The Experts below are selected from a list of 306 Experts worldwide ranked by ideXlab platform
Steve B Jiang - One of the best experts on this subject based on the ideXlab platform.
-
Applications of Machine Learning in Real-Time Tumor Localization
Machine Learning in Computer-Aided Diagnosis, 2012Co-Authors: Steve B JiangAbstract:Recently, machine learning has gained great popularity in many aspects of radiation therapy. In this chapter, the authors will demonstrate the applications of various machine learning techniques in the context of real-time Tumor Localization in lung cancer radiotherapy. These cover a wide range of well established machine learning techniques, including principal component analysis, linear discriminant analysis, artificial neural networks, and support vector machine, etc. Respiratory gating, as a special case of Tumor Localization, will also be discussed. The chapter will demonstrate how domain specific knowledge and prior information can be useful in achieving more accurate and robust Tumor Localization. Future research directions in machine learning that can further improve the accuracy for Tumor Localization are also discussed.
-
3D Tumor Localization through real-time volumetric x-ray imaging for lung cancer radiotherapy
Medical physics, 2011Co-Authors: John H Lewis, Xun Jia, M Folkerts, Chunhua Men, William Y. Song, Steve B JiangAbstract:Recently we have developed an algorithm for reconstructing volumetric images and extracting 3D Tumor motion information from a single x-ray projection. We have demonstrated its feasibility using a digital respiratory phantom with regular breathing patterns. In this work, we present a detailed description and a comprehensive evaluation of the improved algorithm. The algorithm was improved by incorporating respiratory motion prediction. The accuracy and efficiency were then evaluated on 1) a digital respiratory phantom, 2) a physical respiratory phantom, and 3) five lung cancer patients. These evaluation cases include both regular and irregular breathing patterns that are different from the training dataset. For the digital respiratory phantom with regular and irregular breathing, the average 3D Tumor Localization error is less than 1 mm. On an NVIDIA Tesla C1060 GPU card, the average computation time for 3D Tumor Localization from each projection ranges between 0.19 and 0.26 seconds, for both regular and irregular breathing, which is about a 10% improvement over previously reported results. For the physical respiratory phantom, an average Tumor Localization error below 1 mm was achieved with an average computation time of 0.13 and 0.16 seconds on the same GPU card, for regular and irregular breathing, respectively. For the five lung cancer patients, the average Tumor Localization error is below 2 mm in both the axial and tangential directions. The average computation time on the same GPU card ranges between 0.26 and 0.34 seconds.
-
3d Tumor Localization through real time volumetric x ray imaging for lung cancer radiotherapy
Medical Physics, 2011Co-Authors: John H Lewis, Xun Jia, M Folkerts, Chunhua Men, W Song, Steve B JiangAbstract:Purpose : To evaluate an algorithm for real-time 3D Tumor Localization from a single x-ray projection image for lungcancerradiotherapy. Methods : Recently, we have developed an algorithm for reconstructing volumetric images and extracting 3D Tumor motion information from a single x-ray projection [Liet al., Med. Phys. 37, 2822–2826 (2010)]. We have demonstrated its feasibility using a digital respiratory phantom with regular breathing patterns. In this work, we present a detailed description and a comprehensive evaluation of the improved algorithm. The algorithm was improved by incorporating respiratory motion prediction. The accuracy and efficiency of using this algorithm for 3D Tumor Localization were then evaluated on (1) a digital respiratory phantom, (2) a physical respiratory phantom, and (3) five lungcancer patients. These evaluation cases include both regular and irregular breathing patterns that are different from the training dataset. Results : For the digital respiratory phantom with regular and irregular breathing, the average 3D Tumor Localization error is less than 1 mm which does not seem to be affected by amplitude change, period change, or baseline shift. On an NVIDIA Tesla C1060 graphic processing unit (GPU) card, the average computation time for 3D Tumor Localization from each projection ranges between 0.19 and 0.26 s, for both regular and irregular breathing, which is about a 10% improvement over previously reported results. For the physical respiratory phantom, an average Tumor Localization error below 1 mm was achieved with an average computation time of 0.13 and 0.16 s on the same graphic processing unit (GPU) card, for regular and irregular breathing, respectively. For the five lungcancer patients, the average Tumor Localization error is below 2 mm in both the axial and tangential directions. The average computation time on the same GPU card ranges between 0.26 and 0.34 s. Conclusions : Through a comprehensive evaluation of our algorithm, we have established its accuracy in 3D Tumor Localization to be on the order of 1 mm on average and 2 mm at 95 percentile for both digital and physical phantoms, and within 2 mm on average and 4 mm at 95 percentile for lungcancer patients. The results also indicate that the accuracy is not affected by the breathing pattern, be it regular or irregular. High computational efficiency can be achieved on GPU, requiring 0.1–0.3 s for each x-ray projection.
-
TU‐D‐204C‐02: Machine Learning in Real‐time Tumor Localization
Medical Physics, 2010Co-Authors: Steve B JiangAbstract:Real‐time Tumor Localization is an essential step to achieve conformai lungcancerradiotherapy. In practice, Tumors may be localized either directly (e.g., in fluoroscopy) or indirectly by some type of surrogate. Direct Localization with implanted markers poses the risk of pneumothorax to lungcancer patients, while indirect Localization is limited by its accuracy. Direct Localization in fluoroscopy without implanted markers achieves a reasonable tradeoff between accuracy and risk to the patients. However, due to the poor soft tissuecontrast and superposition of different anatomical structures, it is generally very difficult to localize the Tumor in fluoroscopy with conventional methods in computer vision and image processing. Recently, machine learning has gained great popularity in many aspects of radiation therapy. In this lecture, we will demonstrate the applications of various machine learning techniques in the context of real‐time Tumor Localization in lungcancerradiotherapy. These cover a wide range of well established machine learning techniques, including principal component analysis (PCA), artificial neural networks, and support vector regression, etc.. We will distinguish between two different paradigms when applying these techniques: one in a two‐dimensional (2D) framework and the other in a three‐dimensional (3D) framework. In the 2D framework, a (supervised) regression model is built between a parametric representation of the fluoroscopic images. and the Tumor locations of the fluoroscopicimag fluoroscopic images. Because the representation of the fluoroscopic images is implicit in the 2D framework, it is difficult to account for deformational and large translational changes in the Tumor. In the 3D framework, by incorporating the prior information in 4DCT or 4DCBCT, a PCA model is constructed to explicitly represent the entire lung motion in a realistic and efficient way. This model allows one to use a single x‐ray projection image to not only derive the 3D Tumor location, but also reconstruct the corresponding volumetric image of the patient. With the aid of graphics processing units, this computationally intensive task can be achieved within half a second given one projection. This lecture provides an overview of the different machine learning techniques used in real‐time Tumor Localization. A detailed description of these machine learning techniques is also presented. Finally, extensions to the current framework as well as some future directions are discussed. Learning Objectives: 1. Understand the benefits and challenges of real‐time Tumor Localization without implanted markers. 2. Understand the principles of different types of machine learning techniques for real‐time Tumor Localization. 3. Understand how machine learning techniques are used to achieve accurate real‐time Tumor Localization. Conflict of Interest: This work is partially supported by Varian Master Research Agreement.
-
real time volumetric image reconstruction and 3d Tumor Localization based on a single x ray projection image for lung cancer radiotherapy
arXiv: Medical Physics, 2010Co-Authors: Xun Jia, John H Lewis, M Folkerts, Chunhua Men, Steve B JiangAbstract:Purpose: To develop an algorithm for real-time volumetric image reconstruction and 3D Tumor Localization based on a single x-ray projection image for lung cancer radiotherapy. Methods: Given a set of volumetric images of a patient at N breathing phases as the training data, we perform deformable image registration between a reference phase and the other N-1 phases, resulting in N-1 deformation vector fields (DVFs). These DVFs can be represented efficiently by a few eigenvectors and coefficients obtained from principal component analysis (PCA). By varying the PCA coefficients, we can generate new DVFs, which, when applied on the reference image, lead to new volumetric images. We then can reconstruct a volumetric image from a single projection image by optimizing the PCA coefficients such that its computed projection matches the measured one. The 3D location of the Tumor can be derived by applying the inverted DVF on its position in the reference image. Our algorithm was implemented on graphics processing units (GPUs) to achieve real-time efficiency. We generated the training data using a realistic and dynamic mathematical phantom with 10 breathing phases. The testing data were 360 cone beam projections corresponding to one gantry rotation, simulated using the same phantom with a 50% increase in breathing amplitude. Results: The average relative image intensity error of the reconstructed volumetric images is 6.9% +/- 2.4%. The average 3D Tumor Localization error is 0.8 mm +/- 0.5 mm. On an NVIDIA Tesla C1060 GPU card, the average computation time for reconstructing a volumetric image from each projection is 0.24 seconds (range: 0.17 and 0.35 seconds). Conclusions: We have shown the feasibility of reconstructing volumetric images and localizing Tumor positions in 3D in near real-time from a single x-ray image.
J Xiang - One of the best experts on this subject based on the ideXlab platform.
-
combined cd4 th1 effect and lymphotactin transgene expression enhance cd8 tc1 Tumor Localization and therapy
Gene Therapy, 2005Co-Authors: Hui Huang, J Y Yuan, Xulin Guo, J XiangAbstract:Type 1 T cells are the major components in antiTumor immunity. The lack of efficient CD8(+) cytotoxic T (Tc) cell infiltration of Tumors is a major obstacle to adoptive Tc-cell therapy. We have previously demonstrated that adenovirus (AdV)-mediated transgene lymphotactin (Lptn) expression by intraTumoral AdVLptn injection and intravenous CD4(+) helper T (Th) cell transfer can enhance Tc-cell Tumor infiltration and eradication of early stage Tumors (5 mm in diameter). In this study, we generated ovalbumin (OVA)-specific Tc1 and Th1 cells in vitro by incubation of OVA-pulsed dendritic cells with naive T cells from T-cell receptor (TCR) transgenic OT I and OT II mice. We then investigated the potential synergy of Th1 help effect and Lptn transgene expression in Tc1-cell therapy of well-established OVA-expressing EG7 solid Tumors (7 mm in diameter). Our data showed that a combined adoptive T-cell therapy of Th1 (2.5 x 10(6) cells per mouse) and Tc1 (5 x 10(6) cells per mouse) resulted in regression of all eight (100%) transgene Lptn expressed EG7 Tumors, which is significantly higher than four from eight (50%) in AdVLptn/Tc1 group and two from eight (25%) in Tc1/Th1 group (P < 0.05). The amount of transferred Tc1 cells detected in Lptn-expressed Tumors with Th1 treatment is 0.72%, which is significantly higher than those of AdVLptn (0.22%), Th1 (0.41%) and the control AdVpLpA (0.09%) treatment groups (P < 0.05). Enhanced Tc1 Tumor Localization may be derived from the chemotactic effect of Lptn and the proliferative effect of Th1 and Lptn. This novel therapeutic strategy with enhancement of Tc1 Tumor Localization in the therapy of well-established Tumors may become a tool of considerable conceptual interest in the implementation of future clinical objectives.
-
Combined CD4 + Th1 effect and lymphotactin transgene expression enhance CD8 + Tc1 Tumor Localization and therapy
Gene therapy, 2005Co-Authors: Hui Huang, J Y Yuan, Xulin Guo, J XiangAbstract:Type 1 T cells are the major components in antiTumor immunity. The lack of efficient CD8(+) cytotoxic T (Tc) cell infiltration of Tumors is a major obstacle to adoptive Tc-cell therapy. We have previously demonstrated that adenovirus (AdV)-mediated transgene lymphotactin (Lptn) expression by intraTumoral AdVLptn injection and intravenous CD4(+) helper T (Th) cell transfer can enhance Tc-cell Tumor infiltration and eradication of early stage Tumors (5 mm in diameter). In this study, we generated ovalbumin (OVA)-specific Tc1 and Th1 cells in vitro by incubation of OVA-pulsed dendritic cells with naive T cells from T-cell receptor (TCR) transgenic OT I and OT II mice. We then investigated the potential synergy of Th1 help effect and Lptn transgene expression in Tc1-cell therapy of well-established OVA-expressing EG7 solid Tumors (7 mm in diameter). Our data showed that a combined adoptive T-cell therapy of Th1 (2.5 x 10(6) cells per mouse) and Tc1 (5 x 10(6) cells per mouse) resulted in regression of all eight (100%) transgene Lptn expressed EG7 Tumors, which is significantly higher than four from eight (50%) in AdVLptn/Tc1 group and two from eight (25%) in Tc1/Th1 group (P < 0.05). The amount of transferred Tc1 cells detected in Lptn-expressed Tumors with Th1 treatment is 0.72%, which is significantly higher than those of AdVLptn (0.22%), Th1 (0.41%) and the control AdVpLpA (0.09%) treatment groups (P < 0.05). Enhanced Tc1 Tumor Localization may be derived from the chemotactic effect of Lptn and the proliferative effect of Th1 and Lptn. This novel therapeutic strategy with enhancement of Tc1 Tumor Localization in the therapy of well-established Tumors may become a tool of considerable conceptual interest in the implementation of future clinical objectives.
Rajni V. Patel - One of the best experts on this subject based on the ideXlab platform.
-
A Breakthrough in Tumor Localization: Combining Tactile Sensing and Ultrasound to Improve Tumor Localization in Robotics-Assisted Minimally Invasive Surgery
IEEE Robotics & Automation Magazine, 2017Co-Authors: Anish S. Naidu, Michael D Naish, Rajni V. PatelAbstract:Robotics-assisted minimally invasive surgery (RAMIS) helps surgeons to avoid manually palpating organs to locate subsurface Tumors. One solution has been to use ultrasound, but it is not always reliable. Tactile sensing, however, has the potential to augment ultrasound to improve Tumor Localization, but there are no existing RAMIS instruments that employ both modalities.
-
Integration of Force Reflection with Tactile Sensing for Minimally Invasive Robotics-Assisted Tumor Localization
IEEE transactions on haptics, 2013Co-Authors: Ali Talasaz, Rajni V. PatelAbstract:Tactile sensing and force reflection have been the subject of considerable research for Tumor Localization in soft-tissue palpation. The work presented in this paper investigates the relevance of force feedback (presented visually as well as directly) during tactile sensing (presented visually only) for Tumor Localization using an experimental setup close to one that could be applied for real robotics-assisted minimally invasive surgery. The setup is a teleoperated (master-slave) system facilitated with a state-of-the-art minimally invasive probe with a rigidly mounted tactile sensor at the tip and an externally mounted force sensor at the base of the probe. The objective is to capture the tactile information and measure the interaction forces between the probe and tissue during palpation and to explore how they can be integrated to improve the performance of Tumor Localization. To quantitatively explore the effect of force feedback on tactile sensing Tumor Localization, several experiments were conducted by human subjects to locate artificial Tumors embedded in the ex vivo bovine livers. The results show that using tactile sensing in a force-controlled environment can realize, on average, 57 percent decrease in the maximum force and 55 percent decrease in the average force applied to tissue while increasing the Tumor detection accuracy by up to 50 percent compared to the case of using tactile feedback alone. The results also show that while visual presentation of force feedback gives straightforward quantitative measures, improved performance of tactile sensing Tumor Localization is achieved at the expense of longer times for the user. Also, the quickness and intuitive data mapping of direct force feedback makes it more appealing to experienced users.
-
Design of a Minimally Invasive Lung Tumor Localization Device
Volume 3: Renewable Energy Systems; Robotics; Robust Control; Single Track Vehicle Dynamics and Control; Stochastic Models Control and Algorithms in R, 2012Co-Authors: Tomasz P Kurowski, Michael D Naish, Rajni V. Patel, Ana Luisa Trejos, Richard A. MalthanerAbstract:This paper describes the design, analysis, and experimental validation of a novel minimally invasive instrument for lung Tumor Localization. The instrument end effector is a two-degree of freedom lung tissue palpator. It allows for optimal tissue palpation to increase useful sensor feedback by ensuring sensor contact, and prevents tissue damage by uniformly distributing pressure on the tissue. Finite element analysis was used to guide the design process, resulting in a final design that could achieve a factor of safety of 4 for a 20 N force acting on the end effector—the approximate weight of a human lung. Validation experiments were conducted on a prototype instrument to assess its articulation and load-carrying capacity. The end effector design allows for the inclusion of ultrasound, tactile, and kinesthetic sensors. It is expected that this device will form the basis for robotics-assisted palpation and increase the likelihood of positive Tumor Localization.Copyright © 2012 by ASME
-
Design of a Minimally Invasive Lung Tumor Localization Device
Volume 3: Renewable Energy Systems; Robotics; Robust Control; Single Track Vehicle Dynamics and Control; Stochastic Models Control and Algorithms in R, 2012Co-Authors: Tomasz P Kurowski, Michael D Naish, Rajni V. Patel, Ana Luisa Trejos, Richard A. MalthanerAbstract:This paper describes the design, analysis, and experimental validation of a novel minimally invasive instrument for lung Tumor Localization. The instrument end effector is a two-degree of freedom lung tissue palpator. It allows for optimal tissue palpation to increase useful sensor feedback by ensuring sensor contact, and prevents tissue damage by uniformly distributing pressure on the tissue. Finite element analysis was used to guide the design process, resulting in a final design that could achieve a factor of safety of 4 for a 20 N force acting on the end effector—the approximate weight of a human lung. Validation experiments were conducted on a prototype instrument to assess its articulation and load-carrying capacity. The end effector design allows for the inclusion of ultrasound, tactile, and kinesthetic sensors. It is expected that this device will form the basis for robotics-assisted palpation and increase the likelihood of positive Tumor Localization.Copyright © 2012 by ASME
-
Robotic Techniques for Minimally Invasive Tumor Localization
Surgical Robotics, 2010Co-Authors: Michael D Naish, M. T. Perri, Rajni V. Patel, Ana Luisa Trejos, Richard A. MalthanerAbstract:The challenges imposed by Minimally Invasive Surgery (MIS) have been the subject of significant research in the last decade. In the case of cancer surgery, a significant limitation is the inability to effectively palpate the target tissue to localize Tumor nodules for treatment or removal. Current clinical technologies are still limited and Tumor Localization efforts often result in the need to increase the size of the incision to allow finger access for direct palpation. New methods of MIS Tumor Localization under investigation involve restoring the sense of touch, or haptic feedback. The two most commonly investigated modes of haptic perception include kinesthetic and tactile sensing, each with its own advantages and disadvantages. Work in this area includes the development of customized instruments with embedded sensors that aim to solve the problem of limited haptic feedback in MIS. This chapter provides a review of the work to date in the use of kinesthetic and tactile sensing information in MIS for tissue palpation, with the goal of highlighting the benefits and limitations of each mode when used to locate hidden Tumors during MIS.
Hui Huang - One of the best experts on this subject based on the ideXlab platform.
-
combined cd4 th1 effect and lymphotactin transgene expression enhance cd8 tc1 Tumor Localization and therapy
Gene Therapy, 2005Co-Authors: Hui Huang, J Y Yuan, Xulin Guo, J XiangAbstract:Type 1 T cells are the major components in antiTumor immunity. The lack of efficient CD8(+) cytotoxic T (Tc) cell infiltration of Tumors is a major obstacle to adoptive Tc-cell therapy. We have previously demonstrated that adenovirus (AdV)-mediated transgene lymphotactin (Lptn) expression by intraTumoral AdVLptn injection and intravenous CD4(+) helper T (Th) cell transfer can enhance Tc-cell Tumor infiltration and eradication of early stage Tumors (5 mm in diameter). In this study, we generated ovalbumin (OVA)-specific Tc1 and Th1 cells in vitro by incubation of OVA-pulsed dendritic cells with naive T cells from T-cell receptor (TCR) transgenic OT I and OT II mice. We then investigated the potential synergy of Th1 help effect and Lptn transgene expression in Tc1-cell therapy of well-established OVA-expressing EG7 solid Tumors (7 mm in diameter). Our data showed that a combined adoptive T-cell therapy of Th1 (2.5 x 10(6) cells per mouse) and Tc1 (5 x 10(6) cells per mouse) resulted in regression of all eight (100%) transgene Lptn expressed EG7 Tumors, which is significantly higher than four from eight (50%) in AdVLptn/Tc1 group and two from eight (25%) in Tc1/Th1 group (P < 0.05). The amount of transferred Tc1 cells detected in Lptn-expressed Tumors with Th1 treatment is 0.72%, which is significantly higher than those of AdVLptn (0.22%), Th1 (0.41%) and the control AdVpLpA (0.09%) treatment groups (P < 0.05). Enhanced Tc1 Tumor Localization may be derived from the chemotactic effect of Lptn and the proliferative effect of Th1 and Lptn. This novel therapeutic strategy with enhancement of Tc1 Tumor Localization in the therapy of well-established Tumors may become a tool of considerable conceptual interest in the implementation of future clinical objectives.
-
Combined CD4 + Th1 effect and lymphotactin transgene expression enhance CD8 + Tc1 Tumor Localization and therapy
Gene therapy, 2005Co-Authors: Hui Huang, J Y Yuan, Xulin Guo, J XiangAbstract:Type 1 T cells are the major components in antiTumor immunity. The lack of efficient CD8(+) cytotoxic T (Tc) cell infiltration of Tumors is a major obstacle to adoptive Tc-cell therapy. We have previously demonstrated that adenovirus (AdV)-mediated transgene lymphotactin (Lptn) expression by intraTumoral AdVLptn injection and intravenous CD4(+) helper T (Th) cell transfer can enhance Tc-cell Tumor infiltration and eradication of early stage Tumors (5 mm in diameter). In this study, we generated ovalbumin (OVA)-specific Tc1 and Th1 cells in vitro by incubation of OVA-pulsed dendritic cells with naive T cells from T-cell receptor (TCR) transgenic OT I and OT II mice. We then investigated the potential synergy of Th1 help effect and Lptn transgene expression in Tc1-cell therapy of well-established OVA-expressing EG7 solid Tumors (7 mm in diameter). Our data showed that a combined adoptive T-cell therapy of Th1 (2.5 x 10(6) cells per mouse) and Tc1 (5 x 10(6) cells per mouse) resulted in regression of all eight (100%) transgene Lptn expressed EG7 Tumors, which is significantly higher than four from eight (50%) in AdVLptn/Tc1 group and two from eight (25%) in Tc1/Th1 group (P < 0.05). The amount of transferred Tc1 cells detected in Lptn-expressed Tumors with Th1 treatment is 0.72%, which is significantly higher than those of AdVLptn (0.22%), Th1 (0.41%) and the control AdVpLpA (0.09%) treatment groups (P < 0.05). Enhanced Tc1 Tumor Localization may be derived from the chemotactic effect of Lptn and the proliferative effect of Th1 and Lptn. This novel therapeutic strategy with enhancement of Tc1 Tumor Localization in the therapy of well-established Tumors may become a tool of considerable conceptual interest in the implementation of future clinical objectives.
Xun Jia - One of the best experts on this subject based on the ideXlab platform.
-
3D Tumor Localization through real-time volumetric x-ray imaging for lung cancer radiotherapy
Medical physics, 2011Co-Authors: John H Lewis, Xun Jia, M Folkerts, Chunhua Men, William Y. Song, Steve B JiangAbstract:Recently we have developed an algorithm for reconstructing volumetric images and extracting 3D Tumor motion information from a single x-ray projection. We have demonstrated its feasibility using a digital respiratory phantom with regular breathing patterns. In this work, we present a detailed description and a comprehensive evaluation of the improved algorithm. The algorithm was improved by incorporating respiratory motion prediction. The accuracy and efficiency were then evaluated on 1) a digital respiratory phantom, 2) a physical respiratory phantom, and 3) five lung cancer patients. These evaluation cases include both regular and irregular breathing patterns that are different from the training dataset. For the digital respiratory phantom with regular and irregular breathing, the average 3D Tumor Localization error is less than 1 mm. On an NVIDIA Tesla C1060 GPU card, the average computation time for 3D Tumor Localization from each projection ranges between 0.19 and 0.26 seconds, for both regular and irregular breathing, which is about a 10% improvement over previously reported results. For the physical respiratory phantom, an average Tumor Localization error below 1 mm was achieved with an average computation time of 0.13 and 0.16 seconds on the same GPU card, for regular and irregular breathing, respectively. For the five lung cancer patients, the average Tumor Localization error is below 2 mm in both the axial and tangential directions. The average computation time on the same GPU card ranges between 0.26 and 0.34 seconds.
-
3d Tumor Localization through real time volumetric x ray imaging for lung cancer radiotherapy
Medical Physics, 2011Co-Authors: John H Lewis, Xun Jia, M Folkerts, Chunhua Men, W Song, Steve B JiangAbstract:Purpose : To evaluate an algorithm for real-time 3D Tumor Localization from a single x-ray projection image for lungcancerradiotherapy. Methods : Recently, we have developed an algorithm for reconstructing volumetric images and extracting 3D Tumor motion information from a single x-ray projection [Liet al., Med. Phys. 37, 2822–2826 (2010)]. We have demonstrated its feasibility using a digital respiratory phantom with regular breathing patterns. In this work, we present a detailed description and a comprehensive evaluation of the improved algorithm. The algorithm was improved by incorporating respiratory motion prediction. The accuracy and efficiency of using this algorithm for 3D Tumor Localization were then evaluated on (1) a digital respiratory phantom, (2) a physical respiratory phantom, and (3) five lungcancer patients. These evaluation cases include both regular and irregular breathing patterns that are different from the training dataset. Results : For the digital respiratory phantom with regular and irregular breathing, the average 3D Tumor Localization error is less than 1 mm which does not seem to be affected by amplitude change, period change, or baseline shift. On an NVIDIA Tesla C1060 graphic processing unit (GPU) card, the average computation time for 3D Tumor Localization from each projection ranges between 0.19 and 0.26 s, for both regular and irregular breathing, which is about a 10% improvement over previously reported results. For the physical respiratory phantom, an average Tumor Localization error below 1 mm was achieved with an average computation time of 0.13 and 0.16 s on the same graphic processing unit (GPU) card, for regular and irregular breathing, respectively. For the five lungcancer patients, the average Tumor Localization error is below 2 mm in both the axial and tangential directions. The average computation time on the same GPU card ranges between 0.26 and 0.34 s. Conclusions : Through a comprehensive evaluation of our algorithm, we have established its accuracy in 3D Tumor Localization to be on the order of 1 mm on average and 2 mm at 95 percentile for both digital and physical phantoms, and within 2 mm on average and 4 mm at 95 percentile for lungcancer patients. The results also indicate that the accuracy is not affected by the breathing pattern, be it regular or irregular. High computational efficiency can be achieved on GPU, requiring 0.1–0.3 s for each x-ray projection.
-
real time volumetric image reconstruction and 3d Tumor Localization based on a single x ray projection image for lung cancer radiotherapy
arXiv: Medical Physics, 2010Co-Authors: Xun Jia, John H Lewis, M Folkerts, Chunhua Men, Steve B JiangAbstract:Purpose: To develop an algorithm for real-time volumetric image reconstruction and 3D Tumor Localization based on a single x-ray projection image for lung cancer radiotherapy. Methods: Given a set of volumetric images of a patient at N breathing phases as the training data, we perform deformable image registration between a reference phase and the other N-1 phases, resulting in N-1 deformation vector fields (DVFs). These DVFs can be represented efficiently by a few eigenvectors and coefficients obtained from principal component analysis (PCA). By varying the PCA coefficients, we can generate new DVFs, which, when applied on the reference image, lead to new volumetric images. We then can reconstruct a volumetric image from a single projection image by optimizing the PCA coefficients such that its computed projection matches the measured one. The 3D location of the Tumor can be derived by applying the inverted DVF on its position in the reference image. Our algorithm was implemented on graphics processing units (GPUs) to achieve real-time efficiency. We generated the training data using a realistic and dynamic mathematical phantom with 10 breathing phases. The testing data were 360 cone beam projections corresponding to one gantry rotation, simulated using the same phantom with a 50% increase in breathing amplitude. Results: The average relative image intensity error of the reconstructed volumetric images is 6.9% +/- 2.4%. The average 3D Tumor Localization error is 0.8 mm +/- 0.5 mm. On an NVIDIA Tesla C1060 GPU card, the average computation time for reconstructing a volumetric image from each projection is 0.24 seconds (range: 0.17 and 0.35 seconds). Conclusions: We have shown the feasibility of reconstructing volumetric images and localizing Tumor positions in 3D in near real-time from a single x-ray image.
-
MICCAI (3) - Single-projection based volumetric image reconstruction and 3D Tumor Localization in real time for lung cancer radiotherapy
Medical image computing and computer-assisted intervention : MICCAI ... International Conference on Medical Image Computing and Computer-Assisted Inte, 2010Co-Authors: Xun Jia, John H Lewis, M Folkerts, Chunhua Men, Steve B JiangAbstract:We have developed an algorithm for real-time volumetric image reconstruction and 3D Tumor Localization based on a single x-ray projection image. We first parameterize the deformation vector fields (DVF) of lung motion by principal component analysis (PCA). Then we optimize the DVF applied to a reference image by adapting the PCA coefficients such that the simulated projection of the reconstructed image matches the measured projection. The algorithm was tested on a digital phantom as well as patient data. The average relative image reconstruction error and 3D Tumor Localization error for the phantom is 7.5% and 0.9mm, respectively. The Tumor Localization error for patient is ∼2 mm. The computation time of reconstructing one volumetric image from each projection is around 0.2 and 0.3 seconds for phantom and patient, respectively, on an NVIDIA C1060 GPU. Clinical application can potentially lead to accurate 3D Tumor tracking from a single imager.
