Home > Publications database > Feasibility of markerless fluoroscopic real-time tumor detection for adaptive radiotherapy: development and end-to-end testing. |
Journal Article | DKFZ-2020-00738 |
; ; ; ; ;
2020
IOP Publ.
Bristol
This record in other databases:
Please use a persistent id in citations: doi:10.1088/1361-6560/ab8578
Abstract: Respiratory-gated radiotherapy treatments of lung tumors reduce the irradiated normal tissue volume and potentially lower the risk of side effects. However, in clinical routine, the gating signal is usually derived from external markers or other surrogate signals and may not always correlate well with the actual tumor position. This study uses the kV-imaging system of a LINAC in combination with a multiple template matching algorithm for markerless real-time detection of the tumor position in a dynamic anthropomorphic lung phantom. The tumor was realized by a small container filled with polymer dosimetry gel, the so-called gel tumor. A full end-to-end test for a gated treatment was performed and the geometric and dosimetric accuracy was validated. The accuracy of the tumor detection algorithm in SI- direction was found to be (2.3±1.6) mm and the gel tumor was automatically detected in 98 out of 100 images. The measured 3D dose distribution showed a uniform coverage of the gel tumor and comparison with the treatment plan revealed a high 3D γ-passing rate of 86.7 % (3%/3mm). The simulated treatment confirmed the employed margin sizes for residual motion within the gating window and serves as an end-to-end test for a gated treatment based on a markerless fluoroscopic real-time tumor detection.
![]() |
The record appears in these collections: |