The application of deep learning algorithms for medical diagnosis in the real world faces challenges with transparency and interpretability. The labeling of large-scale samples leads to costly investment in developing deep learning algorithms. The application of human prior knowledge is an effective way to solve these problems. Previously, we developed a deep learning system for glaucoma diagnosis based on a large number of samples that had high sensitivity and specificity. However, it is a black box and the specific analytic methods cannot be elucidated. Here, we establish a hierarchical deep learning system based on a small number of samples that comprehensively simulates the diagnostic thinking of human experts. This system can extract the anatomical characteristics of the fundus images, including the optic disc, optic cup, and appearance of the retinal nerve fiber layer to realize automatic diagnosis of glaucoma. In addition, this system is transparent and interpretable, and the intermediate process of prediction can be visualized. Applying this system to three validation datasets of fundus images, we demonstrate performance comparable to that of human experts in diagnosing glaucoma. Moreover, it markedly improves the diagnostic accuracy of ophthalmologists. This system may expedite the screening and diagnosis of glaucoma, resulting in improved clinical outcomes.

No abstract

Objective. Cerebral edema is a common complication of brain tumors in the perioperative period. However, there is currently no reliable and convenient method to evaluate the extent of brain edema. The objective is to explore the effectiveness of serum occludin on predicting the extent of perioperative brain edema and outcome in patients with brain tumors. Methods. This prospective study enrolled 55 patients with brain tumors and 24 healthy controls in Sanbo Brain Hospital from June 2019 through November 2019. Serum occludin levels were measured preoperatively and on postoperative day 1. Peritumoral edema was assessed preoperatively using MRI. Pericavity brain edema on postoperative day 1 was evaluated using CT. Results. Compared with healthy controls, the serum occludin level was higher in patients with brain tumors both preoperatively and postoperatively (P<0.001). The serum occludin level correlated positively with the degree of brain edema preoperatively (r=0.78, P<0.001) and postoperatively (r=0.59, P<0.001). At an optimal cutoff of 3.015 ng/mL, the preoperative serum occludin level discriminated between mild and severe preoperative brain edema with a sensitivity of 90.48% and specificity of 84.62%. At an optimal cutoff value of 3.033 ng/mL, the postoperative serum occludin level distinguished between mild and severe postoperative brain edema with a sensitivity of 97.30% and specificity of 55.56%. Conclusions. The serum occludin level is associated with cerebral edema and could potentially be used as a biomarker for perioperative cerebral edema. This trial is registered with ChiCTR1900023742.