We present the application of deep multi-class classifiers for registration of the pre-radiation image (CBCT) to the treatment planning image (planCT) in Radiation Therapy (RT). We train a multi-class classifier on different classes of displacement between 3D patches of images and use it for registration. As the initial displacement between images might be large, we train multiple classifiers for different resolutions of the data to capture larger displacements in coarser resolutions. We show that having only a few patients, the deep multi-class classifiers enable an accurate and fast rigid registration for CBCT to planCT even with significantly different fields of view. Our work lays the foundation for deformable image registration and prediction of registration uncertainty which can be utilized for adaptive RT.
In contemporary high-dose-rate brachytherapy treatment of superficial tumors, catheters are placed in a wax mould. The creation of current wax models is a difficult and time consuming proces.The irradiation plan can only be computed post-construction and requires a second CT scan. In case no satisfactory dose plan can be created, the mould is discarded and the process is repeated. The objective of this work was to develop an automated method to replace suboptimal wax moulding. We developed a method to design and manufacture moulds that guarantee to yield satisfactory dosimetry. A 3D-printed mould with channels for the catheters designed from the patient’s CT and mounted on a patient-specific thermoplastic mesh mask. The mould planner was implemented as an open-source module in the 3D Slicer platform. Series of test moulds were created to accommodate standard brachytherapy catheters of 1.70mm diameter. A calibration object was used to conclude that tunnels with a diameter of 2.25mm, minimum 12mm radius of curvature, and 1.0mm open channel gave the best fit for this printer/catheter combination. Moulds were created from the CT scan of thermoplastic mesh masks of actual patients. The patient-specific moulds have been visually verified to fit on the thermoplastic meshes. The masks were visually shown to fit onto the thermoplastic meshes, next the resulting dosimetry will have to be compared with treatment plans and dosimetry achieved with conventional wax moulds in order to validate our 3D printed moulds.
Image guidance capability is an important feature of modern radiotherapy linacs, and future cobalt-60 units will be
expected to have similar capabilities. Imaging with the treatment beam is an appealing option, for reasons of simplicity
and cost, but the dose needed to produce cone beam CT (CBCT) images in a Co-60 treatment beam is too high for this
modality to be clinically useful. Digital tomosynthesis (DT) offers a quasi-3D image, of sufficient quality to identify
bony anatomy or fiducial markers, while delivering a much lower dose than CBCT.
A series of experiments were conducted on a prototype Co-60 cone beam imaging system to quantify the resolution,
selectivity, geometric accuracy and contrast sensitivity of Co-60 DT. Although the resolution is severely limited by the
penumbra cast by the ~2 cm diameter source, it is possible to identify high contrast objects on the order of 1 mm in
width, and bony anatomy in anthropomorphic phantoms is clearly recognizable. Low contrast sensitivity down to
electron density differences of 3% is obtained, for uniform features of similar thickness. The conventional shift-and-add
reconstruction algorithm was compared to several variants of the Feldkamp-Davis-Kress filtered backprojection
algorithm result. The Co-60 DT images were obtained with a total dose of 5 to 15 cGy each.
We conclude that Co-60 radiotherapy units upgraded for modern conformal therapy could also incorporate imaging
using filtered backprojection DT in the treatment beam. DT is a versatile and promising modality that would be well
suited to image guidance requirements.
There has been considerable interest in megavoltage CT (MVCT) imaging associated with the development of image guided radiation therapy. It is clear that MVCT can provide good image quality for patient setup verification with soft tissue contrast much better than noted in conventional megavoltage portal imaging. In addition, it has been observed that MVCT images exhibit considerably reduced artifacts surrounding metal implants (e.g., surgical clips, hip implants, dental fillings) compared to conventional diagnostic CT images (kVCT). When encountered, these artifacts greatly limit the usefulness of kVCT images, and a variety of solutions have been proposed to remove the artifacts, but these have met with only partial success. In this paper, we investigate the potential for CT imaging in regions surrounding metal implants using high-energy photons from a Cobalt-60 source and from a 4 MV linear accelerator. MVCT and kVCT images of contrast phantoms and a phantom containing a hip prosthesis are compared and analysed. We show that MVCT scans provide good fidelity for CT number quantification in the high-density regions of the images, and in the regions immediately adjacent to the metal implants. They also provide structural details within the high-density inserts and implants. Calculations will show that practical clinical MVCT imaging, able to detect 3% contrast objects, should be achievable with doses of about 2.5cGy. This suggests that MVCT not only has a role in radiotherapy treatment planning and guidance, but may also be indicated for surgical guidance and follow-up in regions where metal implants cannot be avoided.
Megavoltage computed tomography (MVCT) has been an active area of research and development in image guided radiation therapy. We have been investigating a particular implementation of MVCT in conjunction with studies of the potential for tomotherapy with a Cobalt-60 radiation source. In this paper, we present results comparing MVCT using a Co-60 source and a 4 MV linear accelerator to conventional kVCT imaging. The Co-60 and linac MVCT measurements were obtained with a first generation benchtop CT imager; the KVCT measurements were obtained using a Philips AcQSim CT Simulator). Phantoms containing various inserts ranging in density from air, through lung, soft tissue and bone equivalent materials and extending to high atomic number metals were imaged with the three modalities. The results enable characterization of image artifacts, CT number linearity and beam hardening. The MVCT images have sufficient contrast that soft tissue regions with 2.8% difference in electron density can be visualized. In MVCT, a linear relationship between CT numbers and electron densities extends to materials with Z ≈ 60. In the 4MV CT imaging there is a position dependence of the CT numbers within a uniform water phantom, which is absent in Co-60 CT images, indicating the presence of beam hardening artifacts in the linac MVCT images. The differences between kVCT and MVCT will be discussed considering the variation of the photon interactions dominating the images. Our investigations indicate that MVCT has properties that may potentially extend its utility beyond radiation therapy.