CT图像质量评价体系在肺部疾病筛查中的应用研究

Study on the application of quality evaluation system of CT image in screening of pulmonary disease

目的:建立CT图像质量评价体系,研究图像质量与疾病诊断结果的相关性。方法:选取170例肺部CT检查患者的CT图像资料,按照图像质量排序方法将其分为观察组与对照组,每组85例。对两组患者均采用主观评价,即通过给予评分的方式进行评估和客观评价,即根据人眼的主观视觉系统建立数学模型,并通过具体公式计算图像质量的两种方法,对两组患者图像质量进行评估,统计分析两组图像诊断结果与临床诊治结果的一致性。结果:观察组客观评价与主观评价结果一致性较好,具有统计学意义(r=0.998,P<0.05);图像质量较好的观察组患者,CT图像诊断结果与临床实证结果具有显著的一致性,有统计学意义(r=0.935,P<0.05);图像质量较差的对照组患者,CT图像诊断结果与临床实证结果一致性不佳,无统计学意义(r=0.192,P>0.05)。结论:通过对比噪声比和锐利度两项指标进行图像质量评价的方法,能够有效评估CT图像在肺部疾病诊断中的准确率,本研究建立的图像质量评价体系适用于肺部CT图像的质量评估工作。

Objective： To establish quality evaluation system of CT image so as to research the correlation between image quality and results of disease diagnosis. Methods： Pulmonary CT images of 170 patients were selected and were divided into observation group （85 cases） and control group （85 cases） according to the sequencing method of image quality. The image quality of two groups were evaluated by subjective evaluation （evaluation based on score of observer） and objective evaluation （established mathematical model based on subjective visual system of eyes and image quality was calculated by concrete formulation）. And then the consistency of the results between imaging diagnosis and clinical diagnosis and treatment of different groups was statistically analyzed respectively. Results： Objective evaluation results were significantly consistent with subjective ones in observation group （r=0.998, P〈0.05）. For the patients with better image quality in observation group, the diagnostic results of CT images were significantly consistent with the clinically empirical results （r=0.935, P〈0.05）. While these patients with poorer image quality in control group, the consistency between diagnostic results of CT images and clinically empirical results was not good, and it was no significant （r=0.192, P〉0.05）. Conclusion： The image quality evaluation methods which use contrast-to-noise ratio and sharpness can effectively evaluate accuracy of CT image in the diagnosis of pulmonary diseases. This study shows that the evaluation system of image quality is suitable for quality evaluation of pulmonary CT images.