Daily electronic portal imaging of implanted gold seed fiducials in patients undergoing radiotherapy after radical prostatectomy

Daily electronic portal imaging of implanted gold seed fiducials in patients undergoing radiotherapy after radical prostatectomy

Int. J. Radiation Oncology Biol. Phys., Vol. 67, No. 2, pp. 610 – 619, 2007 Copyright © 2007 Elsevier Inc. Printed in the USA. All rights reserved 036...

688KB Sizes 0 Downloads 48 Views

Int. J. Radiation Oncology Biol. Phys., Vol. 67, No. 2, pp. 610 – 619, 2007 Copyright © 2007 Elsevier Inc. Printed in the USA. All rights reserved 0360-3016/07/$–see front matter

doi:10.1016/j.ijrobp.2006.09.042

PHYSICS CONTRIBUTION

DAILY ELECTRONIC PORTAL IMAGING OF IMPLANTED GOLD SEED FIDUCIALS IN PATIENTS UNDERGOING RADIOTHERAPY AFTER RADICAL PROSTATECTOMY DANIEL C. SCHIFFNER, M.D.,* ALEXANDER R. GOTTSCHALK, M.D., PH.D.,*‡ MICHAEL LOMETTI, M.S.,* MICHELE AUBIN, M.SC.E.E.,* JEAN POULIOT, PH.D.,*‡ JOYCELYN SPEIGHT, M.D., PH.D.,*‡ I-CHOW HSU, M.D.,*‡ KATSUTO SHINOHARA, M.D.,†‡ AND MACK ROACH III, M.D.*‡ Departments of *Radiation Oncology and †Urology and the ‡Comprehensive Cancer Center, University of California, San Francisco, San Francisco, CA Purpose: The aim of this study was to measure interfraction prostate bed motion, setup error, and total positioning error in 10 consecutive patients undergoing postprostatectomy radiotherapy. Methods and Materials: Daily image-guided target localization and alignment using electronic portal imaging of gold seed fiducials implanted into the prostate bed under transrectal ultrasound guidance was used in 10 patients undergoing adjuvant or salvage radiotherapy after prostatectomy. Prostate bed motion, setup error, and total positioning error were measured by analysis of gold seed fiducial location on the daily electronic portal images compared with the digitally reconstructed radiographs from the treatment-planning CT. Results: Mean (ⴞ standard deviation) prostate bed motion was 0.3 ⴞ 0.9 mm, 0.4 ⴞ 2.4 mm, and ⴚ1.1 ⴞ 2.1 mm in the left–right (LR), superior–inferior (SI), and anterior–posterior (AP) axes, respectively. Mean set-up error was 0.1 ⴞ 4.5 mm, 1.1 ⴞ 3.9 mm, and ⴚ0.2 ⴞ 5.1 mm in the LR, SI, and AP axes, respectively. Mean total positioning error was 0.2 ⴞ 4.5 mm, 1.2 ⴞ 5.1 mm, and ⴚ0.3 ⴞ 4.5 mm in the LR, SI, and AP axes, respectively. Total positioning errors >5 mm occurred in 14.1%, 38.7%, and 28.2% of all fractions in the LR, SI, and AP axes, respectively. There was no significant migration of the gold marker seeds. Conclusions: This study validates the use of daily image-guided target localization and alignment using electronic portal imaging of implanted gold seed fiducials as a valuable method to correct for interfraction target motion and to improve precision in the delivery of postprostatectomy radiotherapy. © 2007 Elsevier Inc. Prostate bed motion, Postprostatectomy radiation therapy, Organ motion, Patient setup error, Electronic portal imaging device (EPID).

INTRODUCTION

in the left–right (LR) axis and 2 to 4 mm in both the anterior–posterior (AP) and superior–inferior (SI) axes. Motion in the AP and SI axes has been shown to correlate with variations in rectal and, to a lesser extent, bladder volumes (1, 6 –9). Studies of interfraction patient setup error during the course of definitive RT have found SDs ranging from 1 to 5.5 mm in all axes (2, 5, 10 –17). Total positioning error, or the sum of organ motion and setup error, has been measured with SDs in the range of 2 to 6 mm in all axes (4, 5, 11, 13, 16, 18 –24). To correct for the dosimetric uncertainties that result from these positioning errors, many institutions use im-

In the delivery of external-beam radiation therapy (EBRT) to treat prostate cancer, interfraction target motion caused by both patient setup error and internal organ motion can lead to insufficient dose coverage of the targeted tumor volume, excessive irradiation of adjacent normal structures, and compromised clinical outcomes. Many investigators have studied the characteristics of target organ motion during definitive radiotherapy (RT) delivered to the prostate gland (1–5). These studies have found that the standard deviation (SD) for internal organ motion of the intact prostate is in the order of 1 to 2 mm

CA, February 24 –26, 2006. Supported by the Roy W. Howard Family Fellowship Award (to D.C.S.). Conflict of interest: none. Acknowledgments—We thank Dr. Vivian Weinberg for her statistical input. Received June 20, 2006, and in revised form Sept 20, 2006. Accepted for publication Sept 28, 2006.

Reprint requests to: Alexander R. Gottschalk, M.D., Ph.D., University of California, San Francisco, Department of Radiation Oncology, 1600 Divisadero St., H-1031, San Francisco, CA 94143-1708. Tel: (415) 353-7185; Fax: (415) 353-9883; E-mail: [email protected] Presented at the 47th Annual Meeting of the American Society for Therapeutic Radiology and Oncology (ASTRO), Denver, CO, October 16 –20, 2005; and at the American Society of Clinical Oncology (ASCO) Prostate Cancer Symposium, San Francisco, 610

Prostate bed motion and EPID imaging during postprostatectomy radiotherapy

age-guided target localization and alignment techniques, including ultrasound-based technologies (17–19, 21, 24, 25), cone-beam computed tomography (CT) (26, 27), and Tomotherapy (28). One of the most extensively studied image-guided target localization systems involves the daily use of an electronic portal imaging device (EPID) to visualize radiopaque fiducials implanted into the prostate before the delivery of EBRT. Compensatory couch shifts are then made to align the fiducial markers on the EPID image with their position on the digitally reconstructed radiograph (DRR) from the treatment-planning CT. Based on reports from several institutions, this method has been validated as a cost-effective and clinically valuable means to improve the accuracy and precision of EBRT in the treatment of prostate cancer (4, 5, 10, 15, 29 –31). At the University of California, San Francisco (UCSF), image-guided target localization and alignment using daily EPID imaging of implanted gold seed fiducials has been used since 1997 in patients receiving definitive RT for prostate cancer (29, 31). In addition, investigators at our institution have developed an EPID-based method to correct target positioning errors in patients receiving RT in the postprostatectomy setting. To visualize the postoperative prostate bed and to facilitate target alignment in the delivery of EBRT, gold seed fiducials are implanted into the tissues of the prostate bed transrectally under ultrasound guidance before treatment planning. As in the treatment of the intact prostate gland, daily EPID images are acquired before each fraction. Online compensatory patient shifts are then made to align the gold marker seed positions on the daily portal image with their position on the DRR. Since its inception at UCSF in 2001, more than 70 patients have received postprostatectomy RT using this technique. Although target motion and its correction during definitive RT have been studied extensively, much less is known about these issues in patients undergoing RT after prostatectomy. To explore further this clinically relevant topic, this investigation examined prostate bed motion, setup error, and total positioning error using daily EPID images from 10 consecutive patients who received adjuvant or salvage RT at our institution after implantation of gold seed fiducials into the prostate bed. In addition, migration of the gold marker seeds within the tissues of the prostate bed was measured. Our investigation demonstrates the utility of image-guided target localization and alignment using daily EPID imaging of implanted gold seed fiducials during the delivery of adjuvant and salvage RT after prostatectomy. METHODS AND MATERIALS Gold seed fiducial implantation, patient simulation, and external-beam irradiation Patients underwent transrectal ultrasound (TRUS)– guided implantation of two to three gold seed fiducials into the vesicoure-



D. C. SCHIFFNER et al.

611

thral anastomosis, retrovesicular tissues, and, when present, recurrent tumor nodules. These areas are the most likely sites of local recurrence after prostatectomy (32). The fiducials measured 1.1 mm ⫻ 3 mm and were inserted by a single physician (K.S.) using a 17-gauge needle. This procedure was performed at least 2 weeks before the treatment planning CT to permit the resolution of postprocedure edema. All patients underwent simulation and treatment in the supine position with external immobilization using an alpha cradle from the waist to midthigh as well as a foam block between the ankles. Patients underwent a pelvic CT scan with 3-mm slices, and threedimensional dosimetric planning with Pinnacle software (Philips Medical Systems, Andover, MA) was used. Patients were instructed to have a full bladder and empty rectum at the time of the treatment planning CT and throughout the course of treatment. Before simulation, an enema was administered. The clinical tumor volume (CTV) was contoured as an ellipsoid shape that encompassed the prostatic fossa, vesicourethral anastomosis, bladder neck, retrovesicular space, and periprostatic tissues. Information from surgical and pathology reports, preoperative imaging, and surgical clip location was used to aid in the delineation of the target volumes. To define the planning target volume (PTV), a 15-mm margin was used in the right, left, superior, inferior, and anterior directions. A tighter 7.5-mm posterior margin was used to limit the volume of irradiated rectum. The DRRs were created from the treatment planning CT and served as the reference standards for the daily EPID images. After alignment of the skin tattoos with an in-room laser coordinate system, left lateral and AP electronic portal images were acquired using an amorphous silicon (a-Si) flat panel portal image device (Siemens Oncology Systems, Concord, CA). Compensatory couch shifts were made to align the position of the gold marker seeds in relation to the isocenter on the EPID image with their position in relation to the isocenter on the reference standard DRR. If the couch was shifted, EPID images were repeated to verify target alignment. Online daily EPID image analysis and couch moves were made by the radiation therapists under the remote guidance of the attending physician.

Patient and electronic portal image selection Ten consecutive patients were identified for this retrospective, offline analysis between June 1, 2003, and July 1, 2004. All patients underwent radical retropubic prostatectomy with lymphadenectomy as initial definitive therapy for adenocarcinoma of the prostate. These patients proceeded to receive postprostatectomy RT as an adjuvant therapy (if they were felt to be at high risk for local recurrence postoperatively) or as a salvage therapy (if there was biochemical or clinical evidence of either residual prostate cancer or localized recurrence after surgery). All patients had DRRs as well as paired AP and lateral daily EPID images from at least 10 four-field, whole-pelvic irradiation fractions. The whole-pelvic images were required to reliably identify the key anatomic landmarks of the bony pelvis as well as the implanted gold seed fiducials. Gold marker seeds were clearly identified on every DRR. If no fiducial was visible on either the AP or lateral EPID image for a given fraction because of obstruction by a graticule bead, the treatment fraction was excluded from this study. Overall, a small minority of orthogonal image pairs were excluded because of technical problems involving image acquisition and data storage onto optical discs (n ⫽ 58), as well as AP (n ⫽ 9) and lateral (n ⫽ 20) portal images with no visible fiducial

612

I. J. Radiation Oncology



Biology



Physics

Volume 67, Number 2, 2007

Table 1. Total positioning error: all patients Total fractions (163)

LR

SI

AP

Negative shifts* Positive shifts† Systematic error (mean, mm) Standard deviation (mm) Range (mm) % Shifts ⱕ2 mm % Shifts ⬎5 mm

92 71 0.2 4.5 ⫺16.5 to 22.3 52.8% 14.1%

77 86 1.2 5.1 ⫺9.4 to 12.7 26.4% 38.7%

88 75 ⫺0.3 4.5 ⫺11.6 to 9.9 31.9% 28.2%

Abbreviations: AP ⫽ anterior–posterior axis; LR ⫽ left⫺right axis; SI ⫽ superior–inferior axis. * Right, inferior, and anterior directions. † Left, superior, and posterior directions.

because of obstruction by the graticule beads. In total, 163 orthogonal image pairs were used in this study.

Measurement of total positioning error, prostate bed motion, and setup error Total positioning error was defined as the change in gold seed position in relation to the isocenter on the EPID image compared with the treatment planning DRR. As such, total positioning error comprised the sum of prostate bed motion and setup error. To express this, we devised the equation:

study. The SD was computed to assess day-to-day variations in intermarker distance. In addition, the average of the intermarker distances from the first 3 days of radiation therapy for each patient was calculated and served as an initial reference intermarker distance that was subtracted from the average of the last three intermarker distances from the end of treatment. This absolute difference in intermarker distance served as a measure of fiducial migration over the course of RT.

RESULTS Total Positioning Error ⫽ Prostate Bed Motion ⫹ Setup Error Total positioning error was measured using MATLAB software (MathWorks, Natick, MA) and was calculated as the average of the individual fiducial displacements in relation to the isocenter in the LR, SI, and AP axes. Prostate bed motion was defined as the interfraction motion of the implanted gold seed fiducials in relation to the bony pelvic anatomy on the daily EPID images compared with the DRR. Daily shifts in the location of the bony pelvis reflect patient setup error, and not the physiologic motion of the prostate bed. To correct for setup error and to calculate prostate bed motion, the bony pelvic anatomy on each daily AP and lateral EPID image was carefully aligned with the pelvic bones on the DRR in Adobe Photoshop (Adobe Systems, San Jose, CA) using key anatomic landmarks. Prostate bed motion was measured as the average of the remaining fiducial displacements for each seed in the LR, SI, and AP axes. Setup error was defined as the component of the total positioning error caused by imprecision in patient alignment using skin tattoos. In each treatment fraction, setup error in each axis was calculated using the formula: Setup Error ⫽ Total Positioning Error ⫺ Prostate Bed Motion The precision of this measurement system was found to be 0.9 mm by three repeat measurements of total positioning error, prostate bed motion, and setup error in 10 treatment fractions from a single patient. Systematic error was calculated as the mean target displacement in each axis.

Measurement of gold seed fiducial migration The distances between the gold markers were measured in the 163 orthogonal image pairs for the 10 patients included in this

Total positioning error Interfraction total positioning error, which is the sum of both prostate bed motion and setup error, was measured in 163 orthogonal image pairs from the 10 patients included in this clinical investigation (Table 1). Systematic error ⫾ SD was 0.2 mm ⫾ 4.5 mm, 1.2 mm ⫾ 5.1 mm, and ⫺0.3 mm ⫾ 4.5 mm in the LR, SI, and AP axes, respectively. In the LR, SI, and AP axes, 14.1%, 38.7%, and 28.2% of fractions contained displacements ⬎5 mm, respectively. The range of target displacements was larger in the LR axis than in the SI and AP axes (Fig. 1). Prostate bed motion Interfraction motion of the prostate bed demonstrated a markedly different statistical pattern than was found for total positioning error (Table 2). Systematic error ⫾ SD in the LR, SI, and AP axes was 0.3 mm ⫾ 0.9 mm, 0.4 mm ⫾ 2.4 mm, and ⫺1.1 mm ⫾ 2.1 mm, respectively. The larger SDs in the SI and AP axes suggest greater amounts of day-to-day variation along these planes than in the LR axis. Nonetheless, the SDs for prostate bed motion were small in comparison to those for total positioning error, which indicates minimal day-to-day physiologic movement of this postsurgical anatomic structure. As shown in Fig. 2, the magnitudes of displacement were also quite small, with no shift in any direction ⬎7.5 mm. Again, prostate bed motion was greatest in the SI and AP axes. Nonetheless, the majority of shifts were ⱕ2 mm, and no interfraction prostate bed motion was measured in 16.0%, 16.5%, and 8.0% of fractions in the LR, SI, and AP axes, respectively.

Prostate bed motion and EPID imaging during postprostatectomy radiotherapy



D. C. SCHIFFNER et al.

613

Fig. 1. Total positioning error. (a– c) Histograms representing the aggregate magnitude and frequency of total positioning errors in the left–right (LR), superior–inferior (SI), and anterior–posterior (AP) axes, respectively, for all 10 patients (n ⫽ 163). (d) The relative frequency distribution.

Setup error In comparison with prostate bed motion, daily patient set-up error accounted for larger and more frequent target displacements (Table 3). Systematic error ⫾ SD was 0.1

mm ⫾ 4.5 mm, 1.1 mm ⫾ 3.9 mm, and ⫺0.2 mm ⫾ 5.1 mm in the LR, SI, and AP axes, respectively. The SDs for setup error were similar in magnitude to those found for total positioning error and were much larger than those found for

Table 2. Prostate bed motion: all patients Total fractions (163)

LR

SI

AP

Negative shifts* Positive shifts† No shift Systematic error (mean, mm) Standard deviation (mm) Range (mm) % shifts ⱕ2 mm % shifts ⬎5 mm

48 89 26 0.3 0.9 ⫺1.1 to 5.1 93.3% 0.6%

56 80 27 0.4 2.4 ⫺5.3 to 6.5 56.4% 2.5%

104 46 13 ⫺1.1 2.1 ⫺6.1 to 4.9 62.6% 3.7%

Abbreviations as in Table 1. * Right, inferior, and anterior directions. † Left, superior, and posterior directions.

614

I. J. Radiation Oncology



Biology



Physics

Volume 67, Number 2, 2007

Fig. 2. Prostate bed motion. (a– c) Histograms representing the aggregate magnitude and frequency of prostate bed motion in the left–right (LR), superior–inferior (SI), and anterior–posterior (AP) axes, respectively, for all 10 patients (n ⫽ 163). (d) Relative frequency distribution.

prostate bed motion. The distribution of displacements also demonstrates a greater magnitude and frequency of target shifts because of setup error in comparison with prostate bed motion (Fig. 3). A large proportion of displacements were ⬎5 mm. Only 51.5%, 34.4%, and 31.9% of fractions

had shifts ⱕ2 mm in the LR, SI, and AP axes, respectively. Overall, this statistical pattern was quite similar to that demonstrated for total positioning error, which reflects the significant contribution of setup error in determining the daily total positioning error.

Table 3. Setup error: all patients Total fractions (163)

LR

SI

AP

Negative shifts* Positive shifts† Systematic error (mean, mm) Standard deviation (mm) Range (mm) % shifts ⱕ2 mm % shifts ⬎5 mm

97 66 0.1 4.5 ⫺17.5 to 21.6 51.5% 17.8%

74 89 1.1 3.9 ⫺5.9 to 10.8 34.4% 20.2%

97 66 ⫺0.2 5.1 ⫺9.5 to 14.0 31.9% 31.3%

Abbreviations as in Table 1. * In the right, inferior, and anterior directions. † In the left, superior, and posterior directions.

Prostate bed motion and EPID imaging during postprostatectomy radiotherapy



D. C. SCHIFFNER et al.

615

Fig. 3. Set-up error. (a– c) Histograms representing the aggregate magnitude and frequency of setup error in the left–right (LR), superior–inferior (SI), and anterior–posterior (AP) axes, respectively, for all 10 patients (n ⫽ 163). (d) Relative frequency distribution.

Migration of gold seed fiducials To assess migration of the gold seed fiducials implanted into the prostate bed, the variations in intermarker distance between

fractions and over the course of treatment were measured (Table 4). The interfraction variation in intermarker distance (SD) was very small and ranged from 0.4 mm to 0.9 mm

Table 4. Intermarker distance and gold seed fiducial migration

Patient

A⫺B

B⫺C

A⫺C

Absolute difference in intermarker distance from beginning to end of RT (mm)

1 2 3 4 5 6 7 8 9 10

0.8 0.4 0.7 0.7 0.8 0.8 0.8 0.9 0.6 0.5

⫺ ⫺ 0.5 0.9 ⫺ ⫺ ⫺ ⫺ ⫺ ⫺

⫺ ⫺ 0.6 0.4 ⫺ ⫺ ⫺ ⫺ ⫺ ⫺

0.6 0.2 0.2 0.9 0.4 0.6 0.2 0.4 0.9 0.0

Intermarker distance, SD (mm)*

Abbreviations: RT ⫽ radiotherapy; SD ⫽ standard deviation. * In patients with two implanted fiducials, only one distance (A⫺B) was measurable. In patients with three implanted fiducials (Patients 3 and 4), three distances (A⫺B, B⫺C, and A⫺C) were measured.

616

I. J. Radiation Oncology



Biology



Physics

(mean, 0.7 mm). In addition, the migration of the gold seed fiducials over the course of RT was minimal. The measured differences in intermarker distance between the beginning and conclusion of treatment ranged from 0.0 mm to 0.9 mm (mean, 0.4 mm). None of these values was larger than the tested precision of this measurement system (0.9 mm). DISCUSSION The identification and correction of total positioning errors because of internal organ motion and imprecision in patient setup are critical issues in optimizing the delivery of EBRT. Although many previous studies have examined organ motion, setup error, and total positioning error in patients undergoing definitive RT to treat prostate cancer, very little is known about these parameters in patients receiving adjuvant or salvage RT after prostatectomy. To better interpret our postprostatectomy findings, we reviewed the literature on intact prostate motion during the delivery of definitive RT. In general, the SD for internal organ motion is in the order of 1 to 2 mm in the LR axis and 2 to 4 mm in the AP and SI axes (1–5). In the literature on organ motion, displacements ⬎5 mm are generally considered to have the potential to cause a geographic miss of the CTV margin, if the target shift is not identified and corrected. A rather large proportion of shifts ⬎5 mm has been demonstrated by prior studies to occur primarily in the AP and SI axes. To explain the predominance of displacements along these planes, several studies have confirmed that intact prostate motion is correlated with variations in the volume of the adjacent rectum and, to a lesser extent, the bladder (1, 6 –9). For the 10 patients included in our postprostatectomy study, systematic error ⫾ SD for prostate bed motion was 0.3 mm ⫾ 0.9 mm, 0.4 mm ⫾ 2.4 mm, and ⫺1.1 mm ⫾ 2.1 mm in LR, SI, and AP axes, respectively. Only two prior studies have examined prostate bed motion after prostatectomy. Chinnaiyan et al. (33) studied this outcome in 6 postprostatectomy patients using a three-dimensional, ultrasound-guided system. Prostate bed motion was presented as an average vector length (5 mm ⫾ 3 mm SD). This value was not resolved into its component LR, SI, and AP vectors as was done in the current study, which makes interstudy comparison difficult. Kruse et al. (34) measured prostate bed motion in 6 postprostatectomy patients using daily orthogonal EPID imaging of radiopaque titanium surgical clips. Standard deviations for prostate bed motion were measured to be 1.1 mm, 1.8 mm, and 2.3 mm in the LR, SI, and AP axes, respectively. These results are consistent with the SD values found in our postprostatectomy study. Our findings suggest that prostate bed motion is qualitatively similar to organ motion of the intact prostate gland, with shifts predominantly in the AP and SI axes. As with the intact prostate, these shifts are likely related to variations in rectal and bladder volumes. In comparison with the intact prostate, however, the quantity of day-to-day variation in target position is less for the prostate bed than for the intact

Volume 67, Number 2, 2007

gland. Unlike most studies of intact prostate motion, ⬍5% of fractions in each axis contained displacements ⬎5 mm in the present study. The majority of fractions demonstrated prostate bed shifts ⱕ2 mm. In addition, the standard deviations in our postprostatectomy study are smaller than those found in the vast majority of intact prostate motion studies that we reviewed (1–5). This restricted range of motion is likely explained by the presence of postsurgical adhesions and scarring, which impart reduced elasticity to the postoperative tissues of the prostatic fossa. In addition, during radical prostatectomy the distal bladder neck is pulled inferiorly to the proximal margin of the urethra to form the vesicourethral anastomosis. The taut anatomic tethering at the anastomotic site could also contribute to the restricted interfraction motion of the prostate bed. In addition to prostate bed motion, patient setup error also leads to daily imprecision in target alignment. Several prior investigators have studied the trends and magnitudes of setup error during definitive EBRT delivered to the intact prostate gland (2, 5, 10 –17). Across these studies, the SDs ranged from approximately 1 to 5.5 mm in each axis. In our postprostatectomy study, systematic error ⫾ SD for setup error fell within this range and measured 0.1 mm ⫾ 4.5 mm, 1.1 mm ⫾ 3.9 mm, and ⫺0.2 mm ⫾ 5.1 mm in the LR, SI, and AP axes, respectively. The similarity between setup error outcomes in these two populations is not surprising, as it is unlikely that the precision of patient positioning using skin tattoos would be different among patients who underwent radical prostatectomy compared with those who did not. Total positioning error, which is the sum of daily setup error and interfraction prostate bed motion, can, if uncorrected, lead to dosimetric inaccuracies and potentially compromise clinical outcomes. In our postprostatectomy study, systematic error ⫾ SD for total positioning error was 0.2 mm ⫾ 4.5 mm, 1.2 mm ⫾ 5.1 mm, and ⫺0.3 mm ⫾ 4.5 mm in the LR, SI, and AP axes, respectively. Total positioning errors ⬎5 mm occurred in 14.1%, 38.7%, and 28.2% of fractions in the LR, SI, and AP axes, respectively. Overall, the statistical trends and magnitudes of total positioning error are very similar to those found for setup error. This result reflects a dominant contribution of setup error over prostate bed motion in determining the total positioning error in our study. Several prior investigations have examined total positioning error in the delivery of definitive EBRT to the intact prostate gland (4, 5, 11, 13, 16, 18 –21, 23, 24). The SDs for total positioning error ranged from approximately 2 to 6 mm in each axis. In the vast majority of these studies, target shifts ⬎5 mm were relatively common; and very large displacements, often ⬎15 mm, were observed. In the postprostatectomy setting, Chinnaiyan et al. (33) found that the average total positioning error ⫾ SD in the LR, SI, and AP axes was 3 mm ⫾ 3 mm, 3 mm ⫾ 4 mm, and 5 mm ⫾ 4 mm, respectively. The findings from our study are consistent with the results found for total positioning error in the delivery of EBRT to treat prostate cancer in both the defin-

Prostate bed motion and EPID imaging during postprostatectomy radiotherapy

itive and postprostatectomy settings. In addition, several studies of EBRT delivered to the intact prostate gland have demonstrated that, as in our postprostatectomy patients, setup error dwarfed organ motion as the dominant cause of total positioning error (13, 16, 29, 35). Our findings show that, as with patients undergoing definitive EBRT to the intact prostate gland, target displacements ⬎5 mm occur rather frequently during the delivery of adjuvant and salvage EBRT after radical prostatectomy. If uncorrected, these shifts in target position could result in dosimetric consequences as well as compromised clinical outcomes. The correction of target positioning errors is especially critical in the delivery of EBRT using small prostate bed– only fields or with the use of conformal RT modalities, such as three-dimensional conformal radiotherapy (3DCRT) and intensity-modulated radiotherapy (IMRT). In these settings, there is a heightened dosimetric sensitivity to target motion, and even small shifts of the target volume have the potential to significantly alter the distribution of radiation delivered to the CTV and adjacent normal structures. To address these day-to-day variations in target position, image-guided target localization can be used to correct daily target positioning errors. This daily alignment of the target volume can improve the accuracy and precision of EBRT and permit the reliable use of smaller, more conformal margins surrounding the CTV (2, 26, 30). Several image-guided target localization technologies have been reported in the delivery of definitive RT to the intact prostate, including the use of daily EPID imaging of implanted radiopaque fiducials (4, 5, 10, 15, 29 –31), ultrasound-based techniques (17–19, 21, 24, 25), cone-beam CT (26, 27), and tomotherapy (28). Daily ultrasound-guided target localization using the B-mode Acquisition and Targeting (BAT) system and similar technologies has been used during definitive RT (17–19, 21, 24, 25) as well as in the postprostatectomy setting (33). However, some investigators have found substantial interuser variability in the interpretation of pretreatment ultrasound images (25). This finding suggests that the clinical utility of BAT is highly dependent on the skill of the sonographer. In addition, another study (22) raised concerns about the magnitude of residual target positioning error that remained after ultrasound-based alignment. Image-guided radiotherapy using cone-beam CT (26, 27) and Tomotherapy (28) is also in use. To be implemented, however, technologies such as kilovoltage (kV) cone-beam CT and Tomotherapy require costly modification or replacement of existing equipment. The use of daily electronic portal imaging of gold seed fiducials has been shown by the current study to be a feasible method to achieve target localization and alignment in the delivery of adjuvant or salvage radiotherapy after prostatectomy. In comparison with other methods of imageguided target localization, EPID imaging of gold seed fiducials has several advantages. First, the acquisition and analysis of the pretreatment EPID images can be achieved in a relatively short period, adding only minutes to the treatment fraction (3). In addition, this technique has excellent inter-



D. C. SCHIFFNER et al.

617

observer consistency and requires no modification to the linear accelerator or treatment room. Finally, we have shown that, as with radiopaque markers inserted into the intact prostate gland (31), gold marker seeds implanted into the prostate bed remain fixed within the postoperative tissues and can serve as reliable fiducials to mark the target volume over the course of adjuvant or salvage EBRT after prostatectomy. On the other hand, the acquisition of the daily EPID images does involve the delivery of additional ionizing radiation. In addition, the implantation of the gold seed fiducials under TRUS guidance is an invasive procedure, which must be performed by a skilled clinician. It is also important to note that daily electronic portal imaging can be used without implanted radiopaque fiducials. Patient setup error, which in this study was found to be the dominant cause of total positioning error, can potentially be corrected through the pretreatment visualization of the bony pelvic anatomy alone. We find several problems, however, with the use of EPID imaging to achieve target localization without the daily visualization of the target organ. First, with prostate bed– only irradiation or conformal RT boosts, small treatment fields are used that focus only on the target volume. Our experience shows that the EPID images obtained from these smaller fields exclude the prominent bony pelvic landmarks needed for reliable alignment of the daily portal image with its corresponding DRR. In addition, without fiducials to mark the prostate bed, interfraction internal organ motion cannot be assessed. Without the correction of prostate bed motion, residual target positioning error would remain at the start of the radiation fraction despite the use of ionizing radiation to produce the portal image. Because the imaging of gold seed fiducials permits the simultaneous correction for both setup error and internal prostate bed motion, we believe that the incorporation of implanted radiopaque fiducials maximizes the benefits derived from daily EPID imaging. For this reason, we recommend that the use of this image-guided target localization system be routinely incorporated into the delivery of adjuvant and salvage radiotherapy after prostatectomy. Our clinical investigation addresses many critical issues about target motion in the delivery of postprostatectomy EBRT. Several topics were not addressed by our study that merit further investigation, however. Although our findings suggest that the prostate bed most likely moves along the SI and AP axes in response to bladder and rectal volumes, the dimensions of these adjacent organs were not measured in our study. We also did not examine intrafraction motion of the prostate bed. Future studies could also focus on calculating the dosimetric alterations caused by the target positioning errors measured in this study. In turn, these dosimetric data could be used to determine the changes in tumor control probability and normal tissue complication probability that would have resulted if these target displacements had not been identified and corrected. In addition, studies of clinical outcomes, such as tumor control rates and normal tissue

618

I. J. Radiation Oncology



Biology



Physics

toxicity profiles in patients receiving postprostatectomy RT aided by daily EPID imaging, would help to establish the role of this image-guided target localization system in the delivery of adjuvant and salvage RT after prostatectomy. CONCLUSION As presented in this study, the routine use of daily electronic portal imaging of implanted gold seed fiducials is a feasible method to correct for prostate bed motion, setup error, and total positioning error in the delivery of adjuvant and salvage EBRT after prostatectomy. Although the quantity of internal prostate bed motion was found to be modest, total positioning error, which was largely caused by setup

Volume 67, Number 2, 2007

error, exceeded 5 mm in a relatively large proportion of treatment fractions. This finding is of especially great concern with the use of small-field prostate-bed only RT and conformal treatment planning techniques such as 3D-CRT and IMRT, which are highly sensitive to errors in target positioning. Without the use of target localization and alignment, geographic misses of the target volume and increased radiation exposure to adjacent normal structures might occur, which could result in adverse clinical outcomes. Given the frequency and magnitude of target displacements reported in this investigation, we recommend the use of image-guided target localization and alignment techniques, such as the daily electronic portal imaging of implanted gold seed fiducials, to optimize the accuracy and precision of adjuvant and salvage radiotherapy after prostatectomy.

REFERENCES 1. Langen KM, Jones DT. Organ motion and its management. Int J Radiat Oncol Biol Phys 2001;50:265–278. 2. Schallenkamp JM, Herman MG, Kruse JJ, et al. Prostate position relative to pelvic bony anatomy based on intraprostatic gold markers and electronic portal imaging. Int J Radiat Oncol Biol Phys 2005;63:800 – 811. 3. Chung PW, Haycocks T, Brown T, et al. On-line aSi portal imaging of implanted fiducial markers for the reduction of interfraction error during conformal radiotherapy of prostate carcinoma. Int J Radiat Oncol Biol Phys 2004;60:329 –334. 4. Wu J, Haycocks T, Alasti H, et al. Positioning errors and prostate motion during conformal prostate radiotherapy using on-line isocentre set-up verification and implanted prostate markers. Radiother Oncol 2001;61:127–133. 5. Nederveen A, Lagendijk J, Hofman P. Detection of fiducial gold markers for automatic on-line megavoltage position verification using a marker extraction kernel (MEK). Int J Radiat Oncol Biol Phys 2000;47:1435–1442. 6. Roeske JC, Forman JD, Mesina CF, et al. Evaluation of changes in the size and location 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– 1329. 7. Zelefsky MJ, Crean D, Mageras GS, et al. Quantification and predictors of prostate position variability in 50 patients evaluated with multiple CT scans during conformal radiotherapy. Radiother Oncol 1999;50:225–234. 8. Schild SE, Casale HE, Bellefontaine LP. Movements of the prostate due to rectal and bladder distension: implications for radiotherapy. Med Dosim 1993;18:13–15. 9. Ten Haken RK, Forman JD, Heimburger DK, et al. Treatment planning issues related to prostate movement in response to differential filling of the rectum and bladder. Int J Radiat Oncol Biol Phys 1991;20:1317–1324. 10. Alasti H, Petric MP, Catton CN, et al. Portal imaging for evaluation of daily on-line setup errors and off-line organ motion during conformal irradiation of carcinoma of the prostate. Int J Radiat Oncol Biol Phys 2001;49:869 – 884. 11. Little DJ, Dong L, Levy LB, et al. Use of portal images and BAT ultrasonography to measure setup error and organ motion for prostate IMRT: Implications for treatment margins. Int J Radiat Oncol Biol Phys 2003;56:1218 –1224. 12. Hanley J, Lumley MA, Mageras GS, et al. Measurement of patient positioning errors in three-dimensional conformal radiotherapy of the prostate. Int J Radiat Oncol Biol Phys 1997;37:435– 444.

13. Dawson LA, Mah K, Franssen E, et al. Target position variability throughout prostate radiotherapy. Int J Radiat Oncol Biol Phys 1998;42:1155–1161. 14. Greer PB, Jose CC, Matthews JH. Set-up variation of patients treated with radiotherapy to the prostate measured with an electronic portal imaging device. Australas Radiol 1998;42: 207–212. 15. Vigneault E, Pouliot J, Laverdiere J, et al. Electronic portal imaging device detection of radioopaque markers for the evaluation of prostate position during megavoltage irradiation: A clinical study. Int J Radiat Oncol Biol Phys 1997; 37:205–212. 16. Rudat V, Schraube P, Oetzel D, et al. Combined error of patient positioning variability and prostate motion uncertainty in 3D conformal radiotherapy of localized prostate cancer. Int J Radiat Oncol Biol Phys 1996;35:1027–1034. 17. Lattanzi J, McNeeley S, Donnelly S, et al. Ultrasound-based stereotactic guidance in prostate cancer— quantification of organ motion and set-up errors in external beam radiation therapy. Comput Aided Surg 2000;5:289 –295. 18. Chandra A, Dong L, Huang E, et al. Experience of ultrasoundbased daily prostate localization. Int J Radiat Oncol Biol Phys 2003;56:436 – 447. 19. Lattanzi J, McNeeley S, Hanlon A, et al. Ultrasound-based stereotactic guidance of precision conformal external beam radiation therapy in clinically localized prostate cancer. Urology 2000;55:73–78. 20. Morr J, DiPetrillo T, Tsai JS, et al. Implementation and utility of a daily ultrasound-based localization system with intensitymodulated radiotherapy for prostate cancer. Int J Radiat Oncol Biol Phys 2002;53:1124 –1129. 21. Serago CF, Chungbin SJ, Buskirk SJ, et al. Initial experience with ultrasound localization for positioning prostate cancer patients for external beam radiotherapy. Int J Radiat Oncol Biol Phys 2002;53:1130 –1138. 22. Van den Heuvel F, Powell T, Seppi E, et al. Independent verification of ultrasound based image-guided radiation treatment, using electronic portal imaging and implanted gold markers. Med Phys 2003;30:2878 –2887. 23. Lattanzi J, McNeely S, Hanlon A, et al. Daily CT localization for correcting portal errors in the treatment of prostate cancer. Int J Radiat Oncol Biol Phys 1998;41:1079 –1086. 24. Lattanzi J, McNeeley S, Pinover W, et al. A comparison of daily CT localization to a daily ultrasound-based system in prostate cancer. Int J Radiat Oncol Biol Phys 1999;43:719 – 725.

Prostate bed motion and EPID imaging during postprostatectomy radiotherapy

25. Langen KM, Pouliot J, Anezinos C, et al. Evaluation of ultrasound-based prostate localization for image-guided radiotherapy. Int J Radiat Oncol Biol Phys 2003;57:635– 644. 26. Ghilezan M, Yan D, Liang J, et al. Online image-guided intensity-modulated radiotherapy for prostate cancer: How much improvement can we expect? A theoretical assessment of clinical benefits and potential dose escalation by improving precision and accuracy of radiation delivery. Int J Radiat Oncol Biol Phys 2004;60:1602–1610. 27. Pouliot J, Bani-Hashemi A, Chen J, et al. Low-dose megavoltage cone-beam CT for radiation therapy. Int J Radiat Oncol Biol Phys 2005;61:552–560. 28. Mackie TR, Kapatoes J, Ruchala K, et al. Image guidance for precise conformal radiotherapy. Int J Radiat Oncol Biol Phys 2003;56:89 –105. 29. Millender LE, Aubin M, Pouliot J, et al. Daily electronic portal imaging for morbidly obese men undergoing radiotherapy for localized prostate cancer. Int J Radiat Oncol Biol Phys 2004;59:6 –10. 30. Litzenberg D, Dawson LA, Sandler H, et al. Daily prostate

31.

32.

33. 34.

35.



D. C. SCHIFFNER et al.

619

targeting using implanted radiopaque markers. Int J Radiat Oncol Biol Phys 2002;52:699 –703. Pouliot J, Aubin M, Langen KM, et al. (Non)-migration of radiopaque markers used for on-line localization of the prostate with an electronic portal imaging device. Int J Radiat Oncol Biol Phys 2003;56:862– 866. Connolly JA, Shinohara K, Presti JC Jr., et al. Local recurrence after radical prostatectomy: characteristics in size, location, and relationship to prostate-specific antigen and surgical margins. Urology 1996;47:225–231. Chinnaiyan P, Tomee W, Patel R, et al. 3D-ultrasound guided radiation therapy in the postprostatectomy setting. Technol Cancer Res Treat 2003;2:455– 458. Kruse JJ, Herman MG, Hagness CR, et al. Daily monitoring of patient setup variation and prostatic fossa tissue motion with an electronic portal imaging device [Abstract]. Int J Radiat Oncol Biol Phys 2001;51:387. Antolak JA, Rosen II, Childress CH, et al. Prostate target volume variations during a course of radiotherapy. Int J Radiat Oncol Biol Phys 1998;42:661– 672.