Cycle-generative adversarial network-based bone suppression imaging for highly accurate markerless motion tracking of lung tumors for cyberknife irradiation therapy.

PURPOSE: Lung tumor tracking during stereotactic radiotherapy with the CyberKnife can misrecognize tumor location under conditions where similar patterns exist in the search area. This study aimed to develop a technique for bone signal suppression during kV-x-ray imaging. METHODS: Paired CT images were created with or without bony structures using a 4D extended cardiac-torso phantom (XCAT phantom) in 56 cases. Subsequently, 3020 2D x-ray images were generated. Images with bone were input into cycle-consistent adversarial network (CycleGAN) and the bone suppressed images on the XCAT phantom (BSIphantom ) were created. They were then compared to images without bone using the structural similarity index measure (SSIM) and peak signal-to-noise ratio (PSNR). Next, 1000 non-simulated treatment images from real cases were input into the training model, and bone-suppressed images of the patient (BSIpatient ) were created. Zero means normalized cross correlation (ZNCC) by template matching between each of the actual treatment images and BSIpatient were calculated. RESULTS: BSIphantom values were compared to their paired images without bone of the XCAT phantom test data; SSIM and PSNR were 0.90 ± 0.06 and 24.54 ± 4.48, respectively. It was visually confirmed that only bone was selectively suppressed without significantly affecting tumor visualization. The ZNCC values of the actual treatment images and BSIpatient were 0.763 ± 0.136 and 0.773 ± 0.143, respectively. The BSIpatient showed improved recognition accuracy over the actual treatment images. CONCLUSIONS: The proposed bone suppression imaging technique based on CycleGAN improves image recognition, making it possible to achieve highly accurate motion tracking irradiation.

Effect of a hydrogel spacer on the intrafractional prostate motion during CyberKnife treatment for prostate cancer.

The purpose of this study was to evaluate the effect of a hydrogel spacer on intrafractional prostate motion during CyberKnife treatment. The retrospective study enrolled 24 patients (with the hydrogel spacer = 12, without the hydrogel spacer = 12) with two fiducial markers. Regarding intrafractional prostate motion, the offset values (mm) of three axes (X-axis; superior [+] to inferior [-], Y-axis; right [+] to left [-], Z-axis; posterior [+] to anterior [-]) obtained from fiducial markers position between a digitally reconstructed radiographs images and live images in the Target Locating System were used, and extracted from generated log files. The mean values of the offset and each standard deviation were calculated for each patient, and both the groups were compared. For all the patients, a total of 2204 offset values and timestamps (without the hydrogel spacer group: 1065, with the hydrogel spacer group: 1139) were recorded for the X-, Y-, and Z-axes, respectively. The offset values (mean ± standard deviation) for the X-, Y-, and Z-axes were -0.04 ± 0.92 mm, -0.03 ± 0.97 mm (P = 0.66), 0.02 ± 0.51, -0.02 ± 0.49 mm (P = 0.50), and 0.56 ± 0.97 mm, 0.34 ± 1.07 mm (P = 0.14), in patients inserted without or with the hydrogel spacer, respectively. There was no effect of a hydrogel spacer on the intrafractional prostate motion in the three axes during CyberKnife treatment for prostate cancer.

Comparison of CT artifacts and image recognition of various fiducial markers including two types of thinner fiducial markers for CyberKnife treatment.

AIM: The aim of the study was to evaluate computed tomography (CT) artifacts and image recognition of the CyberKnife system. Regarding fiducial markers, VISICOIL of 0.5 mm × 5.0 mm and 0.75 mm × 5.0 mm, ball-shaped Gold Anchor (GA) of 0.28 mm × 10 mm and 0.28 mm × 20 mm, were compared with the standard cylinder marker of 0.9 mm × 3.0 mm (ACCULOC). BACKGROUND: Recently, various kinds of commercial fiducial markers have been available in CyberKnife treatment. MATERIALS AND METHODS: The CT images of a water equivalent gel with each fiducial marker were acquired for the evaluation of CT artifacts. The evaluation was performed using the standard deviation of Hounsfield Unit (HU) value for a rectangle region near the fiducial marker. Then, to evaluate the image recognition, each fiducial marker was located to overlap in the target locating system (TLS) for the two sites; the vertebral bone and the pubic bone. RESULTS: For CT artifacts, the standard deviations of the VISICOIL of 0.5 mm × 5.0 mm was the smallest. The image recognition of four fiducial markers had a value close to the standard cylinder marker and was feasible for common use, but was slightly poorer when using GA of 0.28 mm × 10 mm in the dynamic conditions. CONCLUSION: Our results indicated that VISICOIL 0.5 × 5.0 mm and the GAs can be used nearly always for CyberKnife treatment in spite of their much thinner needles than those of cylinder types.