Answers are grounded in this paper and cite their sources.
BackgroundMachine learning (ML) models have been proposed to predict risk related to transvenous lead extraction (TLE).ObjectiveWe tested if integrating imaging data into an existing ML model increases its ability to predict major adverse events (MAE: procedure-related major complications and procedure-related deaths) and lengthy procedures (≥100 minutes).MethodsWe hypothesised certain features: i) lead angulation ii) coil percentage inside the superior vena cava (SVC), and iii) number of overlapping leads in the SVC, detected from a pre-TLE plain anterior-posterior (AP) chest x-ray (CXR) would improve prediction of MAE and long procedure times. A deep-learning convolutional neural network was developed to automatically detect these CXR features.Results1050 cases were included, with 24 (2.3%) MAEs. The neural network was able to detect: i) heart border with 100% accuracy, ii) coils: 98% accuracy, iii) acute angle in the right ventricle and SVC: 91% and 70% accuracy respectively. The following features significantly improved MAE prediction: i) ≥50% coil within the SVC, ii) ≥2 overlapping leads in the SVC, and iii) acute lead angulation. Balanced accuracy (0.74 to 0.87), sensitivity (68% to 83%), specificity (72% to 91%), and area under the curve (AUC) (0.767 to 0.962) all improved with imaging biomarkers. Prediction of lengthy procedures also improved: balanced accuracy (0.76 to 0.86), sensitivity (75% to 85%), specificity (63% to 87%), and AUC (0.684 to 0.913).ConclusionRisk prediction tools integrating imaging biomarkers significantly increases the ability of ML models to predict risk of MAE and long procedure time related to TLE.
No takes yet. Share an insight, caveat, or question.
Mehta et al. (2024) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: