|Year : 2018 | Volume
| Issue : 2 | Page : 335-340
Calculation of the contrast of the calcification in digital mammography system: Gate validation
Dooman Arefan1, Alireza Talebpour2, Nasrin Ahmadinejhad3, Alireza Kamali Asl1
1 Department of Radiation Medicine Engineering, Shahid Beheshti University, Tehran, Iran
2 Department of Computer Engineering and Science, Shahid Beheshti University, Tehran, Iran
3 Department of Advanced Diagnostic and Interventional Radiology Research Center, Imam Khomeini Hospital, Tehran University of Medical Sciences, Tehran, Iran
|Date of Web Publication||8-Mar-2018|
Dr. Alireza Talebpour
Department of Electrical and Computer Engineering, Shahid Beheshti University, Evin, Tehran
Source of Support: None, Conflict of Interest: None
Purpose: Validation of the Gate tool in digital mammography image simulation from the viewpoint of image quality (contrast of calcifications).
Materials and Methods: The polymethyl methacrylate (PMMA) phantom containing aluminum foils in different thicknesses is used for measuring the contrast of calcifications in a real system. In this research, the phantom and mammography system have been simulated by the Gate tool with the maximum possible details. The contrast of the aluminum foil in simulations and practical method has been compared with each other and the standard errors in the mean (SEM) for various voltages of X-ray tube, aluminum foil, and PMMA thicknesses have been reported.
Results: Based on the obtained results, by increasing the X-ray tube voltage from 20 to 39 kVp, the image contrast has been decreased in both simulation and practical methods. The minimum and maximum average SEM of the contrast of the aluminum foils among various voltages between two simulations and practical methods for different PMMA thicknesses of 2, 4, and 6 cm have been reported as 0.0105 and 0.0117, 0.0049 and 0.0154, and 0.0037 and 0.0072, respectively.
Discussion: According to the SEM rate reported in this research for calculating the contrast of the aluminum foils in the mammography system based on simulation and practical methods, the capability of the Gate tool for simulating digital mammography system and the images created in it from the viewpoint of image contrast can be confirmed.
Keywords: Gate, image contrast, mammography, phantom, simulation
|How to cite this article:|
Arefan D, Talebpour A, Ahmadinejhad N, Asl AK. Calculation of the contrast of the calcification in digital mammography system: Gate validation. J Can Res Ther 2018;14:335-40
|How to cite this URL:|
Arefan D, Talebpour A, Ahmadinejhad N, Asl AK. Calculation of the contrast of the calcification in digital mammography system: Gate validation. J Can Res Ther [serial online] 2018 [cited 2019 Nov 22];14:335-40. Available from: http://www.cancerjournal.net/text.asp?2018/14/2/335/168967
| > Introduction|| |
Today, lots of studies have been performing for development of digital mammography systems and due to the limited number of inapplicability of the real samples and medical imaging systems, various existing simulating codes and tools are strongly required and every tool and code has its own advantages and disadvantages. Among various tools of medical imaging system, Geant4 (Gate) is one of the most powerful tools for simulating the interactions between particles and materials on Monte Carlo Method. Lots of research were done in simulation of mammography systems using various simulation tools., In the research conducted by Bonifácio et al the X-ray spectrum resulted from mammography system was analyzed for some types of filters, targets and tube voltages using Geant4 toolkit and the similar researches were done by Ng et al and Ay et al,, using ITS codes and MCNP4C codes, respectively. In the study conducted by Grabski et al, the scattered photons in low energies (10–40 kV) were evaluated by Geant4 Code in comparison with the primary photons and the possibility of using this code for simulating a mammography system was confirmed from the viewpoint of primary and scattered photons. In the research conducted by Saito et al, the digital mammography image with low dose was simulated and compared with the real image using MTF of digital mammography system. In this research, the noise resulted from three various noise sources was calculated and was added to the image. In the study conducted by Warren et al, the accuracy of MoCa software in simulating calcifications was analyzed. In this research, the amount of sharpness and contrast of calcifications was compared in real and simulated images and the dependency of optimized digital mammography system, in contrast, was analyzed based on the mass density and size. In the study conducted by Hassan et al, mammography system was simulated on Monte Carlo Method for detecting scattered beam and diagnosing cancer tissue. In the study conducted by Myronakis et al, the amount of normal mean glandular dose in compressed phantom of the breast was calculated by Gate tool and its results were compared with the real results for validating Gate tool in dose calculation and simulation. As mentioned before, lots of research was performed in the field of mammography system simulation based on Monte Carlo Method for various purposes, but there are a few researches in Monte Carlo tool feasibility and possibility of applying this tool in digital mammography system simulation, especially in the feasibility of this tool from the viewpoint of the quality of the created image. Therefore, in this research it has tried to evaluate the accuracy of the Gate tool in digital mammography system simulation and its images from the viewpoint of creating image quality (contrast of calcifications). For tools evaluation, the practical results must be compared with the result of simulation and the error rate must be reported. In this research, Hologic digital mammography system and polymethyl methacrylate (PMMA) phantom containing aluminum foils with various thicknesses was used for calculating image contrast on a practical basis. The X-ray spectrum of digital mammography system based on filter type, target type, focusing cup size, anode angel and kinetic energy of incident electron was calculated using the software has been produced by Institute of Physics and Engineering in Medicine (IPEM), and they were used in system simulation. In part 2, the methods for measuring the required parameters were completely explained in simulation and practical method. In Part 3, the results of simulation and practical measurement was reported and compared with each other and in Part 4 and 5, the feasibility of this tool in simulating digital mammography images from the viewpoint of contrast of calcifications has been discussed and evaluated.
| > Materials and Methods|| |
In this research for accreditation of Gate tool in digital mammography system simulation from the viewpoint of the created image, the contrast of aluminum squares of image resulted from PMMA phantom that contains aluminum foils with various thicknesses was calculated in simulation and practical measurement. Here, we surmised that the object being imaged has similar attenuation characteristics to aluminum foil. This has shown to be valid for calcification cluster.,, After comparing the results, the error rate of simulation tool was calculated and reported. All quantities taken into consideration in practical system and simulation system were calculated between 20 and 39 kVp voltages. The particulars of the digital mammography system are as follows:
- Name: Hologic-Selenia
- Target type: Tungsten (W)
- Target angel: 22°
- Source-image distance: 66 cm
- Electronic beam focusing cup: 0.3 cm
- X-ray window filter (beryllium): 0.8 mm
- External filter (rhodium): 0.05 mm
- Compressor (lexan): 3 mm.
According to digital mammography X-ray tube properties, IPEM software was used for creating an X-ray spectrum in various voltages. A sample of a normalized spectrum is shown in [Figure 1].
|Figure 1: Normalized spectrum created in 30 kVp voltage and anode tungsten and two types of filters, e.g., beryllium (Be) with thickness of 0.8 mm and rhodium (Rh) with thickness of 0.05 mm|
Click here to view
The changes in anode angel in digital mammography system are normally between 9 and 23 degrees. Anode angel used in this research is 22°. The placement geometry of anode to electron beam is shown in [Figure 2].
The amount of focusing cup in various setups of the device differs from 0.1 to 03 mm. In this research, the amount of focusing cup is considered 0.3 mm and the contrast estimation method has been proposed by Edelstein et al, was used for calculating the requested object contrast in the image.
In Equation (1), Mo is the average intensity of pixels in the region of interests (ROI) (requested object), Mb is the average intensity of the pixels out of ROI and next to the requested object (background). All image analyses in this research are performed using MATLAB software for calculating image contrast resulted from practical system and simulation system.
Image contrast calculation in practical method
For calculating the contrast of the calcifications in digital mammography system on a practical basis, the contrast phantom that is designed in this research with the following particulars was used:
- Phantom dimension: 28 cm × 30 cm
- Phantom thicknesses: 2 cm, 4 cm, and 6 cm
- Phantom components: PMMA, aluminum
- Aluminum foil dimension: 33 mm × 33 mm
- Aluminum foil thickness: 0.106 mm, 0.159 mm, and 0.212 mm.
For minimizing the anode heel effect and making beam intensity, uniform in various thicknesses of aluminum foils, the aluminum foils were located at a distance of 4 cm of phantom edge (external edge of breast support) and in one direction. A schema of aluminum foils location in phantom is shown in [Figure 3].
|Figure 3: Schema of aluminum foils location in polymethyl methacrylate phantom|
Click here to view
By locating the requested phantom in digital mammography system with the said particulars and locating compression pedal on phantom surface, the images were produced in 20–39 kVp voltages and the contrast of aluminum squares was calculated in each thickness of aluminum foils and the requested voltages based on Equation (1) and by MATLAB tool.
Image contrast calculation in simulation
In this research, Gate tool was used for simulating digital mammography system with the maximum possible details and creating images from the required phantom. For creating images with the maximum particulars similar to practical system, it is tried that the photon flux emitted from the source in simulation system becomes equal to the photon flux emitted from the source in real system. The photon flux of the considered X-ray tube in this research for various voltages were listed in [Table 1]. For achieving this purpose, it is necessary to create a photon flux about 105 (photon/[mAs.mm2 at 750 mm]) in the source of the simulation system, in such a manner that the creation of such a photon flux in Gate tool requires high processing. For this purpose, in this research, HP system with the following particulars is used for system simulation:
|Table 1: Source photon flux (photons/(mAs.mm2 at 750 mm)) in various voltages|
Click here to view
- Number of CPUs: 24
- Thread(s) per core: 2
- Core(s) per socket: 6
- Socket(s): 2
- NUMA node(s): 2
- CPU: Intel(R) Xeon(R) CPU E5649 at 2.53 GHz, 1600 MHz.
- Memory total: 107255112 kB
- Buffers: 121224 kB
- Cached: 3838236 kB.
For minimizing the processing time, the Gate code created for system simulation was run on a parallel basis on all CPUs by written special C++ code in this research. Digital phantom with the mentioned particulars was located at a distance of 64 cm from the source that is perpendicular to the central axis of photon beam, and compression pedal was placed at the top of the digital phantom. Designed detector that is sensitive to photons of energy range 0–50 KeV was placed at a distance of 66 cm of source and perpendicular to the central axis of photon beam.
| > Results|| |
In practical and simulation method, 60 images in different aluminum and PMMA thicknesses and voltages were generated using amorphous selenium direct digital system and Gate tool. The contrast of aluminum foils were calculated for every PMMA thickness. In order to calculate image contrast according to Equation (1), square ROI from the objects and backgrounds were extracted from phantom images. After analyzing the images resulted from practical and simulated system by MATLAB tool, the contrast of aluminum foils was reported in every voltage and PMMA thickness as shown in [Figure 4], [Figure 5], [Figure 6]. The dashed lines for the contrast of the simulated images are attached with continued lines to make it easier to distinguish these from the points depicting the contrast of the real images of the aluminum foils.
|Figure 4: Local image contrast in various voltages and thicknesses of aluminum foils in simulation and practical systems (polymethyl methacrylate thickness of 2 cm)|
Click here to view
|Figure 5: Local image contrast in various voltages and thicknesses of aluminum foils in simulation and practical systems (polymethyl methacrylate thickness of 4 cm)|
Click here to view
|Figure 6: Local image contrast in various voltages and thicknesses of aluminum foils in simulation and practical systems (polymethyl methacrylate thickness of 6 cm)|
Click here to view
As shown in [Figure 4] (2 cm PMMA), the minimum and maximum of standard errors in the mean (SEM) were equal to 0.0105 and 0.011745, respectively, about [Figure 5] (4 cm PMMA), they were 0.0049 and 0.015455, and about [Figure 6] (6 cm PMMA), they were equal to 0.0037 and 0.0072. From the results, it would appear that the SEM about PMMA thickness of 2 cm was more than others. The cause of this is unknown; however, this is not of concern since the majority of the breast thicknesses are more than 2 cm.
In order to compare the results of two practical and simulation methods visually, the fitted line of contrast versus the voltage changes in every determined thicknesses based on the results was estimated by the regression method in MATLAB software and the results are shown in [Figure 7], [Figure 8], [Figure 9].
|Figure 7: Fitted line of image contrast based on the voltage changes in aluminum foil thicknesses of 106, 159, and 212 μm in two practical and simulation methods and polymethyl methacrylate thickness of 2 cm|
Click here to view
|Figure 8: Fitted line of image contrast based on the voltage changes in aluminum foil thicknesses of 106, 159, and 212 μm in two practical and simulation methods and polymethyl methacrylate thickness of 4 cm|
Click here to view
|Figure 9: Fitted line of image contrast based on the voltage changes in aluminum foil thicknesses of 106, 159, and 212 μm in two practical and simulation methods and polymethyl methacrylate thickness of 6 cm|
Click here to view
| > Discussion|| |
The purpose of this research was the feasibility of Gate simulation tool application in digital mammography system simulation, and especially for creating images of the required object (validation of Gate tool in digital mammography simulation as the contrast of the calcifications). For this purpose, the designed phantom was imaged by two practical and simulation methods in various aluminum foils and PMMA thicknesses and various voltages. The contrast of aluminum foils (calcifications) in every object was separately calculated by MATLAB software, and the results were compared with each other. According to the results, there is good agreement between the contrast of calcifications in real and Gate simulated images. In the study conducted by Warren et al, determined whether the use of MoCa results in simulated calcifications with the correct contrast and sharpness. The general manner discovered by Warren et al, was in agreement with this work. For example, as the tube voltage increased the contrast (aluminum equivalent attenuation coefficient) decreased. As a mentioned in the study conducted by Warren et al, the images of the calcification clusters were extracted using a method described, from images of sliced mastectomy samples acquired at ×5 magnification on a digital specimen cabinet. But in our work, for simulating acquired images, we do not need any specimen cabinet and sliced mastectomy samples.
We compared the image contrast of test objects obtained on a clinical system to the contrast of simulated images of the same test objects (acquired under the same imaging conditions). It was found that the contrast of real and simulated images was equal within the experimental errors.
| > Conclusion|| |
In this research, for digital mammography system scrutinized, it was found that the contrast of calcifications did not disagree significantly between the images obtained with the digital mammography system and the images generated by Gate tool and the ability of the Gate tool in simulating digital mammography system and the images created in it can be confirmed from the viewpoint of contrast of calcifications.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| > References|| |
Bonifácio DA, Murata HM, Moralles M. Monte Carlo Simulation of X-Ray Spectra in Diagnostic Radiology and Mammography Using Geant4; 2005.
Gordon R, Rangayyan RM. Feature enhancement of film mammograms using fixed and adaptive neighborhoods: Correction. Appl Opt 1984;23:2055.
Hess R, Neitzel U. Optimizing image quality and dose for digital radiography of distal pediatric extremities using the contrast-to-noise ratio. Rofo 2012;184:643-9.
Ng KP, Kwok CS, Tang FH. Monte carlo simulation of x-ray spectra in mammography. Phys Med Biol 2000;45:1309-18.
Ay MR, Shahriari M, Sarkar S, Adib M, Zaidi H. Monte carlo simulation of x-ray spectra in diagnostic radiology and mammography using MCNP4C. Phys Med Biol 2004;49:4897-917.
Grabski V, Brandan ME, Ruiz-Trejo C, Villasenor Y. SU-FF-I-34: PSF and S/P in mammography: A validation of simulations using the geant4 code. Med Phys 2005;32:1911.
Saito Y, Sakai M, Fujita N, Kodera Y, editors. Reduction of Patient Dose in Digital Mammography: Simulation of Low-Dose Image Using Computed Radiography System and Flat Panel Detector System. SPIE Medical Imaging; International Society for Optics and Photonics; 2013.
Warren LM, Green FH, Shrestha L, Mackenzie A, Dance DR, Young KC. Validation of simulation of calcifications for observer studies in digital mammography. Phys Med Biol 2013;58:N217-28.
Hassan L, Mac Donald C. SU-EI-64: X-ray coherent scatter mammography simulation. Med Phys 2014;41:145.
Myronakis ME, Zvelebil M, Darambara DG. Normalized mean glandular dose computation from mammography using GATE: A validation study. Phys Med Biol 2013;58:2247-65.
Poludniowski G, Landry G, DeBlois F, Evans PM, Verhaegen F. SpekCalc: A program to calculate photon spectra from tungsten anode x-ray tubes. Phys Med Biol 2009;54:N433-8.
Zanca F, Van Ongeval C, Marshall N, Meylaers T, Michielsen K, Marchal G, et al
. The relationship between the attenuation properties of breast microcalcifications and aluminum. Phys Med Biol 2010;55:1057-68.
Warren LM, Mackenzie A, Cooke J, Given-Wilson RM, Wallis MG, Chakraborty DP, et al
. Effect of image quality on calcification detection in digital mammography. Med Phys 2012;39:3202-13.
Carton AK, Bosmans H, Vandenbroucke D, Souverijns G, Van Ongeval C, Dragusin O,et al
. Quantification of Al-equivalent thickness of just visible microcalcifications in full field digital mammograms. Med Phys 2004;31:2165-76.
Edelstein WA, Bottomley PA, Hart HR, Smith LS. Signal, noise, and contrast in nuclear magnetic resonance (NMR) imaging. J Comput Assist Tomogr 1983;7:391-401.
Warren LM, Mackenzie A, Dance DR, Young KC. Comparison of the x-ray attenuation properties of breast calcifications, aluminium, hydroxyapatite and calcium oxalate. Phys Med Biol 2013;58:N103-13.
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6], [Figure 7], [Figure 8], [Figure 9]