藤田 英雄

The clinical outcomes of percutaneous coronary intervention (PCI) in patients with dialysis are still worse compared with those without dialysis. Among patients with dialysis, those who started dialysis due to diabetic nephropathy (DMN) may have a worse prognosis than those who started dialysis due to non-DMN. This retrospective study aimed to compare the clinical outcomes in dialysis patients who underwent PCI between with and without long-term dialysis due to DMN. We included 303 dialysis patients with PCI. The length of dialysis at the time of PCI was used to stratify the study patients. Patients with DMN and the length of dialysis ≥ 3 years were defined as the long-DMN group (n = 117), and the others were defined as the other group (n = 186). The primary endpoint was the incidence of major adverse cardiac events (MACE), which was defined as a composite of all-cause death, non-fatal myocardial infarction, re-admission for heart failure, and ischemia-driven target vessel revascularization. A total of 165 MACE were observed with the median follow-up of 568 days. The Kaplan-Meier curves showed that MACE was more frequently observed in the long-DMN group than in the other group (p = 0.005). In the multivariate Cox hazard model, long-DMN was significantly associated with MACE (hazard ratio 1.483, 95% confidence interval 1.075-2.046, p = 0.016) after controlling for multiple confounding factors. Among patients with dialysis, the combination of DMN and a long history of dialysis is closely associated with poor clinical outcomes. These patients should be carefully followed up by both cardiologists and nephrologists.

Percutaneous coronary intervention has been developed for patients with coronary artery disease. Calcified lesions are recognized as an unsolved issue where many clinical devices have evolved and some disappeared. Understanding intracoronary imaging of the calcified lesions can help operators to make decisions during the procedure. There are several potential stories of progression of calcification, although a precise mechanism of progression of calcification remains unknown. In the process of a large calcification, it is histologically believed that lipid is replaced by calcification. This process can be observed by intracoronary imaging devices, i.e., intravascular ultrasound and optical coherence tomography. Calcified nodule is a unique type of calcifications. Among the calcified lesions, especially calcified nodule has serious clinical outcomes such as target lesion revascularization (TLR) with stent under-expansion. Additionally, in-stent calcified nodule is a distinctive type of restenosis pattern after stenting to calcified nodule, leading to malignant cycle of repeated TLR. Recently, calcified nodule is divided into two types based on the surface irregularity: (1) eruptive and (2) non-eruptive calcified nodule. Eruptive calcified nodule has higher rate of target vessel revascularization than non-eruptive calcified nodule despite greater stent expansion in eruptive calcified nodule. It is thought that there are differences of component such as the amount of fibrin and the size of calcific nodules between both, although it is common for both to include calcific nodules and fibrin. Histopathological understanding calcified nodule can be helpful to choose the treatment devices during the procedure in the area where there is no correct answer.

Abstract Aims Delayed diagnosis of pulmonary hypertension (PH) is a known cause of poor patient prognosis. We aimed to develop an artificial intelligence (AI) model, using ensemble learning method to detect PH using electrocardiography (ECG), chest X-ray (CXR), and brain natriuretic peptide (BNP), facilitating accurate detection and prompting further examinations. Methods and results We developed a convolutional neural network model using ECG data to predict PH, labelled by ECG from seven institutions. Logistic regression was used for the BNP prediction model. We referenced a CXR deep learning model using ResNet18. Outputs from each of the three models were integrated into a three-layer fully connected multimodal model. Ten cardiologists participated in an interpretation test, detecting PH from patients’ ECG, CXR, and BNP data both with and without the ensemble learning model. The area under the receiver operating characteristic curves of the ECG, CXR, BNP, and ensemble learning model were 0.818 [95% confidence interval (CI), 0.808–0.828], 0.823 (95% CI, 0.780–0.866), 0.724 (95% CI, 0.668–0.780), and 0.872 (95% CI, 0.829–0.915). Cardiologists’ average accuracy rates were 65.0 ± 4.7% for test without AI model and 74.0 ± 2.7% for test with AI model, a statistically significant improvement (P < 0.01). Conclusion Our ensemble learning model improved doctors’ accuracy in detecting PH from ECG, CXR, and BNP examinations. This suggests that earlier and more accurate PH diagnosis is possible, potentially improving patient prognosis.