探讨全栈式自动盆底超声与手动方式获取并测量最小肛提肌裂孔（LH）平面的一致性。

前瞻性选取2020年11月至2021年1月在深圳市第二人民医院进行盆底超声检查的产后6个月内的女性119例为研究对象。应用腹部容积探头经会阴采集患者在静息状态下和最大Valsalva动作下以盆底正中矢状切面为初始切面的三维容积数据，并存储于超声诊断仪中。由2名年轻医师（手动测量年轻医师组）、2名高年资医师（手动测量高年资医师组）分别调取三维容积数据手动获取并测量最小LH平面。由另一名年轻医师（自动测量组）调取三维容积数据利用FSPFU软件一键自动获取并测量最小LH平面，间隔2周后重复测量1次。测量参数为：最小LH的面积、周长、前后径、左右径、左侧肛提肌尿道间隙和右侧肛提肌尿道间隙。记录3组获取并测量最小LH平面所用的时间及测量的结果。采用组间相关系数（ICC）和Bland-Altman图评估3组测量结果的一致性。将3组获得的最小LH轮廓导出，离线计算组间的重合率。

静息状态和Valsalva动作下，均为自动测量组耗时最短[2.67（0.15）s和2.68（0.13）s]，与手动测量高年资医师组和年轻医师组相比，组间差异均有统计学意义（P均<0.001）。自动测量组与手动测量年轻医师组、高年资医师组间3个参数测值的一致性均好，静息状态下ICC范围0.795~0.931、Valsalva动作下ICC范围0.871~0.973（P均<0.001）。手动测量组与自动测量组间的最小LH面积、前后径、左右径的Bland-Altman图显示，大部分散点都在95%一致性界限内，表明一致性良好。自动测量组与手动测量组（年轻医师组、高年资医师组）的最小LH轮廓重合率分别为静息状态下：0.937、0.948，Valsalva状态下：0.934、0.945，自动测量组与手动测量高年资医师组间重合率高于与手动测量低年资医师组重合率，差异有统计学意义（P均<0.05）。

FSPFU软件能够在腹部三维容积数据中快速自动获取并测量最小LH平面，与手动测量结果一致性好，且操作简单、节省时间、测量结果可靠，可作为辅助诊断盆底功能障碍性疾病的有效方法。

To evaluate the consistency of full-stack smart pelvic floor ultrasound (FSPFU) and manual method in obtaining and measuring the minimum levator hiatus (LH) plane.

From November 2020 to January 2021, 119 women within 6 months after delivery who underwent pelvic floor ultrasound examination at Shenzhen Second People's Hospital were prospectively selected as the research subjects. Using the abdominal volume probe to set the midsagittal plane of pelvic floor as the initial plane, the three-dimensional volume data at rest and during the Valsalva maneuver were acquired and stored in the ultrasonic diagnostic instrument. Two young physicians (young physician manual measurement group) and two senior physicians (senior physician manual measurement group) retrieved the three-dimensional volume data to manually obtain and measure the minimum LH plane. Another young physician (automatic measurement group) retrieved the three-dimensional volume data and used the FSPFU software to automatically obtain and measure the minimum LH plane with one click, and the measurement was repeated once after an interval of 2 weeks. The measurement parameters of the minimum LH include area, circumference, anterioposterior diameter, transverse diameter, left-levator-urethra gap, and right-levator urethral gap. The time and results of obtaining and measuring the minimum LH plane were recorded for each group. The interclass correlation coefficient (ICC) and Bland-Altman plot were used to evaluate the consistency among the three groups. The Dice coefficient of the minimal LH contour outlined by the three groups was calculated.

The automatic measurement group took the shortest time both at rest and during the Valsalva maneuver [2.67 (0.15) s and 2.68 (0.13) s, respectively] compared with the two manual measurement groups (P<0.001). The consistency of the three parameters between the automatic measurement group and the two manual measurement groups was good. The ICC ranged from 0.795-0.931 in the resting state and from 0.871-0.973 during the Valsalva maneuver (P<0.001). The Bland Altman plot also showed good consistency among the groups. The Dice coefficient of the minimal LH outlined contours of the automatic group and the two manual measurement groups (young physician group and senior physician group) were 0.937 and 0.948 in the resting state and 0.934 and 0.945 in the Valsalva state, respectively; the Dice coefficient between the automatic measurement group and the senior physician group was higher than that between the automatic measurement group and the young physician group (P<0.05).

FSPFU can quickly and automatically obtain and measure the minimum LH plane based on the three-dimensional volume data of the abdomen. It is simple to operate, has reliable measurement results, and can be used as an effective and time-saving method to assist in the diagnosis of pelvic floor dysfunction diseases.