基于深度学习计算机辅助诊断系统测量脑出血量

目的 探讨基于深度学习计算机辅助诊断系统（DL-CAD）测量脑出血量的应用价值。方法 回顾性分析超急性期、急性期、亚急性期各50例脑出血患者脑部平扫CT资料，分别采用DL-CAD、多田公式法和CT定量法测量脑出血量，记录测量时间，并计算绝对百分误差率（APE）。比较3种方法测量结果、测量时间及APE差异。结果 应用DL-CAD测量超急性期及亚急性期脑出血量结果与CT定量法差异无统计学意义（P均>0.05），而对急性期出血差异有统计学意义（P<0.05）。应用多田公式法测量脑出血结果与CT定量法差异均有统计学意义（P均<0.01）。采用DL-CAD测量不同时期脑出血量，测量时间均显著低于CT定量法及多田公式法（P均<0.01），且APE均低于多田公式法（P均<0.01）。结论 与传统多田公式法相比，DL-CAD测量各期脑出血量准确性更高，且速度更快。

Measuring volume of cerebral hemorrhage based on deep learning computer aided diagnostic system

Objective To explore the application value of deep learning-based computer aided diagnostic system (DL-CAD) in measuring the volume of cerebral hemorrhage. Methods Brain CT data of patients with cerebral hemorrhage in hyperacute, acute and subacute phases (each n=50) were retrospectively analyzed. DL-CAD, TEDA formula and CT quantitative method were used to measure the volume of cerebral hemorrhage, the measuring time was recorded and the absolute probability error (APE) was calculated. The measured results, the measuring time and APE of 3 methods were compared. Results There was no statistical difference between the measuring results of cerebral hemorrhage of hyperacute and subacute phases between DL-CAD and CT quantitative method (both P>0.05), while statistical difference of the acute phase was found (P<0.05). The results of cerebral hemorrhage measured using TEDA formula were statistically different from those of CT quantitative method (all P<0.01). The measuring time of DL-CAD for all periods cerebral hemorrhage were significantly lower than that of CT quantitative method and TEDA formula (all P<0.01). APE of cerebral hemorrhage measured with DL-CAD in all periods were lower than that of TEDA formula (all P<0.01). Conclusion DL-CAD could measure the volume of cerebral hemorrhage in all periods more quickly and accurately than traditional TEDA formula method.