Hadj Henni Ahmed, Martinage Geoffrey, Lebret Lucie, Arhoun Ilias
Radiation Oncology Department, Centre Frederic Joliot, 76000 Rouen, France.
Oncology Department, Clinique Saint Hilaire, 76000 Rouen, France.
Cancers (Basel). 2024 Dec 22;16(24):4267. doi: 10.3390/cancers16244267.
BACKGROUND/OBJECTIVES: An on-board imager on a linear accelerator allows the acquisition of kV-2D images during irradiation. Overlaying specific structures on these images enables the visual verification of movement at regular frequencies. Our aim was to validate this tracking method for the stereotactic treatment of bone metastases.
Shifts in three translational directions were simulated using an anthropomorphic phantom. For these simulated shifts, planar images were acquired at different angles of incidence, with overlaid volumes of interest. A blinded test was then administered to the 18 participants to evaluate their decisions regarding whether to stop treatment. The results considered the experience of the operators. Quantitative analyses were performed on the intra-fractional images of 29 patients.
Participants analyzed each image with an average (standard deviation) decision time of 3.0 s (2.3). For offsets of 0.0, 1.0, 1.5, and 2.0 mm, the results were 78%, 93%, 90%, and 100% for the expert group and 78%, 70%, 79%, and 88% for the less-experienced group. Clinical feedback confirmed this guidance technique and extended it to non-spinal bony metastases. Sudden movements exceeding the 2.0 mm threshold occurred in 3.3% of the analyzed fractions, with a detection rate of 97.8% for vertebral locations. For non-vertebral bone locations, movements exceeding a threshold of 3.0 mm occurred in 3.5% of cases and were detected in 96.5%.
The clinical use of planar OBI and superimposed structures for visual-image guidance in bone stereotactic treatment was validated using an anthropomorphic phantom and clinical feedback.
背景/目的：直线加速器上的机载成像仪可在照射过程中采集千伏二维图像。在这些图像上叠加特定结构能够以固定频率直观地验证运动情况。我们的目的是验证这种跟踪方法在骨转移瘤立体定向治疗中的有效性。
使用人体模型模拟三个平移方向的位移。对于这些模拟位移，在不同入射角下采集平面图像，并叠加感兴趣的体积。然后对18名参与者进行盲法测试，以评估他们关于是否停止治疗的决策。结果考虑了操作人员的经验。对29例患者的分次内图像进行了定量分析。
参与者分析每张图像的平均（标准差）决策时间为3.0秒（2.3）。对于0.0、1.0、1.5和2.0毫米的偏移量，专家组的结果分别为78%、93%、90%和100%，经验较少组的结果分别为78%、70%、79%和88%。临床反馈证实了这种引导技术，并将其扩展到非脊柱骨转移瘤。在3.3%的分析分次中出现了超过2.0毫米阈值的突然运动，椎体部位的检测率为97.8%。对于非椎体骨部位，3.5%的病例出现了超过3.0毫米阈值的运动，检测率为96.5%。
使用人体模型和临床反馈验证了平面机载成像（OBI）和叠加结构在骨立体定向治疗中用于视觉图像引导的临床应用。
