A deep learning for prediction of vascular access quality in hemodialysis patients

Abstract

The assessment of vascular access quality is essential for the effectiveness of treatment in patients with chronic kidney disease (CKD) undergoing hemodialysis. Inadequate and sporadic monitoring of vascular access in these patients can escalate the risks of stenosis and ultimately lead to thrombosis. Regular assessments facilitate early detection and intervention, potentially preventing thrombotic events. Vascular access quality assessment is typically conducted by measuring blood flow velocity through the access vessel, necessitating the installation of signal transducers along the blood conduit. However, the high cost and limited availability of such equipment mean that many CKD patients in Thailand are unable to receive consistent vascular access evaluations. This research introduces the development of a prototype device for non-invasive monitoring of blood flow velocity that is cost-effective and utilizes data-driven scientific methods. Machine learning models, particularly Convolutional Neural Networks (CNNs), are utilized to assess vascular access quality swiftly and efficiently. This research compares the performance of this newly developed device and models with existing techniques and models. Through machine learning methodologies, it has been demonstrated that the models predict vascular access quality with high precision, surpassing existing techniques and models. This study represents a significant advancement in potentially reducing mortality rates among CKD patients. It is anticipated that this prototype device will be broadly implemented for monitoring blood flow velocity in hemodialysis patients at all dialysis facilities, offering a more affordable alternative to the costly equipment currently in use.

????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????? ?????????????????????????????????????????????????????????????????????????????????????? ?????????????????????????????????????????????????????????????????????????????????? ?????????????????????????????????????????????????????????????????????????????????????????????????????????? ??????????????????????????????????????????????????????????????????????????????? ???????????????????????????-????????????????????? ???????????? ??????????????????????????????????????? ?????????????????????????????????????????????????????????? ???????????????????????????????????????? ??????????? ????????????????????????????????????????????????????????????????????????????????????????????????? ?????????????????????????????????? ????????????????????????????????????????????????????????????????????????? ??????????????????????????????????????????????????????????????????????????????????????? ?????????????????????????????????????????????????????????????????????????? ?????????????????????????????????? ?????????????????????? ????? ????????????????????????????????????????????????????? ?????????????????????????????????????????????????????????? ???????????????? ???????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????? ????????????????????????????????????????????????????????????????????