Epiphenomenon or Prognostically Relevant Interventional Target? A Novel Proportionality Framework for Severe Tricuspid Regurgitation

Abstract

Background Tricuspid regurgitation (TR) frequently develops in patients with long‐standing pulmonary hypertension, and both pathologies are associated with increased morbidity and mortality. This study aimed to improve prognostic assessment in patients with severe TR undergoing transcatheter tricuspid valve intervention (TTVI) by relating the extent of TR to pulmonary artery pressures. Methods and Results In this multicenter study, we included 533 patients undergoing TTVI for moderate‐to‐severe or severe TR. The proportionality framework was based on the ratio of tricuspid valve effective regurgitant orifice area to mean pulmonary artery pressure. An optimal threshold for tricuspid valve effective regurgitant orifice area/mean pulmonary artery pressure ratio was derived on 353 patients with regard to 2‐year all‐cause mortality and externally validated on 180 patients. Patients with a tricuspid valve effective regurgitant orifice area/mean pulmonary artery pressure ratio ≤1.25 mm 2 /mm Hg (defining proportionate TR) featured significantly lower 2‐year survival rates after TTVI than patients with disproportionate TR (56.6% versus 69.6%; P =0.005). In contrast with patients with disproportionate TR (n=398), patients with proportionate TR (n=135) showed more pronounced mPAP levels (37.9±9.06 mm Hg versus 27.9±8.17 mm Hg; P <2.2×10 −16 ) and more severely impaired right ventricular function (tricuspid annular plane systolic excursion: 16.0±4.11 versus 17.0±4.64 mm; P =0.012). Moreover, tricuspid valve effective regurgitant orifice area was smaller in patients with proportionate TR when compared with disproportionate TR (0.350±0.105 cm 2 versus 0.770±0.432 cm 2 ; P <2.2×10 −16 ). Importantly, proportionate TR remained a significant predictor for 2‐year mortality after adjusting for demographic and clinical variables (hazard ratio, 1.7; P =0.006). Conclusions The proposed proportionality framework promises to improve future risk stratification and clinical decision‐making by identifying patients who benefit the most from TTVI (disproportionate TR). As a next step, randomized controlled studies with a conservative treatment arm are needed to quantify the net benefit of TTVI in patients with proportionate TR.

Machine Learning Identifies New Predictors on Restenosis Risk after Coronary Artery Stenting in 10,004 Patients with Surveillance Angiography

“…In medicine, ML implies the promise to transform the exponential increase of the amount and complexity of data into clinically usable knowledge. In particular, unsupervised ML has been increasingly used in recent years for in-depth phenotyping to identify subgroups of patients with different clinical characteristics (e.g., high-risk patients) for specific diseases who might particularly benefit from certain treatments [1][2][3][4][5][6][7]. Indeed, 2 of 14 high-resolution 3D or 4D imaging, 'omics'-technologies (genomics, transcriptomics, epigenomics, proteomics, metabolomics), and biometric sensor information may elaborate their full potential in improving risk prediction, diagnostic accuracy, and personalized treatment strategies only based on ML algorithms including most importantly unsupervised (clustering, dimensionality reduction) and supervised (classification, regression) strategies, but also semi-supervised and reinforcement learning [8][9][10][11][12].…”