The Utility of ChatGPT in Diabetic Retinopathy Risk Assessment: A Comparative Study with Clinical Diagnosis.

Purpose: To evaluate the ability of an artificial intelligence (AI) model, ChatGPT, in predicting the diabetic retinopathy (DR) risk. Methods: This retrospective observational study utilized an anonymized dataset of 111 patients with diabetes who underwent a comprehensive eye examination along with clinical and biochemical assessments. Clinical and biochemical data along with and without central subfield thickness (CST) values of the macula from OCT were uploaded to ChatGPT-4, and the response from the ChatGPT was compared to the clinical DR diagnosis made by an ophthalmologist. Results: The study assessed the consistency of responses provided by ChatGPT, yielding an Intraclass Correlation Coefficient (ICC) value of 0.936 (95% CI, 0.913-0.954, p < 0.001) (with CST) and 0.915 (95% CI, 0.706-0.846, p < 0.001) (without CST), both situations indicated excellent reliability. The sensitivity and specificity of ChatGPT in predicting the DR cases were evaluated. The results revealed a sensitivity of 67% with CST and 73% without CST. The specificity was 68% with CST and 54% without CST. However, Cohen's kappa revealed only a fair agreement between ChatGPT predictions and clinical DR status in both situations, with CST (kappa = 0.263, p = 0.005) and without CST (kappa = 0.351, p < 0.001). Conclusion: This study suggests that ChatGPT has the potential of a preliminary DR screening tool with further optimization needed for clinical use.

Extraction of Fetal ECG From Abdominal and Thorax ECG Using a Non-Causal Adaptive Filter Architecture.

Extracting the Electrocardiogram (ECG) of a fetus from the ECG signal of the maternal abdomen is a challenging task due to different artifacts. The paper proposes a N-tap non-causal adaptive filter (NC-AF) that update the weight by considering the N number of past weights and N-1 number of the reference signal and error signal samples after the processing sample number n. Using the maternal abdominal signal as the primary signal and thorax signal as the reference input, the output e(n) is obtained from the mean of N number of errors. The filtering performance of NC-AF was evaluated using the Synthetic dataset and Daisy dataset with the metrics such as correlation coefficient ( γ), peak root mean square difference (PRD), the output signal to noise ratio (SNR), root mean square error (RMSE), and fetal R-peak detection accuracy (FRPDA). The NC-AF provides a maximum correlation coefficient, PRD, SNR, RMSE and FRPDA of 0.9851, 83.04%, 8.52 dB, 0.208 and 97.09% respectively with filter length N=38. The paper also proposes the architecture of NC-AF that can be implemented in hardware like FPGA. Further, the NC-AF was implemented on Virtex-7 FPGA and its performance is evaluated in terms of resource utilization, throughput, and power consumption. For filter length N=38 and wordlength L=24, the maximum performance of the filter can be attained with a power consumption of [Formula: see text] and a maximum clock frequency of 139.47 MHz.

Recent Advancements and Future Prospects on E-Nose Sensors Technology and Machine Learning Approaches for Non-Invasive Diabetes Diagnosis: A Review.

Diabetes mellitus, commonly measured through an invasive process which although is accurate, has manifold drawbacks especially when multiple reading are required at regular intervals. Accordingly, there is a need to develop a dependable non-invasive diabetes detection technique. Recent studies have observed that other human serums such as tears, saliva, urine and breath indicate the presence of glucose in them. These parameters open quite a few ways for non-invasive blood glucose level prediction. The analysis of a persons breath poses as a good non-invasive technique to monitor the glucose levels. It is seen that in breath, there are many bio-markers and monitoring the levels of these bio-markers indicate the possibility of various chronic diseases. Among these bio-markers, acetone a volatile organic compound found in breath has shown a good correlation to the glucose levels present in blood. Therefore, by evaluating the acetone levels in breath samples it is possible to monitor diabetes non-invasively. This paper reviews the various approaches and sensory techniques used to monitor diabetes though human breath samples.

Real-Time Non-Invasive Detection and Classification of Diabetes Using Modified Convolution Neural Network.

Non-invasive diabetes prediction has been gaining prominence over the last decade. Among many human serums evaluated, human breath emerges as a promising option with acetone levels in breath exhibiting a good correlation to blood glucose levels. Such correlation establishes acetone as an acceptable biomarker for diabetes. The most common data analysis strategies to analyze the biomarkers in breath for disease detection use feature extraction and classification algorithms. However, snags such as computational cost and lack of optimal feature selection on application to real-time signals reduce the efficiency of such analysis. This paper explores the use of a one-dimensional (1-D) modified convolution neural network (CNN) algorithm that combines feature extraction and classification techniques. The approach proposed in this paper is found to significantly reduce the limitations associated with using these techniques individually and thereby improving the classifier's performance further. This paper proposes to apply a modified 1-D CNN on real-time breath signals obtained from an array of gas sensors. The experimentation and the performance of the system is carried out and evaluated.