Application Of Transluminal Attenuation Gradient In Arteriosclerosis Obliterans Of Lower Extremity

Purpose:To validate the transluminal attenuation gradient（TAG）and optimized TAG post-processing techniques in the study of arteriosclerosis obliterans（ASO）.Methods:A retrospective collection of 53 cases of ASO patients with complete case data from January2016 to August 2018,totaling 88 femoral arteries is made,and all cases are examined by ultrasound and lower extremity arterial CTA examination in the following 0-12 days（average 4 days）.General information on general ASO patients is collected,including gender,age,height,weight,smoking history,history of diabetes,history of hypertension,total cholesterol（TC）,and triglycerides（TG）.The research object is superficial femoral artery.The arteries are divided into three groups according to the degree of superficial femoral artery stenosis:Group A is with mild stenosis（<50%）;Group B with moderately narrow（50%⁷5%）;Group C a severe stenosis（>75%and≤100%）.Patients with ASO are divided into four phases according to clinical Fontaine staging:stage I（asymptomatic）,stage II（intermittent claudication）,stage III（rest pain）,and stage IV（tissue ulcer,gangrene）.The whole process is divided into three steps.First,two experienced physicians use CT post-processing software to define the opening of the superficial femoral artery（deep midpoint of the inguinal ligament）to the end（the upper edge of the medial and lateral femoral condyles）as the analysis area,and place ROI of the sizes1.0mm²,3.0mm²,and 4.0mm² in the center of the lumen according its transverse image of the corresponding position.Secondly,TAG is measured at intervals of 10mm,20mm,and 30mm on the transverse image,and recorded as TAG,TAG²_1mm,TAG²_4mm,TAG_20mm,and TAG30mm.Thirdly,another analysis area of 20mm above and below the most stenotic part of the superficial femoral artery is designated,another TAG is measured there and recorded as TAG^*_20mm.All data are processed by SPSS 22.0,and the consistency of TAG values of the two physicians is analyzed by Bland-Altman.Kruskal Wallis test is used to compare the statistical differences between different stenosis groups and Fontaine staging.Dunn’s multiple comparisons are used for comparison between groups.The consistency of TAG,TAG²_1mm,TAG²_4mm,TAG_20mm,TAG_30mm and TAG^*_20mm is analyzed by Bland-Altman analysis and Spearman correlation coefficient.The relationship between TAG and PSV is analyzed by Spearman correlation coefficient.Results:（1）No statistically significant difference can be seen in related biochemical indexes between the two groups with different degrees of stenosis（P<0.01）（2）the mean values of TAG determined by different physicians are（-2.481±0.246）HU/10mm and（-2.678±2.520）HU/10mm,the difference of which shows no statistical significance.（P<0.05）The Bland-Altman analysis shows that the mean value of TAG difference determined by two different physicians is 0.197 HU/10mm,and the upper and lower boundary of 95%consistency limit is（-1.410 HU/10mm）to（1.805 HU/10mm）.The linear regression analysis shows a high correlation coefficient（r=0.946,P<0.01）.The difference of TAG among the three groups with different stenosis degrees is statistically significant（P<0.01）,which shows that TAG is positively correlated with corresponding PSV（r=0.5502,P<0.0001）.The difference of TAG between different Fontaine stages shows no statistical significance（P>0.01）.The difference between Stage I and Stage IV is statistically significant（P<0.05）.The mean values of TAG,TAG²_1mm and TAG²_4mm are（-2.580±0.257）HU/10mm,（-2.541±0.254）HU/10mm and（-2.612±0.266）HU/10mm,which shows no statistical significance（F=0.01848,P>0.05）.No statistical differences between TAG and TAG²_1mm and between TAG and TAG²_4mm can be seen（P>0.05）.The Bland-Altman analysis shows that the mean differences between TAG and TAG²_1mm and between TAG and TAG²_4mm are respectively（-0.039;-0.031）HU/10mm,the upper and lower boundaries of the 95%consistency boundary are respectively（-0.142 HU/10mm）to（0.064 HU/10mm）and（-0.542 HU/10mm）to（0.605 HU/10mm）.The linear regression analysis shows a high correlation coefficient.（r=0.999,P<0.0001;r=0.993,P<0.0001）The mean values of TAG_20mm,TAG_30mm and TAG^*_20mm are respectively（-2.491±0.265）HU/10mm,（-2.666±0.298）HU/10mm and（-1.869±2.828）HU/10mm,and the differences between TAG and TAG_20mm（t=0.0286,P>0.05）and between TAG and TAG_30mm（t=0.478,P>0.05）show no statistical significance.The correlation coefficient between TAG and TAG_20mm and between TAG and TAG_30mm are high,and the consistency of TAG between different adjacent ROI intervals is good（r=0.963,P<0.0001;r=0.914,P<0.0001）.The Bland-Altman analysis shows that the mean differences between TAG and TAG_20mm and TAG and TAG_30mm are（0.010;0.185）HU/10mm,the upper and lower bounds of the 95%consistency boundary are respectively（-1.315 HU/10mm）to（1.336 HU/10mm）and（-2.092 HU/10mm）to（2.461 HU/10mm）.There are no significant difference（U=3569,P>0.05）and correlativity（r=-0.1008,P>0.05）between TAG and TAG^*_20mm.Conclusion:The CTA-based TAG is an advanced method to measure the different degrees of stenosis of the superficial femoral artery and is associated with PSV.It has the potential to give more hemodynamic information on the basis of CTA morphological diagnosis of lower extremity arterial stenosis.