skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: SU-G-BRA-05: Application of a Feature-Based Tracking Algorithm to KV X-Ray Fluoroscopic Images Toward Marker-Less Real-Time Tumor Tracking

Abstract

Purpose: To detect target position on kV X-ray fluoroscopic images using a feature-based tracking algorithm, Accelerated-KAZE (AKAZE), for markerless real-time tumor tracking (RTTT). Methods: Twelve lung cancer patients treated with RTTT on the Vero4DRT (Mitsubishi Heavy Industries, Japan, and Brainlab AG, Feldkirchen, Germany) were enrolled in this study. Respiratory tumor movement was greater than 10 mm. Three to five fiducial markers were implanted around the lung tumor transbronchially for each patient. Before beam delivery, external infrared (IR) markers and the fiducial markers were monitored for 20 to 40 s with the IR camera every 16.7 ms and with an orthogonal kV x-ray imaging subsystem every 80 or 160 ms, respectively. Target positions derived from the fiducial markers were determined on the orthogonal kV x-ray images, which were used as the ground truth in this study. Meanwhile, tracking positions were identified by AKAZE. Among a lot of feature points, AKAZE found high-quality feature points through sequential cross-check and distance-check between two consecutive images. Then, these 2D positional data were converted to the 3D positional data by a transformation matrix with a predefined calibration parameter. Root mean square error (RMSE) was calculated to evaluate the difference between 3D tracking and target positions.more » A total of 393 frames was analyzed. The experiment was conducted on a personal computer with 16 GB RAM, Intel Core i7-2600, 3.4 GHz processor. Results: Reproducibility of the target position during the same respiratory phase was 0.6 +/− 0.6 mm (range, 0.1–3.3 mm). Mean +/− SD of the RMSEs was 0.3 +/− 0.2 mm (range, 0.0–1.0 mm). Median computation time per frame was 179 msec (range, 154–247 msec). Conclusion: AKAZE successfully and quickly detected the target position on kV X-ray fluoroscopic images. Initial results indicate that the differences between 3D tracking and target position would be clinically acceptable.« less

Authors:
; ; ; ; ; ;  [1];  [2]
  1. Kyoto University, Graduate School of Medicine, Kyoto (Japan)
  2. Kyoto University, Graduate School of Informatics, Kyoto (Japan)
Publication Date:
OSTI Identifier:
22649293
Resource Type:
Journal Article
Resource Relation:
Journal Name: Medical Physics; Journal Volume: 43; Journal Issue: 6; Other Information: (c) 2016 American Association of Physicists in Medicine; Country of input: International Atomic Energy Agency (IAEA)
Country of Publication:
United States
Language:
English
Subject:
60 APPLIED LIFE SCIENCES; 61 RADIATION PROTECTION AND DOSIMETRY; ALGORITHMS; BIOMEDICAL RADIOGRAPHY; CALCULATION METHODS; DISTANCE; FIDUCIAL MARKERS; GHZ RANGE 01-100; IMAGES; NEOPLASMS; X RADIATION

Citation Formats

Nakamura, M, Matsuo, Y, Mukumoto, N, Iizuka, Y, Yokota, K, Mizowaki, T, Hiraoka, M, and Nakao, M. SU-G-BRA-05: Application of a Feature-Based Tracking Algorithm to KV X-Ray Fluoroscopic Images Toward Marker-Less Real-Time Tumor Tracking. United States: N. p., 2016. Web. doi:10.1118/1.4956929.
Nakamura, M, Matsuo, Y, Mukumoto, N, Iizuka, Y, Yokota, K, Mizowaki, T, Hiraoka, M, & Nakao, M. SU-G-BRA-05: Application of a Feature-Based Tracking Algorithm to KV X-Ray Fluoroscopic Images Toward Marker-Less Real-Time Tumor Tracking. United States. doi:10.1118/1.4956929.
Nakamura, M, Matsuo, Y, Mukumoto, N, Iizuka, Y, Yokota, K, Mizowaki, T, Hiraoka, M, and Nakao, M. 2016. "SU-G-BRA-05: Application of a Feature-Based Tracking Algorithm to KV X-Ray Fluoroscopic Images Toward Marker-Less Real-Time Tumor Tracking". United States. doi:10.1118/1.4956929.
@article{osti_22649293,
title = {SU-G-BRA-05: Application of a Feature-Based Tracking Algorithm to KV X-Ray Fluoroscopic Images Toward Marker-Less Real-Time Tumor Tracking},
author = {Nakamura, M and Matsuo, Y and Mukumoto, N and Iizuka, Y and Yokota, K and Mizowaki, T and Hiraoka, M and Nakao, M},
abstractNote = {Purpose: To detect target position on kV X-ray fluoroscopic images using a feature-based tracking algorithm, Accelerated-KAZE (AKAZE), for markerless real-time tumor tracking (RTTT). Methods: Twelve lung cancer patients treated with RTTT on the Vero4DRT (Mitsubishi Heavy Industries, Japan, and Brainlab AG, Feldkirchen, Germany) were enrolled in this study. Respiratory tumor movement was greater than 10 mm. Three to five fiducial markers were implanted around the lung tumor transbronchially for each patient. Before beam delivery, external infrared (IR) markers and the fiducial markers were monitored for 20 to 40 s with the IR camera every 16.7 ms and with an orthogonal kV x-ray imaging subsystem every 80 or 160 ms, respectively. Target positions derived from the fiducial markers were determined on the orthogonal kV x-ray images, which were used as the ground truth in this study. Meanwhile, tracking positions were identified by AKAZE. Among a lot of feature points, AKAZE found high-quality feature points through sequential cross-check and distance-check between two consecutive images. Then, these 2D positional data were converted to the 3D positional data by a transformation matrix with a predefined calibration parameter. Root mean square error (RMSE) was calculated to evaluate the difference between 3D tracking and target positions. A total of 393 frames was analyzed. The experiment was conducted on a personal computer with 16 GB RAM, Intel Core i7-2600, 3.4 GHz processor. Results: Reproducibility of the target position during the same respiratory phase was 0.6 +/− 0.6 mm (range, 0.1–3.3 mm). Mean +/− SD of the RMSEs was 0.3 +/− 0.2 mm (range, 0.0–1.0 mm). Median computation time per frame was 179 msec (range, 154–247 msec). Conclusion: AKAZE successfully and quickly detected the target position on kV X-ray fluoroscopic images. Initial results indicate that the differences between 3D tracking and target position would be clinically acceptable.},
doi = {10.1118/1.4956929},
journal = {Medical Physics},
number = 6,
volume = 43,
place = {United States},
year = 2016,
month = 6
}
  • Purpose: The new real-time tumor-tracking radiotherapy (RTRT) system was installed in our institution. This system consists of two x-ray tubes and color image intensifiers (I.I.s). The fiducial marker which was implanted near the tumor was tracked using color fluoroscopic images. However, the implantation of the fiducial marker is very invasive. Color fluoroscopic images enable to increase the recognition of the tumor. However, these images were not suitable to track the tumor without fiducial marker. The purpose of this study was to investigate the feasibility of markerless tracking using dual energy colored fluoroscopic images for real-time tumor-tracking radiotherapy system. Methods: Themore » colored fluoroscopic images of static and moving phantom that had the simulated tumor (30 mm diameter sphere) were experimentally acquired using the RTRT system. The programmable respiratory motion phantom was driven using the sinusoidal pattern in cranio-caudal direction (Amplitude: 20 mm, Time: 4 s). The x-ray condition was set to 55 kV, 50 mA and 105 kV, 50 mA for low energy and high energy, respectively. Dual energy images were calculated based on the weighted logarithmic subtraction of high and low energy images of RGB images. The usefulness of dual energy imaging for real-time tracking with an automated template image matching algorithm was investigated. Results: Our proposed dual energy subtraction improve the contrast between tumor and background to suppress the bone structure. For static phantom, our results showed that high tracking accuracy using dual energy subtraction images. For moving phantom, our results showed that good tracking accuracy using dual energy subtraction images. However, tracking accuracy was dependent on tumor position, tumor size and x-ray conditions. Conclusion: We indicated that feasibility of markerless tracking using dual energy fluoroscopic images for real-time tumor-tracking radiotherapy system. Furthermore, it is needed to investigate the tracking accuracy using proposed dual energy subtraction images for clinical cases.« less
  • Purpose: In-treatment tumor localization is critical for the management of tumor motion in lung cancer radiotherapy. Conventional tumor-tracking methods using a kV or MV x-ray projection has limited contrast. To facilitate real-time, marker-less and low-dose in-treatment image tumor tracking, we propose a novel scheme using Compton scatter imaging. This study reports Monte Carlo (MC) simulations on this scheme for the purpose of proof-of-principle. Methods: A slit x-ray beam along the patient superior-inferior (SI) direction is directed to the patient, intersecting the patient lung at a 2D plane containing majority part of the tumor motion trajectory. X-ray photons are scattered duemore » to Compton effect from this plane, which are spatially collimated by, e.g., a pinhole, on one side of the plane and then captured by a detector behind it. The captured image, after correcting for x-ray attenuation and scatter angle variation, reflects the electron density, which allows visualization of the instantaneous anatomy on this plane. We performed MC studies on a phantom and a patient case for the initial test of this proposed method. Results: In the phantom case, the contrast-resolution calculated using tumor/lung as foreground/background for kV fluoroscopy, cone-beam CT, and scattering image were 0.0625, 0.6993, and 0.5290, respectively. In the patient case, tumor motion can be clearly observed in the scatter images. Compared to fluoroscopy, scattering imaging also significantly reduced imaging dose because of its narrower beam design. Conclusion: MC simulation studies demonstrated the potential of the proposed scheme in terms of capturing the instantaneous anatomy of a patient on a 2D plane. Clear visualization of the tumor will probably facilitate ‘marker-less’ and ‘real-time’ tumor tracking with low imaging dose. NIH (1R01CA154747-01, 1R21CA178787-01A1 and 1R21EB017978-01A1)« less
  • Purpose: To propose a computerized framework for localization of anatomical feature points on the patient surface in infrared-ray based range images by using differential geometry (curvature) features. Methods: The general concept was to reconstruct the patient surface by using a mathematical modeling technique for the computation of differential geometry features that characterize the local shapes of the patient surfaces. A region of interest (ROI) was firstly extracted based on a template matching technique applied on amplitude (grayscale) images. The extracted ROI was preprocessed for reducing temporal and spatial noises by using Kalman and bilateral filters, respectively. Next, a smooth patientmore » surface was reconstructed by using a non-uniform rational basis spline (NURBS) model. Finally, differential geometry features, i.e. the shape index and curvedness features were computed for localizing the anatomical feature points. The proposed framework was trained for optimizing shape index and curvedness thresholds and tested on range images of an anthropomorphic head phantom. The range images were acquired by an infrared ray-based time-of-flight (TOF) camera. The localization accuracy was evaluated by measuring the mean of minimum Euclidean distances (MMED) between reference (ground truth) points and the feature points localized by the proposed framework. The evaluation was performed for points localized on convex regions (e.g. apex of nose) and concave regions (e.g. nasofacial sulcus). Results: The proposed framework has localized anatomical feature points on convex and concave anatomical landmarks with MMEDs of 1.91±0.50 mm and 3.70±0.92 mm, respectively. A statistically significant difference was obtained between the feature points on the convex and concave regions (P<0.001). Conclusion: Our study has shown the feasibility of differential geometry features for localization of anatomical feature points on the patient surface in range images. The proposed framework might be useful for tasks involving feature-based image registration in range-image guided radiation therapy.« less
  • Respiration can cause tumor movements in thoracic regions of up to 3 cm. To minimize motion effects several approaches, such as gating and deep inspiration breath hold (DIBH), are still under development. The goal of our study was to develop and evaluate a noninvasive system for gated DIBH (GDIBH) based on external markers. DIBH monitoring was based on an infrared tracking system and an in-house-developed software. The in-house software provided the breathing curve in real time and was used as on-line information for a prototype of a feedback device. Reproducibility and stability of the breath holds were evaluated without andmore » with feedback. Thirty-five patients undergoing stereotactic body radiotherapy (SBRT) performed DIBH maneuvers after each treatment. For 16 patients dynamic imaging sequences on a multislice CT were used to determine the correlation between tumor and external markers. The relative reproducibility of DIBH maneuvers was improved with the feedback device (74.5%{+-}17.1% without versus 93.0%{+-}4.4% with feedback). The correlation between tumor and marker was good (Pearson correlation coefficient 0.83{+-}0.17). The regression slopes showed great intersubject variability but on average the internal margin in a DIBH treatment situation could be theoretically reduced by 3 mm with the feedback device. DIBH monitoring could be realized in a noninvasive manner through external marker tracking. We conclude that reduction of internal margins can be achieved with a feedback system but should be performed with great care due to the individual behavior of target motion.« less
  • Purpose: To propose a simple model to explain the origin of ghost markers in marker-based optical tracking systems (OTS) and to develop retrospective strategies to detect and eliminate ghost markers. Methods: In marker-based OTS, ghost markers are virtual markers created due to the cross-talk between the two camera sensors, which can lead to system execution failure or inaccuracy in patient tracking. As a result, the users have to limit the number of markers and avoid certain marker configurations to reduce the chances of ghost markers. In this work, the authors propose retrospective strategies to detect and eliminate ghost markers. Themore » two camera sensors were treated as mathematical points in space. The authors identified the coplanar within limit (CWL) condition as the necessary condition for ghost marker occurrence. A simple ghost marker detection method was proposed based on the model. Ghost marker elimination was achieved through pattern matching: a ghost marker-free reference set was matched with the optical marker set observed by the OTS; unmatched optical markers were eliminated as either ghost markers or misplaced markers. The pattern matching problem was formulated as a constraint satisfaction problem (using pairwise distances as constraints) and solved with an iterative backtracking algorithm. Wildcard markers were introduced to address missing or misplaced markers. An experiment was designed to measure the sensor positions and the limit for the CWL condition. The ghost marker detection and elimination algorithms were verified with samples collected from a five-marker jig and a nine-marker anthropomorphic phantom, rotated with the treatment couch from −60° to +60°. The accuracy of the pattern matching algorithm was further validated with marker patterns from 40 patients who underwent stereotactic body radiotherapy (SBRT). For this purpose, a synthetic optical marker pattern was created for each patient by introducing ghost markers, marker position uncertainties, and marker displacement. Results: The sensor positions and the limit for the CWL condition were measured with excellent reproducibility (standard deviation ≤ 0.39 mm). The ghost marker detection algorithm had perfect detection accuracy for both the jig (1544 samples) and the anthropomorphic phantom (2045 samples). Pattern matching was successful for all samples from both phantoms as well as the 40 patient marker patterns. Conclusions: The authors proposed a simple model to explain the origin of ghost markers and identified the CWL condition as the necessary condition for ghost marker occurrence. The retrospective ghost marker detection and elimination algorithms guarantee complete ghost marker elimination while providing the users with maximum flexibility in selecting the number of markers and their configuration to meet their clinic needs.« less