universality of the multi-level training strategy. CONCLUSIONS: This study proposed a novel CNN-based architecture, Ψ-Net with multi-level training strategy. It takes less time for training and achives superior performance than previous ICH segmentaion methods.