Department of Radiology Clinical Imaging Physics Group, Duke University Medical Center, Durham, North Carolina, USA.
Carl E. Ravin Advanced Imaging Laboratories, Duke University Medical Center, Durham, North Carolina, USA.
J Appl Clin Med Phys. 2023 Aug;24(8):e14069. doi: 10.1002/acm2.14069. Epub 2023 Jun 30.
Photon-counting computed tomography (PCCT) systems are increasingly available in the U.S. following Food and Drug Administration (FDA) approval of the first clinical PCCT system in Fall 2021. Consequently, there will be a need to incorporate PCCTs into existing fleets of traditional CT systems. The commissioning process of a PCCT was devised by evaluating the degree of agreement between the performance of the PCCT and that of established clinical CT systems. A PCCT system (Siemens NAEOTOM Alpha) was evaluated using the American College of Radiology(ACR) CT phantom (Gammex 464). The phantom was scanned on the system and on a 3rd Generation EID CT system (Siemens Force) at three clinical dose levels. Images were reconstructed across the range of available reconstruction kernels and Iterative Reconstruction (IR) strengths. Two image quality metrics-spatial resolution and noise texture-were calculated using AAPM TG233 software (imQuest), as well as a dose metric to achieve target image noise magnitude of 10 HU. For each pair of EID-PCCT kernel/IR strengths, the difference in metrics were calculated, weighted, and multiplied over all metrics to determine the concordance between systems. IR performance was characterized by comparing relative noise texture and reference dose as a function of IR strength for each system. In general, as kernel "sharpness" increased for each system, spatial resolution, noise spatial frequency, and reference dose increased. For a given kernel, EID reconstruction showed higher spatial resolution compared to PCCT in standard resolution mode. PCCT implementation of IR better preserved noise texture across all strengths compared to the EID, demonstrated by respective 20 and 7% shifts in noise texture from IR "Off" to IR "Max." Overall, the closest match for a given EID reconstruction kernel/IR strength was identified as a PCCT kernel with "sharpness" increased by 1 step and IR strength increased by 1-2 steps. Substantial dose reduction potential of up to 70% was found when targeting a constant noise magnitude.
光子计数计算机断层扫描（PCCT）系统在美国越来越普及，此前 2021 年秋季美国食品和药物管理局（FDA）批准了首个临床 PCCT 系统。因此，将需要将 PCCT 纳入传统 CT 系统的现有设备中。PCCT 的调试过程是通过评估 PCCT 的性能与已建立的临床 CT 系统的性能之间的一致性程度来设计的。使用美国放射学院（ACR）CT 体模（Gammex 464）对 PCCT 系统（西门子 NAEOTOM Alpha）进行了评估。该体模在该系统和第三代 EID CT 系统（西门子 Force）上以三个临床剂量水平进行了扫描。图像在可用重建核和迭代重建（IR）强度范围内进行了重建。使用 AAPM TG233 软件（imQuest）计算了两种图像质量指标-空间分辨率和噪声纹理，以及实现目标图像噪声幅度为 10 HU 的剂量指标。对于每个 EID-PCCT 核/IR 强度对，计算了指标之间的差异，并对所有指标进行加权和相乘，以确定系统之间的一致性。IR 性能通过比较每个系统的相对噪声纹理和参考剂量作为 IR 强度的函数来描述。一般来说，随着每个系统的核“锐度”增加，空间分辨率、噪声空间频率和参考剂量也随之增加。对于给定的核，EID 重建在标准分辨率模式下显示出比 PCCT 更高的空间分辨率。与 EID 相比，PCCT 实现的 IR 在所有强度下更好地保留了噪声纹理，分别从 IR“关闭”到 IR“最大”的噪声纹理变化分别为 20%和 7%。总体而言，对于给定的 EID 重建核/IR 强度，最接近的匹配是核“锐度”增加 1 步，IR 强度增加 1-2 步的 PCCT 核。当目标为恒定噪声幅度时，发现了高达 70%的大幅剂量减少潜力。
