Current conventional 4D Cone Beam Computed Tomography (4DCBCT) imaging is hampered by inconsistent patient breathing that leads to long scan times, reduced image quality and high imaging dose. To address these limitations, Respiratory Motion Guided 4D cone beam computed tomography (RMG-4DCBCT) uses mathematical optimization to adapt the gantry rotation speed and projection acquisition rate in real-time in response to changes in the patient's breathing rate. Here, RMG-4DCBCT is implemented on an Elekta Synergy linear accelerator to determine the minimum achievable imaging dose. 8 patient-measured breathing traces were programmed into a 1D motion stage supporting a 3D-printed anthropomorphic thorax phantom. The respiratory phase and current gantry position were calculated in real-time with the RMG-4DCBCT software, which in turn modulated the gantry rotation speed and suppressed projection acquisition. Specifically, the effect of acquiring 20, 25, 30, 35 and 40 projections/respiratory phase bin RMG scans on scan time and image quality was assessed. Reconstructed image quality was assessed via the contrast-to-noise ratio (CNR) and the Edge Response Width (ERW) metrics. The performance of the system in terms of gantry control accuracy was also assessed via an analysis of the angular separation between adjacent projections. The median CNR increased linearly from 5.90 (20 projections/bin) to 8.39 (40 projections/bin). The ERW did not significantly change from 1.08 mm (20 projections/bin) to 1.07 mm (40 projections/bin), indicating the sharpness is not dependent on the total number of projections acquired. Scan times increased with increasing total projections and slower breathing rates. Across all 40 RMG-4DCBCT scans performed, the average difference in the acquired and desired angular separation between projections was 0.64°. RMG-4DCBCT provides the opportunity to enable fast low-dose 4DCBCT (~70 s, 200 projections), without compromising on current clinical image quality.

中文翻译：

---

使用呼吸运动引导 4DCBCT 最大限度地减少 4DCBCT 成像剂量和扫描时间：临床前研究

当前传统的 4D 锥形束计算机断层扫描 (4DCBCT) 成像受到患者呼吸不一致的阻碍，这会导致扫描时间长、图像质量降低和成像剂量高。为了解决这些限制，呼吸运动引导 4D 锥形束计算机断层扫描 (RMG-4DCBCT) 使用数学优化来实时调整机架旋转速度和投影采集速率，以响应患者呼吸频率的变化。在这里，RMG-4DCBCT 在 Elekta Synergy 直线加速器上实施，以确定可实现的最小成像剂量。8 条患者测量的呼吸轨迹被编程到一个支持 3D 打印拟人胸腔模型的 1D 运动平台中。通过RMG-4DCBCT软件实时计算呼吸相位和当前机架位置，这反过来调制了机架旋转速度并抑制了投影采集。具体而言，评估了获取 20、25、30、35 和 40 个投影/呼吸相位 bin RMG 扫描对扫描时间和图像质量的影响。通过对比度噪声比 (CNR) 和边缘响应宽度 (ERW) 指标评估重建的图像质量。系统在龙门控制精度方面的性能也通过分析相邻投影之间的角间距来评估。CNR 中值从 5.90（20 个投影/bin）线性增加到 8.39（40 个投影/bin）。ERW 从 1.08 mm（20 个投影/bin）到 1.07 mm（40 个投影/bin）没有显着变化，表明清晰度不依赖于获取的投影总数。扫描时间随着总投影的增加和呼吸频率的降低而增加。在执行的所有 40 次 RMG-4DCBCT 扫描中，获得的和所需的投影之间的角间距的平均差异为 0.64°。RMG-4DCBCT 提供了实现快速低剂量 4DCBCT（约 70 秒，200 次投影）的机会，而不会影响当前的临床图像质量。