Jones Christine M, Roth Benjamin, Mercado Ana M, Russell Kathy A, Daskalogiannakis John, Samson Thomas D, Hathaway Ronald R, Smith Andrea, Mackay Donald R, Long Ross E
Division of Plastic Surgery, Penn State Hershey Medical Center, Hershey.
Lancaster Cleft Palate Clinic, Lancaster.
J Craniofac Surg. 2018 Jan;29(1):105-108. doi: 10.1097/SCS.0000000000004102.
This study was conducted to determine if nasolabial appearance is rated with comparable results and reliability on 3-dimensional stereophotogrammetric facial images versus standard clinical photographs (2-dimensional). Twenty-seven consecutively treated patients with repaired complete unilateral cleft lip and palate were selected. Six trained and calibrated raters assessed cropped 2- and 3-dimensional facial images. Nasolabial profile, nasolabial frontal, and vermillion border esthetics were rated with the 5-point scale described by Asher-McDade using the modified Q-sort method. Cropped 3-dimensional images were available for viewing by each rater, allowing for complete rotational control for viewing the images from all aspects. Two- and three-dimensional ratings were done separately and repeated the next day.Interrater reliability scores were good for 2-dimensional (κ = 0.607-0.710) and fair to good for 3-dimensional imaging (κ = 0.374-0.769). Intrarater reliability was good to very good for 2-dimensional (κ = 0.749-0.836) and moderate to good for 3-dimensional imaging (κ = 0.554-0.855). Bland-Altman analysis showed satisfactory agreement of 2- and 3-dimensional scores for nasolabial profile and nasolabial frontal, but more systematic error occurred in the assessment of vermillion border.Although 3-dimensional images may be perceived as more representative of a direct clinical facial evaluation, their use for subjective rating of nasolabial aesthetics was not more reliable than 2-dimensional images in this study. Conventional 2-dimensional images provide acceptable reliability while being readily accessible for most cleft palate centers.
本研究旨在确定在三维立体摄影面部图像与标准临床照片（二维）上，鼻唇外观的评分结果和可靠性是否具有可比性。选取了27例连续接受治疗的单侧完全性唇腭裂修复患者。6名经过培训和校准的评估者对裁剪后的二维和三维面部图像进行评估。采用改良Q分类法，按照阿舍 - 麦克戴德描述的5分制对鼻唇轮廓、鼻唇正面和唇红缘美学进行评分。每位评估者都可以查看裁剪后的三维图像，能够从各个角度对图像进行完全旋转控制以进行查看。二维和三维评分分别进行，并在第二天重复。评估者间信度评分在二维时良好（κ = 0.607 - 0.710），在三维成像时为中等至良好（κ = 0.374 - 0.769）。评估者内信度在二维时为良好至非常好（κ = 0.749 - 0.836），在三维成像时为中等至良好（κ = 0.554 - 0.855）。布兰德 - 奥特曼分析显示，二维和三维评分在鼻唇轮廓和鼻唇正面方面具有令人满意的一致性，但在唇红缘评估中出现了更多的系统误差。尽管三维图像可能被认为更能代表直接的临床面部评估，但在本研究中，其用于鼻唇美学主观评分并不比二维图像更可靠。传统的二维图像在大多数腭裂中心易于获取，同时提供了可接受的可靠性。
