数字乳腺三维断层摄影技术平均腺体剂量与图像质量的关系

【摘要】目的:探究并分析数字乳腺三维断层摄影技术平均腺体剂量（AGD）与图像质量之间的关系。方法:收集在福建医科大学附属第二医院采用乳腺三维断层摄影技术进行检查的236例患者的临床资料，根据患者乳腺厚度的不同，分为4组，每组各59人。采用非参数统计以及相关性分析的方法，对不同厚度乳腺的曝光条件（mAs）和AGD的相关性进行分析。对同一厚度，不同AGD条件下成像的乳腺图片，分析其图像信噪比与AGD之间的关系。结果:当乳腺厚度增加时，曝光条件（mAs）与AGD均随之增加，具有明显的相关性（r=0.977, P<0.05）。对于相同厚度的乳腺图片，AGD越高，图像信噪比越高，具有明显的相关性（厚度≤29 mm:r=0.977, P=0.023；厚度30~45 mm:r=0.994, P=0.006；厚度46~59 mm:r=0.998, P=0.002；厚度≥60mm:r=0.980, P=0.020）。结论:乳腺的厚度越厚，其曝光条件（mAs）与AGD就越大；对同一厚度的乳腺，AGD越高，成像的图像信噪比越高，图像质量越好。

Relationship between average glandular dose in digital breast tomosynthesis and image quality

Abstract: Objective To explore and analyze the relationship between average glandular dose (AGD) in digital breast tomosynthesis and image quality. Methods The clinical data of 236 patients examined by digital breast tomosynthesis in the Second Affiliated Hospital, Fujian Medical University were collected. According to the thickness of mammary gland, the enrolled patients were divided into 4 groups, with 59 patients in each group. Non-parametric statistics and correlation analysis were used to analyze the correlation between the exposure (mAs) for the mammary gland of different thicknesses and AGD. For the mammary gland of the same thickness, the mammograms obtained under different AGD were used for analyzing the relationship between image signal-to-noise ratio and AGD. Results When the thickness of mammary gland increased, the exposure (mAs) and AGD increased, and there was a significant correlation (r=0.977, P<0.05). For the mammogram of the mammary gland of the same thickness, the higher the AGD was, the higher the image signal-to-noise ratio was, with obvious correlations (thickness≤29 mm: r=0.977, P=0.023; thickness of 30-45 mm: r=0.994, P=0.006; thickness of 46-59 mm: r=0.998, P=0.002; thickness≥60 mm: r=0.980, P=0.020). Conclusion The thicker the mammary gland is, the greater the exposure (mAs) and AGD are. For the mammary gland of the same thickness, with a higher AGD, the signal-to-noise ratio of image is higher and image quality is better.