切换至 "中华医学电子期刊资源库"

中华腔镜泌尿外科杂志(电子版) ›› 2025, Vol. 19 ›› Issue (01) : 76 -82. doi: 10.3877/cma.j.issn.1674-3253.2025.01.014

临床研究

深度学习图像重建三低方案在肾动脉血管成像中的应用
张悦1, 张可1, 邓锶锶1, 向青1, 郭亚豪1, 曹键1, 罗涛1, 孟占鳌1,()   
  1. 1.510630 广州,中山大学附属第三医院放射科
  • 收稿日期:2023-06-07 出版日期:2025-02-01
  • 通信作者: 孟占鳌
  • 基金资助:
    广州市科技计划项目(202007030007)中山大学附属第三医院国家自然科学基金培育专项资助项目(2021GZRPYMS06)中山大学附属第三医院“五个五”工程项目(2023WW605)

Application of deep learning-based image reconstruction with three low scheme in renal artery angiography

Yue Zhang1, Ke Zhang1, Sisi Deng1, Qing Xiang1, Yahao Guo1, Jian Cao1, Tao Luo1, Zhan'ao Meng,1()   

  1. 1.Department of Radiology, the Third Affiliated Hospital of Sun Yat-sen University, Guangzhou 510630, China
  • Received:2023-06-07 Published:2025-02-01
  • Corresponding author: Zhan'ao Meng
引用本文:

张悦, 张可, 邓锶锶, 向青, 郭亚豪, 曹键, 罗涛, 孟占鳌. 深度学习图像重建三低方案在肾动脉血管成像中的应用[J/OL]. 中华腔镜泌尿外科杂志(电子版), 2025, 19(01): 76-82.

Yue Zhang, Ke Zhang, Sisi Deng, Qing Xiang, Yahao Guo, Jian Cao, Tao Luo, Zhan'ao Meng. Application of deep learning-based image reconstruction with three low scheme in renal artery angiography[J/OL]. Chinese Journal of Endourology(Electronic Edition), 2025, 19(01): 76-82.

目的

通过深度学习图像重建(DLIR)三低方案(低辐射剂量、低注射速度、低对比剂剂量)与能谱(DECT)低能级方案和自适应统计迭代重建-V(ASIR-V)常规方案进行比较,探讨肾动脉血管成像(CTA)的最佳方案。

方法

研究对象为接受肾动脉CTA检查的90例患者,按扫描方案分为常规方案组(S组30例)、DECT方案组(DE组30例)和三低方案组(L组30例),记录三组对比剂剂量、注射速度、容积CT剂量指数和剂量长度乘积,计算有效剂量。对S组、DE组进行60%ASIR-V重建;对L组进行DLIR-H重建。图像客观评价指标包括腹主动脉、双肾动脉CT值、标准差(SD)值、信噪比(SNR)和对比噪声比(CNR),主观评价由2名临床经验丰富的诊断医师对重建图像进行双盲法评分。三组CT,SD、SNR、CNR值等客观评价参数的分析采用单因素方差分析,主观评分采用Kruskal-Wallis检验,用Kappa检验分析2名放射科医师主观评分的一致性。

结果

L组有效剂量较S组降低了35%,对比剂量降低34%,对比剂注射速度降低30%;较DE组降低42%,对比剂量降低36%,对比剂注射速度降低30%。L组SD值小于S组,SNR、CNR值大于S组(P<0.05),DE组CT值、噪声最大(P<0.05);CNR最大(P<0.05)。主观评分综合2名医师五个因素平均得分L组最高(P<0.05)。2名放射科医师对三组肾动脉的主观图像质量评分亦有较强的一致性(Kappa=0.895,P<0.001;Kappa=0.643,P<0.001;Kappa=0.764,P<0.001)。

结论

三低方案对比DECT方案和常规方案可显著降低辐射剂量,对比剂剂量和注射速度,且获得的组综合图像质量最佳,是肾动脉CTA最佳方案,可以在客观数据没有优势的情况下拥有最高的图像评价。

Objective

The best scheme of renal artery angiography (CTA) was discussed by comparing the three low scheme of deep learning image reconstruction (DLIR) (low radiation dose, low injection rate, low contrast agent dose) with the low energy spectrum (DECT) scheme and the conventional scheme of adaptive statistical iterative Reconstruction -V(ASIR-V).

Methods

Ninety patients receiving renal artery CTA examination were divided into conventional regimen group (S group 30 cases), DECT regimen group (DE group 30 cases) and three low regimen group (L group 30 cases) according to the scanning regimen. The dose of contrast agent, injection speed, volume CT dose index and dose length product of the three groups were recorded, and the effective dose was calculated. 60%ASIR-V reconstruction was performed in groups S and DE. DLR -H reconstruction was carried out for group L. Objective image evaluation indexes included abdominal aorta and double renal artery CT values, standard deviation (SD) values, signal-tonoise ratio (SNR) and contrast-to-noise ratio (CNR). Subjective evaluation was performed by two clinically experienced diagnostic physicians who scored reconstructed images by double-blind method. Objective evaluation parameters such as CT, SD, SNR and CNR values of the three groups were analyzed by one-way analysis of variance, subjective scores were analyzed by Kruskal-Wallis test, and the consistency of subjective scores of two radiologists was analyzed by Kappa test.

Results

Compared with group S, the effective dose in group L was reduced by 35%, the contrast dose by 34%, and the injection speed of contrast agent by 30%.Compared with DE group, the contrast dose was decreased by 42%, the contrast dose was decreased by 36%,and the contrast injection speed was decreased by 30%. SD of L group was smaller than that of S group, SNR and CNR values were larger than that of S group (P<0.05), and the CT value and noise of DE group were the highest (P<0.05), CNR value was the largest (P<0.05). The average score of five factors was the highest in L group (P<0.05). The subjective image quality scores of renal arteries from two radiologists were also consistent(Kappa=0.895, P<0.001; Kappa=0.643, P<0.001; Kappa=0.764, P<0.001).

Conclusions

Three low regimen compared with the low energy spectrum scheme and conventional regimen significantly reduced the radiation dose,the contrast dose, and the injection speed, and it has the best comprehensive image quality, which is the best scheme of renal artery CTA, and can have the highest image evaluation in the absence of advantages of objective data.

表1 90例接受肾动脉CTA检查患者一般数据和辐射剂量比较
图1 90例接受肾动脉CAT检查患者客观参数柱形图 注:实线代表两组间差异无统计学意义,其余差异有统计学意义(P<0.05)
图2 肾动脉CTA最佳对比噪声(CNR)曲线图 注:a为融合图像CNR指标勾画图展示(红色小圆圈);b为40Kev下肾动脉CTA的CNR最大并随能级增高逐渐降低
表2 90例接受肾动脉CTA检查患者图像客观参数比较(±s
图3 接受肾动脉CTA检查的患者肾动脉最大强度投影 注:图a、b、c、d为深度学习重建三低方案组(L组)4例患者,图e、f、g、h为能谱CT组(DECT)患者,图i、j、k、l为常规方案(S组)患者;黄色箭头示主动脉及肾动脉各级血管清晰度、点状噪声L组与DE组相近,较S组有明显优势;红色箭头示分支显示度DE组最弱,L组最佳
表3 90例接受肾动脉CAT检查患者图像主观参数比较(±s
[1]
Rountas C, Vlychou M, Vassiou K, et al. Imaging modalities for renal artery stenosis in suspected renovascular hypertension:prospective intraindividual comparison of color Doppler US, CT angiography, GD-enhanced MR angiography, and digital substraction angiography[J]. Ren Fail, 2007, 29(3): 295-302.
[2]
张文. 老年缺血性肾病的诊断和治疗[J]. 肾脏病与透析肾移植杂志, 2023, 32(4): 346-347.Zhang W. Diagnosis and treatment of elderly ischemic kidney disease[J]. Chin J Nephrol Dial Transplant, 2023, 32(4): 346-347.
[3]
Donaldson JS. Computed tomography angiography for renal artery stenosis in children: a flip flop isn't always bad[J]. Pediatr Radiol,2021, 51(3): 383-384.
[4]
Liu S, Li W, Shi H, et al. Low-dose scanning technology combined with low-concentration contrast material in renal computed tomography angiography (CTA): a preliminary study[J]. Med Sci Monit, 2017, 23: 4351-4359.
[5]
孙顗淼, 张颖. 糖尿病患者急性脑梗死取栓术后发生对比剂肾病的影响因素及预测模型建立[J/OL]. 中华肾病研究电子杂志,2024, 13(4): 188-194.Sun YM, Zhang Y. Influencing factors and prediction modeling of contrast-induced nephropathy after thrombectomy in diabetic patients with acute cerebral infarction[J/OL]. Chin J Kidney Dis Investig Electron Ed, 2024, 13(4): 188-194.
[6]
杨绍汪. 经皮冠状动脉介入治疗急性心肌梗死患者术后发生对比剂肾病的风险模型的构建[J]. 实用医学杂志, 2023, 39(15): 1925-1931.Yang SW. Construction of a risk model of contrast-induced nephropathy after percutaneous coronary intervention for acute myocardial infarction[J]. J Pract Med, 2023, 39(15): 1925-1931.
[7]
Jensen CT, Liu X, Tamm EP, et al. Image quality assessment of abdominal CT by use of new deep learning image reconstruction:initial experience[J]. AJR Am J Roentgenol, 2020, 215(1): 50-57.
[8]
向青, 曹键, 罗涛, 等. 深度学习图像重建算法在80kV管电压下冠状动脉CT血管造影中的应用[J]. 新医学, 2024, 55(9): 685-692.Xiang Q, Cao J, Luo T, et al. Application of CCTA under 80 kV tube voltage based on deep learning image reconstruction algorithm[J].New Med, 2024, 55(9):685-692.
[9]
唐友发, 王秋霞, 张进华. 深度学习重建算法在肠系膜上动脉CT血管成像中的应用评估[J]. 暨南大学学报(自然科学与医学版),2023, 44(3): 316-322. DOI: 10.11778/j.jdxb.20230069.Tang YF, Wang QX, Zhang JH. Evaluation of the application of deep learning reconstruction algorithm in superior mesenteric artery CT angiograpy[J]. J Jinan Univ Nat Sci Med Ed, 2023, 44(3):316-322.
[10]
Koetzier LR, Mastrodicasa D, Szczykutowicz TP, et al. Deep learning image reconstruction for CT: technical principles and clinical prospects[J]. Radiology, 2023, 306(3): e221257.
[11]
Verhagen MV, Dikkers R, de Kleine RH, et al. Assessment of hepatic artery anatomy in pediatric liver transplant recipients: MR angiography versus CT angiography[J]. Pediatr Transplant, 2021,25(4): e14002.
[12]
Zhao XY, Li LL, Song J, et al. Effects of adaptive statistical iterative reconstruction-V technology on the image quality and radiation dose of unenhanced and enhanced CT scans of the piglet abdomen[J].Radiat Res, 2022, 197(2): 157-165.
[13]
Ren Z, Zhang X, Hu Z, et al. Application of adaptive statistical iterative reconstruction-V with combination of 80 kV for reducing radiation dose and improving image quality in renal computed tomography angiography for slim patients[J]. Acad Radiol, 2019,26(11): e324-e332.
[14]
Li Y, Liu X, Zhuang XH, et al. Assessment of low-dose paranasal sinus CT imaging using a new deep learning image reconstruction technique in children compared to adaptive statistical iterative reconstruction V (ASiR-V)[J]. BMC Med Imaging, 2022, 22(1): 106.
[15]
Zhang X, Chen J, Yu N, et al. Reducing contrast medium dose with low photon energy images in renal dual-energy spectral CT angiography and adaptive statistical iterative reconstruction (ASIR)[J]. Br J Radiol, 2021, 94(1120): 20200974.
[16]
Cao J, Lennartz S, Pisuchpen N, et al. Renal lesion characterization by dual-layer dual-energy CT: comparison of virtual and true unenhanced images[J]. AJR Am J Roentgenol, 2022, 219(4): 614-623.
[17]
Mangold D, Salatzki J, Riffel J, et al. Dual-layer spectral CTA for TAVI planning using a split-phase protocol and low-keV virtual monoenergetic images: improved image quality in comparison with single-phase conventional CTA[J]. Rofo, 2022, 194(6): 652-659.
[18]
Gao L, Lv Y, Jin Y, et al. Differential diagnosis of hepatic cancerous nodules and cirrhosis nodules by spectral CT imaging: a feasibility study[J]. Acta Radiol, 2019, 60(12): 1602-1608.
[19]
Willemink MJ, Noël PB. The evolution of image reconstruction for CT-from filtered back projection to artificial intelligence[J]. Eur Radiol, 2019, 29(5): 2185-2195.
[20]
Akagi M, Nakamura Y, Higaki T, et al. Deep learning reconstruction improves image quality of abdominal ultra-high-resolution CT[J].Eur Radiol, 2019, 29(11): 6163-6171.
[21]
Carrascosa P, Leipsic JA, Capunay C, et al. Monochromatic image reconstruction by dual energy imaging allows half iodine load computed tomography coronary angiography[J]. Eur J Radiol, 2015,84(10): 1915-1920.
[22]
Nakamura Y, Higaki T, Tatsugami F, et al. Deep learning-based CT image reconstruction: initial evaluation targeting hypovascular hepatic metastases[J]. Radiol Artif Intell, 2019, 1(6): e180011.
[23]
Greffier J, Hamard A, Pereira F, et al. Image quality and dose reduction opportunity of deep learning image reconstruction algorithm for CT: a phantom study[J]. Eur Radiol, 2020, 30(7): 3951-3959.
[24]
Arndt C, Güttler F, Heinrich A, et al. Deep learning CT image reconstruction in clinical practice[J]. Rofo, 2021, 193(3): 252-261.
[25]
Sun J, Li H, Li J, et al. Improving the image quality of pediatric chest CT angiography with low radiation dose and contrast volume using deep learning image reconstruction[J]. Quant Imaging Med Surg,2021, 11(7): 3051-3058.
[26]
Wu D, Kim K, Li Q. Computationally efficient deep neural network for computed tomography image reconstruction[J]. Med Phys, 2019,46(11): 4763-4776.
[1] 李洋, 蔡金玉, 党晓智, 常婉英, 巨艳, 高毅, 宋宏萍. 基于深度学习的乳腺超声应变弹性图像生成模型的应用研究[J/OL]. 中华医学超声杂志(电子版), 2024, 21(06): 563-570.
[2] 罗刚, 泮思林, 孙玲玉, 李志新, 陈涛涛, 乔思波, 庞善臣. 一种新型语义网络分析模型对室间隔完整型肺动脉闭锁和危重肺动脉瓣狭窄胎儿右心发育不良程度的评价作用[J/OL]. 中华医学超声杂志(电子版), 2024, 21(04): 377-383.
[3] 赫兰, 杨泽堃, 张颖, 王玉东, 陈伟导, 王一同, 申锷. 双输入BCNN-ResNet模型对超声颈动脉斑块稳定性的分类诊断价值[J/OL]. 中华医学超声杂志(电子版), 2024, 21(02): 137-142.
[4] 成汉林, 史中青, 戚占如, 王小贤, 曾子炀, 单淳劼, 钱隼南, 罗守华, 姚静. 基于深度学习的超声心动图动态图像切面识别研究[J/OL]. 中华医学超声杂志(电子版), 2024, 21(02): 128-136.
[5] 孔德铭, 刘铮, 李睿, 钱文伟, 王飞, 蔡道章, 柴伟. 人工智能辅助全髋关节置换三维术前规划准确性评价[J/OL]. 中华关节外科杂志(电子版), 2024, 18(04): 431-438.
[6] 张嘉炜, 王瑞, 张克诚, 易磊, 周增丁. 烧烫伤创面深度智能检测模型P-YOLO的建立及测试效果[J/OL]. 中华损伤与修复杂志(电子版), 2024, 19(05): 379-385.
[7] 叶莉, 杜宇. 深度学习在牙髓根尖周病临床诊疗中的应用[J/OL]. 中华口腔医学研究杂志(电子版), 2024, 18(06): 351-356.
[8] 黄俊龙, 李文双, 李晓阳, 刘柏隆, 陈逸龙, 丘惠平, 周祥福. 基于盆底彩超的人工智能模型在女性压力性尿失禁分度诊断中的应用[J/OL]. 中华腔镜泌尿外科杂志(电子版), 2024, 18(06): 597-605.
[9] 杨龙雨禾, 王跃强, 招云亮, 金溪, 卫娜, 杨智明, 张贵福. 人工智能辅助临床决策在泌尿系肿瘤的应用进展[J/OL]. 中华腔镜泌尿外科杂志(电子版), 2024, 18(02): 178-182.
[10] 赵毅, 李昶田, 唐文博, 白雪婷, 刘荣. 腹腔镜术中超声主胰管自动识别模型的临床应用[J/OL]. 中华腔镜外科杂志(电子版), 2024, 17(05): 290-294.
[11] 尹泽新, 杨继林, 李有尧, 吴美龙, 刘利平. 肝癌微血管侵犯的术前预测研究进展[J/OL]. 中华肝脏外科手术学电子杂志, 2025, 14(01): 128-134.
[12] 陈健, 周静洁, 夏开建, 王甘红, 刘罗杰, 徐晓丹. 基于卷积神经网络实现结直肠息肉的实时检测与自动NICE分型(附视频)[J/OL]. 中华结直肠疾病电子杂志, 2024, 13(03): 217-228.
[13] 潘清, 葛慧青. 基于机械通气波形大数据的人机不同步自动监测方法[J/OL]. 中华重症医学电子杂志, 2024, 10(04): 399-403.
[14] 孙铭远, 褚恒, 徐海滨, 张哲. 人工智能应用于多发性肺结节诊断的研究进展[J/OL]. 中华临床医师杂志(电子版), 2024, 18(08): 785-790.
[15] 林一鑫, 董晶, 贾建文, 黄菊梅, 武军元, 王双坤, 柳云鹏, 汪阳. 基于人工智能分析颈内动脉颅外段迂曲特征及对称性的应用性评价[J/OL]. 中华脑血管病杂志(电子版), 2024, 18(03): 202-209.
阅读次数
全文


摘要


AI


AI小编
你好!我是《中华医学电子期刊资源库》AI小编,有什么可以帮您的吗?