Role of deep learning methods in screening for subcutaneous implantable cardioverter defibrillator in heart failure

Figures

• 

  FIGURE 1. Showing the typical S‐ICD vectors on the left and on the right, the Holter® surface ECG positions. 1 = 1 cm infero‐lateral to the xiphisternum. 2 = 14 cm superior to position 1. 3 = 5th intercostal space, parasternal position. 4 = 6th intercostal space left mid axillary line. 6 = Adjacent to 2. 7 = Adjacent to 4. Holter Channel A records between points 1 and 4 = surrogate of S‐ICD primary vector. Holter Channel B records between points 2 and 3 = surrogate of S‐ICD alternate vector. Holter Channel C records between points 6 and 7 = surrogate of S‐ICD secondary vector. 5 = 5th intercostal space right midclavicular line = neutral electrode. Image prior to annotation © Boston Scientific Corporation or its affiliates.

• 

  FIGURE 2. Box plots for the mean (p < .001) and standard deviation (p = .02) of the T:R ratio over 24 h screening period in the studied subgroups (heart failure patients undergoing diuresis vs healthy volunteers with structurally normal hearts). Leads A, B and C correspond to primary, alternate, and secondary vectors of an S‐ICD respectively.

• 

  FIGURE 3. An example of the T:R ratio fluctuating overtime crossing the S‐ICD screening threshold for the T:R ratio (0.33) on multiple occasions over the 24‐h period. The histogram illustrates the exact number of 10‐s segments at each T:R ratio throughout the 24‐h recording. The above example represents the alternate vector for one of the patients who was recruited to the HF group.

• 

  FIGURE 4. An example of how our tool can help select the most suitable vector for programming the S‐ICD. The analysis of the Holter recording for one of the patient recruited in the HF group: All the 3 leads had acceptable T:R ratios at some stage of the 24‐h recordings, however, while the T:R ratios for leads A and B (corresponding to the primary and alternate vectors) showed significant fluctuations over the 24‐h recording and crossed the screening threshold multiple times, T:R ratio for Lead C (secondary vector) was stable in comparison and did not cross the threshold throughout the 24‐h posing the least risk of TWO.