|Year : 2017 | Volume
| Issue : 6 | Page : 1007-1014
Evaluation of dose calculation accuracy of various algorithms in lung equivalent inhomogeneity: Comparison of calculated data with Gafchromic film measured results
Teerth Raj Verma1, Nirmal K Painuly1, Surendra P Mishra2, Navin Singh1, M. L. B. Bhatt1, Naseem Jamal1, MC Pant1
1 Department of Radiotherapy, King George Medical University UP, Lucknow, Uttart Pradesh, India
2 Department of Radiotherapy, Dr. Ram Manohar Lohia Institute of Medical Sciences, Lucknow, Uttart Pradesh, India
|Date of Web Publication||13-Dec-2017|
Mr. Teerth Raj Verma
Department of Radiotherapy, King George Medical University, Lucknow - 226 003, Uttar Pradesh
Source of Support: None, Conflict of Interest: None
Aim: To evaluate dose calculation accuracy of various algorithms in lung equivalent inhomogeneity comprising tumor within it and comparison with Gafchromic film data.
Materials and Methods: Gafchromic film measured central axis absorbed dose in lung insert (−700 Hounsfield unit [HU]), in racemosa wood cylindrical inhomogeneity (−725 HU) and at three surfaces of tumor (−20 HU) created in cylindrical inhomogeneity, put in the cavity of computerized imaging reference systems (CIRS) thorax phantom were compared with convolution (CON), superposition (SP), fast SP (FSP), and X-ray voxel Monte Carlo (XVMC) algorithms calculated dose using 6 MV beams of field size 2 cm × 2 cm, 3 cm × 3 cm, 4 cm × 4 cm, 5 cm × 5 cm, and 8 cm × 8 cm.
Results: XVMC was in good agreement with film measured results for all selected field sizes except 3 cm × 3 cm. SP under estimated by 5.7% at the center of the lung insert while deviation up to 6% was found at the cent of wood inhomogeneity in 2 cm × 2 cm. Except CON, increase in dose from proximal to the central surface of the tumor and then dose falloff from central to the distal surface for field size 2 cm × 2 cm to 4 cm × 4 cm was recorded. The change in film measured percentage depth dose from 2 cm × 2 cm to 3 cm × 3 cm field sizes was found –8% however for consecutive field size(s) larger than 3 cm × 3 cm this difference was less. CON and FSP produced overestimated results.
Conclusion: Out of four algorithms, XVMC found consistent with measured data. The electronic disequilibrium within and at the interface of inhomogeneity make the accurate dose predictions difficult. These limitations results in deviations from the expected results of the treatments.
Keywords: Computerized imaging reference systems thorax phantom, lung equivalent inhomogeneity, racemosa wood, soft liner, X-ray voxel Monte Carlo
|How to cite this article:|
Verma TR, Painuly NK, Mishra SP, Singh N, Bhatt M, Jamal N, Pant M C. Evaluation of dose calculation accuracy of various algorithms in lung equivalent inhomogeneity: Comparison of calculated data with Gafchromic film measured results. J Can Res Ther 2017;13:1007-14
|How to cite this URL:|
Verma TR, Painuly NK, Mishra SP, Singh N, Bhatt M, Jamal N, Pant M C. Evaluation of dose calculation accuracy of various algorithms in lung equivalent inhomogeneity: Comparison of calculated data with Gafchromic film measured results. J Can Res Ther [serial online] 2017 [cited 2020 Jul 12];13:1007-14. Available from: http://www.cancerjournal.net/text.asp?2017/13/6/1007/168992
| > Introduction|| |
Radiation treatments are becoming increasingly conformal and the pinpoint accuracy of modern radiation treatments have made it possible to fine-tune radiation beams so that they can match the shape and position of a patient's tumor nearly anywhere in the body, however, radiation dosimetry due to the presence of inhomogeneity is still an area of concern. Intensity and scattering characteristics of the photon beam are greatly affected due to the presence of inhomogeneity apart from the loss of electronic equilibrium. Radiotherapy of thorax for such reasons has always been a matter of interest. Lung in thoracic cavity is one of the most heterogeneous organ with density variation and large movement due to inter/intra fractional motion caused by the respiratory motion, that is the potential source of error in radiotherapy. Experimental studies have shown that the presence of low-density inhomogeneous medium such as lung can produce 30% variation in dose.
The modern treatment planning system (TPS) use number of algorithms to generate treatment plans and therefore effective delivery of the dose will depend on computational efficiency of algorithm used in the planning system., The common inhomogeneity correction methods used in TPS include tissue-air ratio (TAR) method, the power law TAR (Batho) method and the equivalent TAR method. However, each one of them has certain limitation. New algorithms have been developed with the emphasis on addressing modeling of charged particle disequilibrium that occur during the deposition of radiation dose in tissue at the interface of two different density mediums.
The present study has been undertaken to evaluate the dose calculation accuracy of various algorithms namely convolution (CON), superposition (SP), fast SP (FSP), and X-ray voxel Monte Carlo (XVMC) in lung equivalent inhomogeneity as well as within the tumor created inside it and comparison of calculated data with Gafchromic film measured results.
| > Materials and Methods|| |
Computerized imaging reference systems three-dimensional phantom
The dynamic thorax phantom (CIRS model 008A; Computerized Imaging Reference Systems Inc., Norfolk, VA, USA) represents an average human thorax in shape, proportion and composition. The phantom contains two lung structures (left-right) having Hounsfield unit (HU) −700 anteriorly at a depth of 3 cm and a spine mimicking the actual human thorax in terms of clinical settings [Figure 1]a. A cylindrical hollow space of diameter 6.4 cm at 5.5 cm depth form the surface of phantom and 2.5 cm from lung structure (anteriorly) is available to accommodate different types of inserts. These inserts can use different types of dose measuring tools such as thermoluminescence detector, Gafchromic films, and MOSFET etc. In the present work, two types of inserts viz., standard lung insert made up of two hemispherical rods of same density and same material as of two lung structures and customized wood cylinder made of Ficus recemoca wood slabs have been used.
|Figure 1: (a) Computerized imaging reference systems phantom with lung insert consisting of (I) and (II) hemisphere: Dose were recorded at proximal (P), central (C) and distal (D) surfaces from Gafchromic film and algorithms, (b) (I) wood slabs with Gafchromic film for dose measurement within the wood cylinder in total eight surfaces, (II) Wood slab cylinder in the cavity of computerized imaging reference systems Thorax phantom; (c) computerized imaging reference systems phantom with wood slab cylinder containing same tumor at different depths (anteriorly) of 6.5 cm, 7.5 cm, and 8.5 cm displaying proximal (P), central (C), and distal (D) surface|
Click here to view
Customized wood slabs insert
Ficus recemoca wood slabs (Goolar, Indian popular name) were used to mimic the inhomogeneity as found in real human lung. A wood slab cylinder (insert) of the same dimension (diameter 6.4 cm) similar to hollow space provided in CIRS phantom was used. The wood slabs of varying width and approximately of the same thickness (along the central line) were used to form the insert. The wood cylinder insert was put in the hollow space provided in the thorax phantom to create lung equivalent inhomogeneity [Figure 1]b.
The Gafchromic EBT ® 3 films (International Specialty Products, NJ, USA) were used in the present work for the measurement of dose (at various depths) across the inserts along central axis of the beam. These high spatial resolution films of uniformity better than ± 3% can be used in dose range 1 cGy–40 Gy. The important feature of the EBT ® 3 films is that it is near tissue equivalent and requires no processing. The film can be handled at room temperature and can withstand up to 70° temperature.
Calibration of Gafchromic film EBT3
The unknown absorbed dose from the exposed film was measured using the calibration curve, that is, curve drawn between measured net optical density from the pixel readings and its corresponding known irradiated radiation dose. In the present study calibration curve was generated by irradiating the Gafchromic films of size 3 cm × 3 cm placed at the depth of maximum dose of solid water slab phantom (PTW, Freiberg, Germany) in full scatter condition. 6 MV photon beam of 10 cm × 10 cm field sizes from linear accelerator (LINAC) (Infinity, Elekta Medical Systems) were used to deliver 30 cGy, 50 cGy, 100 cGy, 150 cGy, 200 cGy, 250 cGy, 300 cGy, 350 cGy, 400 cGy, 450 cGy, and 500 cGy dose. For the delivery of these radiation doses, the monitor units (MUs) were calculated using the data table generated from the radiation field analyzer during the LINAC commissioning. After 24 h of irradiation, each film was scanned in landscape orientation using flatbed scanner (Epson, expression) and absorbed dose was recorded using VeriSoft (version 4.2.1, PTW, Freiberg, Germany). For the use of TPSs in clinical purpose, its performance was verified during its commissioning. Tolerance levels for all fields used in clinical condition were within acceptable limits and were approved by the competent authority.
Study 1: Dose measurements in standard computerized imaging reference systems lung phantom
Experimental setup consists of CIRS thorax phantom, which has a cylindrical insert in the form of two hemispheres of radii 3.2 cm [Figure 1]a. This cylindrical insert has the same density as the two lung structures (HU − 700). The phantom was scanned using CT simulator (Somatom, Siemens Healthcare, USA) and images of the inter slice thickness of 3 mm were taken. These CT images were retrieved in to the XiO (Computerized Medical Systems, USA) and Monaco (V3.2 CMS Inc., St. Louis, MO, USA). TPSs and plans were created using the four algorithms CON, SP, FSP, and XVMC. Dose of 200 cGy was prescribed at the depth of maximum dose applying 6 MV single anterior-posterior fields of size 2 cm × 2 cm, 3 cm × 3 cm, 4 cm × 4 cm, 5 cm × 5 cm, and 8 cm × 8 cm keeping source to phantom surface distance 100 cm and dose was calculated by different algorithms for three surfaces of lung insert namely proximal (P), central (C), and distal (D). Calculation of MUs required for dose delivery of 200 cGy for all the fields was performed from the output data generated by radiation field analyzer at the time of LINAC commissioning. Dose absorbed at P, C, and D surfaces of lung insert were measured using Gafchromic films.
Study 2: Inhomogeneity inserted in the lung phantom
Inhomogeneity in the form of wood cylinder of diameter 6.4 cm and appropriate length [Figure 1]b-I was inserted in the cylindrical cavity provided in the CIRS thorax phantom. Following the similar procedure as in Study 1, algorithms calculated and Gafchromics films measured dose at eight surfaces of wood cylinder as shown in [Figure 1]b-II were recorded.
Study 3: Inhomogeneity containing the tumor in the lung phantom
In racemosa wood cylinder, a cavity of dimension 2 cm × 2 cm × 1 cm was created in two consecutive slabs. These two cavities were filled by tissue soft liner (density – 1 g/cm 3) to create a tumor of dimension 2 cm × 2 cm × 2 cm [Figure 1]c. This material is also used in design of prosthesis used in dentistry. Racemosa wood cylinder with tumor was placed inside the phantom cavity in such a way that the center of the tumor lies along central beam. CIRS thorax phantom containing wood cylinder with the tumor at a depth 6.5 cm (anteriorly) was scanned. It lies along with the same tumor at depths (anteriorly) of 6.5 cm, 7.5 cm, and 8.5 cm. Phantom was scanned again by moving the tumor downward that is, the step of 1 cm so that tumor depth gradually increases to 7.5 cm and 8.5 cm. Using the similar procedure, same dose, and prescription point, algorithm calculated and film measured dose for three surfaces of tumor namely P, C, and D were recorded. The percentage depth dose (PDD) obtained from different algorithms were compared against the film derived PDD for the purpose of evaluation of accuracy of dose calculation algorithms in all the three studies.
| > Results|| |
To validate the Gafchromic film data, it was compared with the 0.6 cc as well as 0.125 cc ionization chamber (PTW, Freiberg, Germany) data in the solid water slab phantom. The agreement between film and ionization chamber data was within 2.1%.
The calculated and measured dose values along the central axis in all the three study groups were recorded for all the fields. These values were normalized to their maximum dose and plotted against the measured film results.
The difference in PDD calculated by different algorithms produced a statistical significance about their performance. ANOVA one-way analysis of means of variance of PDD was used to compare statistically the difference in PDDs calculated by different algorithms and measurements were performed by film in the presence of inhomogeneity. P <0.05 was taken statistically significant. The difference in PDDs produced P = 0.008 for field size 2 cm × 2 cm, P = 0.010 for 3 cm × 3 cm, P = 0.252 for 4 cm × 4 cm, P = 0.518 for 5 cm × 5 cm, and P = 0.408 for 8 cm × 8 cm with level of significance α = 0.05.
XVMC calculated PDD were in good agreement with film derived PDD with maximum deviation of 2.22% (underestimation) compared to film results in case of 2 cm × 2 cm field size at the distal end of CIRS standard insert [Table 1]. It was increasing in nature from the proximal (1.07%) to distal end (2.22%) with 1.69% at the center. XVMC always under estimated compared to film measured data. [Figure 2] shows comparison of calculated and film derived PDD.
|Table 1: Percentage deviation in PDD calculated by various algorithms compared with the film derived PDD at selected points for 2 cm×2 cm and 3 cm×3 cm field sizes|
Click here to view
|Figure 2: Comparison of algorithms produced percentage depth dose with film derived percentage depth dose at three surfaces of lung insert for field sizes (a) 2 cm x 2 cm (b) 3 cm x 3 cm (c) 4 cm x 4 cm (d) 5 cm x 5 cm (e) 8 cm x 8 cm|
Click here to view
For field size 2 cm × 2 cm, the maximum deviation of CON calculated data compared to film measured data was 32% at the proximal end while SP and FSP under estimated by 5.7% and 6.2% at the center of the inset. It was <2% for field size 3 cm × 3 cm and almost negligible for larger fields in XVMC calculated results [Table 2]. The magnitude of this deviation was found decreasing in nature with field size in each of the algorithm with <1.5% deviation (under estimation) in SP and FSP algorithm and 8.96% (overestimation) by CON for field size 8 cm × 8 cm.
|Table 2: Percentage deviation in MC calculated PDD compared with the film derived PDD|
Click here to view
XVMC calculated PDD were in good agreement with film derived PDD for all selected field sizes except for field size 3 cm × 3 cm [Table 3], where there was slightly higher deviation in PDD between dose calculation algorithm XVMC and film derived PDD. The deviation for field size smaller than 4 cm × 4 cm was <2%, while, for larger fields, that is, 5 cm × 5 cm and above it was <1%. For 3 cm × 3 cm, the deviation up to 3.3% in the center and at the distal end while it was around 2% at the proximal end. For field size 3 cm × 3 cm, CON algorithm produced results deviated by as much as 12% at the proximal end while SP and XVMC underestimated the dose by 4% and 2%, respectively. In contrast, the maximum deviation of 7% (underestimation) by SP and FSP algorithms and 6% (overestimation) by CON was found at the center [Figure 3]. Authors also observed that the deviation was more in the center and a distal end for smaller fields while it was more in the center and the proximal end for larger fields [Table 4].
|Table 3: Percentage deviation in PDD calculated by various algorithms compared with the film derived PDD at selected points for 2 cm×2 cm and 3 cm×3 cm field sizes|
Click here to view
|Figure 3: Comparison of algorithms produced percentage depth dose with film derived percentage depth dose at eight surfaces of wood cylindrical inhomogeniety insert for field sizes (a) 2 cm x 2 cm (b) 3 cm x 3 cm (c) 4 cm x 4 cm (d) 5 cm x 5 cm (e) 8 cm x 8 cm|
Click here to view
|Table 4: Percentage deviation in MC calculated PDD compared with the film derived PDD|
Click here to view
PDD values were recorded for proximal, central, and distal surfaces of the tumor in each of the algorithms as well as using the Gafchromic films for field size 2 cm × 2 cm, 3 cm × 3 cm, 4 cm × 4 cm, 5 cm × 5 cm, and 8 cm × 8 cm. The deviation of MC calculated PDD from the film derived PDD was increasing with a depth of the tumor (i.e. highest at 8.5 cm tumor depth) as illustrated in [Figure 4], [Figure 5], [Figure 6]. In case of tumor at depth 6.5 cm, the deviation was <2% except D plane (−2.24%) in 2 cm × 2 cm field size. This was −3.13% atPplane for 2 cm × 2 cm field size and −1.44%, −1.34%, and −1.37% for P, C, and D plane in 3 cm × 3 cm in tumor with depth 8.5 cm. With SP algorithm, deviation up to −6.28%, at D plane of the tumor having depth 6.5 cm in 2 cm × 2 cm, 4.13% at D plane of the tumor with depth 7.5 cm in 3 cm × 3 cm field size and 8.52% atPplane of the tumor with depth 8.5 cm in 2 cm × 2 cm, respectively, was found. FSP showed overestimation trend in different setup with maximum in tumor depth 8.5 cm. CON always found overestimating the PDD. This overestimation was found decreasing (20%, 18%, and 10%) with depth of the tumor. In addition, decrease in PDD with the field size within the same setup was recorded.
|Figure 4: Comparison of algorithms produced percentage depth dose with film derived percentage depth dose of three surfaces of tumor at depth 6.5 cm inside wood cylindrical inhomogeniety for field sizes (a) 2 cm x 2 cm (b) 3 cm x 3 cm (c) 4 cm x 4 cm (d) 5 cm x 5 cm (e) 8 cm x 8 cm|
Click here to view
|Figure 5: Comparison of algorithms produced percentage depth dose with film derived percentage depth dose of three surfaces of tumor at depth 7.5 cm inside wood cylindrical inhomogeneity for field sizes (a) 2 cm x 2 cm (b) 3 cm x 3 cm (c) 4 cm x 4 cm (d) 5 cm x 5 cm (e) 8 cm x 8 cm|
Click here to view
|Figure 6: Comparison of algorithms produced percentage depth dose with film derived percentage depth dose of three surfaces of tumor at depth 8.5 cm inside wood cylindrical inhomogeneity for field sizes (a) 2 cm x 2 cm (b) 3 cm x 3 cm (c) 4 cm x 4 cm (d) 5 cm x 5 cm (e) 8 cm x 8 cm|
Click here to view
| > Discussion|| |
In the present work, authors attempted to verify dose calculation accuracy of different algorithms using 6 MV photon beam including small fields which are commonly used in modern lung cancer treatment modalities such as Intensity Modulated Radiotherapy, CyberKnife, stereotactic radiotherapy etc., however generally larger field size (>4 cm × 4 cm) are used in conventional mode of treatment such as three-dimensional conformal radiotherapy. The small fields facilitate desired dose distribution even in small target volume with lowest dose to nearby normal tissues.,, Before the experiment, verification of beam modeling of TPS for all field sizes (including <4 cm × 4 cm) was done. This was performed by comparing the dose at a point in water slab phantom (PTW, Freiberg, Germany) calculated by algorithms and measured by pin point ion chamber (PTW, Freiberg, Germany). The difference between the two was <2%. The dose calculation accuracy of different algorithms was measured in terms of the deviation from the Gafchromic film measured results. The accuracy of Gafchromic films relies on its spatial resolution (~25 μm) and studies have reported the accuracy of Gafchromic films of the order of the smallest volume chamber (pinpoint) including the small fields., The present study reveals that for 6 MV photon beam maximum contribution of the primary and secondary radiation is dependent upon the field size and the density of the medium. PDD was found increasing with field size. Authors also found that change in film derived PDD from 2 cm × 2 cm to 3 cm × 3 cm field size was –11% (maximum), however for field size(s) larger than 3 cm × 3 cm this difference was less [Table 5]. The effect of change in density on PDDs was also observed as shown in [Table 6]. This is evident from the present study that difference between calculated and measured data was maximum for 2 cm × 2 cm followed by 3 cm × 3 cm both in Study 1 and 2. These two studies comprised of average lung equivalent density of 0.275 g/cm 3 and 0.30 g/cm, 3 respectively. From the results of Studies 1 and 2, it was observed that SP and XVMC underestimated the dose as much as by 6% and 3.3%, respectively, whereas CON overestimated the dose by 32% in small fields. The reason behind these deviation could be the fact that most of the interactions occurring at the therapeutic range produce secondary electrons in addition to the scattered photons of varying ranges according to densities. For small field size particularly in low-density medium electron equilibrium is not achieved in the lateral direction that causes dose reduction in such conditions. The efficiency of the inhomogeneity correction algorithms lies on accounting these events., As far as deviations of SP and CON are concerned, both the algorithms calculate photon energy fluence distribution in similar way but SP uses modified energy deposition kernals to take into account the low densities. This difference becomes more visible in case of small fields due to electronic disequilibrium. In the present study, CON always overestimated the dose because of oversimplified modeling of electron transport as well as overestimating the primary dose. García-Vicente et al. also found overestimation of dose (~10%) by CON algorithm while verifying its accuracy in inhomogeneous phantom using radiographic films and ionization chamber for a 10 cm × 10 cm beam irradiating a mediastinum lung. In a clinical study of lung cancer dose calculation accuracy, Zhao et al. reported higher dose to target calculated by CON based algorithms. The difference between SP and FSP lies in their accuracy and speed. The calculation speed of FSP is about 2–3 times more and 1–2% less accurate compared to SP algorithm. Monte Carlo radiation transport techniques use numerical methods to model the physical processes which govern interactions between radiation and their environment. For calculation of dose, MC uses straightforward simulation of real physical basis unlike to analytic algorithms that use various approximations. In XVMC, the dose is scored in voxels that are defined based on density specification for each voxel individually that makes its results close to measured data as found in the present study. In real clinical settings, lung tumor presents greater challenge for dose delivery as tumor density is invariably higher than the lung density and also lung is surrounded by different density media such as muscle and bone. Dose variations were recorded in the three planes of the tumor (Study 3). Dose build up in the central plane of the tumor was observed. Except CON, for all the algorithms including film measured data, increase in dose from proximal surface to the central surface of the tumor and then dose falloff from the central to distal surface for field size 2 cm × 2 cm, 3 cm × 3 cm, and 4 cm × 4 cm was recorded. For larger fields (>4 cm × 4 cm), the continuous fall off of the dose from proximal to the distal end for the entire three tumor depth was recorded [Figure 4], [Figure 5], [Figure 6]. Takahashi et al. also found same trend of dose build up/down at the tumor/lung interface in real patient stereotactic radiotherapy using XVMC and SP and reported that at the tumor/lung interface both SP and XVMC give similar results. The variation in PDD among three surfaces of tumor was found maximum for 2 cm × 2 cm. In the present study, the variation in film derived PDD between proximal and central surface of the tumor was 9% for tumor at depth 6.5 cm, 6% at tumor depth 7.5 cm and reduced to 2% at tumor depth of 8.5 cm. In contrast to this, decrease in PDD from central plane to distal plane of the tumor was observed though the magnitude of difference was less compared to the difference between proximal and central plane. These build up and build down regions within the tumor can be attributed to the production of large secondary electrons at the proximal (P) surface of tumor which get attenuated by tumor itself. This electronic disequilibrium at the interface makes the accurate dose predictions difficult. Unlike other algorithms, XVMC found consistent with the measured data with maximum deviation up to 3.13% (Study 3). Dobler et al. did a similar study using tumor of diameter 1 cm and 3 cm to assess the accuracy of various algorithms relative to EBT Gafchromic films. For XVMC, they reported deviation up to 3% at the interface of lung/tumor tissue. These results indicate that a dose reduction inside the lung tumor occurs that may compromise the expected results of the treatments. Sethi et al. also reported excellent agreement (<3%) between XVMC, film and ion chamber results measured in heterogeneous lung phantom for 6 MV photon beam. Recently Mesbahi et al. performed a simulation using MCNPX Monte Carlo code to calculate the depth doses in water and low-density medium using 2 cm × 2 cm and 3 cm × 3 cm beam sizes and reported the same type of dose variation within the tumor. The impact of these deviations (under dosing/overdosing) can be pronounced more in real lung due to change in lung density during its exhalation and inhalation and this can be one of the reasons of higher recurrence and lesser survival rate in case of lung cancer as reported in literatures.,
|Table 5: Variation of film derived PDD with field size including P, C and D surfaces of the wood cylinder inhomogeneity|
Click here to view
| > Conclusion|| |
XVMC was found more consistent to predict the dose in lung equivalent inhomogeneity, inside the tumor as well as at the interface for all fields. CON and FSP algorithms should be avoided in low-density medium especially for small fields due to overestimation in results. SP algorithm can be an alternative as this is relatively fast and cost effective compared to XVMC. These results can be utilized in clinical practice as CIRS thorax phantom along with lung equivalent inhomogeneity used in the present study possess the average real patient settings.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| > References|| |
Mesbahi A. The effect of electronic disequilibrium on the received dose by lung in small fields with photon beams: Measurements and Monte Carlo study. Iran J Radiat Res 2008;6:70-6.
Young ME, Kornelsen RO. Dose corrections for low-density tissue inhomogeneities and air channels for 10-MV x rays. Med Phys 1983;10:450-5.
Klein EE, Morrison A, Purdy JA, Graham MV, Matthews J. A volumetric study of measurements and calculations of lung density corrections for 6 and 18 MV photons. Int J Radiat Oncol Biol Phys 1997;37:1163-70.
Arnfield MR, Siantar CH, Siebers J, Garmon P, Cox L, Mohan R. The impact of electron transport on the accuracy of computed dose. Med Phys 2000;27:1266-74.
Künzler T, Fotina I, Stock M, Georg D. Experimental verification of a commercial Monte Carlo-based dose calculation module for high-energy photon beams. Phys Med Biol 2009;54:7363-77.
Andreo P, Burns DT, Hohlfeld K, Huq MS, Kanai T, Laitano F, et al
. Absorbed Dose Determination in External Beam Radiotherapy: An International Code of Practice for Dosimetry Based on Standards of Absorbed Dose to Water. IAEA Technical Report Series No. 398. IAEA Vienna; 2000.
Specht L, Yahalom J, Illidge T, Berthelsen AK, Constine LS, Eich HT. Modern radiation therapy for Hodgkin lymphoma: Field and dose guidelines from the international lymphoma radiation oncology group (ILROG). Int J Radiat Oncol Biol Phys 2014;89:854-62.
Azimi R, Alaei P, Higgins P. The effect of small field output factor measurements on IMRT dosimetry. Med Phys 2012;39:4691-4.
Bassinet C, Huet C, Derreumaux S, Brunet G, Chéa M, Baumann M, et al.
Small fields output factors measurements and correction factors determination for several detectors for a CyberKnife® and linear accelerators equipped with microMLC and circular cones. Med Phys 2013;40:071725.
Hassani H, Nedaie HA, Zahmatkesh MH, Shirani K. A dosimetric study of small photon fields using polymer gel and gafchromic EBT films. Med Dosim 2014;39:102-7.
Ju SG, Han Y, Kum O, Cheong KH, Shin EH, Shin JS, et al.
Comparison of film dosimetry techniques used for quality assurance of intensity modulated radiation therapy. Med Phys 2010;37:2925-33.
Wang L, Yorke E, Desobry G, Chui CS. Dosimetric advantage of using 6 MV over 15 MV photons in conformal therapy of lung cancer: Monte Carlo studies in patient geometries. J Appl Clin Med Phys 2002;3:51-9.
Charland PM, Chetty IJ, Yokoyama S, Fraass BA. Dosimetric comparison of extended dose range film with ionization measurements in water and lung equivalent heterogeneous media exposed to megavoltage photons. J Appl Clin Med Phys 2003;4:25-39.
Papanikolaou N, Battista JJ, Boyer AL, Kappas C, Klein E, Mackie TR, et al
. Tissue inhomogeneity corrections for megavoltage photon beams. (AAPM report No. 85). Madison: Medical Physics Publishing; 2004.
Jones AO, Das IJ. Comparison of inhomogeneity correction algorithms in small photon fields. Med Phys 2005;32:766-76.
Miften M, Wiesmeyer M, Monthofer S, Krippner K. Implementation of FFT convolution and multigrid superposition models in the FOCUS RTP system. Phys Med Biol 2000;45:817-33.
García-Vicente F, Miñambres A, Jerez I, Modolell I, Pérez L, Torres JJ. Experimental validation tests of fast Fourier transform convolution and multigrid superposition algorithms for dose calculation in low-density media. Radiother Oncol 2003;67:239-49.
Zhao Y, Qi G, Yin G, Wang X, Wang P, Li J, et al.
Aclinical study of lung cancer dose calculation accuracy with Monte Carlo simulation. Radiat Oncol 2014;9:287.
Animesh. Advantages of multiple algorithm support in treatment planning system for external beam dose calculations. J Cancer Res Ther 2005;1:12-20.
Chetty IJ, Curran B, Cygler JE, DeMarco JJ, Ezzell G, Faddegon BA, et al.
Report of the AAPM Task Group No 105: Issues associated with clinical implementation of Monte Carlo-based photon and electron external beam treatment planning. Med Phys 2007;34:4818-53.
Takahashi W, Yamashita H, Saotome N, Iwai Y, Sakumi A, Haga A, et al.
Evaluation of heterogeneity dose distributions for stereotactic radiotherapy (SRT): Comparison of commercially available Monte Carlo dose calculation with other algorithms. Radiat Oncol 2012;7:20.
Dobler B, Walter C, Knopf A, Fabri D, Loeschel R, Polednik M, et al.
Optimization of extracranial stereotactic radiation therapy of small lung lesions using accurate dose calculation algorithms. Radiat Oncol 2006;1:45.
Sethi A, Leo P, Kabat C, Cecilio P. Validation of Monte Carlo dose calculation algorithm in heterogeneous medium. Med Phys 2013;40:329.
Mesbahi A, Dadgar H, Ghareh-Aghaji N, Mohammadzadeh M. A Monte Carlo approach to lung dose calculation in small fields used in intensity modulated radiation therapy and stereotactic body radiation therapy. J Cancer Res Ther 2014;10:896-902.
Dela Cruz CS, Tanoue LT, Matthay RA. Lung cancer: Epidemiology, etiology, and prevention. Clin Chest Med 2011;32:605-44.
Zimmermann FB, Molls M, Jeremic B. Treatment of recurrent disease in lung cancer. Semin Surg Oncol 2003;21:122-7.
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5], [Figure 6]
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5], [Table 6]