Multi-scale window transformer for cervical cytopathology image recognition.

Cervical cancer is a major global health issue, particularly in developing countries where access to healthcare is limited. Early detection of pre-cancerous lesions is crucial for successful treatment and reducing mortality rates. However, traditional screening and diagnostic processes require cytopathology doctors to manually interpret a huge number of cells, which is time-consuming, costly, and prone to human experiences. In this paper, we propose a Multi-scale Window Transformer (MWT) for cervical cytopathology image recognition. We design multi-scale window multi-head self-attention (MW-MSA) to simultaneously integrate cell features of different scales. Small window self-attention is used to extract local cell detail features, and large window self-attention aims to integrate features from smaller-scale window attention to achieve window-to-window information interaction. Our design enables long-range feature integration but avoids whole image self-attention (SA) in ViT or twice local window SA in Swin Transformer. We find convolutional feed-forward networks (CFFN) are more efficient than original MLP-based FFN for representing cytopathology images. Our overall model adopts a pyramid architecture. We establish two multi-center cervical cell classification datasets of two-category 192,123 images and four-category 174,138 images. Extensive experiments demonstrate that our MWT outperforms state-of-the-art general classification networks and specialized classifiers for cytopathology images in the internal and external test sets. The results on large-scale datasets prove the effectiveness and generalization of our proposed model. Our work provides a reliable cytopathology image recognition method and helps establish computer-aided screening for cervical cancer. Our code is available at https://github.com/nmyz669/MWT, and our web service tool can be accessed at https://huggingface.co/spaces/nmyz/MWTdemo.

Anatomical Evidence for Parasympathetic Innervation of the Renal Vasculature and Pelvis.

BACKGROUND: The kidneys critically contribute to body homeostasis under the control of the autonomic nerves, which enter the kidney along the renal vasculature. Although the renal sympathetic and sensory nerves have long been confirmed, no significant anatomic evidence exists for renal parasympathetic innervation. METHODS: We identified cholinergic nerve varicosities associated with the renal vasculature and pelvis using various anatomic research methods, including a genetically modified mouse model and immunostaining. Single-cell RNA sequencing (scRNA-Seq) was used to analyze the expression of AChRs in the renal artery and its segmental branches. To assess the origins of parasympathetic projecting nerves of the kidney, we performed retrograde tracing using recombinant adeno-associated virus (AAV) and pseudorabies virus (PRV), followed by imaging of whole brains, spinal cords, and ganglia. RESULTS: We found that cholinergic axons supply the main renal artery, segmental renal artery, and renal pelvis. On the renal artery, the newly discovered cholinergic nerve fibers are separated not only from the sympathetic nerves but also from the sensory nerves. We also found cholinergic ganglion cells within the renal nerve plexus. Moreover, the scRNA-Seq analysis suggested that acetylcholine receptors (AChRs) are expressed in the renal artery and its segmental branches. In addition, retrograde tracing suggested vagus afferents conduct the renal sensory pathway to the nucleus of the solitary tract (NTS), and vagus efferents project to the kidney. CONCLUSIONS: Cholinergic nerves supply renal vasculature and renal pelvis, and a vagal brain-kidney axis is involved in renal innervation.

Brain-wide projection reconstruction of single functionally defined neurons.

Reconstructing axonal projections of single neurons at the whole-brain level is currently a converging goal of the neuroscience community that is fundamental for understanding the logic of information flow in the brain. Thousands of single neurons from different brain regions have recently been morphologically reconstructed, but the corresponding physiological functional features of these reconstructed neurons are unclear. By combining two-photon Ca2+ imaging with targeted single-cell plasmid electroporation, we reconstruct the brain-wide morphologies of single neurons that are defined by a sound-evoked response map in the auditory cortices (AUDs) of awake mice. Long-range interhemispheric projections can be reliably labelled via co-injection with an adeno-associated virus, which enables enhanced expression of indicator protein in the targeted neurons. Here we show that this method avoids the randomness and ambiguity of conventional methods of neuronal morphological reconstruction, offering an avenue for developing a precise one-to-one map of neuronal projection patterns and physiological functional features.

Robust whole slide image analysis for cervical cancer screening using deep learning.

Computer-assisted diagnosis is key for scaling up cervical cancer screening. However, current recognition algorithms perform poorly on whole slide image (WSI) analysis, fail to generalize for diverse staining and imaging, and show sub-optimal clinical-level verification. Here, we develop a progressive lesion cell recognition method combining low- and high-resolution WSIs to recommend lesion cells and a recurrent neural network-based WSI classification model to evaluate the lesion degree of WSIs. We train and validate our WSI analysis system on 3,545 patient-wise WSIs with 79,911 annotations from multiple hospitals and several imaging instruments. On multi-center independent test sets of 1,170 patient-wise WSIs, we achieve 93.5% Specificity and 95.1% Sensitivity for classifying slides, comparing favourably to the average performance of three independent cytopathologists, and obtain 88.5% true positive rate for highlighting the top 10 lesion cells on 447 positive slides. After deployment, our system recognizes a one giga-pixel WSI in about 1.5 min.

UnaG as a reporter in adeno-associated virus-mediated gene transfer for biomedical imaging.

Adeno-associated virus (AAV) is one of the most common gene transfer vectors, but it has a limited capacity. A smaller fluorescent protein is urgently needed since it is more suitable to act as a reporter in AAV. In this study, a bilirubin-dependent reporter smaller than EGFP, termed UnaG, was found to have the ability to label the neurons of a mouse brain as clearly as EGFP without the addition of exogenous bilirubin. We also found that UnaG's pH tolerance is better than that of EGFP; however, its fluorescence recovery after protonated quenching is not as good as that of EGFP. In addition, UnaG preserved its fluorescence better than EGFP in SeeDB clearing. Taken together, this study demonstrates that UnaG can act as a small intrinsically fluorescent reporter in the mouse brain without an additional ligand, thus providing an alternative over EGFP for AAV-mediated neuron labeling in mammals.

PathSRGAN: Multi-Supervised Super-Resolution for Cytopathological Images Using Generative Adversarial Network.

In the cytopathology screening of cervical cancer, high-resolution digital cytopathological slides are critical for the interpretation of lesion cells. However, the acquisition of high-resolution digital slides requires high-end imaging equipment and long scanning time. In the study, we propose a GAN-based progressive multi-supervised super-resolution model called PathSRGAN (pathology super-resolution GAN) to learn the mapping of real low-resolution and high-resolution cytopathological images. With respect to the characteristics of cytopathological images, we design a new two-stage generator architecture with two supervision terms. The generator of the first stage corresponds to a densely-connected U-Net and achieves 4× to 10× super resolution. The generator of the second stage corresponds to a residual-in-residual DenseBlock and achieves 10× to 20× super resolution. The designed generator alleviates the difficulty in learning the mapping from 4× images to 20× images caused by the great numerical aperture difference and generates high quality high-resolution images. We conduct a series of comparison experiments and demonstrate the superiority of PathSRGAN to mainstream CNN-based and GAN-based super-resolution methods in cytopathological images. Simultaneously, the reconstructed high-resolution images by PathSRGAN improve the accuracy of computer-aided diagnosis tasks effectively. It is anticipated that the study will help increase the penetration rate of cytopathology screening in remote and impoverished areas that lack high-end imaging equipment.

From Detection of Individual Metastases to Classification of Lymph Node Status at the Patient Level: The CAMELYON17 Challenge.

Automated detection of cancer metastases in lymph nodes has the potential to improve the assessment of prognosis for patients. To enable fair comparison between the algorithms for this purpose, we set up the CAMELYON17 challenge in conjunction with the IEEE International Symposium on Biomedical Imaging 2017 Conference in Melbourne. Over 300 participants registered on the challenge website, of which 23 teams submitted a total of 37 algorithms before the initial deadline. Participants were provided with 899 whole-slide images (WSIs) for developing their algorithms. The developed algorithms were evaluated based on the test set encompassing 100 patients and 500 WSIs. The evaluation metric used was a quadratic weighted Cohen's kappa. We discuss the algorithmic details of the 10 best pre-conference and two post-conference submissions. All these participants used convolutional neural networks in combination with pre- and postprocessing steps. Algorithms differed mostly in neural network architecture, training strategy, and pre- and postprocessing methodology. Overall, the kappa metric ranged from 0.89 to -0.13 across all submissions. The best results were obtained with pre-trained architectures such as ResNet. Confusion matrix analysis revealed that all participants struggled with reliably identifying isolated tumor cells, the smallest type of metastasis, with detection rates below 40%. Qualitative inspection of the results of the top participants showed categories of false positives, such as nerves or contamination, which could be targeted for further optimization. Last, we show that simple combinations of the top algorithms result in higher kappa metric values than any algorithm individually, with 0.93 for the best combination.

Cell-type-specific and projection-specific brain-wide reconstruction of single neurons.

We developed a dual-adeno-associated-virus expression system that enables strong and sparse labeling of individual neurons with cell-type and projection specificity. We demonstrated its utility for whole-brain reconstruction of midbrain dopamine neurons and striatum-projecting cortical neurons. We further extended the labeling method for rapid reconstruction in cleared thick brain sections and simultaneous dual-color labeling. This labeling system may facilitate the process of generating mesoscale single-neuron projectomes of mammalian brains.

Hematoxylin and eosin staining of intact tissues via delipidation and ultrasound.

Acquiring information on the precise distribution of a tumor is essential to evaluate intratumoral heterogeneity. Conventional hematoxylin and eosin staining, which has been used by pathologists for more than 100 years, is the gold standard of tumor diagnosis. However, it is difficult to stain entire tumor tissues with hematoxylin and eosin and then acquire the three-dimensional distribution of cells in solid tumors due to difficulties in the staining and rinsing. In this paper, we propose a modified hematoxylin and eosin staining method, in which delipidation and ultrasound waves were applied to enhance tissue permeability and accelerate dye diffusion. This improved hematoxylin and eosin staining method is termed iHE (intact tissue hematoxylin and eosin staining). We applied the iHE method to stain intact organs of mice, which were then sectioned and imaged sequentially. The results showed that the whole tissue was stained homogeneously. Combined with micro-optical sectioning tomography (MOST), the iHE method can be used for 3D volume imaging and to evaluate the intratumoral heterogeneity of the entire tumor tissue spatially. Therefore, this method may help to accurately diagnose the invasion stage of tumors and guide clinical treatments.

Rapid imaging of large tissues using high-resolution stage-scanning microscopy.

Rapid and high-resolution imaging of large tissues is essential in biological research, like brain neuron connectivity research and cancer margins imaging. Here a novel stage-scanning confocal microscopy was developed for rapid imaging of large tissues. Line scanning methods and strip imaging strategy were used to increase the imaging speed. The scientific CMOS was used as line detector in sub-array mode and the optical sectioning ability can be easily adjusted by changing the number of line detectors according to different samples. Fluorescent beads imaging showed resolutions of 0.47 µm, 0.56 µm, and 1.56 µm in the X, Y, and Z directions, respectively, with a 40 × objective lens. A 10 × 10 mm(2) coronal plane with enough signal intensity could be imaged in about 88 sec at a sampling resolution of 0.16 µm/pixel. Rapid imaging of mouse brain slices demonstrated the applicability of this system in visualizing neuronal details at high frame rate.