Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
A Finite State Model for Respiratory Motion Analysis in Image Guided Radiation Therapy Huanmei Wu (NU) Gregory Sharp (MGH) Betty Salzberg (NU) David Kaeli (NU) Hiroki Shirato (Japan) Steve B Jiang (MGH) This work is part of CenSSIS, the Center for Subsurface Sensing and Imaging Systems, under the Engineering Research Centers Program of the National Science Foundation (Award # EEC-9986821). REOE 3. Technical Details we have designed and implemented a finite state model that characterizes a breathing cycle, as containing three breathing states: exhale (EX), end-of-exhale (EOE) and inhale (IN). These three states repeat in succession as they do in natural breathing. Any motion that does not fall into any of the above three state will be treated as irregular breathing, a fourth state (IRR). The state transition is guided a finite state automaton as shown in Figure 1. IN EX IN EX IN EX IN EX IN EX IN EX IN AVG - 10 0 0 1 2 3 4 EOE 5 6 7 8 9 EOE EOE EOE EOE EOE -5 EX IN EX IRR IRR IN EX IN EX IN EX IN EX IN -15 Figure 2 Spatio-temporal tumor movements -25 0 2 3 5 7 8 10 12 13 15 17 18 20 22 23 25 27 3 4. Results This model was used to describe the respiratory motion for 23 patients with peak-to-peak motion greater than 7mm. The average RMS error over all patients was less than 1mm and no patient has an error worse than 1.5mm, which is shown in the following figure. EX EOE IN 0.5 Amplitude changes 23 22 21 20 19 18 17 16 15 14 13 12 11 10 9 8 7 6 5 4 3 2 1 0 Frequency changes -3 -9 Raw data -15 predicted value 100ms -21 1 91 181 271 361 451 541 631 721 811 901 991 1081 1171 1261 Time (1/30 sec) Figure 5 Prediction results 4. Future Goals 1.5 1 It provides an easy way for motion prediction using subsequence similarity matching using stream database technology. Figure 5 shows the simple prediction results based on the line segments of the finite state model. New methods to improve the prediction accuracy are under development based on sophisticated statistical analysis of retrieved subsequences 28 Figure 2 Modeling results using the finite state model Regular Figure 4 Frequency patterns 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 SI motion (mm) - 15 23 EX 22 IN 21 EX 20 IN 19 EX 1 18 -5 2 17 0 IN 16 EOE 15 EOE 14 EOE 13 EOE EOE 12 EOE EX 11 EOE 10 EOE 3 7 EOE Shorter EOE state Longer EOE state 6 EOE Our future research includes several areas. One area is modifying the current model to including non-linear states, and more or less states per breathing period. We would like to address cardiac motion in the context of tumor respiratory motion. Another future work is to extend the 1D implementation to 3D implementation. Finally, we are working tumor motion prediction based on the new model we proposed. Strategic Research Plan Overview of the Strategic Research Plan L3 EnviroS4 Civil S5 Bio-Med S2 S3 S1 L2 Validating TestBEDs L1 Fundamental Science R1 Impacting System Level Projects R2 R3 Figure 6 Impact on CenSSIS References Base line shifts Noise Combinations Each of these states corresponds to a natural action: EX is the motion due to lung deflation, EOE is the motion for rest after lung deflation, and IN is the motion due to lung expansion. An example of segmentation into these three states is illustrated in Figure 2. CenSSIS Research and Industrial Collaboration Conference Nov. 2004 4 9 RIN 5 REX 8 IN 4 RIRR RIRR It provides a convenient tool to quantify respiratory motion characteristics, such as patterns of frequency changes and amplitude changes. The following figure shows the frequency patterns based on the finite state model: RIRR RIN 3 5 AVG The goal of our work is to design a quantitative method for describing breathing motion that captures all of the natural characteristics, and can be used for both off-line analysis and on-line monitoring. Our model can be applied to internal or external motion, including internal tumor position, abdominal surface, diaphragm, spirometry, and other surrogates IRR Figure 1 Finite state model for respiratory motion analysis SI Motion (mm) Motion prediction is required by beam tracking and respiratory gating due to system latency and imaging rate REX REX The conformality of delivered dose to thoracic and abdominal lesions is degraded by respiratory motion Most motion compensation methods require an adequate understanding of the motion characteristics RIN RIRR EX SI Motion (mm) Respiratory motion analysis of a moving target is an integral part of effective image guided radiation treatment. IN EX The three state motion model gives convenient ways to summarize and analyze motion characteristics, including offline data analysis and online prediction. 2 REOE 5 2. Challenges and Significance EOE 1 EOE SI mot i on ( mm) Effective image guided radiation treatment of a moving tumor requires adequate information of respiratory motion characteristics. For margin expansion, beam tracking, and respiratory gating, the tumor motion must be quantified for pretreatment planning and monitored on-line. We propose a finite state model for respiratory motion analysis that captures our natural understanding of breathing stages. In this model, a regular breathing cycle is represented by three line segments, exhale, end-of-exhale, and inhale, while abnormal breathing is represented by an irregular breathing state. In addition, we describe an on-line implementation this model in one dimension. We found this model can accurately characterize a wide variety of patient breathing patterns. Our model provides a convenient tool to quantify respiratory motion characteristics, such as patterns of frequency changes and amplitude changes, and can be applied to internal or external motion, including internal tumor position, abdominal surface, diaphragm, spirometry, and other surrogates. 5. Applications SI Motion (mm) I. Abstract Figure 3 Fidelity of our model to raw data 4. Bortfeld T et al 2002 “Effects of intra-fraction motion on IMRT dose delivery: statistical analysis and simulation”, Phys. Med. Biol. 47(13) 2203-20 5. Jiang S B et al 2003 “An experimental investigation on intra-fractional organ motion effects in lung IMRT treatments”, Phys. Med. Biol. 48(12) 1773-84 6. Shirato H et al 2000 “Four-dimensional treatment planning and fluoroscopic real-time tumor tracking radiotherapy”, Int. J. radiat. Oncol. Biol. Phys. 48 1187-95 7. Huanmei Wu et al “Subsequence similarity matching for tumor respiratory motion analysis”, submitted 8. Huanmei Wu et al “A finite state model for respiratory motion analysis in image guided radiation therapy”, submitted to Phys. Med. Biol. Contact: Huanmei Wu, Egan225, Northeastern University, (617) 373 7349, [email protected]