Search

There has been growing scientific interest in the research field of deep learning techniques applied to skin cancer diagnosis in the last decade. Though encouraging data have been globally reported, several discrepancies have been observed in terms of study methodology, result presentations and validation in clinical settings. The present review aimed to screen the scientific literature on the application of DL techniques to dermoscopic melanoma/nevi differential diagnosis and extrapolate those original studies adequately by reporting on a DL model, comparing them among clinicians and/or another DL architecture. The second aim was to examine those studies together according to a standard set of statistical measures, and the third was to provide dermatologists with a comprehensive explanation and definition of the most used artificial intelligence (AI) terms to better/further understand the scientific literature on this topic and, in parallel, to be updated on the newest applications in the medical dermatologic field, along with a historical perspective. After screening nearly 2000 records, a subset of 54 was selected. Comparing the 20 studies reporting on convolutional neural network (CNN)/deep convolutional neural network (DCNN) models, we have a scenario of highly performant DL algorithms, especially in terms of low false positive results, with average values of accuracy (83.99%), sensitivity (77.74%), and specificity (80.61%). Looking at the comparison with diagnoses by clinicians (13 studies), the main difference relies on the specificity values, with a +15.63% increase for the CNN/DCNN models (average specificity of 84.87%) compared to humans (average specificity of 64.24%) with a 14,85% gap in average accuracy; the sensitivity values were comparable (79.77% for DL and 79.78% for humans). To obtain higher diagnostic accuracy and feasibility in clinical practice, rather than in experimental retrospective settings, future DL models should be based on a large dataset integrating dermoscopic images with relevant clinical and anamnestic data that is prospectively tested and adequately compared with physicians.

Abstract Aims There are conflicting data on the benefit of cardiac resynchronization therapy (CRT) in heart failure (HF) patients with permanent atrial fibrillation (AF). We aimed to compare patient outcomes according to the presence or absence of permanent AF at device implantation. Methods and results We retrospectively analysed remote monitoring data from 1141 CRT defibrillators. Propensity score with inverse‐probability weighting method was used to balance AF and sinus rhythm (SR) groups. Analysis endpoints included total mortality, appropriate defibrillation shocks, and CRT percentage. There were 229 patients (20.1%) in the AF group and 912 patients (79.9%) in the SR group. Compared with SR patients, AF patients were older (median age, 77 vs. 72 years, P

Introduction: During the SARS‐CoV‐2 COVID‐19 pandemic, the global health system needed to review important processes involved in daily routines such as outpatient activities within the hospital, including follow‐up visits of implantable cardiac electronic devices (CIEDs) carried out in office. The aim of this study is to describe our 3.5 years of real‐world experience of a full remote CIED follow‐up, evaluate the success rate of remote transmissions, and verify the adopted organizational model. Methods: From April 2020 to November 2023, all patients with an activated and well‐functioning remote monitoring (RM) system and automatic algorithms, like autocapture and autosensing, underwent exclusive RM follow‐up. Unscheduled in‐office visits were only prompted by remote yellow or red alerts. Patients were divided into two groups, based on available technology: Manual Transmission System (MTS) and Automatic Transmission System (ATS). The ATS group, in addition to ensuring a daily transmission of any yellow or red alerts, was checked at least every 15 days to ensure a valid connection. An automatic transmission was scheduled once a year, irrespective of alerts occurred. The MTS group provided a manual transmission every 6 months. Results: One thousand nine hundred thirty‐seven consecutive patients were included in the study. By the end of November 2023, a total of 1409 patients (1192 in the ATS and 217 in the MTS group) were still actively followed by our remote clinic (384 expired, 137 dismissed, 7 transferred). The overall success rate of transmissions with the adopted organizational model was 96.6% in the ATS group (connection index) and 87% in the MTS group. Conventional in‐hospital follow‐up visits decreased by 44%. Total clinic working time, resulting from the sum of the time spent during in‐hospital and remote follow‐up, after an initial increase, was progressively reduced to the actual −25%. Mortality rate for any cause was 7.5% per year in remote follow‐up patients and 8.3% (p=NS) in in‐office patients. In the ATS group, no device malfunctions were notified to our remote clinic, before we had already realized it through appropriate alerts. Conclusions: The available technology makes moving to a 100% remote clinic possible, without overwhelming clinic workflow, safely. Adopting an appropriate organizational model, it is possible to maintain high transmission success rates. The automatic transmissions allow a more frequent control of patients with CIED. [ABSTRACT FROM AUTHOR]

There has been growing scientific interest in the research field of deep learning techniques applied to skin cancer diagnosis in the last decade. Though encouraging data have been globally reported, several discrepancies have been observed in terms of study methodology, result presentations and validation in clinical settings. The present review aimed to screen the scientific literature on the application of DL techniques to dermoscopic melanoma/nevi differential diagnosis and extrapolate those original studies adequately by reporting on a DL model, comparing them among clinicians and/or another DL architecture. The second aim was to examine those studies together according to a standard set of statistical measures, and the third was to provide dermatologists with a comprehensive explanation and definition of the most used artificial intelligence (AI) terms to better/further understand the scientific literature on this topic and, in parallel, to be updated on the newest applications in the medical dermatologic field, along with a historical perspective. After screening nearly 2000 records, a subset of 54 was selected. Comparing the 20 studies reporting on convolutional neural network (CNN)/deep convolutional neural network (DCNN) models, we have a scenario of highly performant DL algorithms, especially in terms of low false positive results, with average values of accuracy (83.99%), sensitivity (77.74%), and specificity (80.61%). Looking at the comparison with diagnoses by clinicians (13 studies), the main difference relies on the specificity values, with a +15.63% increase for the CNN/DCNN models (average specificity of 84.87%) compared to humans (average specificity of 64.24%) with a 14,85% gap in average accuracy; the sensitivity values were comparable (79.77% for DL and 79.78% for humans). To obtain higher diagnostic accuracy and feasibility in clinical practice, rather than in experimental retrospective settings, future DL models should be based on a large dataset integrating dermoscopic images with relevant clinical and anamnestic data that is prospectively tested and adequately compared with physicians. [ABSTRACT FROM AUTHOR]

A 62-year-old man admitted for presyncope presented two symptomatic sustained ventricular tachycardia with right bundle branch morphology and inferior axis suggesting a pathology of the left ventricular lateral wall, the site where Cardiac Magnetic Resonance demonstrated a thinned, hypokinetic segment with fibro-fatty subepicardial infiltration. A very localized Left Dominant Arrhythmogenic Cardiomyopathy was diagnosed and an ICD implanted.

Stroke is a rare but possible complication after atrial fibrillation (AF) ablation. However, its etiopathogenesis is far from being completely characterized. Here we report a case of stroke, with recurrent peripheral embolism after AF ablation procedure. In our patient, an in situ femoral vein thrombosis and iatrogenic atrial septal defect were simultaneously detected. A comprehensive review of multiple pathophysiological mechanisms of stroke in this context is provided. The case underlines the importance of a global evaluation of patients undergoing AF ablation.

Background: Cavo-tricuspid isthmus (CTI) block is currently assessed by coronary sinus (CS) pacing or low lateral and septal atrial pacing. Occasionally, CS catheterization through the femoral route can be difficult to perform or right atrial pacing can be problematic because of catheter instability or saturation of the atrial electrograms recorded near the catheter. Objectives: Our aim was to evaluate the feasibility of assessing cavo-tricuspid isthmus block by means of right ventricular (RV) pacing in patients with ventriculo-atrial conduction, comparing it with CS pacing.Methods: Circumannular activation was analyzed during CS and RV pacing in consecutive patients in sinus rhythm undergoing CTI ablation for typical atrial flutter. Patients without ventriculo-atrial conduction were excluded from the study. The linear lesion was created during RV pacing and split atrial signals on the ablation line were analyzed. CTI block was confirmed by analyzing local electrograms on the line of block and circumannular activation during CS and RV pacing. Results: Out of 31 patients, 20 displayed ventriculo-atrial conduction (64%) and were included in the study. Before ablation, during RV stimulation, the collision front of circumannular activation shifted counterclockwise in contrast with the pattern observed during CS pacing. After ablation, circumannular activation was similar during CS and RV pacing, showing fully descending lateral right atrium activation, even if double potentials registered on the ablation line were less widely split during RV pacing than CS pacing (111±26 ms vs 128±30 , p=0.0001).Conclusions: In patients with ventriculo-atrial conduction, tricuspid annulus activation during CS and RV pacing is similar, before and after CTI ablation. The occurrence of split atrial electrograms separated by an isoelectric interval registered on the line of block can be detected during CS or RV pacing. In patients with difficult CS catheterization via the femoral vein, before trying the subclavian or internal jugular route, if retrograde ventriculo-atrial conduction is present, RV pacing can be an easy trick to assess isthmus block.