Elucidation of the Efficacy of Low Tube Voltage Combined with AIIR in Suppressing Hip Prosthesis CT Artifacts
-
摘要:
目的：探讨低管电压结合人工智能迭代重建（AIIR）在髋关节术后盆腔CT检查中抑制金属伪影的效果。方法：回顾性收集46例在我院接受盆腔CT扫描的髋关节置换患者数据，分为筛查组（80 kVp）和临床组（120 kVp），分别进行常规重建、1~5级AIIR的重建；对图像“蜡像感”和整体质量进行主观评分；勾画ROI并计算ΔCT肌肉、ΔCT脂肪、SD伪影、SD背景、MAI和CNR；对主观和客观指标先进行组内两两对比，再将筛查组最优重建方法与临床组所有重建方法进行对比，并比较其容积CT剂量指数。结果：随着AIIR等级提升，图像整体质量评分提升（均优于常规重建），“蜡像感”降低（AIIR 5级与常规重建差异无统计学意义）。随着AIIR等级提升，筛查组和临床组图像噪声呈上升趋势，CNR呈下降趋势，筛查组MAI呈上升趋势，临床组MAI差异无统计学意义，但均优于常规重建。AIIR等级间ΔCT肌肉、ΔCT脂肪差异无统计学意义，但AIIR使筛查组的该差值明显降低。筛查组容积CT剂量指数明显低于临床组。筛查组最优的AIIR 5级与临床组各重建相比，ΔCT肌肉、ΔCT脂肪差异无统计学意义，其他客观指标均有差异；主观评分与临床组AIIR 5级差异无统计学意义，且都优于其他重建。结论：在髋关节置换术后患者盆腔CT中，5级是最优的AIIR等级，能显著抑制噪声和金属伪影并提升CNR；80 kVp结合5级AIIR，能在辐射剂量低于120 kVp的同时，使图像主观质量达到120 kVp水平。
Abstract:Objective: To investigate the efficacy of low tube voltage combined with artificial intelligence iterative reconstruction (AIIR) in suppressing metal artifacts during pelvic CT scans after hip arthroplasty. Methods: Data from 46 patients with hip replacements who underwent pelvic CT scans at our hospital were retrospectively collected and divided into a screening group (80 kVp) and a clinical group (120 kVp). Conventional reconstruction and AIIR reconstructions at levels 1~5 were performed. Subjective scoring of “wax-like” appearance and overall image quality was conducted. Regions of interest (ROI) were drawn to calculate ΔCT of muscle, ΔCT of fat, SD of artifacts, SD of background, MAI, and CNR. Pairwise comparisons were made within each group, and the optimal reconstruction method within the screening group was compared with all methods applied in the clinical group. Volume CT dose indices were also compared. Results: As the AIIR level increased, the overall image quality score improved (better than conventional reconstruction), and the “wax-like” appearance decreased (no significant difference between AIIR level 5 and conventional reconstruction). With increasing AIIR level, image noise displayed an upward trend in both the screening and clinical groups, while CNR decreased. MAI increased in the screening group, but no significant difference was observed in the clinical group. However, both groups performed better than that obtained from conventional reconstruction. No significant differences were found in ΔCT of muscle and ΔCT of fat between AIIR levels, but AIIR significantly reduced these values in the screening group. The volume CT dose index of the screening group is significantly lower than that of the clinical group. There was no significant difference in ΔCT of muscle or fat between optimal AIIR level 5 in the screening group and all reconstructions in the clinical group, but other objective indicators showed differences. Subjective scores showed no significant difference between the clinical group and AIIR level 5, and all were better than other reconstructions. Conclusion: In pelvic CT scans of patients after hip arthroplasty, AIIR level 5 is the optimal level, and can significantly suppress noise and metal artifacts and improve CNR. Combining 80 kVp with AIIR level 5 reduces the radiation dose below that required with 120 kVp, while maintaining comparable subjective image quality.
-
Keywords:
- CT /
- low dose /
- metal artifacts /
- iterative reconstruction
-
-
![]()
图 1 ROI勾画示意图
注：（a）和（b）分别为筛查组盆腔CT常规重建图像的伪影最重层面、不受伪影影响层面的ROI勾画示意图。
Figure 1. Schematic diagram of ROI delineation
![]()
图 2 “蜡像感”、整体质量评分分布
注：（a）和（b）分别为筛查组、临床组的“蜡像感”评分分布，（c）和（d）分别为筛查组、临床组的图像整体质量评分分布。随着AIIR等级提高，图像“蜡像感”减弱，整体质量评分提高。
Figure 2. Distributions of ‘Wax-like’ appearance and overall quality scores
![]()
图 3 常规重建与AIIR 5级重建图像对比
注：（a）和（b）分别为61岁女患者的80 kVp盆腔CT常规重建、AIIR 5级重建；（c）和（d）分别为66岁女患者的120 kVp盆腔CT常规重建、AIIR 5级重建。AIIR 5级能显著降低图像噪声，减少金属伪影，对低管电压者效果尤为显著。
Figure 3. Comparison of conventional reconstruction with AIIR level 5 reconstruction images
表 1 医师对不同重建的主观评分Kappa值
Table 1 Kappa values of physicians’ subjective scoring of different reconstructions
项目 筛查组 临床组 常规 AIIR 1 AIIR 2 AIIR 3 AIIR 4 AIIR 5 常规 AIIR 1 AIIR 2 AIIR 3 AIIR 4 AIIR 5 “蜡像感”评分Kappa 0.850 0.905 0.722 0.834 0.887 0.768 0.800 0.750 0.818 0.800 0.833 0.846 整体质量评分Kappa 0.866 0.820 0.791 0.751 0.743 0.803 0.800 0.833 0.765 0.833 0.800 0.750 
下载: 导出CSV
表 2 不同重建的客观指标
Table 2 Objective indicators of different reconstruction types
项目 筛查组 临床组 常规 AIIR 1 AIIR 2 AIIR 3 AIIR 4 AIIR 5 常规 AIIR 1 AIIR 2 AIIR 3 AIIR 4 AIIR 5 ΔCT肌肉 51.35
（27.35，77.33）16.94
（7.94，33.74）17.87
（9.05，35.97）17.99
（8.34，35.92）17.98
（8.32，35.89）18.06
（8.24，35.83）21.24
（3.89，59.93）9.76
（5.85，23.53）9.35
（4.95，23.46）9.96
（5.92，23.45）9.98
（5.90，23.41）10.03
（6.02，23.40）ΔCT脂肪 26.52
（10.31，47.64）9.79
（7.74，19.03）9.84
（7.58，18.93）9.89
（7.47，18.90）10.02
（7.40，18.41）10.06
（7.36，18.93）25.45±
22.7514.42±
11.6314.42±
11.6014.39±
11.5514.35±
11.5014.37±
11.52SD伪影 54.89
（45.76，61.55）22.52
（16.28，31.51）23.07
（17.01，32.08）23.49
（17.59，32.15）24.52
（18.34，32.79）25.23
（19.13，33.73）25.24
（21.92，25.91）10.22
（7.64，12.90）10.91
（8.55，13.50）11.95
（9.37，14.06）12.78
（10.06，14.60）13.40
（10.74，15.12）SD背景 21.98
（19.11，24.94）8.41
（5.38，9.94）9.92
（6.72，11.03）11.02
（8.10，12.22）12.31
（9.48，13.41）13.09
（10.51，14.24）11.90±
1.454.82±
1.355.86±
1.236.86±
1.177.82±
1.148.66±
1.18MAI 48.96
（39.99，56.84）20.45
（13.91，30.47）20.78
（14.28，30.73）20.41
（13.57，30.47）21.41
（14.38，30.78）21.69
（14.35，31.27）21.55
（18.44，22.39）8.66
（5.54，12.00）9.00
（5.76，12.17）9.30
（6.00，12.28）9.57
（6.23，12.24）9.82
（6.40，12.20）CNR 7.54
（5.61，9.65）16.47
（11.49，29.03）14.28
（10.47，20.87）12.82
（9.40，17.95）11.44
（8.67，15.65）10.38
（8.00，12.98）10.51±
4.0029.90±
10.3524.38±
6.9021.32±
5.7018.83±
4.6216.88±
3.74注：符合正态分布者以（均数±标准差）表示；不符合正态分布者以（中位数（上下四分位数））表示。
下载: 导出CSV
表 3 筛查组AIIR 5与临床组各重建类型的对比
Table 3 Comparisons of AIIR 5 in the screening group and different reconstruction types in the clinical group
项目 常规 AIIR 1 AIIR 2 AIIR 3 AIIR 4 AIIR 5 “蜡像感”评分 −1.328（0.184） −5.417（ < 0.001） −5.410（ < 0.001） −5.411（ < 0.001） −4.323（ < 0.001） −0.146（0.884） 整体质量评分 −5.485（ < 0.001） −5.328（ < 0.001） −5.357（ < 0.001） −4.477（ < 0.001） −4.207（ < 0.001） −0.647（0.518） ΔCT肌肉 −0.300（0.764） −1.351（0.177） −1.601（0.109） −1.401（0.161） −1.426（0.154） −1.426（0.154） ΔCT脂肪 −1.176（0.240） −0.163（0.871） −0.150（0.881） −0.150（0.881） −0.150（0.881） −0.125（0.900） SD伪影 −0.150（0.881） −4.678（ < 0.001） −4.578（ < 0.001） −4.503（ < 0.001） −4.453（ < 0.001） −4.378（ < 0.001） SD背景 −1.026（0.305） −5.103（ < 0.001） −5.103（ < 0.001） −5.103（ < 0.001） −4.953（ < 0.001） −4.528（ < 0.001） MAI −0.350（0.726） −4.903（ < 0.001） −4.678（ < 0.001） −4.328（ < 0.001） −4.028（ < 0.001） −3.727（ < 0.001） CNR −0.225（0.822） −3.752（ < 0.001） −3.652（ < 0.001） −3.527（ < 0.001） −3.477（0.001） −3.352（ < 0.001） 注：不符合正态分布者以独立样本Mann-Whitney U检验进行比较；检验结果以（Z值（P值））表示。
下载: 导出CSV
-
[1] FERGUSON R J, PALMER A J R, TAYLOR A, et al. Hip replacement[J]. The Lancet, 2018, 392(10158): 1662-1671. DOI: 10.1016/S0140-6736(18)31777-X.
[2] GüNTHER K P, DECKERT S, LüTZNER C, et al. Total hip replacement for osteoarthritis—evidence-based and patient-oriented indications[J]. Deutsches Ä Rzteblatt International, 2021, 118(43): 730.
[3] SELLES M, Van OSCH J A C, MAAS M, et al. Advances in metal artifact reduction in CT images: A review of traditional and novel metal artifact reduction techniques[J]. European Journal of Radiology, 2024, 170: 111276. DOI: 10.1016/j.ejrad.2023.111276.
[4] HOKAMP N G, ECK B, SIEDEK F, et al. Quantification of metal artifacts in computed tomography: Methodological considerations[J]. Quantitative Imaging in Medicine and Surgery, 2020, 10(5): 1033. DOI: 10.21037/qims.2020.04.03.
[5] PUVANASUNTHARARAJAH S, FONTANAROSA D, WILLE M L, et al. The application of metal artifact reduction methods on computed tomography scans for radiotherapy applications: A literature review[J]. Journal of Applied Clinical Medical Physics, 2021, 22(6): 198-223. DOI: 10.1002/acm2.13255.
[6] 张宝平, 李傲, 李宇航, 等. 人工智能迭代重建算法在腹部低剂量增强CT中的临床应用价值[J]. 遵义医科大学学报, 2024, 47(5): 487-494, 501. DOI: 10.14169/j.cnki.zunyixuebao.2024.0066. ZHANG B P, LI A, LI Y H, et al. Clinical application of artificial intelligence iterative reconstruction algorithm in low-dose abdominal contrast-enhanced CT[J]. Journal of Zunyi Medical University, 2024, 47(5): 487-494, 501. DOI: 10.14169/j.cnki.zunyixuebao.2024.0066. (in Chinese).
[7] 曾栋, 张晓瑜, 马思源, 等. 低剂量医学CT成像技术: 过去, 现在与未来[J]. 中国体视学与图像分析, 2022, 27(4): 355-378. DOI: 10.13505/j.1007-1482.2022.27.04.005. ZENG D, ZHANG X Y, MA S Y, et al. Low-dose medical CT imaging: Past, present and future[J]. Chinese Journal of Stereology and Image Analysis, 2022, 27(4): 355-378. DOI: 10.13505/j.1007-1482.2022.27.04.005. (in Chinese).
[8] MOHAMMADINEJAD P, KHANDELWAL A, INOUE A, et al. Utility of an automatic adaptive iterative metal artifact reduction AiMAR algorithm in improving CT imaging of patients with hip prostheses evaluated for suspected bladder malignancy[J]. Abdominal Radiology, 2022, 47(6): 2158-2167. DOI: 10.1007/s00261-022-03475-8.
[9] 游永春, 李万江, 刘洪川, 等. “三低”技术联合人工智能迭代重建算法在主动脉CT血管成像中的临床应用[J]. 四川大学学报(医学版), 2022, 53(4): 676-681. YOU Y C, LI W J, LIU H C, et al. Clinical application of “three-low” technique combined with artificial intelligence iterative reconstruction algorithm in aortic CT angiography[J]. Journal of Sichuan University (Medical Sciences), 2022, 53(4): 676-681. (in Chinese).
[10] SCHRECK J, LAUKAMP K R, NIEHOFF J H, et al. Metal artifact reduction in patients with total hip replacements: Evaluation of clinical photon counting CT using virtual monoenergetic images[J]. European Radiology, 2023, 33(12): 9286-9295. DOI: 10.1007/s00330-023-09879-4.
[11] 李杰, 袁源, 王春杰, 等. 能谱CT去金属伪影(MAR)技术用于减低单髋关节置换物伪影[J]. 中国医学影像技术, 2021, 37(1): 131-135. DOI: 10.13929/j.issn.1003-3289.2021.01.032. LI J, YUAN Y, WANG C J, et al. Energy spectrum CT metal artifacts reduction (MAR) for reducing artifacts of unilateral hip arthroplasty[J]. Chinese Journal of Medical Imaging Technology, 2021, 37(1): 131-135. DOI: 10.13929/j.issn.1003-3289.2021.01.032. (in Chinese).
[12] YANG F, LIANG Y W, SHAO Q, et al. Application value of CT metal artifact correction technology (MAC TM) in CT review after total hip replacement[J]. Journal of Sichuan University (Medical Sciences), 2020, 51(6): 828-833.
[13] RAU A, STRAEHLE J, STEIN T, et al. Photon-counting computed tomography (PC-CT) of the spine: Impact on diagnostic confidence and radiation dose[J]. European Radiology, 2023, 33(8): 5578-5586. DOI: 10.1007/s00330-023-09511-5.
[14] WELLENBERG R H H, HAKVOORT E T, SLUMP C H, et al. Metal artifact reduction techniques in musculoskeletal CT-imaging[J]. European journal of radiology, 2018, 107: 60-69. DOI: 10.1016/j.ejrad.2018.08.010.
[15] KHODARAHMI I, ISAAC A, FISHMAN E K, et al. Metal about the hip and artifact reduction techniques: From basic concepts to advanced imaging[C]//Seminars in Musculoskeletal Radiology. Thieme Medical Publishers, 2019, 23(3): e68-e81.
[16] LAUKAMP K R, LENNARTZ S, NEUHAUS V F, et al. CT metal artifacts in patients with total hip replacements: For artifact reduction monoenergetic reconstructions and post-processing algorithms are both efficient but not similar[J]. European Radiology, 2018, 28: 4524-4533. DOI: 10.1007/s00330-018-5414-2.
[17] KOHYAMA S, YOSHII Y, OKAMOTO Y, et al. Advances in bone joint imaging-metal artifact reduction[J]. Diagnostics, 2022, 12(12): 3079. DOI: 10.3390/diagnostics12123079.
[18] NJITI M M, OSMAN N D, MANSOR M S, et al. Potential of metal artifact reduction (MAR) and deep learning-based reconstruction (DLR) integration in CT metal artifact correction: A review[J]. Radiation Physics and Chemistry, 2024: 111541.
[19] 杨名昊, 韩义成, 刘洪武, 等. 基于深度学习迭代重建算法的肺结节评估效果分析[J]. 医学影像学杂志, 2024, 34(5): 75-77, 86. YANG M H, HAN Y C, LIU H W, et al. Performance analysis of deep learning-based iterative reconstruction algorithms in the evaluation of lung nodules[J]. Journal of Medical Imaging, 2024, 34(5): 75-77, 86. (in Chinese).
[20] 曹建新, 朱卓, 张清, 等. 联合人工智能模型迭代重建和70kV超低管电压大幅度降低肺动脉CTA辐射剂量的研究[J]. 放射学实践, 2024, 39(10): 1361-1366. DOI: 10.13609/j.cnki.1000-0313.2024.10.013. CAO J X, ZHU Z, ZHANG Q, et al. Study of dramatic reducing radiation dose of CT pulmonary angiography using artificial intelligence model-based iterative reconstruction with 70kV ultra-low tube voltage setting[J]. Radiologic Practice, 2024, 39(10): 1361-1366. DOI: 10.13609/j.cnki.1000-0313.2024.10.013. (in Chinese).
计量
- 文章访问数: 43
- HTML全文浏览量: 5
- PDF下载量: 16
