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 impact of the multidisciplinary Endocarditis Team on the management of infective endocarditis.

BACKGROUND: In their latest guidelines for infective endocarditis (IE) (2015), the European Society of Cardiology (ESC) introduced the implementation of the Endocarditis Team (ET) to facilitate the management of IE. This study presents our experiences and the diagnostic and therapeutic impact of the ET on the management of IE. METHODS: From 2016-2020, data of all patients with suspected IE referred to the ET were prospectively collected. The final diagnosis was defined by the ET as either rejected, possible or definite IE. Diagnostic impact was scored as any change in initial diagnosis, the frequency of additional diagnostic tests advised by the ET and any change in diagnosis after these tests. Therapeutic impact was scored as any change in antibiotic therapy or change from conservative to invasive therapy or vice versa. RESULTS: A total of 321 patients (median age 67 [55-77] years, 71% male) were enrolled. The final diagnosis was rejected IE in 47 (15%), possible IE in 34 (11%) and definite IE in 240 (75%) patients. A change of initial diagnosis was seen in 53/321(17%) patients. Additional microbiological tests were advised in 69/321 (21%) patients, and additional imaging tests in 136/321 (42%) patients, which resulted in subsequent change in diagnosis in 23/321 (7%) patients. Any change in antibiotic treatment was advised in 135/321 (42%) patients, and change from initial conservative to additional surgical treatment in 15/321 (5%) patients. CONCLUSION: The ET had a clear impact on the therapeutic policy for patients with suspected IE and is useful in the management of this life-threatening disease. Broad implementation is warranted.

A robust post-insertion method for the preparation of targeted siRNA LNPs.

Targeted delivery of nucleic acids is gaining momentum due to improved efficacy, selectivity, increased circulation time and enhanced tissue retention in target cells. Using nucleic acid-based therapies previously undruggable targets have proven now to be amenable for treatment. Currently, several methods for preparing targeted or labelled delivery vehicles for nucleic acids are based on liposomal formulations. Lipid nanoparticles (LNPs) are structurally different from liposomes and these methods should therefore be evaluated before being translated to siRNA LNPs preparation protocols. Here, we describe a robust and facile method for the preparation of targeted or fluorescently labelled siRNA LNPs. Using a copper free strain-promoted azide-alkyne cycloaddition (SPAAC) we demonstrate that post-insertion of ligand-lipid conjugates into preformed LNPs is superior to direct-surface modification because it preserves the physicochemical parameters of the LNPs. We found that the time point of solvent removal by dialysis is critical and affects the hydrodynamic diameter of the LNPs; post-insertion after dialysis shows the smallest increase in hydrodynamic diameter and polydispersity index (PDI). The post-insertion of ligand-lipid conjugates also proceeded with rapid kinetics and high efficacy over a wide temperature range. Using this optimised protocol, we generated siRNA LNPs containing both targeting and fluorescent tracking ligands allowing us to monitor siRNA LNP uptake kinetics in dependence of the targeting ligand. In aggregate, we describe a robust approach for the generation of targeted and labelled siRNA LNPs that allows their controlled and facile decoration with ligand combinations.

Dual-time-point FDG PET/CT imaging in prosthetic heart valve endocarditis.

PURPOSE: FDG PET/CT has been of increasing interest in the diagnostic workup of prosthetic heart valve endocarditis (PVE). Some reports advocate later imaging time points to improve the diagnostic accuracy for PVE. In this study, we compared standard and late FDG PET/CT images in patients with a clinical suspicion of PVE. MATERIALS AND METHODS: Fourteen scans in 13 patients referred for FDG PET/CT for suspicion of PVE performed at standard (60 min post injection) and late (150 min post injection) time points were scored based on visual interpretation and semi-quantitatively with SUVmax and target-to-background ratio (TBR, defined as [SUVmax valve/SUVmean blood pool]). Final diagnosis was based on surgical findings in all cases of infection (n = 6) and unremarkable follow-up in all others (n = 8). RESULTS: Late images were more prone to false positive interpretation for both visual and semi-quantitative analyses. Visual analysis of the standard images yielded 1 false negative and 1 false positive result. On the late images, no scans were false negative but 5 scans were false positive. CONCLUSION: Late FDG PET/CT imaging for PVE seems prone to false positive results. Therefore, late imaging should be interpreted with caution.

Exposure-survival analyses of pazopanib in renal cell carcinoma and soft tissue sarcoma patients: opportunities for dose optimization.

BACKGROUND: Pazopanib is an angiogenesis inhibitor approved for the treatment of renal cell carcinoma and soft tissue sarcoma. Post hoc analysis of a clinical trial demonstrated a relationship between pazopanib trough concentrations (Cmin) and treatment efficacy. The aim of this study was to explore the pharmacokinetics and exposure-survival relationships of pazopanib in a real-world patient cohort. PATIENTS AND METHODS: Renal cell cancer and soft tissue sarcoma patients who had at least one pazopanib plasma concentration available were included. Using calculated Cmin values and a threshold of > 20 mg/L, univariate and multivariate exposure-survival analyses were performed. RESULTS: Sixty-one patients were included, of which 16.4% were underexposed (mean Cmin < 20 mg/L) using the 800 mg fixed-dosed schedule. In univariate analysis Cmin > 20 mg/L was related to longer progression free survival in renal cell cancer patients (34.1 vs. 12.5 weeks, n = 35, p = 0.027) and the overall population (25.0 vs. 8.8 weeks, n = 61, p = 0.012), but not in the sarcoma subgroup (18.7 vs. 8.8 weeks, n = 26, p = 0.142). In multivariate analysis Cmin > 20 mg/L was associated with hazard ratios of 0.25 (p = 0.021) in renal cancer, 0.12 (p = 0.011) in sarcoma and 0.38 (p = 0.017) in a pooled analysis. CONCLUSION: This study confirms that pazopanib Cmin > 20 mg/L relates to better progression free survival in renal cancer and points towards a similar trend in sarcoma patients. Cmin monitoring of pazopanib can help identify patients with low Cmin for whom individualized treatment at a higher dose may be appropriate.