| 孙其中,孟婷,彭俪颖,韩义成,王锡明.深度学习全模型迭代算法联合60 kVp扫描用于头颈部CT血管成像[J].中国医学影像技术,2025,41(4):520~524 |
| 深度学习全模型迭代算法联合60 kVp扫描用于头颈部CT血管成像 |
| Artificial intelligence iterative reconstruction combined with 60 kVp scanning for craniocervical CT angiography |
| 投稿时间：2024-10-20 修订日期：2025-02-05 |
| DOI：10.13929/j.issn.1003-3289.2025.04.003 |
| 中文关键词: 图像处理,计算机辅助 深度学习 CT血管成像 辐射剂量 |
| 英文关键词:image processing, computer-assisted deep learning computed tomography angiography radiation dosage |
| 基金项目:国家自然科学基金(82271993)。 |
|
| 摘要点击次数: 354 |
| 全文下载次数: 139 |
| 中文摘要: |
| 目的 观察深度学习全模型迭代算法(AIIR)联合60 kVp扫描用于头颈部CT血管成像(CTA)的价值。方法 前瞻性纳入86例疑诊头颈部血管疾病患者,随机对其中43例行常规剂量(120 kVp)、对其余43例行低剂量扫描(60 kVp),记录对比剂用量及辐射剂量。对常规剂量扫描原始图像进行混合迭代重建(HIR)得到A组图像,对低剂量扫描原始图像分别以HIR和AIIR得到B1组、B2组图像。比较3组图像总体质量,显示目标大、小血管及诊断信心的主观评分,以及其目标大血管噪声值(SD)、信噪比(SNR)及对比度噪声比(CNR)。结果 相比常规剂量扫描,低剂量扫描对比剂用量及有效剂量(ED)分别降低36.52%及77.56%。B2组与A组图像质量各项主观评分结果差异均无统计学意义(校正P均＞0.05)且均高于B1组(校正P均＜0.05);除主动脉弓外,B2组中其余各目标大血管SD均低于A组(校正P均＜0.05)。B2组各目标大血管SNR及CNR均高于A组及B1组(校正P均＜0.05),而各大血管SD均低于B1组(校正P均＜0.05)。结论 AIIR联合60 kVp扫描用于头颈CTA可在保证图像质量满足诊断需求的同时降低对比剂用量及辐射剂量。 |
| 英文摘要: |
| Objective To observe the value of artificial intelligence iterative reconstruction (AIIR) combined with 60 kVp scanning in craniocervical CT angiography (CTA). Methods Eighty-six patients with suspected craniocervical vascular diseases were prospectively enrolled and randomly received routine-dose (120 kVp) or low-dose (60 kVp) scanning (each n=43), and contrast dosage and radiation dose were recorded. After scanning, hybrid iterative reconstruction (HIR) was used to reconstruct routine-dose images (group A), while HIR images (group B1) and AIIR images (group B2) were performed to reconstruct low-dose images, respectively. The overall imaging quality, visualization of targeted large and small vessels and diagnostic confidence scores, as well as image noise (SD), signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR) of targeted large vessels were compared among 3 groups. Results Contrast dosage and effective dose (ED) decreased by 36.52% and 77.56% in low-dose scanning compared with those in routine-dose scanning, respectively. No significant difference of subjective scoring of imaging quality indexes was found between groups B2 and A (all adjusted P＞0.05), which were both higher than those in group B1 (all adjusted P＜0.05). SD of target large vessels in group B2 than that of most target large vessels in group A (all adjusted P＜0.05) except for aortic arch. SNR and CNR of target large vessels in group B2 were higher than those in group A and B1 (all adjusted P＜0.05),while SD of target large vessels in group B2 were lower than in group B1. Conclusion AIIR combined with 60 kVp scanning in craniocervical CTA could meet diagnostic requirements of imaging quality and reduce contrast dosage and radiation dose. |
| 查看全文 查看/发表评论 下载PDF阅读器 |
|
|
|
