Lung cancer is the leading cause of cancer mortality, and early detection is key to improving survival. However, there are no reliable blood-based tests currently available for early-stage lung cancer diagnosis. Here, we performed single-cell RNA sequencing of different early-stage lung cancers and found that lipid metabolism was broadly dysregulated in different cell types, with glycerophospholipid metabolism as the most altered lipid metabolism–related pathway. Untargeted lipidomics was carried out in an exploratory cohort of 311 participants. Through support vector machine algorithm-based and mass spectrum–based feature selection, we identified nine lipids (lysophosphatidylcholines 16:0, 18:0, and 20:4; phosphatidylcholines 16:0–18:1, 16:0–18:2, 18:0–18:1, 18:0–18:2, and 16:0–22:6; and triglycerides 16:0–18:1–18:1) as the features most important for early-stage cancer detection. Using these nine features, we developed a liquid chromatography–mass spectrometry (MS)–based targeted assay using multiple reaction monitoring. This target assay achieved 100.00% specificity on an independent validation cohort. In a hospital-based lung cancer screening cohort of 1036 participants examined by low-dose computed tomography and a prospective clinical cohort containing 109 participants, the assay reached more than 90.00% sensitivity and 92.00% specificity. Accordingly, matrix-assisted laser desorption/ionization MS imaging confirmed that the selected lipids were differentially expressed in early-stage lung cancer tissues in situ. This method, designated as Lung Cancer Artificial Intelligence Detector, may be useful for early detection of lung cancer or large-scale screening of high-risk populations for cancer prevention.
Cervical cancer, as one of the most frequently diagnosed cancers worldwide, is curable when detected early. Histopathology images play an important role in precision medicine of the cervical lesions. However, few computer aided algorithms have been explored on cervical histopathology images due to the lack of public datasets. In this article, we release a new cervical histopathology image dataset for automated precancerous diagnosis. Specifically, 100 slides from 71 patients are annotated by three independent pathologists. To show the difficulty of the task, benchmarks are obtained through both fully and weakly supervised learning. Extensive experiments based on typical classification and semantic segmentation networks are carried out to provide strong baselines. In particular, a strategy of assembling classification, segmentation, and pseudo-labeling is proposed to further improve the performance. The Dice coefficient reaches 0.7833, indicating the feasibility of computer aided diagnosis and the effectiveness of our weakly supervised ensemble algorithm. The dataset and evaluation codes are publicly available. To the best of our knowledge, it is the first public cervical histopathology dataset for automated precancerous segmentation. We believe that this work will attract researchers to explore novel algorithms on cervical automated diagnosis, thereby assisting doctors and patients clinically.
Thanks to deep convolutional neural networks (CNNs), Brain Tumor Segmentation (BTS) has made great progresses, while most existing methods are parsed into several stages and cannot be trained end-to-end. In this paper, we propose a novel and light noise suppression network (ENS-Unet) to achieve end-to-end learning without elaborate pre-processing and post-processing. Specifically, Slice-based Normalization is first proposed to enhance the model adaptability, where the impacts of different data distributions between training and test samples are restrained. Additionally, aiming at suppressing noises, Noise Suppression U-net (NS-Unet) is designed to obtain better image representation. Finally, to cope with the unbalanced problem of training data, Batch-based Loss, deriving from the statistical distribution within mini-batch, is adopted to adjust the weights of classes. Extensive experiments on BraTS 2013 and 2015 datasets demonstrate that the proposed method achieves very competitive results with the state-of-the-art approaches.
12 September 2023 in IEEE Journal of Biomedical and Health Informatics
Objective : Electrocardiogram (ECG) signals have wide-ranging applications in various fields, and thus it is crucial to identify clean ECG signals under different sensors and collection scenarios. Despite the availability of a variety of deep learning algorithms for ECG quality assessment, these methods still lack generalization across different datasets, hindering their widespread use. Methods : In this paper, an effective model named Swin Denoising AutoEncoder (SwinDAE) is proposed. Specifically, SwinDAE uses a DAE as the basic architecture, and incorporates a 1D Swin Transformer during the feature learning stage of the encoder and decoder. SwinDAE was first pre-trained on the public PTB-XL dataset after data augmentation, with the supervision of signal reconstruction loss and quality assessment loss. Specially, the waveform component localization loss is proposed in this paper and used for joint supervision, guiding the model to learn key information of signals. The model was then fine-tuned on the finely annotated BUT QDB dataset for quality assessment. Results : SwinDAE achieved 0.02-0.13 mean F1 score improvement on the BUT QDB dataset compared to multiple deep learning methods, and demonstrated applicability on two other datasets. Conclusion : The proposed SwinDAE shows strong generalization ability on different datasets, and surpasses other state-of-the-art deep learning methods on multiple evaluation metrics. In addition, the statistical analysis for SwinDAE prove the significance of the performance and the rationality of the prediction. Significance : SwinDAE can learn the commonality between high-quality ECG signals, exhibiting excellent performance in the application of cross-sensors and cross-collection scenarios.