Diagnosis of diabetes mellitus using high frequency ultrasound and convolutional neural network

Recent Advancements in High-Frequency Ultrasound Applications from Imaging to Microbeam Stimulation

Ultrasound is a versatile and well-established technique using sound waves with frequencies higher than the upper limit of human hearing. Typically, therapeutic and diagnosis ultrasound operate in the frequency range of 500 kHz to 15 MHz with greater depth of penetration into the body. However, to achieve improved spatial resolution, high-frequency ultrasound (>15 MHz) was recently introduced and has shown promise in various fields such as high-resolution imaging for the morphological features of the eye and skin as well as small animal imaging for drug and gene therapy. In addition, high-frequency ultrasound microbeam stimulation has been demonstrated to manipulate single cells or microparticles for the elucidation of physical and functional characteristics of cells with minimal effect on normal cell physiology and activity. Furthermore, integrating machine learning with high-frequency ultrasound enhances diagnostics, including cell classification, cell deformability estimation, and the diagnosis of diabetes and dysnatremia using convolutional neural networks (CNNs). In this paper, current efforts in the use of high-frequency ultrasound from imaging to stimulation as well as the integration of deep learning are reviewed, and potential biomedical and cellular applications are discussed.

Predicting the daily number of patients for allergic diseases using PM10 concentration based on spatiotemporal graph convolutional networks

Air pollution causes and exacerbates allergic diseases including asthma, allergic rhinitis, and atopic dermatitis. Precise prediction of the number of patients afflicted with these diseases and analysis of the environmental conditions that contribute to disease outbreaks play crucial roles in the effective management of hospital services. Therefore, this study aims to predict the daily number of patients with these allergic diseases and determine the impact of particulate matter (PM10) on each disease. To analyze the spatiotemporal correlations between allergic diseases (asthma, atopic dermatitis, and allergic rhinitis) and PM10 concentrations, we propose a multi-variable spatiotemporal graph convolutional network (MST-GCN)-based disease prediction model. Data on the number of patients were collected from the National Health Insurance Service from January 2013 to December 2017, and the PM10 data were collected from Airkorea during the same period. As a result, the proposed disease prediction model showed higher performance (R2 0.87) than the other deep-learning baseline methods. The synergic effect of spatial and temporal analyses improved the prediction performance of the number of patients. The prediction accuracies for allergic rhinitis, asthma, and atopic dermatitis achieved R2 scores of 0.96, 0.92, and 0.86, respectively. In the ablation study of environmental factors, PM10 improved the prediction accuracy by 10.13%, based on the R2 score.

Quantitative blood glucose detection influenced by various factors based on the fusion of photoacoustic temporal spectroscopy with deep convolutional neural networks

In order to efficiently and accurately monitor blood glucose concentration (BGC) synthetically influenced by various factors, quantitative blood glucose in vitro detection was studied using photoacoustic temporal spectroscopy (PTS) combined with a fusion deep neural network (fDNN). Meanwhile, a photoacoustic detection system influenced by five factors was set up, and 625 time-resolved photoacoustic signals of rabbit blood were collected under different influencing factors.In view of the sequence property for temporal signals, a dimension convolutional neural network (1DCNN) was established to extract features containing BGC. Through the parameters optimization and adjusting, the mean square error (MSE) of BGC was 0.51001 mmol/L for 125 testing sets. Then, due to the long-term dependence on temporal signals, a long short-term memory (LSTM) module was connected to enhance the prediction accuracy of BGC. With the optimal LSTM layers, the MSE of BGC decreased to 0.32104 mmol/L. To further improve prediction accuracy, a self-attention mechanism (SAM) module was coupled into and formed an fDNN model, i.e., 1DCNN-SAM-LSTM. The fDNN model not only combines the advantages of temporal expansion of 1DCNN and data long-term memory of LSTM, but also focuses on the learning of more important features of BGC. Comparison results show that the fDNN model outperforms the other six models. The determination coefficient of BGC for the testing set was 0.990, and the MSE reached 0.1432 mmol/L. Results demonstrate that PTS combined with 1DCNN-SAM-LSTM ensures higher accuracy of BGC under the synthetical influence of various factors, as well as greatly enhances the detection efficiency.

Exploring the Diagnostic Value of High-Frequency Ultrasound Technology for Subcutaneous Lipohypertrophy in Diabetes Patients Receiving Insulin Injections

Objective

This study aims to investigate the clinical application value of high-frequency ultrasound technology in diagnosing subcutaneous lipohypertrophy at insulin injection sites in diabetes patients.

Methods

All diabetes patients treated in our hospital from January 2022 to January 2023 were selected as the study subjects. The incidence of subcutaneous lipohypertrophy was calculated at the end of the study period. All patients were screened, and those meeting the inclusion criteria were registered, and basic data were collected. Patients were screened for subcutaneous lipohypertrophy using conventional clinical examination (control group) and high-frequency ultrasound technology (study group).

Results

The study found that the incidence of subcutaneous lipohypertrophy in diabetes patients receiving insulin injections in our hospital from January 2022 to January 2023 was 80.99%. The average longitudinal diameter of subcutaneous lipohypertrophy in these patients was 11.66 (7.56, 21.44) mm, the transverse diameter was 12.04 (8.96, 18.29) mm, depth was (5.62±2.17) mm, and the area was 188.79 (76.85, 331.78) mm². The clinical detection rate in the study group was higher than that in the control group (P<0.05). The quantity of detected sites was greater in the study group compared to the control group (P<0.05).

Conclusion

The incidence of subcutaneous lipohypertrophy in diabetes patients receiving insulin injections is relatively high clinically, and high-frequency ultrasound technology demonstrates significant potential in diagnosis. By providing high-resolution imaging and quantitative data, it effectively improves the clinical detection rate and clarifies symptoms. This technology is likely to become an important auxiliary tool in future diabetes treatment, providing more precise treatment plans for patients.

Design of Preamplifier for Ultrasound Transducers

In diagnostic ultrasound imaging applications, preamplifiers are used as first-stage analog front-end amplifiers for ultrasound transducers because they can amplify weak acoustic signals generated directly by ultrasound transducers. For emerging diagnostic ultrasound imaging applications, different types of preamplifiers with specific design parameters and circuit topologies have been developed, depending on the types of the ultrasound transducer. In particular, the design parameters of the preamplifier, such as the gain, bandwidth, input- or output-referred noise components, and power consumption, have a tradeoff relationship. Guidelines on the detailed design concept, design parameters, and specific circuit design techniques of the preamplifier used for ultrasound transducers are outlined in this paper, aiming to help circuit designers and academic researchers optimize the performance of ultrasound transducers used in the diagnostic ultrasound imaging applications for research directions.