Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Potentials of on-line repositioning based on implanted fiducial markers and electronic portal imaging in prostate cancer radiotherapy

Potentials of on-line repositioning based on implanted fiducial markers and electronic portal... Background: To evaluate the benefit of an on-line correction protocol based on implanted markers and weekly portal imaging in external beam radiotherapy of prostate cancer. To compare the use of bony anatomy versus implanted markers for calculation of setup-error plus/minus prostate movement. To estimate the error reduction (and the corresponding margin reduction) by reducing the total error to 3 mm once a week, three times per week or every treatment day. Methods: 23 patients had three to five, 2.5 mm Ø spherical gold markers transrectally inserted into the prostate before radiotherapy. Verification and correction of treatment position by analysis of orthogonal portal images was performed on a weekly basis. We registered with respect to the bony contours (setup error) and to the marker position (prostate motion) and determined the total error. The systematic and random errors are specified. Positioning correction was applied with a threshold of 5 mm displacement. Results: The systematic error (1 standard deviation [SD]) in left-right (LR), superior-inferior (SI) and anterior-posterior (AP) direction contributes for the setup 1.6 mm, 2.1 mm and 2.4 mm and for prostate motion 1.1 mm, 1.9 mm and 2.3 mm. The random error (1 SD) in LR, SI and AP direction amounts for the setup 2.3 mm, 2.7 mm and 2.7 mm and for motion 1.4 mm, 2.3 mm and 2.7 mm. The resulting total error suggests margins of 7.0 mm (LR), 9.5 mm (SI) and 9.5 mm (AP) between clinical target volume (CTV) and planning target volume (PTV). After correction once a week the margins were lowered to 6.7, 8.2 and 8.7 mm and furthermore down to 4.9, 5.1 and 4.8 mm after correcting every treatment day. Conclusion: Prostate movement relative to adjacent bony anatomy is significant and contributes substantially to the target position variability. Performing on-line setup correction using implanted radioopaque markers and megavoltage radiography results in reduced treatment margins depending on the online imaging protocol (once a week or more frequently). Page 1 of 9 (page number not for citation purposes) Radiation Oncology 2009, 4:13 http://www.ro-journal.com/content/4/1/13 formed weekly for 23 patients with histologically Background There is evidence that dose-escalation in definitive radio- confirmed prostate cancer treated from 1996 to 2000. The majority of patients were treated by a standard irradiation therapy of prostate cancer improves long-term PSA control [1]. One strategy to reduce late side effects is employment regimen in combination with regional hyperthermia in a of gradually smaller radiation field sizes or planning target phase II study as previously described [15]. Informed con- volumes PTV [2]. Tight margins will decrease the volume sent had been obtained from all patients. dose delivered to organs at risk, thus increasing the thera- peutic ratio of tumor control probability versus normal tis- Before treatment planning, three to five spherical gold sue complication probability (TCP/NTCP). On the other (99.9% Au) markers with a diameter of 2.0 mm were hand, this ratio might decline if the clinical target volume inserted transrectally into the prostate of each patient is partially missed by any positioning error not compen- using a modified biopsy needle under ultrasound guid- sated by the specified safety margins [3]. ance and local anaesthesia. Usually three markers were implanted, one into the apex, and two into the superior Retrospective evaluations [4,5] have suggested that ana- lateral parts of the prostate. Gold markers of this size can tomic variations (rectal distension, large rectum) during be visualized using megavoltage beam detector systems of the planning CT in fact reduce the PSA control. A large the first generation. No complications occurred in associ- (distended) rectum during planning can cause a system- ation with the implantation process as reported elsewhere atic error, because it places the prostate more anterior, but [10]. Note that the gold markers presently applied with kV this location might change from fraction to fraction. X-ray tracking systems are < 1 mm in diameter and the Another study did not confirm a correlation between rec- implantation procedure is easier and more feasible. tal and/or bladder distension and errors of prostate posi- tion [6]. Nevertheless, we assume that image-guidance is Each patient underwent a computerized tomography scan crucial and improves the clinical outcome. (CT) (Siemens™, Erlangen, Germany) for treatment plan- ning in treatment position from 2 cm below the ischial An assessment of patient position is based on skeletal tuberosities to the L4/5 interspace obtaining volumetric landmarks imaged by electronic portal imaging devices data at 5 mm slice thickness and at a 5 mm couch transla- (EPID). They are commonly used for the evaluation and tion. In our study, the patients were instructed to fill the correction of set-up deviations [7]. bladder, but no effort was made to control the rectal vol- ume. However, the CT scans were repeated if excessive fill- As documented in a number of studies [8,9], an interfrac- ing of the rectum had been noticed. Patients were tional displacement of the prostate itself can occur during stabilized in supine position with conventional head, radiation therapy fractions relative to the bony structures knee and feet support and no rigid immobilization device of the pelvis. The feasibility of implanting markers for was used. Images were transferred to a workstation localization of the prostate recently has been demon- (Helax™) for anatomic segmentation of targets and organs strated [10,11] and allows to utilize EPIDs to quantify the at risk and conformal dosimetric planning. The PTV was displacement of the target [12,13]. With the improvement defined by a three-dimensional expansion of the CTV by of online imaging quality, pretreatment localization and 8 mm at the prostate-rectum interface and 10 mm in all online protocols allowing positioning corrections with- other directions. External beam radiotherapy was per- out significant delay have gained feasibility [14]. formed by a linear accelerator (Siemens™ Mevatron KD, Erlangen, Germany) with a beam energy of 18 MV using From the comparison of verification protocols during fractions of 1.8 Gy five times weekly up to 68.4 – 72 Gy radiotherapy it is known, that the treatment margins are (38–40 fractions) at the reference point (ICRU-50,16). An institution specific. We performed a prospective study of isocentric 4-field box technique consisting of anterior, patients treated with conformal radiotherapy for prostate posterior and two lateral fields (0°, 180°, 90° and 270°) cancer, analysing both internal organ motion and setup was used in all cases. error with the objective to quantify the variability in pros- tate position. For displacements of bones and markers, All conformal 3D-plans were conventionally simulated statistical data including overall, systematic and random before treatment. Simulator radiographs had been deviations were determined. From the uncorrected and obtained in orthogonal (0°, 90°) projections and served corrected total errors, we calculated the necessary treat- as reference images for the position of bony landmarks ment margins to ensure sufficient target coverage in the and internal markers. majority of cases. On each treatment day, patients were positioned using Patients and methods laser alignment to reference marks (one anterior and two Verification and correction of treatment position by anal- lateral set-up crosses) on their skin. For all patients, ysis of portal images and simulator control films were per- weekly pre-treatment position verification with an EPID Page 2 of 9 (page number not for citation purposes) Radiation Oncology 2009, 4:13 http://www.ro-journal.com/content/4/1/13 system (Siemens Beamview Plus™, Erlangen, Germany) reference images simulator radiographs and, secondly, the was applied [17]. The electronic portal images (EPIs) for corresponding portal images. The AP beam provided data verification were acquired using 6 MV photons for the AP to detect the position of the landmarks and markers in the (0°) and left lateral (90°) fields once a week with 6 – 8 LR and SI direction and the lateral beam for the AP direc- monitor units (MU), each before starting irradiation. The tion and SI direction as well. To identify the position of images were digitally processed (employing high fre- the target m, we used the arithmetic mean of the marker quency filters) to facilitate recognition of bony structures coordinates according to the isocenter (Fig. 2). All meas- and radiopaque markers. On EPIs, the isocenter has been urements were performed by the same author (RG). The made visible by the projection of an isocenter marker (a consistency of the obtained deviations was tested by cor- 1.5 × 3 mm gold pin) located on the reticule. Bony land- relation of the corresponding values in SI direction taken marks and implanted markers were clearly identified on from 0° and 90° projection. The correlation coefficient of almost all portal films (Figure 1). r = 0.86 was satisfactory. The registration procedure takes about 3 minutes cumulating to a total treatment time of For the applied 2D/3D registration method, isocenter, 6–8 minutes on average. bony contours and fiducial markers were drawn from the simulator films on transparent templates for every patient The evaluation procedure and the nomenclature are sum- before irradiation. These templates were then used to marized in Figure 2. Firstly, we determined the vectorial match the reference images (0°, 90°) to the correspond- displacements of the isocenters relative to the bony anat- ing verification images manually. omy of the reference images Δs for j = 1...23 patients and ij i = 1...8 weekly portal images per patient during the radi- An identical scale of the printed portal images and the otherapy course yielding 8 × 23 = 184 setup errors (under- templates was applied to determine the setup errors from lining identifying a vector). Secondly, the differences of the shift Δs of the isocenter (see Figure 2 for definition of the marker positions relative to the isocenters result in the symbols). The components of the vector Δs according to prostate motion Δm . Finally, the total displacement ij the main axes are determined providing the shift in left- (setup error plus organ motion) of the target relative to right (LR) direction (lateral x-axis), anterior-posterior the isocenter is calculated by Δtot = Δs + Δm . ij ij ij (AP) direction (vertical y-axis) and superior-inferior (SI) direction (longitudinal z-axis). For all 184 fractions, mean and standard deviations for all kinds of errors (setup, motion, total) were calculated. We For evaluation and quantification of uncertainties, two analysed the error distributions averaging over all frac- orthogonal sets of 2D projections were available, firstly as tions and patients. image Figure 1 Reference image (simulator film, left) and online image (port film, in the mid) are registered by the method shown on the right Reference image (simulator film, left) and online image (port film, in the mid) are registered by the method shown on the right image. A template containing isocenter, bony contours and radio-opaque markers is traced from the simulator radiograph and positioned on the portal image with isocenters and main axes in coincidence. The isocenter is shifted until the bony contours (setup error) or the implanted markers are in agreement (total error). For the motion error we deter- mine the shift from the setup corrected position to the marker corrected position. The correction method is two-dimensional and performed separately for each projection (0° and 90°). Redundant measurements (in SI direction) are in good correlation (see text). Page 3 of 9 (page number not for citation purposes) Radiation Oncology 2009, 4:13 http://www.ro-journal.com/content/4/1/13 (Σ) and 0.7 times the standard deviation of the random error (σ) to ensure a minimum dose of 95% to the clinical target volume for 90% of the fractions, i.e. allowing signif- icant dose discrepancies in = 10% of sessions. If a position correction was performed (above the action level), we assume a residual error of = 3 mm [19] in all directions for the corrected fraction. Statistical analysis was performed using JMP v7.0 (SAS Institute, Cary, NC, USA). Tests for sub-groups were per- formed using the paired t-Test. Results We performed the analysis for 23 patients with 8 pairs of Basi Figure 2 c definitions of the different error components EPIs per patient, summing up to a total of 368 anterior- Basic definitions of the different error components: posterior and lateral port films in184 fractions. Bony con- The setup error Δs and the motion error Δm can be added tours, implanted markers and isocenter marker were to the total error Δtot. For every patient j = 1...23 and portal clearly visible and evaluable in 96% of cases. All portal image i = 1...8 the setup errors Δs are determined by match- ij images were evaluable with respect to prostate motion ing the bony contours of the portal images (i, j) to the refer- employing the radiopaque markers. We had to replace ence images j (simulator radiographs) according Figure 1. The only 1.8% of portal images due to insufficient identifica- motion errors Δm are determined after these matching pro- ij cedures by subtracting the marker positions of the matched tion of the bony structures. portal images (i, j) and the reference images j. The systematic error for a patient j is defined as the mean of single errors As summarised in Table 1 we analysed all 184 fractions with respect to i = 1...8 portal images. The random error of together and determined the displacements of the iso- patient j is defined as the standard deviation of this series. center relative the bony anatomy (setup error), the dis- Then further statistical evaluation is possible, e.g. the mean placement of the markers relative to the bony structures systematic error for a series of patients (here 23 patients) in (prostate motion) and the displacement of the isocenter Table 2 and the mean standard deviations (mean random relative to the markers (combined or total targeting error). error) in Table 3. Figure 3 shows the measured deviations in a box plot for- mat, indicating mean values, median values and selected percentiles from 10 to 90% (10%, 25%, 75%, 90%) in LR, Then, we determined means and standard deviations SI and AP directions. The observed errors were greatest in from 8 control EPIs for each patient resulting in the same the AP direction, where a range of 13 mm is found for the error types Δs(j), Δm(j) and Δtot(j) for j = 1...23 patients, total deviation of the target (-7 to +6 mm) for 80% of the and analysed the error distributions with respect to the controls. The extremes observed in internal target motion patients. The standard deviations identify the systematic were 8 mm in AP and 7 mm in SI direction. errors Σ(j) for every patient. We calculated the various errors for every patient sepa- Random errors σ(j) for every patient j were calculated as rately (averaging over eight controls) and analysed the standard deviations of the differences Δs(j) - Δs or Δm(j) error distribution for 23 patients (see Table 2). Both, sys- ij - Δm or Δtot(j) - Δtot averaging over i = 1...8 PIs. We can tematic setup and motion errors are in the range of ± 2 ij ij also determine the mean random error for the entire mm (± 1 SD) summing up to a total error of ± 3 mm. group of patients averaging σ(j) over all j = 1...23 patients. For all 8 fractions per patient, the scatter (SD) about the For correction of translational errors before treatment, we individual averages (systematic errors) has been calcu- used an action level of 5 mm, i.e. all errors of 5 and more lated providing the random errors for the different error mm were corrected. The correction was performed on-line types (Table 3). As expected the random errors are larger by repositioning the target according to the internal mark- than the systematic errors and finally amount to total ran- ers, moving the treatment couch manually. To calculate dom errors of ± 4 mm in SI and AP direction, but only ± the minimum required margin width around the clinical 3 mm in LR direction. The extremes can approach 1–1.5 target volume (CTV + margin = PTV), we utilized the pre- cm, but these are rare cases. scription suggested by van Herk [18]. The margin around the clinical target volume (CTV) should be the sum of 2.5 The online protocol was applied at least 8 times per times the standard deviation of the systematic total error patient and, on average, 56% of the controls determined Page 4 of 9 (page number not for citation purposes) Radiation Oncology 2009, 4:13 http://www.ro-journal.com/content/4/1/13 Table 1: Setup error, motion error and total error Mean ± SD [mm] i = 1,..., 8; j = 1,..., 23 Range [mm] Setup error Δs ij Left-right 0.8 ± 2.8 -8/10 Superior-inferior 0.1 ± 3.4 -9/14 Anterior-posterior -1.2 ± 3.6 -15/9 Prostate movement Δm ij Left-right -0.3 ± 1.8 -6/9 Superior-inferior 0.9 ± 2.8 -9/8 Anterior-posterior 0.3 ± 3.5 -10/10 Total error Δtot ij Left-right 0.5 ± 3.5 -10/19 Superior-inferior 0.9 ± 4.4 -14/13 Anterior-posterior -0.8 ± 4.9 -19/14 Measured deviations for all 184 (= 23 × 8) portal imaging controls (averaged over i = 1,..., 8 weekly portal images in j = 1,..., 23 patients). Mean values <...> in all three directions (left-right LR, superior-inferior SI, anterior-posterior AP) ± standard deviations SD and ranges, i.e. extremal deviations in ± orientations of the axes, in LR, SI and AP direction for setup error <Δs >, prostate movement <Δm > and total errror <Δtot > are ij ij ij listed (see Fig. 2 for explanation of symbols). displacements of = 5 mm in at least one direction and had gin of 1 cm around the CTV is sufficient to counterbalance to be corrected. Margin calculations have been performed the set-up and internal motion inaccuracies if a weekly for each of the axes according to the prescription of van portal imaging with online correction is presumed. Grad- Herk [18] as described in section 2 (see Table 4). To con- ual reduction of the errors and derived margins down to a sider for the total targeting error, the margins added to the minimum of 5 mm is obtained if the frequency of online CTV must be as large as 7.0 mm (LR) and 9.5 mm (SI, AP). control is further increased up to a daily correction as summarized in Table 4. After position corrections once a week, these calculated margins reduce to 6.7, 8.2 and 8.7 mm. Therefore, a mar- Discussion Various techniques have been developed to locate the prostate position on-line such as implanted fiducials (detected by X-rays), transabdominal ultrasound [20], electromagnetic tracking [21] and several kinds of in- room CT (e.g. [22]), in particular in conjunction with hel- ical tomotherapy [23]. However, the highest precision is achieved by using intraprostatic markers. The clinical use of implanted gold markers was found to be feasible in our hands. The geometrical center of implanted radiopaque markers characterizes the prostate position. Several groups have investigated the possibility of seeds migration and have found no or only little motion [24,25]. In addition, the reliability of markers for the location of the prostate has been questioned because of interfraction rotation or deformation [26], but these factors leave the prostate dosimetry unaffected [27]. The analysis is standardized so that the interobserver variabil- ity is low. Therefore implanted markers and EPID based 2 izo tion variability an Figure 3 Measured format, showin values (lines in th 5% (bottom of box), 75% ntal line) percenti deviations in g the mean d total error e bo le s x) and th LR, SI plit in (top of b values and A setu e 10% (bla p variability, prostat P o (lower hori directions ck x) and 90% squares)in a , th (upper zontal line), box plot e medi e posi- hor an - Measured deviations in LR, SI and AP directions in a methods are used for targeting in radiotherapy of prostate box plot format, showing the mean values (black cancer with increasing frequency. squares), the median values (lines in the box) and the 10% (lower horizontal line), 25% (bottom of box), Our results provide information about the scatter of target 75% (top of box) and 90% (upper horizontal line) per- positions during radiotherapy. Setup inaccuracies were centile split in setup variability, prostate position var- reviewed by Hurkmans [28]. In his analyses data were iability and total error. obtained from repeated simulations, from EPID studies Page 5 of 9 (page number not for citation purposes) Radiation Oncology 2009, 4:13 http://www.ro-journal.com/content/4/1/13 Table 2: Systematic errors Mean ± SD [mm] j = 1,..., 23 Range [mm] j = 1,..., 23 Systematic setup error <Δs> Left-right 0.8 ± 1.6 -2.9/3.9 Superior-inferior 0.1 ± 2.1 -2.7/6.2 Anterior-posterior -1.2 ± 2.4 -5/4.1 Systematic prostate movement <Δm> Left-right -0.3 ± 1.1 -2.9/2.7 Superior-inferior 0.9 ± 1.9 -3.1/4.7 Anterior-posterior 0.3 ± 2.3 -4.3/5.0 Systematic total error <Δtot> Left-right 0.5 ± 2.0 -2.6/6.6 Superior-inferior 0.9 ± 2.7 -4.3/7.3 Anterior-posterior -0.8 ± 2.6 -6.7/4.0 Analysis of the mean vectorial setup errors per patient (after averaging over i = 1,..., 8 weekly controls) (1/8)Σ Δs =: <Δs> , the mean vectorial i = 1...8 ij j prostate movements (1/8)Σ Δm =: <Δm> and the mean vectorial total errors (1/8)Σ Δtot =: <Δtot> . After statistical analysis with i = 1...8 ij j i = 1...8 ij j respect to j = 1 to 23 patients we achieve the population based (mean) vectorial systematic error defined as (1/23)Σ <Δs> , (1/23)Σ j = 1...23 j j = <Δm> and (1/23)Σ <Δtot> splitted in the RL, SI and AP direction and the standard deviation SD from these errors (see Fig. 2 for 1...23 j j = 1...23 j explanation of symbols). Also the range (minimum to maximum) is listed (right row). and from repeated CT scans. The standard deviations of the the prostate at the time of treatment can be visualized setup errors ranged from 1 to 4 mm, which is in accordance with a variety of techniques, and differences in measure- with our results. We also found standard deviations below ment techniques make it difficult to compare the results 4 mm. Analysis of the contributions to the total targeting of published studies. In summary, the SDs of the prostate error indicates, that the setup errors cause approximately motion range in the LR direction from 0.7 to 1.9 mm, in one half of the entire target position variability and offers a SI from 1.7 to 3.6 and for AP from 1.5 to 4.0 mm. We potential improvement in total target positioning. measured for prostate motion in RL, SI and AP standard deviations of 1.8, 2.8 and 3.5 mm, even though some The prostate position can move relative to the skeleton extremes of motion were registered in a few patients (table [4]. An overview of interfraction prostate motion studies 1). Thus, our results are in general agreement with litera- was presented in a paper by Langen [29]. The position of ture [30-34]. Table 3: Random errors Mean [mm] j = 1...23 Range [mm] j = 1...23 Random setup error Left-right 2.3 -7.9/7.4 Superior-inferior 2.7 -8.6/7.8 Anterior-posterior 2.7 -12.1/5.4 Random prostate movement Left-right 1.4 -5.5/6.3 Superior-inferior 2.3 -5.9/8.6 Anterior-posterior 2.7 -7.0/8.0 Random total error Left-right 2.9 -8.0/6.6 Superior-inferior 3.9 -11.5/8.2 Anterior-posterior 4.3 -14.9/11.2 The standard deviations of measured deviations Δs , Δm , Δtot around the systematic errors (of each patient j = 1...23) <Δs> , <Δm> , <Δtot> after ij ij ij j j j 2 2 averaging over i = 1,..., 8 weekly controls provide the random errors sqrt(1/7Σ (<Δs> - Δs ) ), sqrt(1/7Σ (<Δm > - Δm ) ), sqrt(1/7Σ i = 1..8 j ij i = 1..8 i j ij i = (<Δtot > - Δtot ) ) for patient j = 1,..., 23. They are splitted into the RL, SI and AP direction for every error type, i.e. setup error, prostate 1..8 i j ij movement and total error (see Fig. 1 for explanation of symbols). The mean random errors after averaging over 23 patients and their range, i.e. maximum deviations in both orientations of the axes, are given in the table. Page 6 of 9 (page number not for citation purposes) Radiation Oncology 2009, 4:13 http://www.ro-journal.com/content/4/1/13 Table 4: Estimation of margins. Random σ and systematic Σ error [mm] Margin [mm] Direction LR SI AP LR SI AP ΣσΣσΣσ No correction 2.0 2.9 2.7 3.9 2.6 4.3 7.0 9.5 9.5 Correction 1×/week 1.9 2.8 2.3 3.5 2.4 3.9 6.7 8.2 8.7 Correction 3×/week 1.6 2.5 1.8 3.0 1.7 3.3 5.8 6.6 7.7 Correction 5×/week 1.4 2.0 1.4 2.3 1.3 2.2 4.9 5.1 4.8 For our patient group the standard deviations of systematic errors (Σ) and random errors (σ) are listed in LR, SI and AP directions. Then the margins from CTV to PTV are calculated in all three directions according to the prescription 2.5 × Σ + 0.7 × σ [14] in order to consider these errors. The errors are summarized in the first line as derived from the weekly portal images without any correction. In the following lines the variation is reduced by controling and correcting the position with increasing frequency per week (once, three times, every day). Correction is conducted if the threshold of 5 mm is crossed in any direction. After correction we assume a residual error of 3 mm in the respective direction. Under these circumstances the margins can be diminished from approximately 10 mm to 5 mm. We found the largest errors, for both, setup as well as pros- respectively. We found for setup, prostate location varia- tate motion, in the AP direction, followed by SI and LR tion and combined error in general larger random errors directions in accordance with the series of Beaulieu and than systemic errors, obviously due to the reduction of others [14,29,35]. Along the lateral axis the prostate is systematic errors by the weekly performed corrections. confined within the pelvis and published data show only small deviations in this direction. In our study, the distri- In our study we calculated necessary CTV-PTV margins bution of organ motion and setup errors for translation is (without correction) of 7.0 to 9.5 mm (RL, SI and AP in the range of the published values [36], e.g. 90% of the direction) according Table 4. Similar margins (without observed displacements were 7 mm or less. correction) are reported by Kupelian [38] with 10, 10 and 12 mm, McNair [40] with 5, 7.5 and 11 mm and van den Interfraction position variation of the prostate as a source Heuvel [41] with 9.5, 8.6 and 10 mm. of treatment error is mainly caused by variable fillings of the bladder and/or rectum that displace the prostate According to the formula given in Section 2 to estimate mainly in SI and AP direction as shown by magnetic reso- the margin between CTV and PTV [18], systematic errors nance imaging of the pelvis [27]. Patient instructions have the largest impact on the size of PTV margins. There- attempt to prepare rectal and bladder distension in a fore, offline correction protocols attempt to determine standardized way before treatment. This may reduce the and correct the systematic error. They have the advantage frequency of large prostate movements, but does not elim- to be effective despite a low imaging frequency. Different inate the motion error [21]. There is even an intrafrac- offline protocols have been successful implemented into tional motion of 1–3 mm on average [37] and after initial clinical practice [42,43]. On the other hand, Litzenberg positioning the displacement of the prostate gland [44] figured out, that because of changes in patient's setup increases with elapsed time. This matter raises concerns characteristics off-line protocols, especially those directed with regard to correction for misalignments [38] and the to localize the prostate using markers did not show any treatment time of 20–30 minutes per session using novel significant benefit in reducing the total error of implanted techniques i.e. intensity modulated radiotherapy, tomo- fiducial gold markers in 10 prostate cancer patients in therapy etc., which will induce a new intrafraction errors. comparison to daily online position correction. For the Recently published analyses of this issue indicate that a 3- same reasons, applying these methods directly to the mm planning target margin is in most cases sufficient to implanted markers also gave larger residual errors than account for intrafractional motion [39]. expected. It may be difficult to identify patients who would benefit from off-line protocols and those who may Both uncertainties, setup error and target motion can be require daily on-line corrections [44]. split into random and systematic deviations. The system- atic component of setup error is largely caused by the sys- Evaluating their possible benefit, on-line correction pro- tematic error inherent to the use of a reference image tocols have the potential to reduce both systematic and obtained by use of the planning-CT. The random compo- random errors, but at the expense of increasing treatment nent of the setup error is mainly caused by uncertainties time per fraction. As expected, systematic errors are effec- from utilisation of skin markers, while the random error tively reduced with increasing imaging frequency [38]. of target position is mainly caused by organ movement, After one weekly online correction and 5 mm action level, Page 7 of 9 (page number not for citation purposes) Radiation Oncology 2009, 4:13 http://www.ro-journal.com/content/4/1/13 Late rectal bleeding after conformal radiotherapy of pros- we found margins of about 7 to 9 mm. These margins can tate cancer. II. Volume effects and dose-volume histograms. be further reduced to a minimum of 5 mm by increasing Int J Radiat Oncol Biol Phys 2001, 49:685-98. the control frequency (Table 4). Kupelian [38] calculated 4. de Courvoisier R, Tucker SL, Dong L, et al.: Increased risk of bio- chemical and local failure in patients with distended rectum treatment margins for 8 different potential non-daily on the planning CT for prostate cancer radiotherapy. Int J imaging strategies, among them low-workload weekly Radiat Oncol Biol Phys 2005, 62:965-973. 5. Heemsbergen WD, Hoogeman MS, Witte MG, Peeters ST, Incrocci protocols. For a weekly online protocol with 3-mm L, Lebesque JV: Increased risk of biochemical and clinical fail- threshold, he found margins of 8, 8 and 6 mm (LR, SI and ure for prostate patients with a large rectum at radiotherapy AP), which agrees quite well with our results. planning: Results from the Dutch Trial of 68 Gy versus 78 Gy. Int J Radiat Oncol Biol Phys 2007, 67:1418-1424. 6. Kim PJ, Parthasarathy A, Ho J, Tam B, Gee H, Lee JW: Impact of A daily positioning correction is feasible under routine rectal distension in IGRT for prostate cancer. Int J Radiat Oncol conditions employing the new generation of linear accel- Biol Phys 2007:S715. 7. Chen J, Lee RJ, Handrahan D, Sause WT: Intensity-modulated erators with image guidance (on-board imaging or x-ray radiotherapy using implanted fiducial markers with daily tracking). Using these techniques the residual error can be portal imaging: assessment of prostate organ motion. Int J Radiat Oncol Biol Phys 2007, 68:912-9. further decreased below 3 mm and the required safety 8. Schallenkamp JM, Herman MG, Kruse JJ, Pisansky TM: Prostate margin is reduced down to 3 mm (unpublished data). An position relative to pelvic bony anatomy based on intrapros- accuracy of only 5 mm is achieved using megavoltage CT tatic gold markers and electronic portal imaging. Int J Radiat Oncol Biol Phys 2005, 63:800-11. without intraprostatic markers [38]. 9. Lattanzi J, McNeeley S, Pinover W, Horwitz E, Das I, Schultheiss TE, Hanks GE: A comparison of daily CT localization to a daily ultrasound-based system in prostate cancer. Int J Radiat Oncol Conclusion Biol Phys 1999, 43:719-725. In summary, correction of setup errors alone is not suffi- 10. Shirato H, Harada T, Harabayashi T, Hida K, Endo H, Kitamura K, cient because target motion contributes significantly to Onimaru R, Yamazaki K, Kurauchi N, Shimizu T, Shinohara N, Matsu- shita M, Dosaka-Akita H, Miyasaka K: Feasibility of insertion/ positioning inaccuracies. The implantation of gold mark- implantation of 2.0-mm-diameter gold internal fiducial ers for a correction protocol was feasible in our study. A markers for precise setup and real-time tumor tracking in weekly on-line setup verification employing these radio- radiotherapy. Int J Radiat Oncol Biol Phys 2003, 56:240-7. 11. Henry AM, Wilkinson C, Wylie JP, Logue JP, Price P, Khoo VS: paque markers and megavoltage radiography results in Trans-perineal implantation of radio-opaque treatment ver- CTV-PTV margins of 7 to 8.5 mm. More effort can further- ification markers into the prostate: an assessment of proce- dure related morbidity, patient acceptability and accuracy. more decrease these margins. A correction of three times Radiother Oncol 2004, 73:57-9. per week leads to margins of 6 to 7.5 mm, and daily cor- 12. Litzenberg D, Dawson LA, Sandler H, Sanda MG, McShan DL, Ten- rections can further reduce the margin down to 5 mm. Haken RK, Lam KL, Brock KK, Balter JM: Daily prostate targeting using implanted radiopaque markers. Int J Radiat Oncol Biol Phys 2002, 52:699-703. Competing interests 13. Herman MG, Pisansky TM, Kruse JJ, Prisciandaro JI, Davis BJ, Geyer The authors declare that they have no competing interests. SM, King BF: Daily on-line positioning of the prostate for three-dimensional conformal radiotherapy (3D-CRT) using an electronic portal imaging device (EPID). Int J Radiat Oncol Authors' contributions Biol Phys 2003, 57:1131-40. 14. Beaulieu L, Girouard LM, Aubin S, Aubry JF, Brouard L, Roy-Lacroix RG analysed the simulator films and portal images, deter- L, Dumont J, Tremblay D, Laverdière J, Vigneault E: Performing mined the errors, performed the statistical analysis and daily prostate targeting with a standard V-EPID and an auto- drafted the manuscript. PW initiated the study, treated the mated radio-opaque marker detection algorithm. Radiother Oncol 2004, 73:61-64. patients, formulated the mathematical background and 15. Tilly W, Gellermann J, Graf R, Hildebrandt B, Weissbach L, Budach V, revised the first draft of the manuscript. VB participated in Felix R, Wust P: Regional hyperthermia in conjunction with designing the study and approved the treatment concepts. definitive radiotherapy against recurrent or locally advanced prostate cancer T3 pN0 M0. Strahlenther Onkol 2005, 181:35-41. DB coordinated the recruitment of patients and data 16. International Commission on Radiation Units and Measurements: acquisition. All authors participated in the critical discus- Prescribing, recording and reporting photon beam therapy. ICRU Report 50. Bethesda, MD; 1993. sion of the data and their statistical analysis. All authors 17. Gagel B, Schramm O, Harms W, Mulhern A, Wenz F, van Kampen M, improved the manuscript and approved the final version. Wannenmacher M, Eble MJ: The electronic portal imaging sys- tem Siemens Beamview Plus versus the conventional verifi- cation films CEA-TVS and DuPont COL-7. A critical Acknowledgements appraisal of visual image quality. Strahlenther Onkol 2002, The authors thank the Lieselotte-Beutel-Stiftung for the valuable support of 178:446-52. the prostate center. 18. Van Herk M, Remeijer P, Rasch C, Lebesque JV: The probability of correct target dosage: dose-population histograms for deriv- ing treatment margins in radiotherapy. Int J Radiat Oncol Biol References Phys 2000, 47:1121-35. 1. Kuban DA, Tucker SL, Dong L, Starkschall G, Huang EH, Cheung MR, 19. Poulsen PR, Muren LP, Høyer M: Residual set-up errors and mar- Lee AK, Pollack A: Long-term results of the M. D. Anderson gins in on-line image-guided prostate localization in radio- randomized dose-escalation trial for prostate cancer. Int J therapy. Radiother Oncol 2007, 85:201-6. Radiat Oncol Biol Phys 2008, 70:67-74. 20. Scarbrough TJ, Golden NM, Ting JY, Fuller CD, Wong A, Kupelian PA, 2. Zelefsky MJ, Fuks Z, Leibel SA: Intensity-modulated radiation Thomas CR Jr: Comparison of ultrasound and implanted seed therapy for prostate cancer. Semin Radiat Oncol 2002, 12:229-37. marker localisation methods: Implications for image-guided 3. Jackson A, Skwarchuk MW, Zelefsky MJ, Cowen DM, Venkatraman radiotherapy. Int J Radiat Oncol Biol Phys 2006, 65:378-387. ES, Levegrun S, Burman CM, Kutcher GJ, Fuks Z, Liebel SA, Ling CC: Page 8 of 9 (page number not for citation purposes) Radiation Oncology 2009, 4:13 http://www.ro-journal.com/content/4/1/13 21. Langen KM, Willoughby TR, Meeks SL, Santhanam A, Cunningham A, the potential benefit of online and offline verification proto- Levine L, Kupelian PA: Observations on real-time prostate cols for prostate radiotherapy. Int J Radiat Oncol Biol Phys 2008, gland motion using electromagnetic tracking. Int J Radiat Oncol 71:41-50. Biol Phys 2008, 71:1084-90. 41. Heuvel F Van den, Fugazzi J, Seppi E, Forman JD: Clinical applica- 22. Langen KM, Zhang Y, Andrews RD, Hurley ME, Meeks SL, Poole DO, tion of a repositioning scheme, using gold markers and elec- Willoughby TR, Kupelian PA: Initial experience with megavolt- tronic portal imaging. Radiother Oncol 2006, 79:94-100. age (MV) CT guidance for daily prostate alignments. Int J 42. De Boer HC, Heijmen BJ: A protocol for the reduction of sys- Radiat Oncol Biol Phys 2005, 62:1517-1524. tematic patient setup errors with minimal portal imaging 23. Kupelian PA, Lee C, Langen KM, Zeidan OA, Manon RR, Willoughby workload. Int J Radiat Oncol Biol Phys 2001, 50:1350-65. TR, Meeks SL: Evaluation of image-guidance strategies in the 43. de Boer HC, van Os MJ, Jansen PP, Heijmen BJ: Application of the treatment of localized prostate cancer. Int J Radiat Oncol Biol No Action Level (NAL) protocol to correct for prostate Phys 2008, 70:1151-1157. motion based on electronic portal imaging of implanted 24. Pouliot J, Aubin M, Langen KM, Liu YM, Pickett B, Shinohara K, Roach markers. Int J Radiat Oncol Biol Phys 2005, 61:969-983. M 3rd: (Non)-migration of radio-opaque prostate markers. 44. Litzenberg DW, Balter JM, Lam KL, Sandler HM, Ten Haken RK: Ret- Int J Radiat Oncol Biol Phys 2003, 56:862-6. rospective analysis of prostate cancer patients with 25. Poggi MM, Gant DA, Litzenberg D, Dawson LA, Sandler H, Sanda MG, implanted gold markers using off-line and adaptive therapy McShan DL, Ten-Haken RK, Lam KL, Brock KK, Balter JM: Marker protocols. Int J Radiat Oncol Biol Phys 2005, 63:123-33. seed migration. Int J Radiat Oncol Biol Phys 2002, 52:699-703. 26. Ghilezan MJ, Jaffray DA, Sieverdsen JH: Prostate gland motion assessed with cine-magnetic resonance imaging (cine-MRI). Int J Radiat Oncol Biol Phys 2005, 62:406-17. 27. Kupelian PA, Langen KM, Zeidan OA: Daily variations in delivered doses in patients treated with radiotherapy for localized prostate cancer. Int J Radiat Oncol Biol Phys 2006, 66:876-82. 28. Hurkmans CW, Remeijer P, Lebesque JV, Mijnheer BJ: Set-up veri- fication using portal imaging: review of current clinical prac- tice. Radiother Oncol 2001, 58:105-20. 29. Langen KM, Jones DTL: Organ Motion and its Management. Int J Radiat Oncol Biol Phys 2001, 50:265-78. 30. Crook JM, Raymond Y, Salhani D, Yang H, Esche B: Prostate motion during standard radiotherapy as assessed by fiducial markers. Radiother Oncol 1995, 37:35-42. 31. Nederveen AJ, Dehnad H, Heide UA van der, van Moorselaar RJ, Hof- man P, Lagendijk JJ: Comparison of megavoltage position veri- fication for prostate irradiation based on bony anatomy and implanted fiducials. Radiother Oncol 2003, 68:81-8. 32. Chung PW, Haycocks T, Brown T, Cambridge Z, Kelly V, Alasti H, Jaf- fray DA, Catton CN: On-line aSi portal imaging of implanted fiducial markers for the reduction of interfraction error dur- ing conformal radiotherapy of prostate carcinoma. Int J Radiat Oncol Biol Phys 2004, 60:329-34. 33. Dehnad H, Nederveen AJ, Heide UA van der, van Moorselaar RJ, Hof- man P, Lagendijk JJ: Clinical feasibility study for the use of implanted gold seeds in the prostate as reliable positioning markers during megavoltage irradiation. Radiother Oncol 2003, 67:295-302. 34. Wu J, Haycocks T, Alasti H, Ottewell G, Middlemiss N, Abdolell M, Warde P, Toi A, Catton C: Positioning errors and prostate motion during conformal prostate radiotherapy using on- line isocenter set-up verification and implanted prostate markers. Radiother Oncol 2001, 61:127-33. 35. Vigneault E, Pouliot J, Laverdiere J, Roy J, Dorion M: Electronic por- tal imaging device detection of radioopaque markers for the evaluation of prostate position during megavoltage irradia- tion: A clinical study. Int J Radiat Oncol Biol Phys 1997, 37:205-12. 36. Roeske JC, Forman JD, Mesina CF, He T, Pelizzari A, Fontenla E, Vijay- akumar S, Chen GT: Evaluation of changes in the size and loca- tion of the prostate, seminal vesicles, bladder, and rectum during a course of external beam radiation therapy. Int J Radiat Oncol Biol Phys 1995, 33:1321-9. 37. Smitsmans MH, Pos FJ, de Bois J, Heemsbergen WD, Sonke JJ, Leb- Publish with Bio Med Central and every esque JV, van Herk M: The influence of a dietary protocol on scientist can read your work free of charge cone beam CT-guided radiotherapy for prostate cancer patients. Int J Radiat Oncol Biol Phys 2008, 71:1279-86. "BioMed Central will be the most significant development for 38. Kupelian PA, Langen KM, Willoughby TR, Zeidan OA, Meeks SL: disseminating the results of biomedical researc h in our lifetime." Image-guided radiotherapy for localized prostate cancer: Sir Paul Nurse, Cancer Research UK Treating a moving target. Semin Radiat Oncol 2008, 18:58-66. 39. Melancon AD, O'Daniel JC, Zhang L, Kudchadker RJ, Kuban DA, Lee Your research papers will be: AK, Cheung RM, de Crevoisier R, Tucker SL, Newhauser WD, available free of charge to the entire biomedical community Mohan R, Dong L: Is a 3-mm intrafractional margin sufficient for daily image-guided intensity-modulated radiation ther- peer reviewed and published immediately upon acceptance apy of prostate cancer? Radiother Oncol 2007, 85:251-9. cited in PubMed and archived on PubMed Central 40. McNair HA, Hansen VN, Parker CC, Evans PM, Norman A, Miles E, Harris EJ, Del-Acroix L, Smith E, Keane R, Khoo VS, Thompson AC, yours — you keep the copyright Dearnaley DP: A comparison of the use of bony anatomy and BioMedcentral internal markers for offline verification and an evaluation of Submit your manuscript here: http://www.biomedcentral.com/info/publishing_adv.asp Page 9 of 9 (page number not for citation purposes) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Radiation Oncology Springer Journals

Potentials of on-line repositioning based on implanted fiducial markers and electronic portal imaging in prostate cancer radiotherapy

Loading next page...
 
/lp/springer-journals/potentials-of-on-line-repositioning-based-on-implanted-fiducial-jbNNLYqKCi
Publisher
Springer Journals
Copyright
Copyright © 2009 by Graf et al; licensee BioMed Central Ltd.
Subject
Medicine & Public Health; Oncology; Radiotherapy
eISSN
1748-717X
DOI
10.1186/1748-717X-4-13
pmid
19397824
Publisher site
See Article on Publisher Site

Abstract

Background: To evaluate the benefit of an on-line correction protocol based on implanted markers and weekly portal imaging in external beam radiotherapy of prostate cancer. To compare the use of bony anatomy versus implanted markers for calculation of setup-error plus/minus prostate movement. To estimate the error reduction (and the corresponding margin reduction) by reducing the total error to 3 mm once a week, three times per week or every treatment day. Methods: 23 patients had three to five, 2.5 mm Ø spherical gold markers transrectally inserted into the prostate before radiotherapy. Verification and correction of treatment position by analysis of orthogonal portal images was performed on a weekly basis. We registered with respect to the bony contours (setup error) and to the marker position (prostate motion) and determined the total error. The systematic and random errors are specified. Positioning correction was applied with a threshold of 5 mm displacement. Results: The systematic error (1 standard deviation [SD]) in left-right (LR), superior-inferior (SI) and anterior-posterior (AP) direction contributes for the setup 1.6 mm, 2.1 mm and 2.4 mm and for prostate motion 1.1 mm, 1.9 mm and 2.3 mm. The random error (1 SD) in LR, SI and AP direction amounts for the setup 2.3 mm, 2.7 mm and 2.7 mm and for motion 1.4 mm, 2.3 mm and 2.7 mm. The resulting total error suggests margins of 7.0 mm (LR), 9.5 mm (SI) and 9.5 mm (AP) between clinical target volume (CTV) and planning target volume (PTV). After correction once a week the margins were lowered to 6.7, 8.2 and 8.7 mm and furthermore down to 4.9, 5.1 and 4.8 mm after correcting every treatment day. Conclusion: Prostate movement relative to adjacent bony anatomy is significant and contributes substantially to the target position variability. Performing on-line setup correction using implanted radioopaque markers and megavoltage radiography results in reduced treatment margins depending on the online imaging protocol (once a week or more frequently). Page 1 of 9 (page number not for citation purposes) Radiation Oncology 2009, 4:13 http://www.ro-journal.com/content/4/1/13 formed weekly for 23 patients with histologically Background There is evidence that dose-escalation in definitive radio- confirmed prostate cancer treated from 1996 to 2000. The majority of patients were treated by a standard irradiation therapy of prostate cancer improves long-term PSA control [1]. One strategy to reduce late side effects is employment regimen in combination with regional hyperthermia in a of gradually smaller radiation field sizes or planning target phase II study as previously described [15]. Informed con- volumes PTV [2]. Tight margins will decrease the volume sent had been obtained from all patients. dose delivered to organs at risk, thus increasing the thera- peutic ratio of tumor control probability versus normal tis- Before treatment planning, three to five spherical gold sue complication probability (TCP/NTCP). On the other (99.9% Au) markers with a diameter of 2.0 mm were hand, this ratio might decline if the clinical target volume inserted transrectally into the prostate of each patient is partially missed by any positioning error not compen- using a modified biopsy needle under ultrasound guid- sated by the specified safety margins [3]. ance and local anaesthesia. Usually three markers were implanted, one into the apex, and two into the superior Retrospective evaluations [4,5] have suggested that ana- lateral parts of the prostate. Gold markers of this size can tomic variations (rectal distension, large rectum) during be visualized using megavoltage beam detector systems of the planning CT in fact reduce the PSA control. A large the first generation. No complications occurred in associ- (distended) rectum during planning can cause a system- ation with the implantation process as reported elsewhere atic error, because it places the prostate more anterior, but [10]. Note that the gold markers presently applied with kV this location might change from fraction to fraction. X-ray tracking systems are < 1 mm in diameter and the Another study did not confirm a correlation between rec- implantation procedure is easier and more feasible. tal and/or bladder distension and errors of prostate posi- tion [6]. Nevertheless, we assume that image-guidance is Each patient underwent a computerized tomography scan crucial and improves the clinical outcome. (CT) (Siemens™, Erlangen, Germany) for treatment plan- ning in treatment position from 2 cm below the ischial An assessment of patient position is based on skeletal tuberosities to the L4/5 interspace obtaining volumetric landmarks imaged by electronic portal imaging devices data at 5 mm slice thickness and at a 5 mm couch transla- (EPID). They are commonly used for the evaluation and tion. In our study, the patients were instructed to fill the correction of set-up deviations [7]. bladder, but no effort was made to control the rectal vol- ume. However, the CT scans were repeated if excessive fill- As documented in a number of studies [8,9], an interfrac- ing of the rectum had been noticed. Patients were tional displacement of the prostate itself can occur during stabilized in supine position with conventional head, radiation therapy fractions relative to the bony structures knee and feet support and no rigid immobilization device of the pelvis. The feasibility of implanting markers for was used. Images were transferred to a workstation localization of the prostate recently has been demon- (Helax™) for anatomic segmentation of targets and organs strated [10,11] and allows to utilize EPIDs to quantify the at risk and conformal dosimetric planning. The PTV was displacement of the target [12,13]. With the improvement defined by a three-dimensional expansion of the CTV by of online imaging quality, pretreatment localization and 8 mm at the prostate-rectum interface and 10 mm in all online protocols allowing positioning corrections with- other directions. External beam radiotherapy was per- out significant delay have gained feasibility [14]. formed by a linear accelerator (Siemens™ Mevatron KD, Erlangen, Germany) with a beam energy of 18 MV using From the comparison of verification protocols during fractions of 1.8 Gy five times weekly up to 68.4 – 72 Gy radiotherapy it is known, that the treatment margins are (38–40 fractions) at the reference point (ICRU-50,16). An institution specific. We performed a prospective study of isocentric 4-field box technique consisting of anterior, patients treated with conformal radiotherapy for prostate posterior and two lateral fields (0°, 180°, 90° and 270°) cancer, analysing both internal organ motion and setup was used in all cases. error with the objective to quantify the variability in pros- tate position. For displacements of bones and markers, All conformal 3D-plans were conventionally simulated statistical data including overall, systematic and random before treatment. Simulator radiographs had been deviations were determined. From the uncorrected and obtained in orthogonal (0°, 90°) projections and served corrected total errors, we calculated the necessary treat- as reference images for the position of bony landmarks ment margins to ensure sufficient target coverage in the and internal markers. majority of cases. On each treatment day, patients were positioned using Patients and methods laser alignment to reference marks (one anterior and two Verification and correction of treatment position by anal- lateral set-up crosses) on their skin. For all patients, ysis of portal images and simulator control films were per- weekly pre-treatment position verification with an EPID Page 2 of 9 (page number not for citation purposes) Radiation Oncology 2009, 4:13 http://www.ro-journal.com/content/4/1/13 system (Siemens Beamview Plus™, Erlangen, Germany) reference images simulator radiographs and, secondly, the was applied [17]. The electronic portal images (EPIs) for corresponding portal images. The AP beam provided data verification were acquired using 6 MV photons for the AP to detect the position of the landmarks and markers in the (0°) and left lateral (90°) fields once a week with 6 – 8 LR and SI direction and the lateral beam for the AP direc- monitor units (MU), each before starting irradiation. The tion and SI direction as well. To identify the position of images were digitally processed (employing high fre- the target m, we used the arithmetic mean of the marker quency filters) to facilitate recognition of bony structures coordinates according to the isocenter (Fig. 2). All meas- and radiopaque markers. On EPIs, the isocenter has been urements were performed by the same author (RG). The made visible by the projection of an isocenter marker (a consistency of the obtained deviations was tested by cor- 1.5 × 3 mm gold pin) located on the reticule. Bony land- relation of the corresponding values in SI direction taken marks and implanted markers were clearly identified on from 0° and 90° projection. The correlation coefficient of almost all portal films (Figure 1). r = 0.86 was satisfactory. The registration procedure takes about 3 minutes cumulating to a total treatment time of For the applied 2D/3D registration method, isocenter, 6–8 minutes on average. bony contours and fiducial markers were drawn from the simulator films on transparent templates for every patient The evaluation procedure and the nomenclature are sum- before irradiation. These templates were then used to marized in Figure 2. Firstly, we determined the vectorial match the reference images (0°, 90°) to the correspond- displacements of the isocenters relative to the bony anat- ing verification images manually. omy of the reference images Δs for j = 1...23 patients and ij i = 1...8 weekly portal images per patient during the radi- An identical scale of the printed portal images and the otherapy course yielding 8 × 23 = 184 setup errors (under- templates was applied to determine the setup errors from lining identifying a vector). Secondly, the differences of the shift Δs of the isocenter (see Figure 2 for definition of the marker positions relative to the isocenters result in the symbols). The components of the vector Δs according to prostate motion Δm . Finally, the total displacement ij the main axes are determined providing the shift in left- (setup error plus organ motion) of the target relative to right (LR) direction (lateral x-axis), anterior-posterior the isocenter is calculated by Δtot = Δs + Δm . ij ij ij (AP) direction (vertical y-axis) and superior-inferior (SI) direction (longitudinal z-axis). For all 184 fractions, mean and standard deviations for all kinds of errors (setup, motion, total) were calculated. We For evaluation and quantification of uncertainties, two analysed the error distributions averaging over all frac- orthogonal sets of 2D projections were available, firstly as tions and patients. image Figure 1 Reference image (simulator film, left) and online image (port film, in the mid) are registered by the method shown on the right Reference image (simulator film, left) and online image (port film, in the mid) are registered by the method shown on the right image. A template containing isocenter, bony contours and radio-opaque markers is traced from the simulator radiograph and positioned on the portal image with isocenters and main axes in coincidence. The isocenter is shifted until the bony contours (setup error) or the implanted markers are in agreement (total error). For the motion error we deter- mine the shift from the setup corrected position to the marker corrected position. The correction method is two-dimensional and performed separately for each projection (0° and 90°). Redundant measurements (in SI direction) are in good correlation (see text). Page 3 of 9 (page number not for citation purposes) Radiation Oncology 2009, 4:13 http://www.ro-journal.com/content/4/1/13 (Σ) and 0.7 times the standard deviation of the random error (σ) to ensure a minimum dose of 95% to the clinical target volume for 90% of the fractions, i.e. allowing signif- icant dose discrepancies in = 10% of sessions. If a position correction was performed (above the action level), we assume a residual error of = 3 mm [19] in all directions for the corrected fraction. Statistical analysis was performed using JMP v7.0 (SAS Institute, Cary, NC, USA). Tests for sub-groups were per- formed using the paired t-Test. Results We performed the analysis for 23 patients with 8 pairs of Basi Figure 2 c definitions of the different error components EPIs per patient, summing up to a total of 368 anterior- Basic definitions of the different error components: posterior and lateral port films in184 fractions. Bony con- The setup error Δs and the motion error Δm can be added tours, implanted markers and isocenter marker were to the total error Δtot. For every patient j = 1...23 and portal clearly visible and evaluable in 96% of cases. All portal image i = 1...8 the setup errors Δs are determined by match- ij images were evaluable with respect to prostate motion ing the bony contours of the portal images (i, j) to the refer- employing the radiopaque markers. We had to replace ence images j (simulator radiographs) according Figure 1. The only 1.8% of portal images due to insufficient identifica- motion errors Δm are determined after these matching pro- ij cedures by subtracting the marker positions of the matched tion of the bony structures. portal images (i, j) and the reference images j. The systematic error for a patient j is defined as the mean of single errors As summarised in Table 1 we analysed all 184 fractions with respect to i = 1...8 portal images. The random error of together and determined the displacements of the iso- patient j is defined as the standard deviation of this series. center relative the bony anatomy (setup error), the dis- Then further statistical evaluation is possible, e.g. the mean placement of the markers relative to the bony structures systematic error for a series of patients (here 23 patients) in (prostate motion) and the displacement of the isocenter Table 2 and the mean standard deviations (mean random relative to the markers (combined or total targeting error). error) in Table 3. Figure 3 shows the measured deviations in a box plot for- mat, indicating mean values, median values and selected percentiles from 10 to 90% (10%, 25%, 75%, 90%) in LR, Then, we determined means and standard deviations SI and AP directions. The observed errors were greatest in from 8 control EPIs for each patient resulting in the same the AP direction, where a range of 13 mm is found for the error types Δs(j), Δm(j) and Δtot(j) for j = 1...23 patients, total deviation of the target (-7 to +6 mm) for 80% of the and analysed the error distributions with respect to the controls. The extremes observed in internal target motion patients. The standard deviations identify the systematic were 8 mm in AP and 7 mm in SI direction. errors Σ(j) for every patient. We calculated the various errors for every patient sepa- Random errors σ(j) for every patient j were calculated as rately (averaging over eight controls) and analysed the standard deviations of the differences Δs(j) - Δs or Δm(j) error distribution for 23 patients (see Table 2). Both, sys- ij - Δm or Δtot(j) - Δtot averaging over i = 1...8 PIs. We can tematic setup and motion errors are in the range of ± 2 ij ij also determine the mean random error for the entire mm (± 1 SD) summing up to a total error of ± 3 mm. group of patients averaging σ(j) over all j = 1...23 patients. For all 8 fractions per patient, the scatter (SD) about the For correction of translational errors before treatment, we individual averages (systematic errors) has been calcu- used an action level of 5 mm, i.e. all errors of 5 and more lated providing the random errors for the different error mm were corrected. The correction was performed on-line types (Table 3). As expected the random errors are larger by repositioning the target according to the internal mark- than the systematic errors and finally amount to total ran- ers, moving the treatment couch manually. To calculate dom errors of ± 4 mm in SI and AP direction, but only ± the minimum required margin width around the clinical 3 mm in LR direction. The extremes can approach 1–1.5 target volume (CTV + margin = PTV), we utilized the pre- cm, but these are rare cases. scription suggested by van Herk [18]. The margin around the clinical target volume (CTV) should be the sum of 2.5 The online protocol was applied at least 8 times per times the standard deviation of the systematic total error patient and, on average, 56% of the controls determined Page 4 of 9 (page number not for citation purposes) Radiation Oncology 2009, 4:13 http://www.ro-journal.com/content/4/1/13 Table 1: Setup error, motion error and total error Mean ± SD [mm] i = 1,..., 8; j = 1,..., 23 Range [mm] Setup error Δs ij Left-right 0.8 ± 2.8 -8/10 Superior-inferior 0.1 ± 3.4 -9/14 Anterior-posterior -1.2 ± 3.6 -15/9 Prostate movement Δm ij Left-right -0.3 ± 1.8 -6/9 Superior-inferior 0.9 ± 2.8 -9/8 Anterior-posterior 0.3 ± 3.5 -10/10 Total error Δtot ij Left-right 0.5 ± 3.5 -10/19 Superior-inferior 0.9 ± 4.4 -14/13 Anterior-posterior -0.8 ± 4.9 -19/14 Measured deviations for all 184 (= 23 × 8) portal imaging controls (averaged over i = 1,..., 8 weekly portal images in j = 1,..., 23 patients). Mean values <...> in all three directions (left-right LR, superior-inferior SI, anterior-posterior AP) ± standard deviations SD and ranges, i.e. extremal deviations in ± orientations of the axes, in LR, SI and AP direction for setup error <Δs >, prostate movement <Δm > and total errror <Δtot > are ij ij ij listed (see Fig. 2 for explanation of symbols). displacements of = 5 mm in at least one direction and had gin of 1 cm around the CTV is sufficient to counterbalance to be corrected. Margin calculations have been performed the set-up and internal motion inaccuracies if a weekly for each of the axes according to the prescription of van portal imaging with online correction is presumed. Grad- Herk [18] as described in section 2 (see Table 4). To con- ual reduction of the errors and derived margins down to a sider for the total targeting error, the margins added to the minimum of 5 mm is obtained if the frequency of online CTV must be as large as 7.0 mm (LR) and 9.5 mm (SI, AP). control is further increased up to a daily correction as summarized in Table 4. After position corrections once a week, these calculated margins reduce to 6.7, 8.2 and 8.7 mm. Therefore, a mar- Discussion Various techniques have been developed to locate the prostate position on-line such as implanted fiducials (detected by X-rays), transabdominal ultrasound [20], electromagnetic tracking [21] and several kinds of in- room CT (e.g. [22]), in particular in conjunction with hel- ical tomotherapy [23]. However, the highest precision is achieved by using intraprostatic markers. The clinical use of implanted gold markers was found to be feasible in our hands. The geometrical center of implanted radiopaque markers characterizes the prostate position. Several groups have investigated the possibility of seeds migration and have found no or only little motion [24,25]. In addition, the reliability of markers for the location of the prostate has been questioned because of interfraction rotation or deformation [26], but these factors leave the prostate dosimetry unaffected [27]. The analysis is standardized so that the interobserver variabil- ity is low. Therefore implanted markers and EPID based 2 izo tion variability an Figure 3 Measured format, showin values (lines in th 5% (bottom of box), 75% ntal line) percenti deviations in g the mean d total error e bo le s x) and th LR, SI plit in (top of b values and A setu e 10% (bla p variability, prostat P o (lower hori directions ck x) and 90% squares)in a , th (upper zontal line), box plot e medi e posi- hor an - Measured deviations in LR, SI and AP directions in a methods are used for targeting in radiotherapy of prostate box plot format, showing the mean values (black cancer with increasing frequency. squares), the median values (lines in the box) and the 10% (lower horizontal line), 25% (bottom of box), Our results provide information about the scatter of target 75% (top of box) and 90% (upper horizontal line) per- positions during radiotherapy. Setup inaccuracies were centile split in setup variability, prostate position var- reviewed by Hurkmans [28]. In his analyses data were iability and total error. obtained from repeated simulations, from EPID studies Page 5 of 9 (page number not for citation purposes) Radiation Oncology 2009, 4:13 http://www.ro-journal.com/content/4/1/13 Table 2: Systematic errors Mean ± SD [mm] j = 1,..., 23 Range [mm] j = 1,..., 23 Systematic setup error <Δs> Left-right 0.8 ± 1.6 -2.9/3.9 Superior-inferior 0.1 ± 2.1 -2.7/6.2 Anterior-posterior -1.2 ± 2.4 -5/4.1 Systematic prostate movement <Δm> Left-right -0.3 ± 1.1 -2.9/2.7 Superior-inferior 0.9 ± 1.9 -3.1/4.7 Anterior-posterior 0.3 ± 2.3 -4.3/5.0 Systematic total error <Δtot> Left-right 0.5 ± 2.0 -2.6/6.6 Superior-inferior 0.9 ± 2.7 -4.3/7.3 Anterior-posterior -0.8 ± 2.6 -6.7/4.0 Analysis of the mean vectorial setup errors per patient (after averaging over i = 1,..., 8 weekly controls) (1/8)Σ Δs =: <Δs> , the mean vectorial i = 1...8 ij j prostate movements (1/8)Σ Δm =: <Δm> and the mean vectorial total errors (1/8)Σ Δtot =: <Δtot> . After statistical analysis with i = 1...8 ij j i = 1...8 ij j respect to j = 1 to 23 patients we achieve the population based (mean) vectorial systematic error defined as (1/23)Σ <Δs> , (1/23)Σ j = 1...23 j j = <Δm> and (1/23)Σ <Δtot> splitted in the RL, SI and AP direction and the standard deviation SD from these errors (see Fig. 2 for 1...23 j j = 1...23 j explanation of symbols). Also the range (minimum to maximum) is listed (right row). and from repeated CT scans. The standard deviations of the the prostate at the time of treatment can be visualized setup errors ranged from 1 to 4 mm, which is in accordance with a variety of techniques, and differences in measure- with our results. We also found standard deviations below ment techniques make it difficult to compare the results 4 mm. Analysis of the contributions to the total targeting of published studies. In summary, the SDs of the prostate error indicates, that the setup errors cause approximately motion range in the LR direction from 0.7 to 1.9 mm, in one half of the entire target position variability and offers a SI from 1.7 to 3.6 and for AP from 1.5 to 4.0 mm. We potential improvement in total target positioning. measured for prostate motion in RL, SI and AP standard deviations of 1.8, 2.8 and 3.5 mm, even though some The prostate position can move relative to the skeleton extremes of motion were registered in a few patients (table [4]. An overview of interfraction prostate motion studies 1). Thus, our results are in general agreement with litera- was presented in a paper by Langen [29]. The position of ture [30-34]. Table 3: Random errors Mean [mm] j = 1...23 Range [mm] j = 1...23 Random setup error Left-right 2.3 -7.9/7.4 Superior-inferior 2.7 -8.6/7.8 Anterior-posterior 2.7 -12.1/5.4 Random prostate movement Left-right 1.4 -5.5/6.3 Superior-inferior 2.3 -5.9/8.6 Anterior-posterior 2.7 -7.0/8.0 Random total error Left-right 2.9 -8.0/6.6 Superior-inferior 3.9 -11.5/8.2 Anterior-posterior 4.3 -14.9/11.2 The standard deviations of measured deviations Δs , Δm , Δtot around the systematic errors (of each patient j = 1...23) <Δs> , <Δm> , <Δtot> after ij ij ij j j j 2 2 averaging over i = 1,..., 8 weekly controls provide the random errors sqrt(1/7Σ (<Δs> - Δs ) ), sqrt(1/7Σ (<Δm > - Δm ) ), sqrt(1/7Σ i = 1..8 j ij i = 1..8 i j ij i = (<Δtot > - Δtot ) ) for patient j = 1,..., 23. They are splitted into the RL, SI and AP direction for every error type, i.e. setup error, prostate 1..8 i j ij movement and total error (see Fig. 1 for explanation of symbols). The mean random errors after averaging over 23 patients and their range, i.e. maximum deviations in both orientations of the axes, are given in the table. Page 6 of 9 (page number not for citation purposes) Radiation Oncology 2009, 4:13 http://www.ro-journal.com/content/4/1/13 Table 4: Estimation of margins. Random σ and systematic Σ error [mm] Margin [mm] Direction LR SI AP LR SI AP ΣσΣσΣσ No correction 2.0 2.9 2.7 3.9 2.6 4.3 7.0 9.5 9.5 Correction 1×/week 1.9 2.8 2.3 3.5 2.4 3.9 6.7 8.2 8.7 Correction 3×/week 1.6 2.5 1.8 3.0 1.7 3.3 5.8 6.6 7.7 Correction 5×/week 1.4 2.0 1.4 2.3 1.3 2.2 4.9 5.1 4.8 For our patient group the standard deviations of systematic errors (Σ) and random errors (σ) are listed in LR, SI and AP directions. Then the margins from CTV to PTV are calculated in all three directions according to the prescription 2.5 × Σ + 0.7 × σ [14] in order to consider these errors. The errors are summarized in the first line as derived from the weekly portal images without any correction. In the following lines the variation is reduced by controling and correcting the position with increasing frequency per week (once, three times, every day). Correction is conducted if the threshold of 5 mm is crossed in any direction. After correction we assume a residual error of 3 mm in the respective direction. Under these circumstances the margins can be diminished from approximately 10 mm to 5 mm. We found the largest errors, for both, setup as well as pros- respectively. We found for setup, prostate location varia- tate motion, in the AP direction, followed by SI and LR tion and combined error in general larger random errors directions in accordance with the series of Beaulieu and than systemic errors, obviously due to the reduction of others [14,29,35]. Along the lateral axis the prostate is systematic errors by the weekly performed corrections. confined within the pelvis and published data show only small deviations in this direction. In our study, the distri- In our study we calculated necessary CTV-PTV margins bution of organ motion and setup errors for translation is (without correction) of 7.0 to 9.5 mm (RL, SI and AP in the range of the published values [36], e.g. 90% of the direction) according Table 4. Similar margins (without observed displacements were 7 mm or less. correction) are reported by Kupelian [38] with 10, 10 and 12 mm, McNair [40] with 5, 7.5 and 11 mm and van den Interfraction position variation of the prostate as a source Heuvel [41] with 9.5, 8.6 and 10 mm. of treatment error is mainly caused by variable fillings of the bladder and/or rectum that displace the prostate According to the formula given in Section 2 to estimate mainly in SI and AP direction as shown by magnetic reso- the margin between CTV and PTV [18], systematic errors nance imaging of the pelvis [27]. Patient instructions have the largest impact on the size of PTV margins. There- attempt to prepare rectal and bladder distension in a fore, offline correction protocols attempt to determine standardized way before treatment. This may reduce the and correct the systematic error. They have the advantage frequency of large prostate movements, but does not elim- to be effective despite a low imaging frequency. Different inate the motion error [21]. There is even an intrafrac- offline protocols have been successful implemented into tional motion of 1–3 mm on average [37] and after initial clinical practice [42,43]. On the other hand, Litzenberg positioning the displacement of the prostate gland [44] figured out, that because of changes in patient's setup increases with elapsed time. This matter raises concerns characteristics off-line protocols, especially those directed with regard to correction for misalignments [38] and the to localize the prostate using markers did not show any treatment time of 20–30 minutes per session using novel significant benefit in reducing the total error of implanted techniques i.e. intensity modulated radiotherapy, tomo- fiducial gold markers in 10 prostate cancer patients in therapy etc., which will induce a new intrafraction errors. comparison to daily online position correction. For the Recently published analyses of this issue indicate that a 3- same reasons, applying these methods directly to the mm planning target margin is in most cases sufficient to implanted markers also gave larger residual errors than account for intrafractional motion [39]. expected. It may be difficult to identify patients who would benefit from off-line protocols and those who may Both uncertainties, setup error and target motion can be require daily on-line corrections [44]. split into random and systematic deviations. The system- atic component of setup error is largely caused by the sys- Evaluating their possible benefit, on-line correction pro- tematic error inherent to the use of a reference image tocols have the potential to reduce both systematic and obtained by use of the planning-CT. The random compo- random errors, but at the expense of increasing treatment nent of the setup error is mainly caused by uncertainties time per fraction. As expected, systematic errors are effec- from utilisation of skin markers, while the random error tively reduced with increasing imaging frequency [38]. of target position is mainly caused by organ movement, After one weekly online correction and 5 mm action level, Page 7 of 9 (page number not for citation purposes) Radiation Oncology 2009, 4:13 http://www.ro-journal.com/content/4/1/13 Late rectal bleeding after conformal radiotherapy of pros- we found margins of about 7 to 9 mm. These margins can tate cancer. II. Volume effects and dose-volume histograms. be further reduced to a minimum of 5 mm by increasing Int J Radiat Oncol Biol Phys 2001, 49:685-98. the control frequency (Table 4). Kupelian [38] calculated 4. de Courvoisier R, Tucker SL, Dong L, et al.: Increased risk of bio- chemical and local failure in patients with distended rectum treatment margins for 8 different potential non-daily on the planning CT for prostate cancer radiotherapy. Int J imaging strategies, among them low-workload weekly Radiat Oncol Biol Phys 2005, 62:965-973. 5. Heemsbergen WD, Hoogeman MS, Witte MG, Peeters ST, Incrocci protocols. For a weekly online protocol with 3-mm L, Lebesque JV: Increased risk of biochemical and clinical fail- threshold, he found margins of 8, 8 and 6 mm (LR, SI and ure for prostate patients with a large rectum at radiotherapy AP), which agrees quite well with our results. planning: Results from the Dutch Trial of 68 Gy versus 78 Gy. Int J Radiat Oncol Biol Phys 2007, 67:1418-1424. 6. Kim PJ, Parthasarathy A, Ho J, Tam B, Gee H, Lee JW: Impact of A daily positioning correction is feasible under routine rectal distension in IGRT for prostate cancer. Int J Radiat Oncol conditions employing the new generation of linear accel- Biol Phys 2007:S715. 7. Chen J, Lee RJ, Handrahan D, Sause WT: Intensity-modulated erators with image guidance (on-board imaging or x-ray radiotherapy using implanted fiducial markers with daily tracking). Using these techniques the residual error can be portal imaging: assessment of prostate organ motion. Int J Radiat Oncol Biol Phys 2007, 68:912-9. further decreased below 3 mm and the required safety 8. Schallenkamp JM, Herman MG, Kruse JJ, Pisansky TM: Prostate margin is reduced down to 3 mm (unpublished data). An position relative to pelvic bony anatomy based on intrapros- accuracy of only 5 mm is achieved using megavoltage CT tatic gold markers and electronic portal imaging. Int J Radiat Oncol Biol Phys 2005, 63:800-11. without intraprostatic markers [38]. 9. Lattanzi J, McNeeley S, Pinover W, Horwitz E, Das I, Schultheiss TE, Hanks GE: A comparison of daily CT localization to a daily ultrasound-based system in prostate cancer. Int J Radiat Oncol Conclusion Biol Phys 1999, 43:719-725. In summary, correction of setup errors alone is not suffi- 10. Shirato H, Harada T, Harabayashi T, Hida K, Endo H, Kitamura K, cient because target motion contributes significantly to Onimaru R, Yamazaki K, Kurauchi N, Shimizu T, Shinohara N, Matsu- shita M, Dosaka-Akita H, Miyasaka K: Feasibility of insertion/ positioning inaccuracies. The implantation of gold mark- implantation of 2.0-mm-diameter gold internal fiducial ers for a correction protocol was feasible in our study. A markers for precise setup and real-time tumor tracking in weekly on-line setup verification employing these radio- radiotherapy. Int J Radiat Oncol Biol Phys 2003, 56:240-7. 11. Henry AM, Wilkinson C, Wylie JP, Logue JP, Price P, Khoo VS: paque markers and megavoltage radiography results in Trans-perineal implantation of radio-opaque treatment ver- CTV-PTV margins of 7 to 8.5 mm. More effort can further- ification markers into the prostate: an assessment of proce- dure related morbidity, patient acceptability and accuracy. more decrease these margins. A correction of three times Radiother Oncol 2004, 73:57-9. per week leads to margins of 6 to 7.5 mm, and daily cor- 12. Litzenberg D, Dawson LA, Sandler H, Sanda MG, McShan DL, Ten- rections can further reduce the margin down to 5 mm. Haken RK, Lam KL, Brock KK, Balter JM: Daily prostate targeting using implanted radiopaque markers. Int J Radiat Oncol Biol Phys 2002, 52:699-703. Competing interests 13. Herman MG, Pisansky TM, Kruse JJ, Prisciandaro JI, Davis BJ, Geyer The authors declare that they have no competing interests. SM, King BF: Daily on-line positioning of the prostate for three-dimensional conformal radiotherapy (3D-CRT) using an electronic portal imaging device (EPID). Int J Radiat Oncol Authors' contributions Biol Phys 2003, 57:1131-40. 14. Beaulieu L, Girouard LM, Aubin S, Aubry JF, Brouard L, Roy-Lacroix RG analysed the simulator films and portal images, deter- L, Dumont J, Tremblay D, Laverdière J, Vigneault E: Performing mined the errors, performed the statistical analysis and daily prostate targeting with a standard V-EPID and an auto- drafted the manuscript. PW initiated the study, treated the mated radio-opaque marker detection algorithm. Radiother Oncol 2004, 73:61-64. patients, formulated the mathematical background and 15. Tilly W, Gellermann J, Graf R, Hildebrandt B, Weissbach L, Budach V, revised the first draft of the manuscript. VB participated in Felix R, Wust P: Regional hyperthermia in conjunction with designing the study and approved the treatment concepts. definitive radiotherapy against recurrent or locally advanced prostate cancer T3 pN0 M0. Strahlenther Onkol 2005, 181:35-41. DB coordinated the recruitment of patients and data 16. International Commission on Radiation Units and Measurements: acquisition. All authors participated in the critical discus- Prescribing, recording and reporting photon beam therapy. ICRU Report 50. Bethesda, MD; 1993. sion of the data and their statistical analysis. All authors 17. Gagel B, Schramm O, Harms W, Mulhern A, Wenz F, van Kampen M, improved the manuscript and approved the final version. Wannenmacher M, Eble MJ: The electronic portal imaging sys- tem Siemens Beamview Plus versus the conventional verifi- cation films CEA-TVS and DuPont COL-7. A critical Acknowledgements appraisal of visual image quality. Strahlenther Onkol 2002, The authors thank the Lieselotte-Beutel-Stiftung for the valuable support of 178:446-52. the prostate center. 18. Van Herk M, Remeijer P, Rasch C, Lebesque JV: The probability of correct target dosage: dose-population histograms for deriv- ing treatment margins in radiotherapy. Int J Radiat Oncol Biol References Phys 2000, 47:1121-35. 1. Kuban DA, Tucker SL, Dong L, Starkschall G, Huang EH, Cheung MR, 19. Poulsen PR, Muren LP, Høyer M: Residual set-up errors and mar- Lee AK, Pollack A: Long-term results of the M. D. Anderson gins in on-line image-guided prostate localization in radio- randomized dose-escalation trial for prostate cancer. Int J therapy. Radiother Oncol 2007, 85:201-6. Radiat Oncol Biol Phys 2008, 70:67-74. 20. Scarbrough TJ, Golden NM, Ting JY, Fuller CD, Wong A, Kupelian PA, 2. Zelefsky MJ, Fuks Z, Leibel SA: Intensity-modulated radiation Thomas CR Jr: Comparison of ultrasound and implanted seed therapy for prostate cancer. Semin Radiat Oncol 2002, 12:229-37. marker localisation methods: Implications for image-guided 3. Jackson A, Skwarchuk MW, Zelefsky MJ, Cowen DM, Venkatraman radiotherapy. Int J Radiat Oncol Biol Phys 2006, 65:378-387. ES, Levegrun S, Burman CM, Kutcher GJ, Fuks Z, Liebel SA, Ling CC: Page 8 of 9 (page number not for citation purposes) Radiation Oncology 2009, 4:13 http://www.ro-journal.com/content/4/1/13 21. Langen KM, Willoughby TR, Meeks SL, Santhanam A, Cunningham A, the potential benefit of online and offline verification proto- Levine L, Kupelian PA: Observations on real-time prostate cols for prostate radiotherapy. Int J Radiat Oncol Biol Phys 2008, gland motion using electromagnetic tracking. Int J Radiat Oncol 71:41-50. Biol Phys 2008, 71:1084-90. 41. Heuvel F Van den, Fugazzi J, Seppi E, Forman JD: Clinical applica- 22. Langen KM, Zhang Y, Andrews RD, Hurley ME, Meeks SL, Poole DO, tion of a repositioning scheme, using gold markers and elec- Willoughby TR, Kupelian PA: Initial experience with megavolt- tronic portal imaging. Radiother Oncol 2006, 79:94-100. age (MV) CT guidance for daily prostate alignments. Int J 42. De Boer HC, Heijmen BJ: A protocol for the reduction of sys- Radiat Oncol Biol Phys 2005, 62:1517-1524. tematic patient setup errors with minimal portal imaging 23. Kupelian PA, Lee C, Langen KM, Zeidan OA, Manon RR, Willoughby workload. Int J Radiat Oncol Biol Phys 2001, 50:1350-65. TR, Meeks SL: Evaluation of image-guidance strategies in the 43. de Boer HC, van Os MJ, Jansen PP, Heijmen BJ: Application of the treatment of localized prostate cancer. Int J Radiat Oncol Biol No Action Level (NAL) protocol to correct for prostate Phys 2008, 70:1151-1157. motion based on electronic portal imaging of implanted 24. Pouliot J, Aubin M, Langen KM, Liu YM, Pickett B, Shinohara K, Roach markers. Int J Radiat Oncol Biol Phys 2005, 61:969-983. M 3rd: (Non)-migration of radio-opaque prostate markers. 44. Litzenberg DW, Balter JM, Lam KL, Sandler HM, Ten Haken RK: Ret- Int J Radiat Oncol Biol Phys 2003, 56:862-6. rospective analysis of prostate cancer patients with 25. Poggi MM, Gant DA, Litzenberg D, Dawson LA, Sandler H, Sanda MG, implanted gold markers using off-line and adaptive therapy McShan DL, Ten-Haken RK, Lam KL, Brock KK, Balter JM: Marker protocols. Int J Radiat Oncol Biol Phys 2005, 63:123-33. seed migration. Int J Radiat Oncol Biol Phys 2002, 52:699-703. 26. Ghilezan MJ, Jaffray DA, Sieverdsen JH: Prostate gland motion assessed with cine-magnetic resonance imaging (cine-MRI). Int J Radiat Oncol Biol Phys 2005, 62:406-17. 27. Kupelian PA, Langen KM, Zeidan OA: Daily variations in delivered doses in patients treated with radiotherapy for localized prostate cancer. Int J Radiat Oncol Biol Phys 2006, 66:876-82. 28. Hurkmans CW, Remeijer P, Lebesque JV, Mijnheer BJ: Set-up veri- fication using portal imaging: review of current clinical prac- tice. Radiother Oncol 2001, 58:105-20. 29. Langen KM, Jones DTL: Organ Motion and its Management. Int J Radiat Oncol Biol Phys 2001, 50:265-78. 30. Crook JM, Raymond Y, Salhani D, Yang H, Esche B: Prostate motion during standard radiotherapy as assessed by fiducial markers. Radiother Oncol 1995, 37:35-42. 31. Nederveen AJ, Dehnad H, Heide UA van der, van Moorselaar RJ, Hof- man P, Lagendijk JJ: Comparison of megavoltage position veri- fication for prostate irradiation based on bony anatomy and implanted fiducials. Radiother Oncol 2003, 68:81-8. 32. Chung PW, Haycocks T, Brown T, Cambridge Z, Kelly V, Alasti H, Jaf- fray DA, Catton CN: On-line aSi portal imaging of implanted fiducial markers for the reduction of interfraction error dur- ing conformal radiotherapy of prostate carcinoma. Int J Radiat Oncol Biol Phys 2004, 60:329-34. 33. Dehnad H, Nederveen AJ, Heide UA van der, van Moorselaar RJ, Hof- man P, Lagendijk JJ: Clinical feasibility study for the use of implanted gold seeds in the prostate as reliable positioning markers during megavoltage irradiation. Radiother Oncol 2003, 67:295-302. 34. Wu J, Haycocks T, Alasti H, Ottewell G, Middlemiss N, Abdolell M, Warde P, Toi A, Catton C: Positioning errors and prostate motion during conformal prostate radiotherapy using on- line isocenter set-up verification and implanted prostate markers. Radiother Oncol 2001, 61:127-33. 35. Vigneault E, Pouliot J, Laverdiere J, Roy J, Dorion M: Electronic por- tal imaging device detection of radioopaque markers for the evaluation of prostate position during megavoltage irradia- tion: A clinical study. Int J Radiat Oncol Biol Phys 1997, 37:205-12. 36. Roeske JC, Forman JD, Mesina CF, He T, Pelizzari A, Fontenla E, Vijay- akumar S, Chen GT: Evaluation of changes in the size and loca- tion of the prostate, seminal vesicles, bladder, and rectum during a course of external beam radiation therapy. Int J Radiat Oncol Biol Phys 1995, 33:1321-9. 37. Smitsmans MH, Pos FJ, de Bois J, Heemsbergen WD, Sonke JJ, Leb- Publish with Bio Med Central and every esque JV, van Herk M: The influence of a dietary protocol on scientist can read your work free of charge cone beam CT-guided radiotherapy for prostate cancer patients. Int J Radiat Oncol Biol Phys 2008, 71:1279-86. "BioMed Central will be the most significant development for 38. Kupelian PA, Langen KM, Willoughby TR, Zeidan OA, Meeks SL: disseminating the results of biomedical researc h in our lifetime." Image-guided radiotherapy for localized prostate cancer: Sir Paul Nurse, Cancer Research UK Treating a moving target. Semin Radiat Oncol 2008, 18:58-66. 39. Melancon AD, O'Daniel JC, Zhang L, Kudchadker RJ, Kuban DA, Lee Your research papers will be: AK, Cheung RM, de Crevoisier R, Tucker SL, Newhauser WD, available free of charge to the entire biomedical community Mohan R, Dong L: Is a 3-mm intrafractional margin sufficient for daily image-guided intensity-modulated radiation ther- peer reviewed and published immediately upon acceptance apy of prostate cancer? Radiother Oncol 2007, 85:251-9. cited in PubMed and archived on PubMed Central 40. McNair HA, Hansen VN, Parker CC, Evans PM, Norman A, Miles E, Harris EJ, Del-Acroix L, Smith E, Keane R, Khoo VS, Thompson AC, yours — you keep the copyright Dearnaley DP: A comparison of the use of bony anatomy and BioMedcentral internal markers for offline verification and an evaluation of Submit your manuscript here: http://www.biomedcentral.com/info/publishing_adv.asp Page 9 of 9 (page number not for citation purposes)

Journal

Radiation OncologySpringer Journals

Published: Apr 27, 2009

References