原始数据迭代重建算法在胃癌术前分期诊断中的应用价值研究

Value of SAFIRE for staging gastric cancer

目的 ：比较原始数据迭代重建算法（sinogram affirmed iterative reconstruction algorithm,SAFIRE）与传统滤波反投影（filtered back projection,FBP）重建算法在胃癌术前CT分期中的图像质量。方法：选取40例经手术病理证实为进展期胃癌的患者,所有患者术前均行128层双源炫速CT扫描并分期。将原始数据通过SAFIRE的1~5级及FBP算法分别进行重建,获得5 mm层厚的CT图像。测量并比较所得图像的胃壁病灶、肝脏、右侧竖脊肌、皮下脂肪及腹主动脉的噪声和CT值、病灶的对比噪声比（contrast-to-noise ratio,CNR）和信噪比（signal-to-noise ratio,SNR）及腹主动脉SNR,并对胃癌术前分期的图像质量进行主观评分。结果：FBP重建算法与SAFIRE（1~5级）间各部位所测得的平均CT值差异无统计学意义（P〉0.05）。FBP重建图像中病灶的平均噪声最高,而在SAFIRE图像中,随着1~5等级的增高,噪声逐渐降低,CNR及SNR也随之降低。主观评分中,FBP重建图像的评分最高,而SAFIRE图像随等级的升高,评分逐渐降低;其中,FBP重建算法与SAFIRE S1、SAFIRE S2与S3的主观评分结果间差异均无统计学意义（P〉0.05）。结论：SAFIRE技术应用于胃癌术前分期检查中,虽能降低图像噪声,提高图像的CNR、SNR,但相比常规FBP重建算法并不能提高进展期胃癌术前分期的准确率。在相同扫描条件下,更推荐应用FBP重建算法或SAFIRE的S1图像来对进展期胃癌进行分期诊断。

Objective： To compare the value of sonogram affirmed iterative reconstruction algorithm（SAFIRE） and filtered back projection（FBP） in CT scanning for the staging of gastric cancer. Methods： Forty patients with pathologically confirmed advanced gastric cancer were enrolled. All received CT scanning for preoperative staging of gastric cancer using the dual-source flash CT. The raw data were reconstructed with SAFIRE（grade 1-5） and FBP of slice thickness 5 mm.The image noise and CT value of the gastric lesion, liver, the right erector spinae muscle, subcutaneous fat and abdominal aorta, the CNR（contrast-to-noise ratio, CNR） and SNR（signal-to-noise ratio, SNR） of the lesion, and the SNR of abdominal aorta were statistically analyzed. Subjective image quality and rating of preoperative gastric cancer staging were also eval-uated. Results： There was no significant difference in average CT values measured between FBP and SAFIRE grade1-5（P〉0.05）. The average noise of lesions for FBP was the highest. With the increase of SAFIRE grade from 1-5, the noise, CNR and SNR decreased gradually. In subjective rating, the score of FBP images was the highest, and the score of SAFIRE decreased gradually with increase of grade level; the score of FBP and S1, S2 and S3 were not statistically diffe-rent（P〉0.05）. Conclusions： Although the use of SAFIRE in preoperative gastric cancer staging could decrease the noise,CNR and SNR of whole image, yet it could not improve the diagnostic accuracy of preoperative gastric cancer staging when compared with FBP. Under conditions of the same scan protocol, FBP or S1 images are preferred for staging of advanced gastric cancer.