Dahl, Jordi

Abstract Aims Aortic stenosis (AS) is a common and progressive disease, which, if left untreated, results in increased morbidity and mortality. Monitoring and follow-up care can be challenging due to significant variability in disease progression. This study aimed to develop machine learning models to predict the risks of disease progression and mortality in patients with mild AS. Methods and results A comprehensive database including 9611 patients with serial transthoracic echocardiograms was collected from a single institution across three clinical sites. The data set included parameters from echocardiograms, electrocardiograms, laboratory values, and diagnosis codes. Data from a single clinical site were preserved as an independent test group. Machine learning models were trained to identify progression to severe stenosis and all-cause mortality and tested in their performance for endpoints at 2 and 5 years. In the independent test group, the AS progression model differentiated those with progression to severe AS within 2 and 5 years with an area under the curve (AUC) of 0.86 for both. The feature of greatest importance was aortic valve mean gradient, followed by other valve haemodynamic measurements including valve area and dimensionless index. The mortality model identified those with mortality within 2 and 5 years with an AUC of 0.84 and 0.87, respectively. Smaller reduced-input validation models had similarly robust findings. Conclusion Machine learning models can be used in patients with mild AS to identify those at high risk of disease progression and mortality. Implementation of such models may facilitate real-time, patient-specific follow-up recommendations.

Objective To determine whether echocardiography-derived left ventricular filling pressure influences survival in patients with severe aortic stenosis (AS) undergoing aortic valve replacement (AVR). Methods We retrospectively reviewed 1383 consecutive patients with severe AS, normal ejection fraction and interpretable filling pressure undergoing AVR. Left ventricular filling pressure was determined according to current guidelines using mitral inflow, mitral annular tissue Doppler, estimated right ventricular systolic pressure and left atrial volume index. Cox proportional hazards regression was used to assess the influence of various parameters on mortality. Results Age was 75±10 years and 552 (40%) were female. Left ventricular filling pressure was normal in 325 (23%), indeterminate in 463 (33%) and increased in 595 (43%). Mean follow-up was 7.3±3.7 years, and mortality was 1.2%, 4.2% and 18.9% at 30 days and 1 and 5 years, respectively. Compared with patients with normal filling pressure, patients with increased filling pressure were older (78±9 vs 70±12, p<0.001), more often female (45% vs 35%, p=0.002) and were more likely to have New York Heart Association class III–IV symptoms (35% vs 24%, p=0.004), coronary artery disease (55% vs 42%, p<0.001) and concentric left ventricular hypertrophy (63% vs 37%, p<0.001). After correction for other factors, increased left ventricular filling pressure remained an independent predictor of mortality after successful AVR (adjusted HR 1.45 (95% CI 1.16 to 1.81), p=0.005). Conclusions Preoperative increased left ventricular filling pressure is common in patients with AS undergoing AVR and has important prognostic implications, regardless of symptom status. Future prospective studies should consider whether patients with increased filling pressure would benefit from earlier operation.