膝关节骨髓水肿的识别:双能CT与MRI诊断准确性的比较
这项前瞻性研究纳入33例连续患者,在6天内行双能CT(80和150 kV)与MRI检查,以MRI为参考标准评估双能CT诊断膝关节骨髓水肿的准确性。MRI在25/33例(75.7%)中显示骨髓水肿;定性分析敏感度92.9%、特异度92.9%(读者1)和88.2%、93.9%(读者2),以-15 HU为阈值定量分析敏感度84.7%、特异度93.6%。结论认为双能CT能准确识别膝关节骨髓水肿。
要点
- 双能CT以-15 HU为阈值识别膝关节骨髓水肿,敏感度84.7%、特异度93.6%。
- 定性分区分析敏感度88.2%–92.9%、特异度92.9%–93.9%,观察者间一致性k=0.793。
- MRI显示骨髓水肿者密度均值3.6±25.3 HU,无骨髓水肿者-72.2±45.1 HU(p<0.001)。
核心结论
- 双能CT能准确识别膝关节骨髓水肿。
研究设计与样本
- 前瞻性IRB批准研究,纳入33例连续患者(男20例、女13例,平均年龄52.2岁),在6天内行双能CT(80和150 kV)与MRI检查。
- 以MRI为参考标准,用受试者工作特征曲线(ROC)和曲线下面积(AUC)计算准确性,用k统计量计算观察者间和观察者内一致性,p<0.05为有统计学意义。
- 由两名有经验的放射科医生对双能CT图像进行定性和定量评估。
关键影像表现/征象
- MRI在25/33例(75.7%)中显示骨髓水肿。
- 定量分析中,阳性病例密度均值为3.6±25.3 HU,阴性病例为-72.2±45.1 HU,两者差异有统计学意义(p<0.001)。
诊断标准与关键数值
- 定性分区分析:读者1敏感度92.9%、特异度92.9%、阳性预测值78.2%、阴性预测值97.9%、准确性92.9%;读者2敏感度88.2%、特异度93.9%、阳性预测值79.8%、阴性预测值96.6%、准确性92.6%。
- 观察者间一致性为中等(k=0.793),观察者内一致性接近完美(k=0.844)。
- 以-15 HU为阈值识别骨髓水肿,双能CT敏感度84.7%、特异度93.6%、阳性预测值78.2%、阴性预测值95.7%、准确性91.6%。
临床应用与阅片要点
- 原文未据此给出推荐。
局限性
- 原文未报告局限性。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
英文原文(摘要)· 点击展开
BACKGROUND: To assess the diagnostic accuracy of dual-energy computed tomography (DECT) in diagnosing bone marrow edema (BME) of the knee in traumatic and non-traumatic patients.
METHODS: This prospective IRB approved study included 33 consecutive patients (20 males, 13 females; mean age of 52.2 years) evaluated with DECT (80 and 150 kV) and MRI within 6 days. Two experienced radiologists qualitatively and quantitatively evaluated DECT images. The accuracy values were calculated by using receiver operator curves (ROC) and area under the curve (AUC), using MRI as the reference standard. Inter-observer and intra-observer agreement were calculated with k-statistics. A p < 0.05 was considered statistically significant.
RESULTS: MRI depicted BME in 25/33 patients (75.7%). The sensitivity, specificity, PPV, NPV, and accuracy of per-partition qualitative analysis were 92.9, 92.9, 78.2, 97.9, and 92.9%, for reader 1, and 88.2, 93.9, 79.8, 96.6, and 92.6%, for reader 2, respectively. The inter-observer agreement was substantial (k = 0.793) and the intra-observer agreement was near-perfect (k = 0.844). At the quantitative analysis, a significant difference (p < 0.001) was depicted between the density values of positive (mean 3.6 ± 25.3 HU) and negative cases (mean - 72.2 ± 45.1 HU). By using - 15 HU cutoff to identify BME, sensitivity, specificity, PPV, NPV, and accuracy of DECT were 84.7, 93.6, 78.2, 95.7, and 91.6%, respectively.
CONCLUSION: DECT can accurately identify BME of the knee.
引用
- Foti G, Mantovani W, Faccioli N, et al. Identification of bone marrow edema of the knee: diagnostic accuracy of dual-energy CT in comparison with MRI. Radiol Med. 2020. PMID: 32840730. doi:10.1007/s11547-020-01267-y
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。