单色能量CTA成像结合锐利迭代重建技术评价支架术后再狭窄的效果

Application of CTA with monoenergetic imaging combined with iterative reconstruction using a sharp convolution kernel in the evaluation of in-stent restenosis

摘要:

目的：研究下肢CTA单色能量成像结合中等锐利迭代重建（B60f，SAFIRE）技术在下肢血管支架植入术后再狭窄（in-stent restenosis，ISR）诊断中的应用价值。方法：对72例（男37，女35）股动脉植入支架的患者行下肢动脉CT血管造影（comput－er tomography angiography，CTA）。25例采用双能量扫描，中等锐利迭代重建(B60f SAFIRE)，分离出2组（70 keV锐，80 keV锐）单色能量图像。47例采用100 kV扫描，其中23例（100 kV标组）标准平滑迭代重建（50% ASIR），24例（100 kV锐组）中等锐利迭代重建（B60f SAFIRE）。比较图像噪声（SD）、信噪比（SNR）、对比噪声比（CNR）、支架-血管密度差（SVD）、测量-实际支架壁厚差（MAD），结合血管超声或DSA，评价ISR的诊断准确率。结果：①不同重建同能级：SD，100 kV标<100 kV锐，SNR、CNR之间无统计学差异（P>0.05）。②不同能级同重建比较：SD，70 keV锐<80 keV锐<100 kV锐，各组间差异有统计学意义（P<0.05）。腔内SNR和CNR：100 kV锐<70 keV锐<80 keV锐，组间差异有统计学意义（P<0.05）。③支架-血管密度差（SVD）：70 keV锐<80 keV锐<100 kV锐<100 kV标，测量-真实支架壁厚差（MAD）：80 keV锐<100 kV锐<70 keV锐<100 kV标，组间差异有统计学意义（P<0.05）。④主观图像质量评分：70 keV锐<100 kV标<100 kV锐<80 keV锐（P<0.05），支架内再狭窄的诊断准确率：70 keV锐（71.04%）<100 kV标（86.19%）<100 kV锐（91.27%）<80 keV锐（95.8%），（χ2=9.95，P=0.019），?资=0.743（P<0.05）。结论：80 keV锐单色能量结合中等锐利迭代重建（B60f SAFIRE）成像有助于优化支架血管CTA的图像质量，提高ISR的诊断准确率。

Abstract:

Objective：To study the practical value of computed tomography angiography（CTA） with monoenergetic imaging combined with iterative reconstruction using a medium-sharp convolution kernel（B60f，SAFIRE） in the diagnosis of in-stent restenosis（ISR） of the lower extremities. Methods：CTA of the lower extremity arteries was performed in 72 patients（37 males and 35 females） with a stent implanted in the femoral artery. Among the above patients，25 underwent dual-energy scanning combined with iterative recon－struction using a medium-sharp convolution kernel（B60f，SAFIRE） and were subdivided into two monoenergetic imaging groups（70 keVsharp and 80 keVsharp）；47 underwent 100-kV scanning and were subdivided into 100 kVstandard group（23 cases，using adaptive statistical iterative reconstruction[50% ASIR]） and 100 kVsharp group（24 cases，using iterative reconstruction with a medium-sharp convolution kernel[B60f，SAFIRE]）. These groups were com－pared for image noise（standard deviation[SD]），signal-to-noise ratio（SNR），contrast-to-noise ratio（CNR），stent-vessel deviation（SVD），and measured-actual deviation（MAD）；the diagnostic accuracy for ISR was evaluated by taking into ac－count the results of vascular ultrasound or digital subtraction angiography. Results：In groups of the same energy level and processed with different reconstruction methods，the image noise（SD） was significantly lower in the 100 kVstandard group than in the 100 kVsharp group；there were no significant differences between groups in SNR or CNR（P>0.05）. Between groups of different energy levels and processed with the same reconstruction method，there were significant differences in image noise（SD）（70 keVsharp<80 keVsharp<100 kVsharp，P<0.05） and intraluminal SNR and CNR（100 kV sharp<70 keVsharp<80 keVsharp，P<0.05）. There were significant differences between groups in SVD（70 keVsharp<80 keVsharp<100 kVsharp<100 kVstandard，P<0.05），MAD（80 keVsharp<100 kVsharp<70 keVsharp<100 kVstandard，P<0.05），subjective image quality score（70 keVsharp<100 kVs－tandard<100 kVsharp<80 keVsharp，P<0.05），and diagnostic accuracy for ISR（70 keVsharp（71.04%）<100 kVstandard（86.19%）<100 kVsharp（91.27%）<80 keVsharp（95.8%），（ χ2=9.95，P=0.019），κ=0.743，P<0.05）. Conclusion：Monoenergetic imaging combined with iterative reconstruc－tion using a medium-sharp convolution kernel（B60f，SAFIRE） in the 80 keVsharp group can help optimize the image quality of CTA for stent-implanted vessels and improve the diagnostic accuracy for ISR.