Radiology Unit, National Institute for Infectious Diseases "L. Spallanzani" IRCCS, Via Portuense 292, 00149, Rome, Italy.
Clinical Epidemiology Unit, National Institute for Infectious Diseases "L. Spallanzani" IRCCS, Via Portuense 292, 00149, Rome, Italy.
Radiol Med. 2020 May;125(5):451-460. doi: 10.1007/s11547-020-01139-5. Epub 2020 Feb 11.
To evaluate the effect of dose reduction with iterative reconstruction (IR) on image quality of chest CT scan comparing two protocols.
Fifty-nine patients were enrolled. The two CT protocols were applied using Iterative Reconstruction (ASIR™) 40% but different noise indexes, recording dose-length product (DLP) and volume computed tomography dose index (CTDI). The subjective IQ was rated based on the distinction of anatomic details using a 4-point Likert scale based on the European Guidelines on Quality Criteria for CT. For each patient, two single CTs, at enrollment (group 1) and at follow-up after lowering the dose (group 2), were evaluated by two radiologists evaluating, for each examination, five different lung regions (central zone-CZ; peripheral zone-PZ; sub-pleural region-SPR; centrilobular region-CLR; and apical zone-AZ). An inter-observer agreement was expressed by weighted Cohen's kappa statistics (k) and intra-individual differences of subjective image analysis through visual grading characteristic (VGC) analysis.
An average 50.4% reduction in CTDI and 51.5% reduction in DLP delivered were observed using the dose-reduced protocol. An agreement between observers evaluating group 1 CTs was perfect (100%) and moderate to good in group 2 examinations (k-Cohen ranging from 0.56 for PZ and AZ to 0.70 for SPR). In the VGC analysis, image quality ratings were significantly better for group 1 than group 2 scans for all regions (AUC ranging from 0.56 for CZ to 0.62). However, disagreement was limited to a score 4 (excellent)-to-score 3 (good) IQ transition; apart from a single case in PZ, both the observers scored the IQ at follow-up as 2 (sufficient) starting from a score 4 (excellent).
Dose reduction achieved in the follow-up CT scans, although a lower IQ still allows a good diagnostic confidence.
评估迭代重建（IR）降低剂量对两种方案的胸部 CT 扫描图像质量的影响。
共纳入 59 例患者。使用迭代重建（ASIR™）40%，但不同噪声指数，记录剂量长度乘积（DLP）和容积 CT 剂量指数（CTDI），应用两种 CT 方案。使用欧洲 CT 质量标准指南中的 4 分李克特量表基于解剖细节的区分对主观 IQ 进行评分。对于每位患者，在入组时（第 1 组）和降低剂量后（第 2 组）进行两次单次 CT 检查，两名放射科医生对每组检查的五个不同肺区（中央区-CZ；周围区-PZ；胸膜下区-SPR；小叶中心区-CLR；和尖区-AZ）进行评估。观察者间的一致性通过加权 Cohen 的 kappa 统计（k）和视觉分级特征（VGC）分析评估主观图像分析的个体内差异来表示。
使用降低剂量的方案，平均 CTDI 降低 50.4%，DLP 降低 51.5%。评估第 1 组 CT 的观察者之间的一致性为完美（100%），而评估第 2 组检查的观察者之间的一致性为中度至良好（k-Cohen 范围从 PZ 和 AZ 的 0.56 到 SPR 的 0.70）。在 VGC 分析中，与第 2 组相比，第 1 组的所有区域的图像质量评分均显著更好（CZ 的 AUC 范围从 0.56 到 0.62）。然而，分歧仅限于从评分 4（优秀）到评分 3（良好）的 IQ 转换；除了 PZ 中的一个单独病例外，两位观察者均从评分 4（优秀）开始将随访时的 IQ 评分记为 2（足够）。
尽管降低 IQ，但在随访 CT 扫描中仍能实现剂量降低，仍能获得良好的诊断信心。
