Daily Isocenter Correction With Electromagnetic-Based Localization Improves Target Coverage and Rectal Sparing During Prostate Radiotherapy

Daily Isocenter Correction With Electromagnetic-Based Localization Improves Target Coverage and Rectal Sparing During Prostate Radiotherapy

Int. J. Radiation Oncology Biol. Phys., Vol. 76, No. 4, pp. 1092–1099, 2010 Copyright Ó 2010 Elsevier Inc. Printed in the USA. All rights reserved 036...

820KB Sizes 0 Downloads 40 Views

Int. J. Radiation Oncology Biol. Phys., Vol. 76, No. 4, pp. 1092–1099, 2010 Copyright Ó 2010 Elsevier Inc. Printed in the USA. All rights reserved 0360-3016/10/$–see front matter

doi:10.1016/j.ijrobp.2009.03.036

CLINICAL INVESTIGATION

Prostate

DAILY ISOCENTER CORRECTION WITH ELECTROMAGNETIC-BASED LOCALIZATION IMPROVES TARGET COVERAGE AND RECTAL SPARING DURING PROSTATE RADIOTHERAPY RAMJI RAMASWAMY RAJENDRAN, M.D., PH.D.,* JOHN P. PLASTARAS, M.D., PH.D.,* ROSEMARIE MICK, M.S.,y DIANE MCMICHAEL KOHLER, C.M.D.,* ALIREZA KASSAEE, PH.D.,* AND NEHA VAPIWALA, M.D.* * Department of Radiation Oncology, Hospital of the University of Pennsylvania, Philadelphia, PA; and y Department of Biostatistics and Epidemiology, University of Pennsylvania School of Medicine, Philadelphia, PA Purpose: To evaluate dosimetric consequences of daily isocenter correction during prostate cancer radiation therapy using the Calypso 4D localization system. Methods and Materials: Data were analyzed from 28 patients with electromagnetic transponders implanted in their prostates for daily target localization and tracking. Treatment planning isocenters were recorded based on the values of the vertical, longitudinal, and lateral axes. Isocenter location obtained via alignment with skin tattoos was compared with that obtained via the electromagnetic localization system. Daily isocenter shifts, based on the isocenter location differences between the two alignment methods in each spatial axis, were calculated for each patient over their entire course. The mean isocenter shifts were used to determine dosimetric consequences of treatment based on skin tattoo alignments alone. Results: The mean ± SD of the percentages of treatment days with shifts beyond + = 0.5 cm for vertical, longitudinal and lateral shifts were 62% + = 28%, 35% + = 26%, and 38% + =21%, respectively. If daily electromagnetic localization was not used, the excess in prescribed dose delivered to 70% of the rectum was 10 Gy and the deficit in prescribed dose delivered to 95% of the planning target volume was 10 Gy. The mean isocenter shift was not associated with the volumes of the prostate, rectum, or bladder, or with patient body mass index. Conclusions: Daily isocenter localization can reduce the treatment dose to the rectum. Correcting for this variability could lead to improved dose delivery, reduced side effects, and potentially improved treatment outcomes. Ó 2010 Elsevier Inc. Prostate cancer, Image-guided radiation therapy, Daily target localization, Organ motion, Dosimetry.

Prostate cancer is the most commonly diagnosed cancer in men in the United States, with over 180,000 cases detected annually (1). The treatment options for early stage prostate cancer include radical prostatectomy, high dose brachytherapy, and high dose external beam radiation therapy (RT). High dose ($78 Gy) definitive radiation is the standard as a result of several dose escalation trials, which also revealed the tolerances of normal tissues in the treatment field. Recently reported data from the Proton Radiation Oncology Group 9509 study comparing 70.2 Gy vs. 79.2 Gy delivered using combined photon–proton treatment show a 5-year freedom from biochemical failure of 78.8% vs. 91.3% for the increased dose (2). Although this study showed no differences in acute or

late normal tissue toxicity, many other trials have shown increased side effects from dose escalation for prostate cancer radiation. At the MD Anderson Cancer Center, patients were treated with either 70 Gy or 78 Gy, and the higher dose increased the 8.7-year biochemical recurrence-free survival from 59% to 78%, but also increased the Grade 2 or higher rectal toxicity from 13% to 26% (3). The Medical Research Council–RT01 randomized patients to 64 Gy or 74 Gy with conformal external beam RT and also showed an improvement in progression-free survival (60% vs. 71%), but there was an increased incidence of late bowel toxicity (24% vs. 33%) at 5 years (4). The Dutch study randomized patients to 68 Gy or 78 Gy, which improved biochemical progression free survival (bPFS) (54% vs. 64%) (5), but increased late

Reprint requests to: Neha Vapiwala, M.D., Hospital of the University of Pennsylvania, Department of Radiation Oncology, 2 Donner, 3400 Spruce Street, Philadelphia, PA 19104. Tel: (215) 662-2428; Fax (215) 349-5445; E-mail: vapiwala@xrt. upenn.edu

Presented at the 50th Annual Meeting of the American Society for Radiation Oncology, Boston, MA, September, 19–23, 2008. Conflict of interest: none. Received Dec 21, 2008, and in revised form March 13, 2009. Accepted for publication March 15, 2009.

INTRODUCTION

1092

Daily isocenter correction improves target coverage d R. R. RAJENDRAN et al.

gastrointestinal (GI) toxicity which was significant for late rectal bleeding requiring cauterization (6). The differences in rectal toxicity were worse for patients with a history of abdominal surgery and was directly proportional to the length of rectum treated, specifically the volume receiving 55 Gy and 65 Gy (7). Furthermore, this group showed that the presence of acute rectal side effects is a significant predictor of late rectal toxicity (6). Acute and late toxicity from prostate RT can be reduced to some extent by the use of intensity-modulated radiotherapy (IMRT) techniques that focus a high dose to the prostate while decreasing dose to the bladder and rectum (8). A retrospective study of patients treated between 66 and 81 Gy showed that, compared with 3D-CRT, use of IMRT significantly decreases the risk of GI toxicities (13–5%) (9). With higher doses being delivered with increased conformality, it has become critical that the isocenter of the prostate treatment volume be placed with precision in daily set-up (10). It has been known for many years that prostate movement is an issue for treatment planning. There is significant motion of the prostate as the bladder anteriorly and the rectum posteriorly are filled to different volumes on a daily basis (11). There are two types of motion that affect target verification: interfraction motion of a target from day-to-day, and intrafraction movement during the treatment itself. Correction for these types of target movement can be achieved in different ways. Interfraction prostate position from daily set-up variation has been addressed by various techniques including ultrasound, kV or MV X-ray imaging with and without gold fiducial placement, and cone beam or diagnostic CT imaging (12–15). Litzenberg et al. suggested that implanted markers are superior to skin markers and can be used to reduce anterior–posterior margins by about half (13). Day-to-day variation in patient set-up has been studied using the location of implanted fiducials relative to skin tattoos (16). In this study, shifts were made if the localization was off by 3 mm or greater, which occurred 90% of the time. Movement of the prostate during a treatment (intrafraction motion) is thought to be caused predominantly by contraction/relaxation of the pelvic floor and by rectal gas (17, 18). Immobilization of the prostate with endorectal balloons can be used to reduce prostate motion and positional variability during an entire course of therapy (19). However, this typically requires bowel preparation and can cause patient discomfort. The electromagnetic localization system (Calypso 4D Localization System–Calypso System, Calypso Medical, Seattle, WA) based on implanted transponders allows three-dimensional interfractional positioning and intrafractional tracking of the target isocenter. Operation and accuracy of the electromagnetic localization and tracking system have been described previously (20). This system allows real-time tracking of the prostate during RT, as well as objective daily measurements of pretreatment isocenter shifts that are automatically recorded before each fraction (21, 22). With highly conformal radiation, errors in isocenter localization may result in major changes in dosimetry. The accurately recorded daily shifts using an electromagnetic localization system can be used to readily determine the impact of daily isocenter ver-

1093

ification on target dosimetry. Correcting for this variability could lead to improved dose delivery, reduced side effects, and potentially improved treatment outcomes. MATERIALS AND METHODS Data collection Patients with pathologically confirmed early-stage adenocarcinoma of the prostate were retrospectively examined at the Hospital of the University of Pennsylvania. This study is a single institution, institutional review board–approved, retrospective study. Patients were enrolled sequentially from May 10, 2007, to April 10, 2008. Our data set included over 1100 separate shifts in 28 patients over an average of 40 pretreatment set-up time points. Patients were treated with 7920 cGy in 44 fractions of 180 cGy each using IMRT with electromagnetic target localization and tracking. The major limitation for use of this electromagnetic localization system is that the distance in the central sagittal plane between the prostate and the anterior body surface must be <17 cm to achieve both target localization and tracking. A distance of 17 to 23 cm permits localization only, whereas a distance >23 cm precludes use of this system. Patients with prostate–anterior body surface separations <23 cm each had three transponders implanted in their prostates using transrectal ultrasound guidance. Transponders were implanted in the right base, left base, and apex of the prostate within the prostatic capsule and off midline to avoid loss through the urethra. Prophylactic oral antibiotics were given to prevent infection. The planning treatment volume (PTV) was defined as the volume encompassing the prostate and proximal two-thirds of the seminal vesicles with 6 mm posterior margin and 1 cm margin in the other directions. The volumes of the prostate, rectum (defined as the entire volume of the rectum contoured 1.5 cm superior and inferior to the PTV), and bladder, were measured for each patient using the treatment planning CT scan. Body mass index (BMI) was calculated from measurements of height and weight at initial consultation. Computed tomography simulation was performed 4 or more days after implantation. Patients were placed in the supine position with a foot box designed to immobilize the hip and knee joints. The CT images were performed with a slice thickness of 1.0 mm through the transponders implanted in the prostate. Patients were encouraged to have a full bladder at the time of simulation and for each treatment. Before each fraction of RT, patients were first positioned according to the skin tattoos that were placed at simulation and that correlated to the isocenter based on the geometric center of the transponders. This location was set as the origin. Patients were then repositioned to the isocenter based on the electromagnetic localization system, which generates exact measurements on the vertical, longitudinal, and lateral axes of the locational shifts from the origin (skin tattoo alignment). The distance and direction of the daily shift on each axis, if any, were recorded over the entire course of treatment. ‘‘Isocenter shift’’ was defined as the distance (in centimeters) between the isocenter location by skin tattoo–based alignment compared with the isocenter location by the electromagnetic system. The mean isocenter shift was calculated for each patient and was employed to estimate the location of the isocenter that would have been used if daily electromagnetic localization was not used. With respect to target tracking, per departmental protocol, intrafraction movement of >5 mm in any axis for >10 s consecutively triggers an alarm that alerts the therapist to stop treatment and to reposition the patient. A daily quality assurance check was performed to test the system for a positioning accuracy of <2 mm. Pan et al. demonstrated that

I. J. Radiation Oncology d Biology d Physics

1094

Volume 76, Number 4, 2010

changes in prostate volume can alter transponder position during a course of treatment (23). We used the residual index to measure intertransponder pair distances so that discrepancies in these distances could be reconciled if needed, but we found no appreciable changes in our study group. On the few occasions when equipment malfunction prevented use of the system, skin marks alone were used for isocenter localization and a shift was not recorded. Some patients had more shifts than the number of treatment fractions because these patients had treatment interruptions because of intrafractional motion, causing treatment interruption during a fraction and relocalization of the isocenter. After the determination of the average isocenter shift for each patient, it was possible to reproduce a dose distribution that would estimate the dose that would have been delivered if only the skin tattoo–based alignment had been used. We then calculated the dose differences to various organs between the skin tattoo–based treatment plan and the electromagnetic localization treatment plan.

Statistical methods Descriptive statistics were used to summarize distributions of variables. Summary statistics, such as means, standard deviations (SD), and ranges, were computed for continuous variables. Frequencies and percentages were computed for categorical variables. Graphical methods, including scatter and box plots, were used extensively to examine distributions of variables and to guide data transformations if warranted. The mean isocenter shifts in the vertical, longitudinal, and lateral axes were plotted and examined for each individual patient. Variations in time trends were noted among patients. To model the relationship of isocenter shift over the treatment period, a random-effects regression model was used. The random effects model adjusts for within-patient correlation among the repeated isocenter shift measurements and allows patient-specific intercepts (24). The model takes the following form: Yij ¼ B0 þ B1 T þ bi þ eij th

in which Yij is the j isocentric shift measurement for the ith patient. The number of isocentric shift measurements varied among patients, such that j = 1,.,ni . B0 and B1 are the population estimates of intercept and slope over time (T), respectively; bi is the estimate of the random effect for intercept; and eij is the random error term. The random effects regression was performed using the xtreg command in Stata/ SE 10.0 (Stata Corporation, College Station, TX). Within-patient comparisons of bladder and rectum dose at 50% and 70% volume and PTV dose at 95% volume with and without isocenter correction were performed by Student’s paired t test. Pearson’s correlation was used to assess the magnitude of linear correlation between the mean isocenter shift over the treatment period in the vertical, longitudinal, and lateral axes and the volumes of bladder, rectum, and PTV, the BMI, and the differences in bladder and rectum dose at 50% and 70% volume and PTV dose at 95% volume with and without isocenter correction. All tests were two-sided and conducted at a 5% significance level. Student’s paired t tests and Pearson correlations were performed in SPSS 15.0 (SPSS Inc., Chicago, IL).

RESULTS Isocenter shifts over time The 28 study patients contributed an average of 40 isocenter shift measurements (range, 28–46) over the treatment period. Figure 1a–c display the isocenter shift data for all 28 patients and the regression lines estimated by the random

Fig 1. (a–c) Daily isocenter shifts for 28 individual patients over a treatment period of up to 84 days. Isocenter shift is defined as difference between the estimated skin tattoo–based location compared with an electromagnetic localization (a) Vertical isocenter shift data with random-effects regression equation: Vertical shift = 0.5582 + 0.0044*day (solid line) (b) Longitudinal isocenter shift data with random effects regression equation: Longitudinal shift = 0.2963 + 0.0024*day (solid line) (c) Lateral isocenter shift data with random effects regression equation: Lateral shift = 0.0793 + 0.00023*day (solid line).

Daily isocenter correction improves target coverage d R. R. RAJENDRAN et al.

Coefficient Vertical Intercept Slope* Longitudinal Intercept Slope* Lateral Intercept Slope*

SE

z Statistic

p Value

0.56 0.0044

0.10 0.00084

5.65 5.24

<0.001 <0.001

0.30 0.0024

0.071 0.00050

4.17 4.86

<0.001 <0.001

0.076 0.00091

1.05 0.26

0.079 0.00023

0.30 0.80

a MEAN ISOCENTER SHIFT (cm)

Table 1. Random-effects linear regression of isocenter shift over the treatment period

1095

VERTICAL 2.0 1.5 1.0 0.5 0.0 -0.5

PATIENT

* Change in isocenter shift (in cm) per day on treatment.

100

Percent of Treatment Days with Isocenter Shifts Beyond +/- 0.5 cm

90

LONGITUDINAL 1.0 0.5 0.0 -0.5 -1.0 -1.5

PATIENT

c MEAN ISOCENTER SHIFT (cm)

Percentage of treatment days associated with significant isocenter shifts In examination of the isocenter shifts, we noted that a substantial percentage of patients had isocenter shifts of more than 0.5 cm in absolute value (less than 0.5 cm or greater than 0.5 cm). We calculated the percentage of treatment

MEAN ISOCENTER SHIFT (cm)

b effects models. Modeling isocenter shifts over the treatment period by random effects regression revealed a small but statistically significant trend over time in the vertical (anterior–posterior) shift (p < 0.001). On average, an increase in vertical shift of 0.0044 cm per day was estimated by the model, which would lead to a 0.26-cm shift over a 60-day treatment period (Table 1). There was also a statistically significant but even smaller trend over time in the longitudinal (craniocaudal) shift (p < 0.001). On average, an increase in longitudinal shift of 0.0024 cm per day was estimated by the model, which would lead to a 0.14-cm shift over a 60-day treatment course. There was no trend over time in the lateral (right–left) shift (p = 0.80).

LATERAL 1.2 1.0 0.8 0.6 0.4 0.2 0.0 -0.2 -0.4 -0.6

PATIENT

80

Fig. 3. Waterfall plots showing range of isocenter shifts in the vertical (A, anterior–posterior), longitudinal (B, craniocaudal), and lateral (C, right–left) directions among 28 patients.

70 60 50 40 30 20 10 0

Vertical

Longitudinal Spatial Axis

Lateral

Fig. 2. Boxplots of the percentage of treatment days associated with isocenter shifts beyond 0.5cm for each spatial axis. The lower and upper edges of the box and horizontal line within the box indicate the 25th, 75th, and 50th percentiles, respectively.

days associated with isocenter shifts beyond + = 0.5 cm for each spatial axis (Fig. 2). The mean + SD (range) of the per= 0.5 cm for centages of treatment days with shifts beyond + = vertical, longitudinal, and lateral shifts were as follows: 62% + = 28% (3% to 95%), 35% + = 26% (4% to 90%) and 21% (9% to 80%), respectively. Of the 28 patients, 38% + = 12 (43%) had vertical isocenter shifts beyond + = 0.5 cm for 80% or more of their treatments. To test whether radiation therapists became increasingly reliant on the electromagnetic system instead of skin marks over the first several months of its implementation, we examined the frequency of large isocenter shifts (+ = 0.5 cm) with respect to patient study number.

1096

I. J. Radiation Oncology d Biology d Physics

Volume 76, Number 4, 2010

Fig. 4. (a) Representative axial image of dose delivered to planning target volume using isocenter set by electromagnetic localization beacons at simulation (Left). White arrows illustrate the posterior movement of dose caused by using the average isocenter set by skin tattoos during a course of radiation therapy (Right) (b). Representative dose–volume histogram (DVH) based on planning isocenter (triangles) compared with DVH based on the mean isocenter determined by skin tattoos during a course of radiation therapy (squares) for the clinical target volume (red), planning target volume (blue), rectum (brown), and bladder (orange).

This study number marked consecutive patients as they started treatment, although their courses did overlap. There was no correlation between study number and the frequency of large shifts in any spatial axis (data not shown), suggesting that the relatively large shifts observed did not result from changing vigilance when aligning to skin marks.

Mean daily shifts Combining data from all patients, the mean + = SD (range) isocenter shifts over the treatment period in the vertical, longitudinal, and lateral spatial axes were 0.71 + = 0.74 cm (1.69 to 3.14 cm), 0.22 + = = 0.49 cm (1.76 to 2.12 cm) and 0.08 + 0.68 cm (4.50 to 3.56 cm), respectively. Waterfall plots are

Table 2. Dosimetric consequences of isocenter shifts on the dose to 95% of the planning target volume (PTV) and 70% of the rectum

95% PTV 70% Rectum

Electromagnetic localization + SD isocenter, mean dose = (range)

Skin tattoo alignment isocenter, + SD (range) mean dose =

+ 1.3 Gy ( 79.3–73.1 Gy) 76.7 = + 3.5 Gy (65.9–51.6 Gy) 60.7 =

+ 9.0 Gy (77.1–49.6 Gy) 66.9 = + 12.1 Gy (83.1–43.4 Gy) 70.5 =

Daily isocenter correction improves target coverage d R. R. RAJENDRAN et al.

Table 3. Comparison of dose differences for bladder, rectal, and PTV volumes with or without isocenter shifts Mean difference in dose* (cGy) SD (cGy) t Statistic p Value Bladder Dose at 50% volume Dose at 70% volume Rectal Dose at 50% volume Dose at 70% volume PTV Dose at 95% volume

41 304

831 1069

0.26 1.51

0.80 0.14

1447 1086

1374 1111

5.57 5.17

<0.001 <0.001

1006

888

5.99

<0.001

* Value without isocenter shift minus value with isocenter shift.

presented to show the range of mean isocenter shifts in the vertical, longitudinal, and lateral directions amongst the 28 patients (Fig. 3a–c). Dosimetric implications Averaged over the entire treatment course, the variation in daily isocenter localization using skin tattoos alone could have led to significant changes in dose to bladder, rectum, and prostate. To examine this issue further, we compared the dose that was planned during simulation (i.e., with the isocenter based on an electromagnetic localization system) vs. the dose that would have been delivered based on skin tattoo– based alignment alone. An example axial slice of the CT scan with an overlaid dose distribution diagram of the PTV is shown in Fig. 4a. The dose–volume histogram for the same patient using the isocenter based on skin tattoos and the isocenter based on the electromagnetic localization system is shown in Fig. 4b. The difference in the dose calculations at different volumes was most evident as a decreased dose to the PTV and an increased dose to the rectum (Table 2). The mean + = SD (range) planned dose delivered using the true isocenter with electromagnetic localization was 76.7 + = 1.3 Gy (79.33.5 Gy (65.9–51.6 73.1 Gy) to 95% of the PTV and 60.7 + = Gy) to 70% of the rectum. The mean + SD (range) dose using =

1097

the isocenter alignment with skin tattoos only would have resulted in 66.9 + = 9.0 Gy (77.1–49.6 Gy) to 95% of the PTV and 70.5 + = 12.1 Gy (83.1–43.4 Gy) to 70% of the rectum. Variation in dose to organs with and without isocenter shift Within-patient comparisons revealed that rectal doses at 50% and 70% volume and PTV dose at 95% volume were highly statistically significantly different with and without isocenter shift (p < 0.001 paired t test), as shown in Table 3. Rectal doses at 50% and 70% volume were lower with, as compared to without, isocenter shift. The PTV dose at 95% volume was lower without, as compared to with, isocenter shift. The mean differences in dose were approximately 10 Gy (Figure 5). Little difference was observed for bladder doses at 50% or 70% volume, with mean differences in dose of approximately 40 and 300 cGy, respectively. Correlation of isocenter shift and variation in dose to organs The average vertical isocenter shift over the treatment period for each patient was significantly correlated with the patient’s difference in rectal doses at 50% and 70% volume and PTV at 95% volume, with or without isocenter shift (p # 0.001, Table 4). In addition, a patient’s average longitudinal isocenter shift over the treatment period was significantly correlated with the patient’s difference in bladder doses at 50% and 70% volume with or without isocenter shift (p < 0.001). There were no significant correlations of bladder, rectum and PTV volumes or the patient’s BMI, with the increased dose to the rectum and decreased dose to the PTV (data not shown). DISCUSSION This study quantifies the impact of daily isocenter localization on dose to target and normal tissues during a course of prostate radiotherapy in clinically significant terms. The methods used here are especially unique in that they involve precise measurements of daily isocenter shifts acquired through an electromagnetic target localization system.

Table 4. Correlations of dose to rectum, bladder and PTV with average isocenter shift in lateral, longitudinal and vertical axes Difference in Doses With and Without Isocenter Shift Rectal dose at Isocenter shift Vertical Pearson correlation, r p Longitudinal Pearson correlation, r p Lateral Pearson correlation, r p Abbreviation: Vol = volume.

PTV dose at

Bladder dose at

50% Vol

70% Vol

50% Vol

70% Vol

95% Vol

0.71 <0.001

0.59 0.001

0.29 0.13

0.38 0.05

0.64 <0.001

0.21 0.29

0.34 0.08

0.75 <.001

0.69 <.001

0.13 0.52

0.12 0.54

0.16 0.42

0.07 0.71

0.02 0.93

0.11 0.57

I. J. Radiation Oncology d Biology d Physics

1098

a

8500

Rectum Dose at 70% Volume (cGy)

8000 7500 7000 6500 6000 5500 5000 4500 4000

LOCALIZATION SYSTEM

SKIN TATTOOS

b 8500

PTV dose at 95% Volume (cGy)

8000 7500 7000 6500 6000 5500 5000 4500 4000

LOCALIZATION SYSTEM

SKIN TATTOOS

Fig. 5. (a) Mean  SD (range) dose to 70% of the rectum volume, with and without isocenter correction. Skin tattoo–based estimated dose: 70.5  12.1 Gy (range, 43.4–83.1 Gy) vs. electromagnetic localization-planned dose: 60.7  3.5 Gy (range, 51.6–65.9 Gy) (b). Mean  SD (range) dose to 95% of the planning target volume with and without isocenter correction. Skin tattoo–based estimated dose: 66.9  9.0 Gy (range, 77.1–49.6 Gy) vs. electromagnetic localization–planned dose 76.7  1.3 Gy (range, 79.3–73.1 Gy).

Patients require shifts of >0.5 cm for more than 60% of treatment fractions. Furthermore, our data demonstrate a very gradual posterior drift of the prostate toward the sacrum and an inferior drift toward the feet that occurs over the 9-week treatment period. In comparing the original dose distributions (which assume noncorrected skin-based localization only) to the revised dose distributions created using the image-guided isocenters, we found that the former resulted in a 10 Gy dose increase to the rectum and a 10 Gy dose decrease delivered to the PTV. These findings highlight the necessity for daily prostate localization when using high dose, highly conformal external beam radiation therapy. The magnitude of these dosimetric advantages with using daily isocenter localization is surprisingly large. Admittedly,

Volume 76, Number 4, 2010

these dramatic differences in daily shifts could have been exaggerated because of an over-reliance on the electromagnetic localization system. As therapists become increasingly comfortable with and reliant upon a localization system that is used to correct the skin marks and direct the final patient set-up just before treatment, there may be a tendency to become more careless with the skin marks over time. Thus it was reassuring to find that there were no statistically significant trends toward larger shifts (i.e., percentage of treatment days with shifts beyond + = 0.5 cm) with successive patients treated with this localization system. This lack of a systematic confounding effect improves our confidence in the validity of the dosimetric implications of our data. As for the increasing shifts that occurred preferentially in the vertical axis, these are explained primarily by the gradual posterior drift of the prostate seen in our study patients over the course of radiation, a phenomenon that has previously been described. The drift in the inferior direction was even more gradual. These drifts are likely caused by one or more factors, including progressive relaxation of patients’ pelvic floor muscles as they adjust to treatment, as well as possible decrease in the rectal volume because of increased stooling as treatment proceeds. We then sought to investigate potential factors that might help to identify specific patient subsets who are most likely to have major discrepancies in daily set-up if treated without using daily isocenter correction, and thus the greatest variations in planned vs. delivered dose to isocenter and normal tissues. In theory, these patients may have modifiable factors to help attenuate the potential discrepancies and to reduce the risk of clinically inferior outcomes. Volumes of prostate, bladder, and rectum did not predict for larger or smaller dosimetric discrepancies. We also analyzed the effect of patient BMI, as longer distances from the isocenter to the external surface of the patient may contribute to greater set-up variation with the skin marks, as well as the electromagnetic radiofrequency–based system (25). Furthermore, a larger body habitus with significant pannus could possibly alter the position of skin marks relative to the prostate isocenter (14). However, our results did not find a correlation with BMI and dosimetric discrepancies. These data further support the use of daily isocenter correction for all prostate cancer patients undergoing definitive high-dose external beam radiation. Several studies have now established the biochemical benefit of radiation dose escalation in prostate cancer, although this achievement has come at the cost of greater toxicity. It is possible that this increased toxicity is a reflection of not only the higher total dose, but also the inclusion of more rectum in the fields over the course of treatment. This latter exacerbating factor may be especially relevant during those final fractions of treatment, for which our data indicate the greatest need for on-board correction of skin-based isocenter localization. The use of technologies such as electromagnetic transponders for daily target localization, and potentially for real-time prostate tracking, will presumably become increasingly important as definitive prostate radiation treatment regimens move ever more toward hypofractionated schedules and/or extreme conformality.

Daily isocenter correction improves target coverage d R. R. RAJENDRAN et al.

1099

REFERENCES 1. Jemal A, Siegel R, Ward E, et al. Cancer statistics, 2008. CA Cancer J Clin 2008;58:71–96. 2. Zietman AL. Correction: Inaccurate analysis and results in a study of radiation therapy in adenocarcinoma of the prostate. J Am Med Assoc 2008;299:898–899. 3. Kuban DA, Tucker SL, Dong L, et al. Long-term results of the M.D. Anderson Randomized Dose-Escalation Trial for Prostate Cancer. Int J Radiat Oncol Biol Phys 2008;70:67–74. 4. Dearnaley DP, Sydes MR, Graham JD, et al. Escalated-dose vs. standard-dose conformal radiotherapy in prostate cancer: First results from the MRC RT01 randomised controlled trial. Lancet Oncol 2007;8:475–487. 5. Peeters ST, Heemsbergen WD, Koper PC, et al. Dose-response in radiotherapy for localized prostate cancer: Results of the Dutch Multicenter Randomized Phase III trial comparing 68 Gy of radiotherapy with 78 Gy. J Clin Oncol 2006;24:1990–1996. 6. Peeters ST, Heemsbergen WD, van Putten WL, et al. Acute and late complications after radiotherapy for prostate cancer: Results of a multicenter randomized trial comparing 68 Gy to 78 Gy. Int J Radiat Oncol Biol Phys 2005;61:1019–1034. 7. Peeters ST, Hoogeman MS, Heemsbergen WD, et al. Volume and hormonal effects for acute side effects of rectum and bladder during conformal radiotherapy for prostate cancer. Int J Radiat Oncol Biol Phys 2005;63:1142–1152. 8. Al-Mamgani A, Heemsbergen WD, Peeters ST, et al. Role of intensity-modulated radiotherapy in reducing toxicity in dose escalation for localized prostate cancer. Int J Radiat Oncol Biol Phys 2009;73:685–691. 9. Zelefsky MJ, Chan H, Hunt M, et al. Long-term outcome of high dose intensity modulated radiation therapy for patients with clinically localized prostate cancer. J Urol 2006;176: 1415–1419. 10. Zhu SY, Mizowaki T, Norihisa Y, et al. Comparisons of the impact of systematic uncertainties in patient setup and prostate motion on doses to the target among different plans for definitive external-beam radiotherapy for prostate cancer. Int J Clin Oncol 2008;13:54–61. 11. Kupelian PA, Langen KM, Willoughby TR, et al. Image-guided radiotherapy for localized prostate cancer: Treating a moving target. Semin Radiat Oncol 2008;18:58–66. 12. Bylund KC, Bayouth JE, Smith MC, et al. Analysis of interfraction prostate motion using megavoltage cone beam computed tomography. Int J Radiat Oncol Biol Phys 2008;72:949–956. 13. Litzenberg DW, Balter JM, Hadley SW, et al. Influence of intrafraction motion on margins for prostate radiotherapy. Int J Radiat Oncol Biol Phys 2006;65:548–553.

14. 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. 15. Scarbrough TJ, Golden NM, Ting JY, et al. Comparison of ultrasound and implanted seed marker prostate localization methods: Implications for image-guided radiotherapy. Int J Radiat Oncol Biol Phys 2006;65:378–387. 16. Chen J, Lee RJ, Handrahan D, et al. Intensity-modulated radiotherapy using implanted fiducial markers with daily portal imaging: Assessment of prostate organ motion. Int J Radiat Oncol Biol Phys 2007;68:912–919. 17. Ghilezan MJ, Jaffray DA, Siewerdsen JH, et al. Prostate gland motion assessed with cine-magnetic resonance imaging (cineMRI). Int J Radiat Oncol Biol Phys 2005;62:406–417. 18. Mah D, Freedman G, Milestone B, et al. Measurement of intrafractional prostate motion using magnetic resonance imaging. Int J Radiat Oncol Biol Phys 2002;54:568–575. 19. El-Bassiouni M, Davis JB, El-Attar I, et al. Target motion variability and on-line positioning accuracy during external-beam radiation therapy of prostate cancer with an endorectal balloon device. Strahlenther Onkol 2006;182:531–536. 20. Kupelian P, Willoughby T, Mahadevan A, et al. Multi-institutional clinical experience with the Calypso System in localization and continuous, real-time monitoring of the prostate gland during external radiotherapy. Int J Radiat Oncol Biol Phys 2007;67:1088–1098. 21. Langen KM, Willoughby TR, Meeks SL, et al. Observations on real-time prostate gland motion using electromagnetic tracking. Int J Radiat Oncol Biol Phys 2008;71:1084–1090. 22. Willoughby TR, Kupelian PA, Pouliot J, et al. Target localization and real-time tracking using the Calypso 4D localization system in patients with localized prostate cancer. Int J Radiat Oncol Biol Phys 2006;65:528–534. 23. Pan CC, Sandler HM, Levine LL, et al. Targeting the Prostate with External Beam Radiotherapy: Prostate Volume Changes and their Relationship to Implantable Markers. ASCO 2005 Prostate Cancer Symposium. Orlando, FL; 2005. 24. Diggle PJ, Liang KY, Zeger SL. Analysis of Longitudinal Data. Oxford, UK: Oxford University Press; 1994. 25. Stroup SP, Cullen J, Auge BK, et al. Effect of obesity on prostate-specific antigen recurrence after radiation therapy for localized prostate cancer as measured by the 2006 Radiation Therapy Oncology Group-American Society for Therapeutic Radiation and Oncology (RTOG-ASTRO) Phoenix consensus definition. Cancer 2007;110:1003–1009.