Search

Introduction Techniques exist which allow surface ECGs to be reconstructed from reduced lead sets. We aimed to reconstruct an 8-lead ECG from two independent S-ICD sensing electrodes vectors as proof of this principle. Methods Participants with ICDs ( N =61) underwent 3minute ECGs using a TMSi Porti7 multi-channel signal recorder (TMS international, The Netherlands) with electrodes in the standard S-ICD and 12-lead positions. Participants were randomised to either a training ( N =31) or validation ( N =30) group. The transformation used was a linear combination of the 2 independent S-ICD vectors to each of the 8 independent leads of the 12-lead ECG, with coefficients selected that minimized the root mean square error (RMSE) between recorded and derived ECGs when applied to the training group. The transformation was then applied to the validation group and agreement between the recorded and derived lead pairs was measured by Pearson correlation coefficient ( r ) and normalised RMSE (NRMSE). Results In total, 27 patients with complete data sets were included in the validation set consisting of 57,888 data points from 216 full lead sets. The distribution of the r and NRMSE were skewed. Mean r =0.770 (SE 0.024), median r =0.925. NRMSE mean=0.233 (SE 0.015) median=0.171. Conclusions We have demonstrated that the reconstruction of an 8-lead ECG from two S-ICD vectors is possible. If perfected, the ability to generate accurate multi-lead surface ECG data from an S-ICD would potentially allow recording and review of clinical arrhythmias at follow-up.

The article contains data pertaining to the reconstruction of an 8-lead ECG from 2 subcutaneous implantable cardioverter defibrillator vectors. The location of electrodes on the precordium required for the data collection are detailed; the flow chart for patient selection and exclusion is shown; the summary data of the root mean square error (RMSE) (in microvolts) and Pearson r for the ECG transformation all cases and the pearson correlation for all the leads measured and reconstructed leads are also shown. Detailed background, methodology and discussion can be found in the linked research article.

INTRODUCTION AND OBJECTIVES:Several nomograms have been developed to predict prostate cancer upgrading however very little artificial intelligence (AI) tools are available for this purpose. Aim of ...

Background: Data from reduced lead ECG configurations can be transformed to reconstruct a 12-lead ECG. The sensing electrodes of a subcutaneous ICD (S-ICD) could provide a suitable basis for such a system. Reconstruction of a 12-lead ECG from conventional S-ICD vectors has not previously been demonstrated. Hypothesis: A 12-lead ECG signal can be accurately derived from S-ICD electrode vectors. Methods: Participants with ICDs (N=61) underwent 3 minute ECGs with electrodes in the standard S-ICD and 12-lead positions. Participants were randomised to either a training (N=31) or validation (N=30) group. Two participants and one further lead were subsequently excluded from the validation group due to technical faults. The transformation was a linear combination of the 2 independent S-ICD vectors to each of the 8 independent leads of the 12-lead ECG, with coefficients selected that minimized the root mean square error (RMSE) between recorded and derived ECGs when applied to the training group. The transformation was then applied to the validation group and agreement between the recorded and derived lead pairs was measured by Pearson correlation coefficient (r) and normalised RMSE (NRMSE). Results: In total, 335 lead pairs were compared (figure1). Distribution of r and NRMSE were skewed. Mean r = 0.731 (SE 0.021), Median r = 0.903 (SE 0.015). NRMSE Mean = 0.227 (SE 0.011) Median = 0.167 (SE 0.008). 252/335 (75.2%) of all pairs had an r > 0.7 (indicating very strong positive correlation). Figure 1: Example 12-lead ECG, recorded (blue line) and derived (red line). Participant randomly selected from top 20% by mean Pearson correlation coefficient. Conclusions: We have demonstrated reconstruction of a 12-lead ECG from S-ICD vectors. The agreement between derived and measured signals was generally excellent and comparable with other widely accepted reduced lead systems. If perfected, the ability to generate 12-lead data from an S-ICD could help clinicians to localise and treat arrhythmias.

We consider the task of segmentation of images of mosaics, where the goal is to segment the image in such a way that each region corresponds exactly to one tile of the mosaic. We propose to use a recent deep learning technique based on a kind of convolutional neural networks, called U-net, that proved to be effective in segmentation tasks. Our method includes a preprocessing phase that allows to learn a U-net despite the scarcity of labeled data, which reflects the peculiarity of the task, in which manual annotation is, in general, costly. We experimentally evaluate our method and compare it against the few other methods for mosaic images segmentation using a set of performance indexes, previously proposed for this task, computed using 11 images of real mosaics. In our results, U-net compares favorably with previous methods. Interestingly, the considered methods make errors of different kinds, consistently with the fact that they are based on different assumptions and techniques. This finding suggests that combining different approaches might lead to an even more effective segmentation.