Prediction of cardiac autonomic neuropathy using a machine learning model in patients with diabetes

Artificial intelligence-based classification of cardiac autonomic neuropathy from retinal fundus images in patients with diabetes: The Silesia Diabetes Heart Study

BACKGROUND

Cardiac autonomic neuropathy (CAN) in diabetes mellitus (DM) is independently associated with cardiovascular (CV) events and CV death. Diagnosis of this complication of DM is time-consuming and not routinely performed in the clinical practice, in contrast to fundus retinal imaging which is accessible and routinely performed. Whether artificial intelligence (AI) utilizing retinal images collected through diabetic eye screening can provide an efficient diagnostic method for CAN is unknown.

METHODS

This was a single center, observational study in a cohort of patients with DM as a part of the Cardiovascular Disease in Patients with Diabetes: The Silesia Diabetes-Heart Project (NCT05626413). To diagnose CAN, we used standard CV autonomic reflex tests. In this analysis we implemented AI-based deep learning techniques with non-mydriatic 5-field color fundus imaging to identify patients with CAN. Two experiments have been developed utilizing Multiple Instance Learning and primarily ResNet 18 as the backbone network. Models underwent training and validation prior to testing on an unseen image set.

RESULTS

In an analysis of 2275 retinal images from 229 patients, the ResNet 18 backbone model demonstrated robust diagnostic capabilities in the binary classification of CAN, correctly identifying 93% of CAN cases and 89% of non-CAN cases within the test set. The model achieved an area under the receiver operating characteristic curve (AUCROC) of 0.87 (95% CI 0.74-0.97). For distinguishing between definite or severe stages of CAN (dsCAN), the ResNet 18 model accurately classified 78% of dsCAN cases and 93% of cases without dsCAN, with an AUCROC of 0.94 (95% CI 0.86-1.00). An alternate backbone model, ResWide 50, showed enhanced sensitivity at 89% for dsCAN, but with a marginally lower AUCROC of 0.91 (95% CI 0.73-1.00).

CONCLUSIONS

AI-based algorithms utilising retinal images can differentiate with high accuracy patients with CAN. AI analysis of fundus images to detect CAN may be implemented in routine clinical practice to identify patients at the highest CV risk.

TRIAL REGISTRATION

This is a part of the Silesia Diabetes-Heart Project (Clinical-Trials.gov Identifier: NCT05626413).

Artificial intelligence-enhanced electrocardiogram analysis for identifying cardiac autonomic neuropathy in patients with diabetes

AIM

To develop and employ machine learning (ML) algorithms to analyse electrocardiograms (ECGs) for the diagnosis of cardiac autonomic neuropathy (CAN).

MATERIALS AND METHODS

We used motif and discord extraction techniques, alongside long short-term memory networks, to analyse 12-lead, 10-s ECG tracings to detect CAN in patients with diabetes. The performance of these methods with the support vector machine classification model was evaluated using 10-fold cross validation with the following metrics: accuracy, precision, recall, F1 score, and area under the receiver-operating characteristic curve (AUC).

RESULTS

Among 205 patients (mean age 54 ± 17 years, 54% female), 100 were diagnosed with CAN, including 38 with definite or severe CAN (dsCAN) and 62 with early CAN (eCAN). The best model performance for dsCAN classification was achieved using both motifs and discords, with an accuracy of 0.92, an F1 score of 0.92, a recall at 0.94, a precision of 0.91, and an excellent AUC of 0.93 (95% confidence interval [CI] 0.91-0.94). For the detection of any stage of CAN, the approach combining motifs and discords yielded the best results, with an accuracy of 0.65, F1 score of 0.68, a recall of 0.75, a precision of 0.68, and an AUC of 0.68 (95% CI 0.54-0.81).

CONCLUSION

Our study highlights the potential of using ML techniques, particularly motifs and discords, to effectively detect dsCAN in patients with diabetes. This approach could be applied in large-scale screening of CAN, particularly to identify definite/severe CAN where cardiovascular risk factor modification may be initiated.

Features characterising cardiac autonomic neuropathy in diabetes using ensembled classification

OBJECTIVE

Using supervised machine learning to classify the severity of cardiovascular autonomic neuropathy (CAN). The aims were 1) to investigate which features contribute to characterising CAN 2) to generate an ensembled set of features that best describes the variation in CAN classification.

METHODS

Eighty-two features from demographic, beat-to-beat, biochemical, and inflammation were obtained from 204 people with diabetes and used in three machine-learning-classifiers, these are: support vector machine, decision tree, and random forest. All data were ensembled using a weighted mean of the features from each classifier.

RESULTS

The 10 most important features derived from the domains: Beat-to-beat, inflammation markers, disease-duration, and age.

CONCLUSIONS

Beat-to-beat measures associate with CAN as diagnosis is mainly based on cardiac reflex responses, disease-duration and age are also related to CAN development throughout disease progression. The inflammation markers may reflect the underlying disease process, and therefore, new treatment modalities targeting systemic low-grade inflammation should potentially be tested to prevent the development of CAN.

SIGNIFICANCE

Cardiac reflex responses should be monitored closely to diagnose and classify severity levels of CAN accurately. Standard clinical biochemical analytes, such as glycaemic level, lipidic level, or kidney function were not included in the ten most important features. Beat-to-beat measures accounted for approximately 60% of the features in the ensembled data.

The burden of type 2 diabetes in Australia during the period 1990-2019: Findings from the global burden of disease study

AIMS

To describe morbidity and mortality trends of type 2 diabetes in Australia, from 1990 to 2019, compared with similar sociodemographic index (SDI) countries.

METHODS

Australia-specific Global Burden of Diseases data were used to estimate age-standardised, age-specific, and sex-specific rates for prevalence, years lived with disability (YLDs), years of life lost (YLLs), disability-adjusted life years (DALYs), and deaths due to type 2 diabetes between 1990 and 2019. Australian data were compared with 14 similar SDI countries.

RESULTS

Type 2 diabetes increased in Australia between 1990 and 2019. The age-standardised prevalence increased from 1,985 [95% Confidence Interval (CI): 1,786.7-2195.3] per 100,000 population, to 3,429 [95% CI 3,053.3-3,853.7]. Cases tripled, from 379,532 [342,465-419,475] to 1,307,261 [1,165,522-1,461,180]. The age-standardised death rates doubled, from 2,098 [1,953-2,203] per 100,000, to 4,122 [3,617-4,512]. DALYs doubled, from 70,348 [59,187-83,500] to 169,763 [129,792-216,150], with increases seen in YLDs and YLLs. Men displayed higher rates. Compared to similar SDI countries, Australia ranked 4th in terms of burden for type 2 diabetes.

CONCLUSIONS

The burden of type 2 diabetes in Australia has increased considerably over three decades. There is an urgent need to prioritise resource allocation for prevention programs, screening initiatives to facilitate early detection, and effective and accessible management strategies for the large proportion of the population impacted by type 2 diabetes.

Consensus on Recommended Functions of a Smart Home System to Improve Self-Management Behaviors in People With Heart Failure: A Modified Delphi Approach

Background

Smart home systems could enhance clinical and self-management of chronic heart failure by supporting health monitoring and remote support, but evidence to guide the design of smart home system functionalities is lacking.

Objective

To identify consensus-based recommendations for functions of a smart home system that could augment clinical and self-management for people living with chronic heart failure in the community.

Methods

Healthcare professionals caring for people living with chronic heart failure participated in a two-round modified Delphi survey and a consensus workshop. Thirty survey items spanning eight chronic health failure categories were derived from international guidelines for the management of heart failure. In survey Round 1, participants rated the importance of all items using a 9-point Liket scale and suggested new functions to support people with chronic heart failure in their homes using a smart home system. The Likert scale scores ranged from 0 (not important) to 9 (very important) and scores were categorized into three groups: 1–3 = not important, 4–6 = important, and 7–9 = very important. Consensus agreement was defined a priori as ≥70% of respondents rating a score of ≥7 and ≤ 15% rating a score ≤ 3. In survey Round 2, panel members re-rated items where consensus was not reached, and rated the new items proposed in earlier round. Panel members were invited to an online consensus workshop to discuss items that had not reached consensus after Round 2 and agree on a set of recommendations for a smart home system.

Results

In Round 1, 15 experts agreed 24/30 items were “very important”, and suggested six new items. In Round 2, experts agreed 2/6 original items and 6/6 new items were “very important”. During the consensus workshop, experts endorsed 2/4 remaining items. Finally, the expert panel recommended 34 items as “very important” for a smart home system including, healthy eating, body weight and fluid intake, physical activity and sedentary behavior, heart failure symptoms, tobacco cessation and alcohol reduction, medication adherence, physiological monitoring, interaction with healthcare professionals, and mental health among others.

Conclusion

A panel of healthcare professional experts recommended 34-item core functions in smart home systems designed to support people with chronic heart failure for self-management and clinical support. Results of this study will help researchers to co-design and protyping solutions with consumers and healthcare providers to achieve these core functions to improve self-management and clinical outcomes in people with chronic heart failure.

Wearable cuffless blood pressure monitoring devices: a systematic review and meta-analysis

Aims

High blood pressure (BP) is the commonest modifiable cardiovascular risk factor, yet its monitoring remains problematic. Wearable cuffless BP devices offer potential solutions; however, little is known about their validity and utility. We aimed to systematically review the validity, features and clinical use of wearable cuffless BP devices.

Methods and results

We searched MEDLINE, Embase, IEEE Xplore and the Cochrane Database till December 2019 for studies that reported validating cuffless BP devices. We extracted information about study characteristics, device features, validation processes, and clinical applications. Devices were classified according to their functions and features. We defined devices with a mean systolic BP (SBP) and diastolic BP (DBP) biases of <5 mmHg as valid as a consensus. Our definition of validity did not include assessment of device measurement precision, which is assessed by standard deviation of the mean difference-a critical component of ISO protocol validation criteria. Study quality was assessed using the Quality Assessment of Diagnostic Accuracy Studies version 2 tool. A random-effects model meta-analysis was performed to summarise the mean biases for SBP and DBP across studies. Of the 430 studies identified, 16 studies (15 devices, 974 participants) were selected. The majority of devices (81.3%) used photoplethysmography to estimate BP against a reference device; other technologies included tonometry, auscultation and electrocardiogram. In addition to BP and heart rate, some devices also measured night-time BP (n = 5), sleep monitoring (n = 3), oxygen saturation (n = 3), temperature (n = 2) and electrocardiogram (n = 3). Eight devices showed mean biases of <5 mmHg for SBP and DBP compared with a reference device and three devices were commercially available. The meta-analysis showed no statistically significant differences between the wearable and reference devices for SBP (pooled mean difference = 3.42 mmHg, 95% CI: -2.17, 9.01, I² 95.4%) and DBP (pooled mean = 1.16 mmHg, 95% CI: -1.26, 3.58, I² 87.1%).

Conclusion

Several cuffless BP devices are currently available using different technologies, offering the potential for continuous BP monitoring. The variation in standards and validation protocols limited the comparability of findings across studies and the identification of the most accurate device. Challenges such as validation using standard protocols and in real-life settings must be overcome before they can be recommended for uptake into clinical practice.