Integrating Echocardiography Parameters With Explainable Artificial Intelligence for Data-Driven Clustering of Primary Mitral Regurgitation Phenotypes.

BACKGROUND: Primary mitral regurgitation (MR) is a heterogeneous clinical disease requiring integration of echocardiographic parameters using guideline-driven recommendations to identify severe disease. OBJECTIVES: The purpose of this preliminary study was to explore novel data-driven approaches to delineate phenotypes of MR severity that benefit from surgery. METHODS: The authors used unsupervised and supervised machine learning and explainable artificial intelligence (AI) to integrate 24 echocardiographic parameters in 400 primary MR subjects from France (n = 243; development cohort) and Canada (n = 157; validation cohort) followed up during a median time of 3.2 years (IQR: 1.3-5.3 years) and 6.8 (IQR: 4.0-8.5 years), respectively. The authors compared the phenogroups' incremental prognostic value over conventional MR profiles and for the primary endpoint of all-cause mortality incorporating time-to-mitral valve repair/replacement surgery as a covariate for survival analysis (time-dependent exposure). RESULTS: High-severity (HS) phenogroups from the French cohort (HS: n = 117; low-severity [LS]: n = 126) and the Canadian cohort (HS: n = 87; LS: n = 70) showed improved event-free survival in surgical HS subjects over nonsurgical subjects (P = 0.047 and P = 0.020, respectively). A similar benefit of surgery was not seen in the LS phenogroup in both cohorts (P = 0.70 and P = 0.50, respectively). Phenogrouping showed incremental prognostic value in conventionally severe or moderate-severe MR subjects (Harrell C statistic improvement; P = 0.480; and categorical net reclassification improvement; P = 0.002). Explainable AI specified how each echocardiographic parameter contributed to phenogroup distribution. CONCLUSIONS: Novel data-driven phenogrouping and explainable AI aided in improved integration of echocardiographic data to identify patients with primary MR and improved event-free survival after mitral valve repair/replacement surgery.

Frequency and predictors for chronic thromboembolic pulmonary hypertension after a first unprovoked pulmonary embolism: Results from PADIS studies.

BACKGROUND: Chronic thromboembolic pulmonary hypertension (CTEPH) is a life-threatening complication of a pulmonary embolism (PE) whose incidence and predictors are not precisely determined. OBJECTIVE: To determine the frequency and predictors for CTEPH after a first unprovoked PE. PATIENTS/METHODS: In a randomized trial comparing an additional 18-month warfarin versus placebo in patients after a first unprovoked PE initially treated with vitamin K antagonist for 6 months, we applied recommended CTEPH screening strategies through an 8-year follow-up to determine cumulative incidence of CTEPH. CTEPH predictors were estimated using Cox models. Pulmonary vascular obstruction (PVO) and systolic pulmonary arterial pressure (sPAP) at PE diagnosis and 6 months were studied by receiver operating curves analysis. All CTEPH cases and whether they were incident or prevalent were adjudicated. RESULTS: During a median follow-up of 8.7 years, nine CTEPH cases were diagnosed among 371 patients, with a cumulative incidence of 2.8% (95% confidence interval [CI] 0.95-4.64), and of 1.31% (95% CI 0.01-2.60) after exclusion of five cases adjudicated as prevalent. At PE diagnosis, PVO > 45% and sPAP > 56 mmHg were associated with CTEPH with a hazard ratio (HR) of 33.00 (95% CI 1.64-667.00, p = .02) and 12.50 (95% CI 2.10-74.80, p < .01), respectively. Age > 65 years, lupus anticoagulant antibodies and non-O blood groups were also predictive of CTEPH. PVO > 14% and sPAP > 34 mmHg at 6 months were associated with CTEPH (HR 63.90 [95% CI 3.11-1310.00, p < .01]and HR 17.2 [95% CI 2.75-108, p < .01]). CONCLUSION: After a first unprovoked PE, CTEPH cumulative incidence was 2.8% during an 8-year follow-up. PVO and sPAP at PE diagnosis and at 6 months were the main predictors for CTEPH diagnosis.

Evaluation of Venous Thromboembolism Recurrence Scores in an Unprovoked Pulmonary Embolism Population: A Post-hoc Analysis of the PADIS-PE trial.

BACKGROUND: We aimed to validate the Men Continue and HERDOO2 (HERDOO2), D-dimer, age, sex, hormonal therapy (DASH), and updated Vienna recurrent venous thromboembolism prediction models in a population composed entirely of first unprovoked pulmonary embolism, and to analyze the impact of the addition of the pulmonary vascular obstruction index (PVOI) on score accuracy. METHODS: Analyses were based on the double-blind, randomized PADIS-PE trial, which included 371 unprovoked pulmonary embolism patients initially treated for 6 months, successively randomized to receive an additional 18 months of warfarin or placebo, and subsequently followed-up for 2 years. RESULTS: The HERDOO2, DASH, and updated Vienna scores displayed C-statistics of 0.61 (95% CI 0.54-0.68), 0.60 (95% CI 0.53-0.66), and 0.58 (95% CI 0.51-0.66), respectively. Only the HERDOO2 score identified low recurrence risk patients (<3%/year) after anticoagulation was stopped. When added to either of the prediction models, PVOI measured at pulmonary embolism diagnosis, after 6 months of anticoagulation, or both, improved scores' C-statistics between +0.06 and +0.11 points and consistently led to identifying at least 50% of patients who experienced recurrence but in whom the scores would have indicated against extended anticoagulation. CONCLUSIONS: In patients with a first unprovoked pulmonary embolism, the HERDOO2 score is able to identify patients with a low recurrence risk after treatment discontinuation. Addition of PVOI improves accuracy of all scores. CLINICAL TRIALS REGISTRATION: URL: http://www.controlled-trials.com. Unique identifier: NCT00740883.

Predictors for Residual Pulmonary Vascular Obstruction after Unprovoked Pulmonary Embolism: Implications for Clinical Practice-The PADIS-PE Trial.

BACKGROUND: We aimed to identify risk factors for residual pulmonary vascular obstruction after a first unprovoked pulmonary embolism (PE). METHODS: Analyses were based on data from the double-blind randomized "PADIS-PE" trial that included 371 patients with a first unprovoked PE initially treated during 6 uninterrupted months; all patients underwent baseline ventilation-perfusion lung scanning at inclusion (i.e., after 6 months of anticoagulation). Each patient's pulmonary vascular obstruction indexes (PVOIs) at PE diagnosis and at inclusion were centrally assessed. RESULTS: Among the 371 included patients, residual PVOI was available in 356 patients, and 150 (42.1%) patients had PVOI ≥ 5%. At multivariable analysis, age > 65 years (odds ratio [OR], 2.81, 95% confidence interval [CI], 1.58-5.00), PVOI ≥ 25% at PE diagnosis (OR, 3.53, 95% CI, 1.94-6.41), elevated factor VIII (OR, 3.89, 95% CI, 1.41-10.8), and chronic respiratory disease (OR, 2.18, 95% CI, 1.11-4.26) were independent predictors for residual PVOI ≥ 5%. Patients with ≥ 1 of these factors represented 94.5% (123 patients) of all patients with residual PVOI ≥ 5%. CONCLUSION: Six months after a first unprovoked PE, age > 65 years, PVOI ≥ 25% at PE diagnosis, elevated factor VIII, or chronic respiratory disease were found to be independent predictors for residual pulmonary vascular obstruction. CLINICAL TRIALS REGISTRATION: URL: http://www.controlled-trials.com. Unique identifier: NCT00740883.

Risk factors for recurrent venous thromboembolism after unprovoked pulmonary embolism: the PADIS-PE randomised trial.

We aimed to identify risk factors for recurrent venous thromboembolism (VTE) after unprovoked pulmonary embolism.Analyses were based on the double-blind randomised PADIS-PE trial, which included 371 patients with a first unprovoked pulmonary embolism initially treated during 6 months who were randomised to receive an additional 18 months of warfarin or placebo and followed up for 2 years after study treatment discontinuation. All patients had ventilation/perfusion lung scan at inclusion (i.e. at 6 months of anticoagulation).During a median follow-up of 41 months, recurrent VTE occurred in 67 out of 371 patients (6.8 events per 100 person-years). In main multivariate analysis, the hazard ratio for recurrence was 3.65 (95% CI 1.33-9.99) for age 50-65 years, 4.70 (95% CI 1.78-12.40) for age >65 years, 2.06 (95% CI 1.14-3.72) for patients with pulmonary vascular obstruction index (PVOI) ≥5% at 6 months and 2.38 (95% CI 1.15-4.89) for patients with antiphospholipid antibodies. When considering that PVOI at 6 months would not be available in practice, PVOI ≥40% at pulmonary embolism diagnosis (present in 40% of patients) was also associated with a 2-fold increased risk of recurrence.After a first unprovoked pulmonary embolism, age, PVOI at pulmonary embolism diagnosis or after 6 months of anticoagulation and antiphospholipid antibodies were found to be independent predictors for recurrence.

The utilisation of the cardiovascular automated radiation reduction X-ray system (CARS) in the cardiac catheterisation laboratory aids in the reduction of the patient radiation dose.

AIMS: The radiation exposure resulting from cardiovascular procedures may increase the risk of cancer, and/or cause skin injury. Whether the novel cardiovascular automated radiation reduction X-ray system (CARS) can help reduce the patient radiation dose in daily clinical practice remains unknown. The aim of this study was to evaluate the reduction in patient radiation dose with the use of CARS in the cardiac catheterisation laboratory (CCL). METHODS AND RESULTS: This study retrospectively analysed 1,403 consecutives patients who underwent a cardiac catheterisation with coronary angiography (CA) and/or a percutaneous coronary intervention (PCI) in the Brest University Hospital over the course of one year. Patient radiation doses (dose area product and air kerma) were collected and compared between the CCL with (new CCL) and without (control CCL) CARS. Additionally, the patient radiation doses according to femoral versus radial access, procedural complexity and body mass index were compared. The radiation lesion position on the skin was assessed by automatically optimising the X-ray source to image distance (SID) and subsequently generating a radiation Dose-Map for those procedures exceeding 3 Gray of exposure. Overall, 447 patients underwent procedures in the control CCL and 956 in the new CCL. Baseline patient and procedural characteristics were similar between the two groups, with the exception of male gender and primary PCI, which were more prevalent in the new CCL group. Compared to the control CCL, the utilisation of the CARS in the new CCL resulted in a reduction of dose area product by 46% for CA, 56% for PCI alone and 54% for CA and PCI during the same procedure. Of note, radial access generated a higher radiation dose than femoral access (p<0.001). In this study, seven patients had an air kerma exceeding 3 Gray; however, only one patient had a skin dose greater than 3 Gray. CONCLUSIONS: The utilisation of the CARS resulted in a significant reduction in patient radiation doses compared to the control equipment. A real-time Dose-Map may help the operator change the projection during complex procedures to reduce the patient skin dose.