2024年 · 第04卷第04期

The healthcare industry is significantly transforming toward digital and smart healthcare. Blockchain, as an emerging distributed collaborative paradigm, offers a promising solution for ensuring trustworthiness and high availability of services in the evolving healthcare sector. This study aimed to provide a comprehensive survey of blockchain-based applications in smart healthcare. We first present the real-world blockchain use cases in smart healthcare and related fields, outlining the motivations for this study. Next, we review the cutting-edge blockchain applications in various domains, including health data sharing, public health management, drug supply chains, insurance claims, and the Internet-of-Medical-Things. A detailed analysis of several blockchain-based healthcare data sharing scenarios is also included. The findings illustrate the diverse applications of blockchain technology in enhancing healthcare systems, along with a detailed examination of the challenges related to technical implementation and adoption. We discussed the challenges encountered in blockchain integration in smart healthcare and propose potential solutions to guide future research in this area.

Statistical shape models (SSMs) are effective for image processing and analysis and have been used in various medical fields, including face recognition and cranial bone recognition. In orthopedics, SSMs are being used in numerous applications, such as automated segmentation of medical images, preoperative planning, intraoperative navigation combined with robotics, simulation reconstruction of defects, human biomechanics research, description of anatomical shape changes, and prosthesis design. This review highlighted the wide range of applications while acknowledging the diversity of methodologies and techniques encompassed by SSMs, including Gaussian process models and nonlinear solutions. In addition, the available software packages for constructing shape models, such as Scalismo, ShapeWorks, and Deformetrica, were discussed.

This paper gave attention to two issues that arise because of the growth in the use of health data by general practitioners (GPs) and general practices in England. The issues were (a) the use and commercialisation of patients’ personal health data; and(b) the propensity of GPs and general practice staff, in utilising those data, to see patients as fragmented bodies rather than as 'whole persons’. The paper included attention to the computing needs of general practice from the 1960s and notes the period of growth in GP computer use during the 1990s. The implications of recent increased computer use by GPs and general practices, as a contributor to the further scientification of health, were then explored. Significant is the fact that the paper finds consciousness, from the 1970s, of the two issues. Their importance was emphasised as the momentum increases around the utilisation and sharing of patient data. Related concerns about data privacy and confidentiality are highlighted. In this context, the paper recommended that further research be undertaken with urgency to explore the ways that caring relationships (that have been a hallmark of the work of GPs) can be safeguarded.

Alterations in glucose metabolism, especially the postprandial glucose response (PPGR), are crucial contributors to metabolic dysfunction, which underlies the pathogenesis of metabolic syndrome. Personalized low-glycemic diets have shown promise in reducing postprandial glucose spikes. However, current methods such as invasive continuous glucose monitoring (CGM) or multi-omics data integration to assess PPGR have limitations, including cost and invasiveness that hinder the widespread adoption of these methods in primary disease prevention. Our aim was to assess machine learning algorithms for predicting individual PPGR using non-invasive wearable devices, thereby, circumventing the limitations associated with the existing approaches. By identifying the most accurate model, we sought to provide a more accessible and efficient method for managing glucose metabolic dysfunction.

This data-driven analysis used the experimental dataset from the SENSE (" Systemische Ernährungsmedizin" ) study. Healthy participants used an Empatica E4 wristband and Abbott Freestyle Libre 3 CGM for 10 days. Blood volume pulse, electrodermal activity, heart rate, skin temperature, and the corresponding CGM values were measured. Subsequently, four data-driven deep learning (DL) models-convolutional neural network, lightweight transformer, long short-term memory with attention, and Bi-directional LSTM (BiLSTM) were implemented and compared to determine the potential of DL in predicting interstitial glucose levels without involving food and activity logs.

The proposed BiLSTM achieved the best interstitial glucose prediction performance, with an average root mean squared error of 13.42 mg/dL, an average mean absolute percentage error of 0.12, and only 3.01% values falling within area D in Clarke error grid analysis, incorporating the leave-one-out cross-validation strategy for a five-minute prediction horizon.

The findings of this study may demonstrate the feasibility of transferring knowledge gained from invasive glucose monitoring devices to non-invasive approaches. Furthermore, it could emphasize the promising prospects of combining DL with wearable technologies to predict glucose levels in healthy individuals.

Vertebral segmentation in computed tomography (CT) images remains an essential issue in medical image analysis, stemming from the variability in vertebral shapes, high complex deformations, and the inherent ambiguity in CT scans. The purpose of this study was to develop advanced methods to effectively address this challenging task.

We proposed an attention-driven asymmetric convolution deep learning (AACDL) framework for vertebral segmentation. Specifically, our approach involved constructing a novel asymmetric convolutional U-shaped deep learning architecture to enhance the feature extraction capabilities by increasing its depth for capturing richer spatial details. Further, we constructed a pyramid global context module that integrates global context information through pyramid pooling to boost segmentation accuracy particularly in smaller anatomical regions. Sequential channel and spatial attention mechanisms were also implemented within the network to enable it to automatically concentrate on learning the most salient features and regions across different dimensions.

The performance precision of our network was rigorously assessed using a suite of four benchmark metrics: the dice coefficient, mean intersection over union (mIoU), precision rate, and F1-score. When compared against the ground truth, our model delivered outstanding scores, attaining a dice coefficient of 82.79%, an mIoU of 90.72%, a precision rate of 90.19%, and an F1-score of 90.09%, each reflecting the commendable accuracy and reliability of our network’s segmentation output.

The proposed AACDL method might successfully realize accurate segmentation of vertebral CT images, thereby demonstrating significant potential for clinical applications with its robust performance metrics. Its ability to handle the complexities associated with vertebral segmentation may pave the way for enhanced diagnostic and treatment planning processes in healthcare settings.

Colorectal cancer is a prevalent and deadly disease worldwide, posing significant diagnostic challenges. Traditional histopathologic image classification is often inefficient and subjective. Although some histopathologists use computer-aided diagnosis to improve efficiency, these methods depend heavily on extensive data and specific annotations, limiting their development. To address these challenges, this paper proposes a method based on few-shot learning.

This study introduced a few-shot learning approach that combines transfer learning and contrastive learning to classify colorectal cancer histopathology images into benign and malignant categories. The model comprises modules for feature extraction, dimensionality reduction, and classification, trained using a combination of contrast loss and cross-entropy loss. In this paper, we detailed the setup of hyperparameters: n-way, k-shot, β, and the creation of support, query, and test datasets.

Our method achieved over 98% accuracy on a query dataset with 35 samples per category using only 10 training samples per category. We documented the model’s loss, accuracy, and the confusion matrix of the results. Additionally, we employed the t-SNE algorithm to analyze and assess the model’s classification performance.

The proposed model may demonstrate significant advantages in accuracy and minimal data dependency, performing robustly across all tested n-way, k-shot scenarios. It consistently achieved over 93% accuracy on comprehensive test datasets, including 1916 samples, confirming its high classification accuracy and strong generalization capability. Our research could advance the use of few-shot learning in medical diagnostics and also lays the groundwork for extending it to deal with rare, difficult-to-diagnose cases.

尊敬的编辑：