多模型自适应统计迭代重建算法在颅脑CT扫描中最优权重组合的研究OA

Objective To explore the optimal combination of pre-and post-weights for adaptive statistical iterative reconstruction-V(ASIR-V)in cranial computed tomography(CT)examinations.Methods Ninety patients undergoing cranial CT plain scans were divided into groups A,B,and C(30 cases each)using the random number table method.Pre-ASIR-V weights of 20％,30％,and 40％were preset for groups A,B,and C before scanning,respectively.After scanning,image reconstruction was performed with post-ASIR-V weights of 50％,60％,70％,and 80％,yielding subgroups A5-A8,B5-B8,and C5-C8.Radiation dose parameters(DLP,CTDIvol),effective dose(ED),signal-to-noise ratio(SNR),and contrast-to-noise ratio(CNR)were recorded.Two radiologists conducted subjective image evaluation using a 5-point scale in a double-blinded manner.Results In objective evaluation,ED decreased linearly with increasing pre-ASIR-V weight(regression equation:y=-0.33x+2.71,r2=0.918,P<0.001).SNR and CNR in all subgroups increased with higher post-ASIR-V weights(P<0.001).Subjective scores showed a trend of first increasing then decreasing;subgroups A6,B7,and C7 had the highest scores at their respective pre-ASIR-V levels,with subgroup B7 scoring the highest overall.Conclusion At a noise index of 3 and tube voltage of 120 kV,using 30％pre-ASIR-V and 70％post-ASIR-V weights in cranial CT reduces noise,improves image quality,and significantly lowers patient radiation dose.