View clinical trials related to Deep Learning.
Filter by:Breast cancer related lymphedema (BCRL) is the most common complication after breast cancer surgery, which brings a heavy psychological and spiritual burden to patients. For a long time, the diagnosis and treatment of lymphedema has been a difficult point in domestic and foreign research. To a large extent, it is because most of the patients who come to see a doctor have already developed obvious lymphedema, and the internal lymphatic vessels have undergone pathological remodeling[1] Therefore, it is particularly important to detect early lymphedema and intervene in time through the use of sensitive screening tools. Indocyanine green (ICG) lymphangiography is a relatively new method, which can display superficial lymph flow in real time and quickly, and will not be affected by radioactivity [7]. In 2007, indocyanine green lymphography was used for the first time to evaluate the function of superficial lymphatic vessels. In 2011, Japanese scholars found skin reflux signs based on ICG lymphography data of 20 patients with lymphedema after breast cancer surgery, and they were roughly divided into three types according to their severity: splash, star cluster, and diffuse (Figure 1) [8]. Later, in 2016, a prospective study involving 196 people affirmed the value of ICG lymphography in the early diagnosis of lymphedema, and made the images of ICG lymphography more specific stages 0-5 [9], but The staging is still based on the three types of skin reflux symptoms found in a small sample clinical study in 2011, which is not completely applicable in actual clinical applications. In addition, when abnormal skin reflux symptoms appear on ICG lymphangiography, the pathophysiological changes that occur in the body lack research and exploration. Therefore, this research hopes to refine the image features of ICG lymphography through machine learning (deep learning), and establish a PKUPH model for diagnosing early lymphedema by staging the image features.
The objective of this study is to establish deep learning (DL) algorithm to automatically classify multi-diseases from fundus photography and differentiate major vision-threatening conditions and other retinal abnormalities. The effectiveness and accuracy of the established algorithm will be evaluated in community derived dataset.