Helical Tomotherapy Qualitative dose Delivery Verification

Transcription

1 University of Tennessee, Knoxville Trace: Tennessee Research and Creative Exchange University of Tennessee Honors Thesis Projects University of Tennessee Honors Program Helical Tomotherapy Qualitative dose Delivery Verification Bart Daniel Lynch University of Tennessee - Knoxville Follow this and additional works at: Recommended Citation Lynch, Bart Daniel, "Helical Tomotherapy Qualitative dose Delivery Verification" (2003). University of Tennessee Honors Thesis Projects. This is brought to you for free and open access by the University of Tennessee Honors Program at Trace: Tennessee Research and Creative Exchange. It has been accepted for inclusion in University of Tennessee Honors Thesis Projects by an authorized administrator of Trace: Tennessee Research and Creative Exchange. For more information, please contact trace@utk.edu.

2 UNIVERSITY HONORS PROGRAM SENIOR PROJECT - APPROVAL Name: Bart D. Lynch College: College of Engineering Department: Nuclear Engineering Faculty Mentor: Laurence Miller PROJECT TITLE: Helical Tomotherapy Qualitative Dose Delivery Verification I have reviewed this completed senior honors thesis with this student and certify that it is a project commensurate with honors level undergraduate research in this field. Signed: ~1J'~,~?J?t,ttl,Faculty Mentor l/. v Date: /.-77lt~!;~;Z C{):2 Comments (Optional):

3 UNIVERSITY HONORS PROGRAM SENIOR PROJECT - PROSPECTUS Name: Bart D. Lynch College: College of Engineering Department: Nuclear Engineering Faculty Mentor: Laurence Miller PROJECT TITLE: Helical Tomotherapy Qualitative Dose Delivery Verification PROJECT DESCRIPTION (Attach not more than one additional page, if necessary): Imaging has been used in radiation therapy to help minimize dose to nonnal tissue while treating a tumor volume. Radiographic film planes are successfully used to measure fluence distributions for IMRT and static beam therapies. However, the advent ofhelical tomotherapy treatments necessitates the institution ofsome sort of cylindrical phantom for comparing fluence distributions. An Archimedean spiral machined into a block ofsolid water can accept film, but some sort ofverification needs to take place between a film study and the output from the treatment-planning computer. Matlab was used to write a program capable of extracting a spiral from a 3D matrix of data. The code was verified by first inserting an image, and then extracting it using a second program. The two images were compared visually and by an image difference and showed good agreement. Projected completion date: May 8, 2003 Signed: I have discussed this research proposal with this student and agree to serve in an advisory role, as faculty mentor, and to certify the acceptability of the completed project. ~ ~J '()(P Signed: ~4~,;z /??t<~~l, Faculty Mentor

4 Helical Tomotherapy Qualitative Dose Delivery Verification Bart Lynch Mentors Dr. Stephen Mahan Dr. Lawrence Miller Submitted to: University of Tennessee Honors Program May 8, 2003

5 ACKNOWLEDGEMENTS I would like to thank Dr. Stephen Mahan ofthe Thompson Cancer Survival Clinic for his help and patience as I completed this project. Furthermore, I would like to thank Dr. Laurence Miller of the UT Department ofnuclear Engineering for serving as my faculty mentor. ABSTRACT Imaging has been used in radiation therapy to help minimize dose to normal tissue while treating a tumor volume. Radiographic film planes are successfully used to measure fluence distributions for IMRT and static beam therapies. However, the advent ofhelical tomotherapy treatments necessitates the institution ofsome sort of cylindrical phantom for comparing fluence distributions. An Archimedean spiral machined into a block ofsolid water can accept film, but some sort of verification needs to take place between a film study and the output from the treatment-planning computer. Matlab was used to write a program capable of extracting a spiral from a 3D matrix of data. The code was verified by first inserting an image, and then extracting it using a second program. The two images were compared visually and by an image difference and showed good agreement.

6 I INTRODUCTION The use of images in radiation therapy is an accepted practice in cancer treatment planning and dose delivery verification. Image guided therapy helps to improve the accuracy ofradiotherapy procedures, in which the ultimate goal is to treat a tumor volume while sparing normal tissue as much as possible. Physicists today use planes of radiographic film to verify fluence distributions of particles (photons or electrons) as they pass through a medium (typically solid water). While these are acceptable for traditional radiotherapy techniques including static beam therapy and intensity modulated radiotherapy, helical tomotherapy presents a unique challenge because it does not involve static radiation beams. In a typical tomotherapy treatment, one would expect the machine gantry (and thus the radiation beam itself) to rotate completely around the patient or target. Researchers at the University of Wisconsin have developed a means by which film can be inserted into a cylindrical solid water phantom to measure the energy fluence associated with a tomotherapy treatment. In implementation of this quality assurance procedure, one must still compare the film with something, typically output from the treatment-planning computer. The goal of this project was to write a Matlab program which was capable of extracting a particular plane of data from a three dimensional matrix of data. Verification of the program was accomplished by inserting an image into a data matrix and then extracting it using the same conditions used to insert it. Both visual and image difference analysis were performed to verify success of the program.

7 II MATERIALS AND METHODS Archimedean Spiral An Archimedean spiral is defined by a point moving with constant velocity v along a rod that rotates with a constant angular velocity mabout one ofits end points.. Mathematically, this can be written in polar coordinates as: p=ale>1 where p == the radius ofthe spiral E> == angle subtended by the spiral a = vim The time dependence of a is eliminated because both the point and the rod begin and end moving at the same time. Thus, a is the ratio between the rod length r and the total angle subtended by the spiral. The radius ofthe spiral is directly proportional to the angle subtended by the spiral; as the angle ofthe spiral increases, so does the radius. In tenus of rectangular coordinates, the spiral can be parameterized as: x(e» = a E> cos(e» y(e» = a E> sin(e» The spiral is shown in Figure 1.

8 Fig. 1. An Archimedean spiral generated by Matlab is shown. The image was generated as part ofthe code validity / verification process. Cylindrical Sold Water Phantom Tomotherapy Inc. provided Thompson Cancer Survival Center with a cylindrical solid water phantom that had been machined to create an Archimedean spiral cavity for placement of radiographic film. The arc length ofthe cavity is approximately 42 cm while the total angle subtended by the spiral is approximately radians. The effective spiral radius is 7 cm. Matlab Code Matlab was used to extract the curve from a rectangular three-dimensional matrix of data, based on the number ofimage slices, slice width and height, and the particular geometry ofthe spiral. The code is found in Appendix A. Because both the start and stop points and angles of the spiral were known (from the machined cylinder itself), the program calculated an (x,y) ordered pair at equal distances along the spiral from a given value of Qand a. A two-dimensional grayscale image was used to qualitatively test the ability ofthe code to remove the correct information from the cube. A grayscale image

9 was chosen because images from the patient planning system are also grayscale. However, one could associate a color map with the image if one so chose. The program converts each image into a representative matrix, with each image pixel having a grayscale value between 0 and 255. The image data was inserted into a 3D zero matrix by the program, while a second program extracted the data using the same conditions used to insert the image. The extracted data was then converted back into an image for visual analysis. III RESULTS Fig. 2. The image on the left is the original grayscale image while the image on the right is the extracted data. Cause of image negation is unknown. The figures above show the image both before insertion and after extraction from the data matrix. Any size discrepancy that exists between the above images is a result of formatting for the paper and is not reflective of an actual size difference between the two

10 tmages. It is uncertain why the program converted the extracted data into a negative image of the original. Furthermore, blocking in the horse's forelegs and underbelly are probably attributed to a grayscale saturation limit that the original image exceeds. An image difference was calculated, and the results are shown in the figure below (Fig 3). Fig. 3. Difference image computed between the original and extracted images. IV DISCUSSION The extracted image shows good visual agreement with the original image. While the image has been converted to a negative, further analysis showed that it seems to be a result ofthe image writing function and not a property ofthe extracted image data themselves. Furthermore, the difference image only shows disagreement in areas of image saturation.

11 V CONCLUSIONS Matlab can be used successfully in extracting a particular plane ofdata from a three dimensional matrix ofdata. Because tomotherapy utilizes a helical treatment beam, it is necessary to compare a helical plane ofdata with helical information from the treatment planning computers. Thus, initial steps involved in the process would be to verify that programming software can be used to accurately extract information in a particular way. Ultimately, a user would input image slices from a treatment planning system, and the code would generate a 2D spiral from the 3D matrix data. The 2D image could then be compared with a film study to verify the accuracy of the planning machine.

12 REFERENCES Paliwal B, Tome W, Richardson S, Mackie T. (2000). A spiral phantom for IMRT and tomotherapy treatment delivery verification. Med Phys 27(11), Mackie T.R. et al. (2002). Image guidance for precise conformal radiotherapy. International Journal ofradiation Oncology*Biology*Physics 56(1), Chang J, Magera G, Chui C, Ling C, Lutz W. (1999) Relative profile and dose verification of intensity-modulated radiation therapy. International Journal of Radiation Oncology*Biology*Physics 47(1),

13 APPENDIX A Matlab Code %% SPlRALBUILDER.M clear all; %%% Curve Creation %%% Define variables r 7; em min ; % angle at which begins (pi/2) phi max 12.26; % angle subtended by image_length = 450; % s width = 315; pixels a = r/(phi_max/(2 ) ) ; % ratio btwn v and omega % Generate block, number string blockl = zeros(256,256,image width); spirall = imread('c:\my Documents\Bart\Honors project\horse2.jpg', 'jpg'); phi = phi min: ((phi_max-phi min)/(image length-i)) : maxi - x = round(a * * cos (phi) + 256/2); x coord rounded to nearest integer y = round(a * phi * sin (phi) + 256/2) ; % y coord rounded to nearest for z = l:image_width for i 1: length block1(x(i),y(i),z) end end spiral1(i,z); imwrite(block1(:, :,1), 'C:\My Documents\Bart\Honors ect\spiralhorsel.jpg', 'jp9')i

14 %%% SPIRALEXTRACTOR.M % Curve Extraction block2 = blockl; = zeros(image_length,image_width); % Extract curve for z = l:image width for i = end end l:image_length (i,z) = block2(x(i),y(i),z); imwrite(,'c:\my Documents\Bart\Honors ect\rehorse.jpg', 'jpg');