- Biomedical Optics & Medical Imaging
- Defense & Security
- Electronic Imaging & Signal Processing
- Illumination & Displays
- Lasers & Sources
- Micro/Nano Lithography
- Optical Design & Engineering
- Optoelectronics & Communications
- Remote Sensing
- Sensing & Measurement
- Solar & Alternative Energy
- Sign up for Newsroom E-Alerts
- Information for:
Biomedical Optics & Medical Imaging
Dynamic-data-driven systems aid patient-specific cancer therapy
Feedback control systems combine models and experimental data to guide thermal treatments of tumors.
6 October 2008, SPIE Newsroom. DOI: 10.1117/2.1200810.1220
Recent developments in computational science provide a powerful methodology for planning and optimizing thermal therapy delivery in cancer treatments.1–3 The use of feedback control systems that integrate computer models with field measurement systems can greatly increase the fidelity and predictive power of computer simulations by allowing real-time calibration of the models. Such systems are collectively called dynamic data driven application systems (DDDAS), and can theoretically deliver predictive simulations of unprecedented accuracy and reliability.4
Figure 1. The communication architecture used to control a surgical procedure between two sites separated by 150mi.. The computational arena is managed by the Institute for Computational Engineering and Sciences. Computational models running on high performance computing systems at the Texas Advanced Computing Center in Austin interact continually with thermal imaging data acquired at M.D. Anderson Cancer Center in Houston to provide feedback and control. Treatment visualizations are rendered in Austin and forwarded to Houston.
A canonical DDDAS, designed to predict and control the outcome of laser therapies for prostate cancer, is being studied at The University of Texas at Austin (UTA). As this project focuses on the delivery of minimally invasive thermal therapies designed to provide local treatment of focal disease, an extensive list of technologies and new methodologies are inherent to the system. These include magnetic resonance thermal imaging (MRTI), computer visualization, laser optics, high-speed networks, nonlinear dynamic bioheat transfer models of heterogeneous tissue, adaptive meshing, high-performance parallel computing, cell damage and heat-shock protein models, inverse analysis, calibration, model validation, signal processing, optimal control algorithms, and error estimation and control.
The fundamental hypothesis underlying the creation of this system is that cancer cells can be eradicated when subjected to sufficient thermal doses. Such heat sources can be supplied through interstitial laser fibers inserted into the infected tissue. The ability to predict the evolution of the temperature field generated by such a heat source can be exploited to optimize the location of the fiber and the power of the laser to minimize damage to healthy tissue while maximizing damage to tumor cells. The development of calibrated models that accurately simulate bioheat transfer in heterogeneous tissue is thus a key component of the system.5–7
Computer prediction can be used to accurately control the affected tissue response through a collection of imaging based measurements about how the complex physiological system is responding to the surgery and to make changes in the treatment plan based on an intelligent understanding of the physiological pathways and feedback mechanisms. An overview of the control system is available online (see video 8).
The core simulation tool embeds experimental real-time thermal imaging data within a nonlinear Pennes bioheat transfer model.9 The thermal imaging data is provided by an MRTI device that follows the actual therapy at the imaging laboratory of the M.D. Anderson Cancer Center (MDACC) in Houston, Texas. The MRTI technology developed over the past decade provides the 3D temperature field in the living tissue as a function of position and time,10 and the measured temperatures are used to calibrate the bioheat transfer model on a patient-specific basis. The optimal control system is guided by simulations performed at the computational modeling arena within the Institute for Computational Engineering and Sciences (ICES) in Austin, Texas, with high bandwidth network connections to computers at the Texas Advanced Computing Center.
The control system, in its entirety, has been tested on in vivo canine prostate tissue. Handling of animals was in accordance with an Institutional Animal Care and Use Committee approved protocol. The goal of this treatment was to demonstrate that the computational model could control the bioheat transfer and heat a region of 1.2cm in diameter to 60°C. The entire duration of the treatment was 18min.
Figure 2. Image of the control system applied to in vivo laser treatment of canine prostate cancer. As a result of the real-time response imaging, the laser tip was moved 5mm from its initial position (upper left window). The color scale shown at the left is from 35–65°C.
The treatment protocol is divided into four stages. During the first stage, the biological domain is pulse heated and MRTI thermal image data is acquired for the heating as well as the cooling. The second stage of the treatment accounts for the time span of the calculations that use the imaging data for model calibration. The third stage accounts for the time delay to re-compute the patient-specific optimal heating protocol. In the fourth and final stage, the optimal laser control parameters are applied to the biological domain. For this particular trial, the perfusion, thermal conductivity, absorption coefficient, and laser position were calibrated. Results of the real-time calibration computations moved the laser tip 5mm from the initial estimated position (See Figure 2). Cutlines illustrating the temperature as a function of distance are taken through the thermal imaging data and finite element predictions for comparison. The cutlines show good agreement to within 5°C at the peak between the desired treatment plan, the computational prediction, and the experimentally measured MRTI temperature field. The temperature measured in the thermal images at MDACC was controlled by supercomputers at UTA located over 150mi. away.
Current results produced by the DDDAS support the proposition that robust simulation methodologies, based on calibrated nonlinear models of bioheat transfer in heterogeneous tissue, can be designed to interact with clinical thermal imaging modalities to provide real-time control of laser therapy for prostate cancer. Given the increasing ability of diagnostic radiology to detect diseases earlier and earlier, the impact of minimally invasive approaches to surgery such as thermal therapy will play a large role in safely and efficiently treating cancer with minimal impact on the patient. This work demonstrates that the interaction of computer modeling and simulation with medical technologies can dramatically improve cancer therapies and enhance and prolong the life of cancer patients.
J. Tinsley Oden, David Fuentes, Jon Bass
University of Texas at Austin
Dept. of Mechanical Engineering
University of Texas at San Antonio
San Antonio, TX
1. D. Fuentes, J. T. Oden, K. R. Diller, J. Hazle, A. Elliott, A. Shetty, R. J. Stafford, Computational modeling and real-time control of patient-specific laser treatment cancer, Ann. BME., Submitted for publication, 2008.
2. K. R. Diller, J. T. Oden, C. Bajaj, J. C. Browne, J. Hazle, I. Babuška, J. Bass, L. Bidaut, L. Demkowicz, A. Elliott, Y. Feng, D. Fuentes, S. Goswami, A. Hawkins, S. Khoshnevis, B. Kwon, S. Prudhomme, R. J. Stafford, Advances in Numerical Heat Transfer, Taylor & Francis Group, 2008. vol. 3: Numerical implementation of bioheat models and equations, ch. 9: Computational infrastructure for the real-time patient-specific treatment of cancer
3. J. T. Oden, K. R. Diller, C. Bajaj, J. C. Browne, J. Hazle, I. Babuška, J. Bass, L. Demkowicz, Y. Feng, D. Fuentes, S. Prudhomme, M. N. Rylander, R. J. Stafford, Y. Zhang, Dynamic data-driven finite element models for laser treatment of prostate cancer, Num. Meth. PDE, 23, no. 4, pp. 904-922, 2007.
6. Y. Feng, D. Fuentes, A. Hawkins, J. Bass, M. N. Rylander, A. Elliott, A. Shetty, R. J. Stafford, J. T. Oden, Nanoshell-mediated laser surgery simulation for prostate cancer treatment, Engineering with Computers, Accepted for publication, 2007.
10. R. J. Stafford, R. E. Price, C. J. Diederich, M. Kangasniemi, L. E. Olsson, J. D. Hazle, Interleaved echo-planar imaging for fast multiplanar magnetic resonance temperature imaging of ultrasound thermal ablation therapy, J. Magnetic Res. Imag. 20, no. 4, pp. 706-714, 2004.