Using machine learning to improve the diagnostic accuracy of the modified Duke/ESC 2015 criteria in patients with suspected prosthetic valve endocarditis - a proof of concept study.

INTRODUCTION: Prosthetic valve endocarditis (PVE) is a serious complication of prosthetic valve implantation, with an estimated yearly incidence of at least 0.4-1.0%. The Duke criteria and subsequent modifications have been developed as a diagnostic framework for infective endocarditis (IE) in clinical studies. However, their sensitivity and specificity are limited, especially for PVE. Furthermore, their most recent versions (ESC2015 and ESC2023) include advanced imaging modalities, e.g., cardiac CTA and [18F]FDG PET/CT as major criteria. However, despite these significant changes, the weighing system using major and minor criteria has remained unchanged. This may have introduced bias to the diagnostic set of criteria. Here, we aimed to evaluate and improve the predictive value of the modified Duke/ESC 2015 (MDE2015) criteria by using machine learning algorithms. METHODS: In this proof-of-concept study, we used data of a well-defined retrospective multicentre cohort of 160 patients evaluated for suspected PVE. Four machine learning algorithms were compared to the prediction of the diagnosis according to the MDE2015 criteria: Lasso logistic regression, decision tree with gradient boosting (XGBoost), decision tree without gradient boosting, and a model combining predictions of these (ensemble learning). All models used the same features that also constitute the MDE2015 criteria. The final diagnosis of PVE, based on endocarditis team consensus using all available clinical information, including surgical findings whenever performed, and with at least 1 year follow up, was used as the composite gold standard. RESULTS: The diagnostic performance of the MDE2015 criteria varied depending on how the category of 'possible' PVE cases were handled. Considering these cases as positive for PVE, sensitivity and specificity were 0.96 and 0.60, respectively. Whereas treating these cases as negative, sensitivity and specificity were 0.74 and 0.98, respectively. Combining the approaches of considering possible endocarditis as positive and as negative for ROC-analysis resulted in an excellent AUC of 0.917. For the machine learning models, the sensitivity and specificity were as follows: logistic regression, 0.92 and 0.85; XGBoost, 0.90 and 0.85; decision trees, 0.88 and 0.86; and ensemble learning, 0.91 and 0.85, respectively. The resulting AUCs were, in the same order: 0.938, 0.937, 0.930, and 0.941, respectively. DISCUSSION: In this proof-of-concept study, machine learning algorithms achieved improved diagnostic performance compared to the major/minor weighing system as used in the MDE2015 criteria. Moreover, these models provide quantifiable certainty levels of the diagnosis, potentially enhancing interpretability for clinicians. Additionally, they allow for easy incorporation of new and/or refined criteria, such as the individual weight of advanced imaging modalities such as CTA or [18F]FDG PET/CT. These promising preliminary findings warrant further studies for validation, ideally in a prospective cohort encompassing the full spectrum of patients with suspected IE.

The value of 18F-FDG PET/CT for the diagnosis of device-related infections in patients with a left ventricular assist device: a systematic review and meta-analysis.

BACKGROUND: Left ventricular assist devices (LVADs) are increasingly used for the treatment of advanced heart failure. LVADs improve quality of life and decrease mortality, but the driveline carries substantial risk for major infections. These device-related LVAD and driveline infections are difficult to diagnose with conventional imaging. We reviewed and analysed the current literature on the additive value of 18F-fluorodeoxyglucose positron emission tomography combined with computed tomography (FDG-PET/CT) imaging for the diagnosis of LVAD-related infections." MATERIALS/METHODS: We performed a systematic literature review using several databases from their inception until the 31st of December, 2019. Studies investigating the diagnostic performance of FDG-PET/CT in patients with suspected LVAD infection were retrieved. After a bias risk assessment using QUADAS-2, a study-aggregate meta-analysis was performed on a per examination-based analysis. RESULTS: A total of 10 studies were included in the systematic review, eight of which were also eligible for study-aggregate meta-analysis. For the meta-analysis, a total of 256 FDG-PET/CT scans, examining pump/pocket and/or driveline infection, were acquired in 230 patients. Pooled sensitivity of FDG-PET/CT was 0.95 (95% confidence interval (CI) 0.89-0.97) and pooled specificity was 0.91 (95% CI 0.54-0.99) for the diagnosis of device-related infection. For pump/pocket infection, sensitivity and specificity of FDG-PET/CT were 0.97 (95%CI 0.69-1.00) and 0.93 (95%CI 0.64-0.99), respectively. For driveline infection, sensitivity and specificity were 0.96 (95%CI 0.88-0.99) and 0.99 (95%CI 0.13-1.00) respectively. Significant heterogeneity existed across studies for specificity, mostly caused by differences in scan procedures. Predefined criteria for suspicion of LVAD and/or driveline infection were lacking in all included studies. CONCLUSIONS: FDG-PET/CT is a valuable tool for assessment of device-related infection in LVAD patients, with high sensitivity and high, albeit variable, specificity. Standardization of FDG-PET/CT procedures and criteria for suspected device-related LVAD infections are needed for consistent reporting of FDG-PET/CT scans.

Hemodynamic performance of stented and stentless aortic bioprostheses.

BACKGROUND: This study compares the hemodynamic performance of stented and stentless bioprostheses used for aortic valve replacement in patients with aortic stenosis and small aortic root. METHODS: Between 1995 and 1998, 37 patients with a 21-mm aortic annulus (group 1) underwent aortic valve replacement with either a 21-mm Edwards Perimount or a 23-mm St. Jude Toronto bioprosthesis whereas 47 patients with a 23-mm aortic annulus (group 2) received either a 23-mm Medtronic Mosaic or a 25-mm Edwards Prima bioprosthesis. In each group mean and peak gradients, effective orifice area index, and left ventricular mass index were compared during follow-up. RESULTS: Group 1 patients showed a significant reduction of mean (p < 0.001) and peak gradients (p = 0.001) during follow-up, more evident for St. Jude Toronto versus Edwards Perimount (p = 0.02 and p = 0.05, respectively). Group 2 patients showed a significant reduction of mean and peak gradients (p < 0.001), more evident for Edwards Prima versus Medtronic Mosaic (p < 0.001 and p = 0.07, respectively). Effective orifice area index significantly increased only in group 1 (p = 0.005). Left ventricular mass index significantly decreased in all patients regardless of the type of valve (p < 0.001). Patients with Edwards Prima showed a trend to a higher regression of left ventricular mass index versus Medtronic Mosaic recipients (p = 0.07). CONCLUSIONS: After aortic valve replacement, stented and stentless bioprostheses exhibited similar results with a more evident hemodynamic improvement during follow-up in the stentless valves. Stented bioprostheses of new generation, however, may parallel the hemodynamic performance of stentless valves and appear to be a valid alternative for aortic valve replacement in elderly patients with a small aortic annulus.

Evaluation of valve-related complications in patients with Sorin Bicarbon prosthesis: a seven-year experience.

BACKGROUND AND AIM OF THE STUDY: The study aim was to evaluate the medium-term results of aortic valve replacement (AVR), mitral valve replacement (MVR) and double valve replacement (DVR) with the Sorin Bicarbon prosthesis. METHODS: A total of 990 patients (568 men, 422 women; mean age 60+/-10 years; range: 20-86 years) was reviewed who received 1,108 Sorin Bicarbon prostheses between 1992 and 1998 at three institutions. AVR was performed in 541 patients (55%), MVR in 330 (33%) and DVR in 119 (12%). Concomitant procedures, mainly coronary artery grafting, were performed in 222 patients (22%). Follow up was 98% complete; total cumulative follow up was 3,091 patient-years. RESULTS: Hospital mortality was 3.7% (n = 37). There were 49 late deaths; actuarial survival at seven years was 88+/-2% after AVR, 86+/-5% after MVR, and 78+/-8% after DVR. At last follow up, 915 survivors were in NYHA functional class I or II. At seven years, actuarial freedom from valve-related deaths, valve thrombosis, embolism and bleeding respectively was 96+/-1%, 99+/-1%, 93+/-2% and 91+/-3% after AVR; 97+/-2%, 97+/-3%, 90+/-3% and 86+/-7% after MVR; and 92+/-6%, 98+/-1%, 64+/-2% and 82+/-2% after DVR. Reoperation was required in 20 patients (due to valve thrombosis in six, endocarditis in five and periprosthetic leak in nine). At seven years, actuarial freedom from reoperation was 97+/-1%, 96+/-2% and 84+/-9% after AVR, MVR and DVR, respectively; actuarial freedom from endocarditis was 99+/-1%, 99+/-1% and 95+/-5%. Nine patients experienced a nonstructural valve dysfunction (all periprosthetic leak), while no cases of structural failure were observed. CONCLUSION: The Sorin Bicarbon prosthesis has shown good medium-term results with regard to clinical improvement, and low incidence of valve-related complications. Thus, it appears to be a reliable valve substitute when the use of a mechanical prosthesis is indicated.