Computed Tomography as a Source of Electron Density Information for Radiation Treatment Planning

Transcription

1 Strahlentherapie und Onkologie Original Article Computed Tomography as a Source of Electron Density Information for Radiation Treatment Planning Witold Skrzyński 1, Sylwia Zielińska-Da browska 2, Marta Wachowicz 2, Wioletta Ślusarczyk-Kacprzyk 1, Paweł F. Kukołowicz 2, Wojciech Bulski 1 Purpose: To evaluate the performance of computed tomography (CT) systems of various designs as a source of electron density ( ) data for treatment planning of radiation therapy. Material and Methods: Dependence of CT numbers on relative electron density of tissue-equivalent materials (HU- relationship) was measured for several general-purpose CT systems (single-slice, multislice, wide-bore multislice), for radiotherapy simulators with a single-slice CT and kv CBCT (cone-beam CT) options, as well as for linear accelerators with kv and MV CBCT systems. Electron density phantoms of four sizes were used. Measurement data were compared with the standard HU- relationships predefined in two commercial treatment-planning systems (TPS). Results: The HU- relationships obtained with all of the general-purpose CT scanners operating at voltages close to 120 kv were very similar to each other and close to those predefined in TPS. Some dependency of HU values on tube voltage was observed for bone- equivalent materials. For a given tube voltage, differences in results obtained for different phantoms were larger than those obtained for different CT scanners. For radiotherapy simulators and for kv CBCT systems, the information on was much less precise because of poor uniformity of images. For MV CBCT, the results were significantly different than for kv systems due to the differing energy spectrum of the beam. Conclusion: The HU- relationships predefined in TPS can be used for general-purpose CT systems operating at voltages close to 120 kv. For nontypical imaging systems (e.g., CBCT), the relationship can be significantly different and, therefore, it should always be measured and carefully analyzed before using CT data for treatment planning. Key Words: Radiotherapy Computed tomography Electron density Strahlenther Onkol 2010;186:32733 DOI /s CT-Systeme als Datenquelle der Elektronendichte in Bestrahlungsplanungssystemen Ziel: Vergleich verschiedener Computertomographie-(CT-)Systeme zur Bestimmung der Elektronendichte ( ) für die Bestrahlungsplanung. Material und Methodik: Die Relation des CT-Werts zur Elektronendichte wurde an verschiedenen modernen CT-Scannern ( single-slice, multislice, wide-bore multislice ) ermittelt, für die Therapiesimulatoren mit einem single-slice -CT und kv-cbct-( cone-beam -CT-)Optionen sowie für Linearbeschleuniger mit kv- und MV-CBCT-Systemen. Vier unterschiedlich große Phantome zweier Hersteller wurden zur Messung der Elektronendichte benutzt. Die Messdaten wurden mit den Standardumrechnungsformeln zweier marktüblicher Therapieplanungssysteme (TPS) verglichen. Ergebnisse: Die HU- -Relationen, die in allen modernen CT-Systemen vorhanden sind, waren untereinander sehr ähnlich, ebenso wie zu den vorgegebenen Relationen in den TPS. Einige Abweichungen der HU-Werte in Abhängigkeit von der Röhrenspannung wurden bei knochenäquivalentem Material beobachtet. Bei vorgegebener Röhrenspannung wurden bei den verschiedenen Phantomen größere Differenzen gemessen als in den verschiedenen CT-Scannern. Weniger exakt waren die Informationen über mit den Therapiesimulatoren und KV-CBCT-Systemen aufgrund der mäßigen Uniformität der Bilder. Die Ergebnisse des MV-CBCT unterschieden sich aufgrund des unterschiedlichen Energiespektrums der Röntgenstrahlen signifikant von denen der kv-systeme. Schlussfolgerung: Die im TPS vorgegebene HU- -Relation kann bei modernen CT-Systemen mit einer Röhrenspannung im Bereich von 120 kv genutzt werden. Signifikant unterschiedlich dagegen ist die Relation bei nichttypischen Bildsystemen (z.b. CBCT). Deshalb sollte bei solchen Systemen immer gemessen und sorgfältig analysiert werden, bevor die CT-Daten für die Therapieplanung herangezogen werden. Schlüsselwörter: Radiotherapie Computertomographie Elektronendichte 1 Medical Physics Department, Center of Oncology, Warsaw, Poland, 2 Medical Physics Department, Holycross Cancer Center, Kielce, Poland. Received: September 4, 2009; accepted: March 5, 2010 Published Online: May 17, 2010 Strahlenther Onkol 2010 Nr. 6 Urban & Vogel 327

2 Introduction Calculation of dose distribution within the treated volume is an essential step of contemporary treatment planning in radiotherapy. Many factors have an influence on the dose distribution, heterogeneity of the patient s body being one of them. Data characterizing each patient are therefore needed for the calculations. X-ray computed tomography (CT) has been used as a basic source of such data for over 30 years now [17], and is used as a base for treatment planning even in less common radiotherapy techniques, such as helical tomotherapy [22] or radiotherapy with proton beams [5]. Other imaging modalities are sometimes used as an addition to CT, as they may offer better visualization of target volume (e.g., magnetic resonance imaging [18], positron emission tomography [1], or both of them [26]). Nevertheless, the role of X-ray CT is fundamental, as it provides information on the attenuation of radiation by the patient s tissues in a form of CT numbers, expressed in Hounsfield units (HU) as in the following equation: HU tissue = [(μ tissue μ water ) / μ water ] 1,000, where μ is the linear attenuation coefficient of water and of the tissue. It is known that precise calculation of dose distribution in radiotherapy can be performed on the basis of knowledge of the electron density of the tissues [20]. Treatment-planning systems (TPS) usually convert HU values to (relative electron density, normalized to water) by means of the predefined relationship between the two quantities, e.g., one given by Knöös et al. [11]. In some TPS the relation is fixed, in others the user is allowed to change it. It should be remembered that Hounsfield numbers for a given tissue depend on the quality of the X-ray beam; therefore, the values can differ between scanners. Even for a single scanner, CT numbers for the same tissue depend on the kv setting and on beam filtration [3, 4]. As a result, the dose calculated by TPS can change by as much as 2% if CT scans are obtained using 80 kv instead of 130 kv while using the same HU- relationship [7]. The HU- relationships can be measured with the use of phantoms with tissue-equivalent materials [3], i.e., materials that have an atomic composition similar to human tissues [8]. Data obtained with such phantoms for the particular CT scanner operating at a particular kv can then be introduced into the TPS to make the calculations more precise. However, different manufacturers of commercial electron density phantoms use different tissue-equivalent materials. It is known that solid (resin-based) bone-equivalent materials give systematically lower HU values than water solutions of CaCl 2 of the same electron density [24]. Even for a single CT unit and a single scanning protocol differences of in can be observed depending on the choice of phantom (all resin-based), leading to differences of the calculated dose in the order of 12% [6]. HU values observed for a given material also depend on the dimensions of the phantom [6, 9] and on the positioning of the phantom, especially the presence and the type of patient support [4] or location on/off axis [9]. Also, a change in position of the tissue-equivalent insert in the phantom can lead to a difference of up to 80 HU [3]. Such differences may be attributed to differences in beam quality caused by different filtration at different depths in the phantom (beam hardening effect). Many of the described dependencies have also been confirmed in Monte Carlo simulations [19]. Some uncertainty of data seems therefore to be unavoidable, especially as patients also differ between themselves in dimensions and in the composition of their tissues. It is suggested that one universal HU- relationship can be used for all CT scanners operating at typical voltages ( kv), leading to dose calculation errors not greater than 1% [24]. These errors are comparable to those caused, e.g., by use of uncorrected contrast-enhanced CT scans, which leads to change in dose calculation of 1% on average (3% maximum) in the lung [2], or of 0.67% on average (1.8% maximum) in the brain [28]. Larger errors in dose calculations can, however, be expected if applying standard predefined HU- relationships to data obtained with nontypical CT scanners, e.g., cone-beam CT (CBCT) systems installed onto radiotherapy linear accelerators or radiotherapy simulators. It is known, that such systems are more prone to inaccuracies of HU values than conventional CT systems because of the higher effect of X-ray scatter associated with cone-beam geometry [21]. The use of kv CBCT data for dose calculation could introduce errors of 3%, partly because of nonuniformity of CBCT images [12, 23]. In the case of MV CBCT, acceptable accuracy of calculated doses can be obtained, if the images are uniformity-corrected and if density calibration is done [15, 25]. The aim of this study was to evaluate the performance of CT systems of various designs as a source of data for treatment planning of radiation therapy, and to compare the standard HU- relationships predefined in TPS with the actual relationships obtained by measurements. Material and Methods All the CT systems included in the study are listed in Table 1. Three of them are general-purpose CT scanners and are routinely used as a source of electron density data, the other four are designed as radiotherapy verification tools. The systems present a wide range of designs, from single-slice X-ray CT to megavoltage CBCT. Electron density phantoms of four sizes made by two manufacturers were used in the measurements as listed in Table 2. For each phantom, several tissue-equivalent inserts were available, covering a wide range of tissue densities and compositions (lungs, soft tissues, bones). RMI 465 is a typical electron density phantom. RMI 463 was designed for general quality assurance and can accommodate only three 328 Strahlenther Onkol 2010 Nr. 6

3 Table 1. List of computed tomography (CT) systems used in the study. Systems marked with + in the TP (treatment planning) column are routinely used as a source of electron density data for radiation treatment planning. Tabelle 1. Liste der in dieser Studie verwendeten Computertomographie-(CT-)Systeme. Die in der Spalte TP (Therapieplanung) mit + markierten Systeme dienen in der Routine als Datenquelle für die Elektronendichte an Bestrahlungsplanungssystemen. Name Details TP GE HiSpeed DX/i General-purpose single-slice CT, third generation + Siemens Somatom AR General-purpose multislice CT, third generation + Siemens Somatom Sensation Open Varian Ximatron Nucletron Simulix Evolution Varian OBI Siemens MVision General purpose multislice CT, third generation (large gantry bore 82 cm) kv CT as an option of radiotherapy simulator (single slice, part of image intensifier used as a detector) kv CBCT with flat-panel detector, as an option of radiotherapy simulator kv CBCT with flat-panel detector, installed on linear accelerator MV CBCT with flat-panel detector, installed on linear accelerator Table 2. List of electron density phantoms used in the study. Tabelle 2. Übersicht der Phantome, an denen die Elektronendichte in dieser Studie gemessen wurde. Phantom Dimensions Description RMI cm (body) phantom, 16 inserts ( 2.8 cm each): tissue-equivalent inserts, water, four solid-water rods in various locations (for uniformity check), titanium RMI 463 CIRS 062 (inner section) CIRS 062 (both sections) 16 cm (head) 18 cm (head) cm (body, elliptic) Quality assurance phantom, can simultaneously accommodate up to three inserts from RMI 465 phantom, eight tissue-equivalent inserts ( 3.05 cm each) and water phantom, consists of inner and outer sections, each section with identical set of eight tissue-equivalent inserts tissue-equivalent inserts simultaneously. The phantom has some internal structures designed for assessment of image quality, however, they are not located very close to the inserts and they do not cause artifacts in the image. We assumed that the presence of the structures does not significantly influence HU readings of tissue-equivalent inserts. The CIRS 062 electron density phantom consists of inner and outer sections. The inner section alone can be used to simulate a patient s head. When used together, the two sections simulate a patient s torso. The relationships measured for different CT units were compared with each other and with default relationships implemented in two commercial TPS, namely Oncentra MasterPlan [16] and Varian Cadplan [27]. Criteria proposed by + ESTRO [13] were adopted, i.e., it was assumed that the values of calculated by TPS should not differ from the known true values by more than 0.05 for < 1.5 and by more than 0.1 for > 1.5. Larger differences were treated as significant. The influence of scan parameters (e.g., kv) and of the choice of phantom on the results was also evaluated. Results General-Purpose CT Scanners Figure 1 presents results obtained with RMI 465 phantom for three general-purpose X-ray CT scanners operating at various kv settings. For between 0 and 1 (air, lungs, soft tissues), all datasets did not seem to differ, while for > 1 (bones), the results were dependent on kv setting, and for a given kv, they were different between scanners. For each scanner operating at the most common voltage setting (120 kv), the dependence of the results on the choice of other parameters was also investigated with RMI 465. The parameters included tube current, slice width, imaging mode (axial/spiral), pitch in spiral mode, position of the inserts within the phantom, and positioning of the phantom in the gantry (i.e., on-axis or few centimeters off-axis). The range of HU values obtained for each material for the GE HiSpeed unit was generally smaller than 20 HU, only for high-density bones ( > 1.4) it reached 50 HU. For both Siemens units, larger differences were observed, reaching 5060 HU for lung tissue and HU for high-density bones. Figure 2 presents the dependence of the results obtained for the GE HiSpeed scanner on the choice of phantom and field of view (FOV; all the other parameters remaining the same). Similar measurements were also done for the Siemens Somatom Sensation Open unit and similar dependencies were Figure 1. Results of measurements with RMI 465 phantom for three general-purpose X-ray CT scanners (GE HiSpeed, Siemens Somatom AR, Siemens Somatom Sensation Open) operating at three selected kv settings. Abbildung 1. Messergebnisse mit drei Routine-CT-Scannern (GE HiSpeed, Siemens Somatom AR, Siemens Somatom Sensation Open) am RMI-465-Phantom bei drei vorgewählten kv-stufen. Strahlenther Onkol 2010 Nr

4 Figure 2. Results of measurements with all available phantoms for GE HiSpeed scanner operating at 120 kv. Relationships implemented in two TPS shown for comparison. Abbildung 2. Messergebnisse mit allen verfügbaren Phantomen des GE-HiSpeed-Scanners bei 120 kv. Die Relationen wurden in zwei TPS implementiert und zum Vergleich dargestellt. Table 3. Variability of Hounsfield (HU) values for three general-purpose computed tomography (CT) scanners. Tabelle 3. Variabilität der HU-Werte (Hounsfield-Einheiten) für drei Allzweck-Computertomographie-(CT-)Scanner. Altered parameter Constant parameters Range of observed HU values (maximumminimum) Lung Soft tissue Bone kv CT scanner, phantom Phantom CT scanner, 120 kv Protocol CT scanner, 120 kv, phantom Scanner 80 kv, phantom Scanner 140 kv, phantom Scanner 120 kv, phantom observed. For the GE unit, the HU values were dependent on the FOV itself, even for the same phantom and tube voltage. This can be explained as, for that particular scanner, different beam filtration is automatically chosen for large FOV rather than for a small one, resulting in different beam quality. Table 3 presents data on the variability of the HU values for three general-purpose CT scanners. The largest source of the variability is kv setting. Should this be kept constant at 120 kv, the largest remaining source of the variability is the choice of phantom and of the FOV. Differences caused by change of other parameters, or even by change of the scanner, seem to be less significant. Radiotherapy Simulator with Image Intensifier Measurements on a Varian Ximatron radiotherapy simulator with CT option were done for RMI 465 phantom only (Figure 3). Significant nonuniformity was visible in the image as a white ring, and CT numbers for four identical solid-water Figure 3. Results of measurements with RMI 465 phantom for Varian Ximatron radiotherapy simulator operating as single-slice CT at 120 kv. Relationships implemented in two TPS shown for comparison. Abbildung 3. Messergebnisse mit dem RMI-465-Phantom des Radiotherapiesimulators Varian Ximatron als Einzelschicht-CT bei 120 kv. Die Relationen wurden in zwei TPS implementiert und zum Vergleich dargestellt. inserts placed in different positions of the phantom ranged from 208 HU to +13 HU (for general-purpose CT scanners the values would be identical within a few HU). kv CBCT Measurements on two kv CBCT systems were performed with CIRS 062 phantom (Figure 4). For Varian OBI and inner section of the phantom ( head ), the results were almost in agreement with the relationships predefined in TPS, significant differences were observed only for lung-equivalent materials. Some dependency on dimension of FOV was observed, which could be explained as an effect of differences in acquisition geometry and beam quality for small FOV (Ø 25 cm and smaller) a full-fan acquisition was used, while for larger FOV half-fan geometry and a different bow-tie filter was used. For both sections of CIRS 062 phantom placed together ( torso ), the results were significantly different. HU values for two identical tissue-equivalent inserts simultaneously placed in outer and inner sections of the phantom differed by more than 300 HU. Another kv CBCT system included in the study was Nucletron Simulix Evolution radiotherapy simulator. The results did not agree with the predefined relationships and were also dependent on the dimension of the phantom. Nonuniformity of the images of the CIRS torso phantom was visible, similarly as described for Varian OBI. MV CBCT Figure 5 presents results for Siemens MVision MV CBCT and RMI 463 phantom. The relationship differed from those obtained for general-purpose CT scanners, especially for materials of high electron density. Some dependence on choice parameters (MU, reconstruction kernel) was observed. 330 Strahlenther Onkol 2010 Nr. 6

5 titanium was 4,000, which is basically the maximum HU value used in that CT unit. Figure 4. Results of measurements with CIRS 062 phantom on two kv CBCT systems: Varian OBI (125 kv) and Nucletron Simulix Evolution (100 kv). Relationships implemented in two TPS shown for comparison. Abbildung 4. Messergebnisse mit dem CIRS-062-Phantom an zwei kv-cbct-systemen: Varian OBI (125 kv) und Nucletron Simulix Evolution (100 kv). Die Relationen wurden in zwei TPS implementiert und zum Vergleich dargestellt. Figure 5. Results of measurements with RMI 463 phantom for Siemens MVision MV CBCT system operating at 6 MV. Relationships implemented in two TPS shown for comparison. Abbildung 5. Messergebnisse mit dem RMI-463-Phantom des MV-CBCT-Systems Siemens MVision bei 6 MV. Die Relationen wurden in zwei TPS implementiert und zum Vergleich dargestellt. Handling of High-Density Materials (Titanium) Despite some visible artifacts, small-diameter (ca. 1 cm) titanium insert in the RMI 465 phantom did not significantly disturb the HU values for tissue-equivalent materials on general-purpose scanners. The result could possibly be different for a larger volume of high-density material (e.g., phantom simulating hip prosthesis with metal alloys). For MV CBCT, the titanium did not introduce any artifacts. It is worth noting that for the GE HiSpeed scanner, the HU value obtained for Discussion Table 4 presents maximum differences between real values of (as given by the manufacturers of the phantoms) and calculated from the measured CT numbers with use of unmodified Cadplan calibration curve. The criteria suggested by ESTRO [13] were generally fulfilled for general-purpose CT scanners operating at 120 kv or 140 kv and TPS using predefined relationships. Only for some specific combinations of scanner, phantom, and parameters, the differences for materials of high density (bone with of 1.47) reached This was outside the tolerance, as the differences should not exceed Anyway, it should be remembered that for slightly higher ( 1.5), higher differences are allowed (0.10). It should also be noted, that for the same scanners and the same voltages (but different phantom, or different settings) the differences were lower than All the results obtained for general-purpose CT scanners operating at 120 kv or 140 kv fall within tolerance levels proposed by Kilby et al. [10], which are based on calculation of impact of inaccuracies on dose calculations. The results suggest that relationships implemented in TPS can generally be used for general-purpose CT systems operating at voltages close to 120 kv. Data measured with electron density phantom can, of course, be used to modify the relationship (provided that the TPS allows it), however, some uncertainty of data will remain as there is no obvious way to eliminate dependencies on size and shape of the phantom (or patient). The results for the radiotherapy simulator were not very different from those obtained for general-purpose CT scanners, however, the information on was much less precise because of the decidedly worse uniformity of images. Also for two kv CBCT systems, very serious nonuniformity of images was observed for large phantoms (and patients, e.g., in pelvic cases). All three systems could probably be used as a Table 4. Maximum differences between electron density calculated from the measured computed tomography (CT) numbers (with use of unmodified Cadplan calibration curve) and true values of electron density. Tabelle 4. Maximale Unterschiede zwischen Elektronendichte, berechnet aus den gemessenen Computertomographie-(CT-)Zahlen (unter Nutzung der unmodifizierten Cadplan-Eichkurve) und den wahren Werten der Elektronendichte. System Maximum error of calculated Lung Soft tissue Bone General-purpose CT, 80 kv General-purpose CT, 120/140 kv Varian Ximatron Nucletron Simulix Evolution Varian OBI Siemens MVision Strahlenther Onkol 2010 Nr

6 source for data. However, in all cases some work should be done first to correct the nonuniformity of the images. For MV CBCT, the results were significantly different than for kv systems due to the different energy spectrum of the beam. MV CT numbers could be potentially a good source of data for treatment planning, as they are closely correlated with and represent attenuation of the therapeutic beam [25]. Smooth handling of high-density materials (such as titanium) is another advantage. On the other hand, image fusion with other imaging modalities could be necessary to obtain anatomic data for treatment planning, because of limited visualization of soft tissue in MV CBCT images (as compared to kv CT). However, it is known that image quality can be substantially improved by correction of uniformity and by optimization of system settings [14]. Unfortunately, it was not possible to measure the HU- relation for each CT scanner for a full range of selectable parameters and with use of all phantoms. The most extensive measurements were performed for general-purpose CT scanners. The results showed that the predefined relationships can be used for those scanners, even if some variability of the results on choice of phantoms was observed for bone-equivalent materials. For nontypical CT scanners (radiotherapy simulators, CBCT systems), less extensive evaluations were done, in some cases with only one phantom, or with only one set of acquisition parameters. Nevertheless, this was enough to show that inaccuracies of values are obviously larger than those observed for general-purpose CT scanners, and that they occur in the whole range of electron densities. Conclusion The results obtained for two TPS showed that the HU- relationships predefined in TPS can be used for general-purpose CT systems operating at voltages close to 120 kv. We, how ever, advise to check the accuracy of values calculated by TPS for the particular CT scanner and scanning protocol. For nontypical imaging systems (e.g., CBCT), the HU- relationship can differ significantly from the predefined ones and, therefore, it should be measured and carefully analyzed before using CT data for treatment planning. Some uncertainty of data is always unavoidable, as even for general-purpose CT scanners, the HU values for a given tissue can differ depending on the dimensions of the scanned object, either phantom or patient. References 1. Bral S, Vinh-Hung V, Everaert H. The use of molecular imaging to evaluate radiation fields in the adjuvant setting of breast cancer. A feasibility study. Strahlenther Onkol 2008;184: Burridge NA, Rowbottom CG, Burt PA. Effect of contrast-enhanced CT scans on heterogeneity corrected dose computations in the lung. J Appl Clin Med Phys 2006;7: Constantinou C, Harrington JC, DeWerd LA. An electron density calibration phantom for CT-based treatment planning computers. Med Phys 1992;19: Cozzi L, Fogliata A, Buffa F, Bieri S. Dosimetric impact of computed tomography calibration on a commercial treatment planning system for external radiation therapy. Radiother Oncol 1998;48: Georg D, Georg P, Hillbrand M, et al. Assessment of improved organ at risk sparing for advanced cervix carcinoma utilizing precision radiotherapy techniques. Strahlenther Onkol 2008;184: Goerlitz E, Albers D, Cremers F, Schmidt R. Influence of the CT calibration process on the dose computation in radiotherapy treatment planning. Proceedings of the International Conference on Quality Assurance and New Techniques in Radiation Medicine, Vienna, November 1315, 2006: Guan H, Fang-Fang Y, Kim JH. Accuracy of inhomogeneity correction in photon radiotherapy from CT scans with different settings. Phys Med Biol 2002;47:N International Commission on Radiation Units and Measurements. ICRU report 44: tissue substitutes in radiation dosimetry and measurement. Bethesda: ICRU, Jessen KA. Application of CT in treatment planning. Proceedings of the College on Medical Physics: Radiation Protection and Imaging Technologies; Trieste, Sep 523, Kilby W, Sage J, Rabett V. Tolerance levels for quality assurance of electron density values generated from CT in radiotherapy treatment planning. Phys Med Biol 2002;47: Knöös T, Nilsson M, Ahlgren L. A method for conversion of Hounsfield number to electron density and prediction of macroscopic pair production cross section. Radiother Oncol 1986;5: Lo T, Yang Y, Schreibmann E, et al. Mapping electron density distribution from planning CT to cone-beam CT (CBCT): a novel strategy for accurate dose calculation based on CBCT. Int J Radiat Oncol Biol Phys 2005;63: S Mijnheer B, Olszewska A, Fiorino C, et al. Quality assurance of treatment planning systems practical examples for non-imrt photon beams. ESTRO booklet nr 7. Brussels: ESTRO, Morin O, Aubry JF, Aubin M, et al. Physical performance and image optimization of megavoltage cone-beam CT. Med Phys 2009;36: Morin O, Chen J, Aubin M, et al. Dose calculation using megavoltage cone-beam CT. Int J Radiat Oncol Biol Phys 2007;67: Nucletron. Oncentra MasterPlan v1.5 SP1 physics reference manual ENG-02. Veenendaal: Nucletron, Parker RP, Hobday PA, Cassell KJ. The direct use of CT numbers in radiotherapy dosage calculations for inhomogeneous media. Phys Med Biol 1979;24: Pech M, Mohnike K, Wieners G, et al. Radiotherapy of liver metastases. Comparison of target volumes and dose-volume histograms employing CT- or MRI-based treatment planning. Strahlenther Onkol 2008;184: Qamhiyeh S. A Monte Carlo study of the accuracy of CT-numbers for range calculations in carbon ion therapy. PhD Thesis, Ruperto-Carola University of Heidelberg, Seco J, Evans PM. Assessing the effect of electron density in photon dose calculations. Med Phys 2006;33: Siewerdsen JH, Jaffray DA. Cone-beam computed tomography with a flat-panel imager: magnitude and effects of x-ray scatter. Med Phys 2001;28: Sterzing F, Schubert K, Sroka-Perez, et al. Helical tomotherapy. Experiences of the first 150 patients in Heidelberg. Strahlenther Onkol 2008;184: Sua Y, Fang-Fang Y. Dosimetric feasibility of cone-beam CT-based treatment planning compared to CT-based treatment planning. Int J Radiat Oncol Biol Phys 2006;66: Thomas SJ. Relative electron density calibration of CT scanners for radiotherapy treatment planning. Br J Radiol 1999;72: Strahlenther Onkol 2010 Nr. 6

7 25. Thomas THM, Devakumar D, Purnima S, et al. The adaptation of megavoltage cone beam CT for use in standard radiotherapy treatment planning. Phys Med Biol 2009;54: Vandecasteele K, De Neve W, De Gersem W. Intensity-modulated arc therapy with simultaneous integrated boost in the treatment of primary irresectable cervical cancer. Treatment planning, quality control, and clinical implementation. Strahlenther Onkol 2009;185: Varian Oncology Systems. Cadplan external beam modeling physics, manual. Palo Alto: Varian Oncology Systems, Zabel-du Bois A, Ackermann B, Hauswald H, et al. Influence of intravenous contrast agent on dose calculation in 3-D treatment planning for radiosurgery of cerebral arteriovenous malformations. Strahlenther Onkol 2009;185: Address for Correspondence Witold Skrzyński Medical Physics Department Center of Oncology Roentgena Warsaw Poland Phone/Fax (+48/22) w.skrzynski@zfm.coi.pl Strahlenther Onkol 2010 Nr