CT Research and Collaboration, Siemens Healthcare, Tokyo, Japan.
Siemens Healthcare GmbH, Forchheim, Germany.
Invest Radiol. 2018 Nov;53(11):673-680. doi: 10.1097/RLI.0000000000000491.
The aim of this study was to assess the ability of third-generation dual-source dual-energy computed tomography to quantify cisplatin concentration using a 3-material decomposition algorithm in an experimental phantom.
Fifteen agarose-based phantoms containing various concentrations of iodine (0, 1.0, 2.0 mg I/mL) and cisplatin (0, 0.5, 1.0, 2.0, 3.0 mg Pt/mL) were scanned using third-generation dual-source dual-energy computed tomography at 80 kV and 150 kV with tin prefiltration. A cisplatin map was generated using the cisplatin-specific 3-material decomposition algorithm to differentiate cisplatin from iodine and agar. The computed tomography (CT) values at 80 kV, 150 kV, mixed 120 kV, and the cisplatin map were measured. Interobserver variabilities for the CT measurements on the cisplatin map were assessed using interclass correlation coefficients. Correlation between the CT values and titrated cisplatin concentrations was correlated using Spearman rank correlation analysis. To assess the influence of iodine, linear regression lines for the CT values on the cisplatin map and titrated cisplatin concentrations were compared using an analysis of covariance.
Interobserver agreement revealed almost perfect agreements (interclass correlation coefficients = 0.941-0.995). Significant and excellent positive correlations were observed between the CT values on the cisplatin map and titrated cisplatin concentrations (ρ = 0.980, P < 0.001 for all). The cisplatin map could identify the lowest cisplatin concentration of 0.5 mg Pt/mL in the presence of iodine. The iodine concentration had no significant effect on the CT measurements on the cisplatin map (P = 0.297, adjusted R = 0.993).
The cisplatin map generated from the 3-material decomposition algorithm allows quantification of a cisplatin concentration in an experimental phantom, independent of co-present iodine.
本研究旨在评估第三代双源双能计算机断层扫描（CT）使用三物质分解算法通过实验性体模定量顺铂浓度的能力。
15 个基于琼脂糖的体模，含有不同浓度的碘（0、1.0、2.0mgI/mL）和顺铂（0、0.5、1.0、2.0、3.0mgPt/mL），使用第三代双源双能 CT 在 80kV 和 150kV 下进行扫描，采用锡预滤。使用顺铂特异性三物质分解算法生成顺铂图谱，以将顺铂与碘和琼脂糖区分开来。测量 80kV、150kV、混合 120kV 和顺铂图谱的 CT 值。使用组内相关系数评估顺铂图谱 CT 值的观察者间变异性。使用 Spearman 秩相关分析评估 CT 值与滴定顺铂浓度之间的相关性。为了评估碘的影响，使用协方差分析比较顺铂图谱和滴定顺铂浓度的 CT 值的线性回归线。
观察者间一致性显示出几乎完美的一致性（组内相关系数=0.941-0.995）。顺铂图谱上的 CT 值与滴定顺铂浓度之间存在显著的、极好的正相关关系（ρ=0.980，P<0.001）。在存在碘的情况下，顺铂图谱可以识别出最低浓度为 0.5mgPt/mL 的顺铂。碘浓度对顺铂图谱上的 CT 值没有显著影响（P=0.297，调整后的 R=0.993）。
三物质分解算法生成的顺铂图谱允许在实验性体模中定量顺铂浓度，与共存的碘无关。
