Impact of setup uncertainty in the dosimetry of

Impact of setup uncertainty in the dosimetry of

(Parte 1 de 2)


PATIENTS TREATED WITH PEACOCK/IMRT Presented at the 87th Scientific Assembly and Annual Meeting of Radiological Society of North America, Chicago, IL, November 25–30, 2001.

E. BRIAN BUTLER, M.D., and WALTER H. GRANT, I, PH.D. Department of Veterans Affairs Medical Center and Baylor College of Medicine, Houston, TX

Abstract—The purpose of this paper was to assess the effect of setup uncertainty on dosimetry of prostate, seminal vesicles, bladder, rectum, and colon in prostate cancer patients treated with Peacock intensity-modulated radiation therapy (IMRT). Ten patients underwent computed tomography (CT) scans using the “prostate box” for external, and an “endorectal balloon” for target imobilization devices, and treatment plans were generated (T1). A maximum of 5-m setup error was chosen to model dosimetric effects. Isodose lines from the T1 treatment plan were then superimposed on each patient’s CT anatomy shifted by 5 m toward the cephalad and caudal direction, generating 2 more dosimetric plans (H1 and H2, respectively). Average mean doses ranged from 74.5 to 74.92 Gy for prostate and 73.65 to 74.94 Gy for seminal vesicles. Average percent target volume below 70 Gy increased significantly for seminal vesicles, from 0.53% to 6.26%, but minimally for prostate, from 2.08% to 4.4%. Dose statistics adhered to prescription limits for normal tissues. Setup uncertainty had minimum impact on target dose escalation and normal tissue dosing. The impact of target dose inhomogeneity is currently evaluated in clinical studies. © 2005 American Association of Medical Dosimetrists.

Key Words: Prostate cancer, IMRT, Setup error.

Radiation therapy is a standard therapeutic technique for early and locally-advanced prostate cancer. Conventional techniques have been associated with high failure rates, as they do not allow the delivery of higher curative tumor doses without overdosing the surrounding normal tissues. Dose escalation to tumor is necessary to overcome tumor clone resistance and intracellular repair of radiation-induced damage.1 The limitations of conventional therapy have been overcome with the development of three-dimensional conformal radiotherapy2–5 and, especially, with intensity modulated radiation therapy (IMRT).6,7 IMRT conforms radiation to the shape of the target while minimizing exposure of surrounding critical structures.8–1 Such approach allows for safer tumor dose escalation and promises to widen the therapeutic window of the treatment by improving tumor control and decreasing treatment-related complications.

At Baylor College of Medicine and at the Houston

Veterans Affairs Medical Center (VAMC), prostate immobilization is achieved by an endorectal balloon12–15 inflated with 100 c of air. The balloon pushes the prostate toward the pubic symphysis and the posterior rectal wall away from the prostate and, therefore, from the area targeted by high radiation doses.16–18 Our group has previously reported that the maximum prostate displacement in the presence of an endorectal balloon during radiotherapy occurs in the superior-inferior direction with a standard deviation of 1.78 m.12,13 By using a special external immobilization device, our group has also reported that the standard deviation of setup errors is 3.5 m and from daily portal films, most of the setup errors observed were closer to 5 m.19 Dosimetric data resulting from setup variations are lacking.19

The present study was undertaken to assess the effect of setup uncertainty on the dosimetry of the prostate, seminal vesicles, and surrounding normal tissues. It focuses on modeling the range of dosimetric variations, based on a maximum setup uncertainty of 5 m. The objective of this study is to report the impact of these variations on target dose escalation and on normal tissue avoidance, as these may greatly influence tumor control and normal tissue complication probability.

Ten prostate cancer patients treated with IMRT at the Houston VAMC were included in the present study. The treatment planning was performed with the Peacock NOMOS system consisting of the multileaf intensitymodulating collimator (MIMiC™) and the Corvus™ (version 3.0 rev. 1) treatment planning system.12,20–2

Reprint requests to: Salahuddin Ahmad, Ph.D., Department of

Radiation Oncology, University of Oklahoma HSC, Everett Tower, 1200 North Everett Drive, Room B603, Oklahoma City, OK 73104

Medical Dosimetry, Vol. 30, No. 1, p. 1-7, 2005

Copyright © 2005 American Association of Medical Dosimetrists

Printed in the USA. All rights reserved 0958-3947/05/$–see front matter

External immobilization

The external immobilization with the Peacock/

IMRT system is achieved with a device called the “prostate box.” Details about this technique have been published elsewhere.12,13,19 In summary, it consists of a wooden box-like frame that supports the fiducial plates for target alignment and a beanbag that molds over the shape of the patient (Fig. 1). The beanbag is a commercially available vinyl bag (Soule) filled with 5 liters of Styrofoam beads. By sequentially introducing and evacuating air with a vacuum pump, the beads move and the bag molds and solidifies, providing a rigid impression of the patient’s body. Treatment planning is performed with the patient in the prone position and the fiducial plates are used to ensure that the central axis of the beam is tangential to the top of the pubic symphysis.

Organ immobilization

Prostate immobilization is achieved with the use of an inflated endorectal balloon. A nonlatex endorectal catheter (Flexi-Cuff™, EZEM) covered by a condom is inserted into the rectum before each treatment and the inflatable balloon is then filled with 100 c of air (Fig. 2).

CT scan technique

All 10 patients underwent computerized tomography (CT) planning in the prone position. CT scans were taken from just above the top of the bladder to the level of the anal verge with the endorectal balloon in the rectum. Anteroposterior and lateral scout films were obtained to ensure that the pubic symphysis was aligned with the central axis, defined by the wired center axis of the fiducial plates (Fig. 3).

During the course of treatment, the patient’s external treatment position was secured by 3 horizontal lines marked on the beanbag, and on the patient’s lower legs (Fig.1). For patient position verification throughout the treatment, daily lateral port films were obtained for the first week, followed by weekly lateral port films for all patients. Comparisons between portal and scout films demonstrated maximum patient setup deviations of 4.6 m in the superior and 3.9 m in the inferior axis (data not shown). For the purposes of this study, a setup error of 5 m was thus chosen to model dosimetric effects resulting from setup uncertainties.

Computerized treatment planning

The plan (T1) used for patient treatment was performed using the delineated anatomy from CT and dosevolume histograms (DVHs) and isodose lines were generated for this plan. Two hybrid plans (H1 and H2) were then generated by superimposing each patient’s isodose lines from treatment plan T1 on CT anatomy shifted by 5 m superiorly and inferiorly, respectively, to the setup

Fig. 1. The “prostate box” is the external immobilization device used for patient treatments with the Peacock/IMRT technique. It supports the moldable beanbag and affixes to the fiducial alignment system.

Fig. 2. Inflated endorectal balloon for prostate immobilization.

Fig. 3. Sagittal CT scout film depicting the endorectal balloon as well as the alignment of the center of beam axis with the top of the pubic symphysis.

Medical Dosimetry Volume 30, Number 1, 20052

point. These hybrid plans are used to recalculate the dosimetry if shifts are to occur due to setup error.

The prescribed radiation dose to target (prostate and seminal vesicles) was 70 Gy in 35 fractions, dose prescribed to the 83.4% isodose line (range 81–86.9%). According to the prescription, only 15% of rectum and colon volumes are allowed to exceed 68 Gy, and 3% of bladder volume may exceed 65 Gy. Clinically acceptable percent volumes of rectum and colon above 68 Gy may rise up to 20% to ensure optimal target dosing, in cases with difficult setup or complex target anatomy. The planning target volume (PTV) encompassed the CTV plus a margin of 5 m for the prostate and 3 m for the seminal vesicles. The DVHs of plans T1, H1, and H2 were compared for: (1) target, bladder, rectum, and colon mean doses; (2) target minimum doses; (3) percent target volume receiving below 70 Gy; (4) percent normal tissue volumes exceeding prescription limits; and (5) target equivalent uniform doses (EUDs).23

For plan T1, the average mean dose to prostate, seminal vesicles, rectum, colon, and bladder was 74.84, 74.94, 3.23, 31.59, and 19.58 Gy, respectively. Among all dosimetric plans (T1, H1, H2), the average mean doses ranged from 74.5 to 74.92 Gy for prostate, and from 73.65 to 74.94 Gy (Table 1) for seminal vesicles.

Average minimum target doses ranged significantly for prostate, from 50.85 to 63.51 Gy, and from 65.04 to 69.73 Gy for seminal vesicles. The average percent volumes receiving below 70 Gy for seminal vesicles increased from 0.53% to 6.26% while for the prostate, the observed increase was from 2.08% to 4.4% (Table 2).

For normal tissues, the average mean doses for rectum ranged from 32.6 to 34.20 Gy, for colon from 29.4 to 31.95 Gy, and for bladder from 14.84 to 24.69 Gy. In all plans, the average percent volume of rectum and colon receiving above the prescription limit of 68 Gy were below 15.2% and 1.9% , respectively. The average percent volume of bladder receiving above 65 Gy ranged from 3.1% to 13.32% among all plans but remained well below the limit of 3%, as defined by the prescription (Table 3).

All dosimetric plans were also compared on the basis of the equivalent uniform dose (EUD) method that calculates the biological dose effect from the entire DVH data. Among all plans, the mean EUD for the seminal vesicles ranged from 71.8 to 74.72 Gy, whereas for

Table 1. The mean doses for prostate and seminal vesicles for each of the 10 studied patients using all dosimetric plans

Prostate Seminal Vesicles

Patient T1 H1 H2 T1 H1 H2

Table 2. The percent prostate and seminal vesicles volume receiving a dose below 70 Gy for each of the 10 studied patients using all dosimetric plans

Prostate Seminal Vesicles

Patient T1 H1 H2 T1 H1 H2

Setup uncertainty in prostate cancer treated with IMRT ● S. AHMAD et al. 3

prostate, it ranged from 65.62 to 73.86 Gy (Table 4). These target EUD changes were observed with shifts in the superior direction only. A regression method was used to evaluate the possible correlation between target EUD values for all 30 plans (3 plans per patient, 10 patients) and their mean and minimum doses as well as the percent target volumes below goal. It was observed that EUD doses correlated highly only with the minimum doses for both the prostate and seminal vesicles, with coefficients of determination (squared value of correlation coefficient R) of 0.932 and 0.704, respectively. Notably, EUD doses below 72.5 Gy corresponded to minimum doses below 61 Gy for prostate and below 65 Gy for seminal vesicles (Fig.4).

This study assesses the dosimetric effect of both external and organ immobilization uncertainties on targets and normal tissues in prostate cancer patients treated with Peacock/IMRT using an endorectal balloon for prostate immobilization. It was designed to model and describe the range of dosimetric variations resulting from a maximum clinically acceptable setup error of 5 m. Measurement and understanding of these dosimetric variations may prove useful in optimizing inhomogeneous plans and in developing data-evidenced guidelines for the evaluation and selection of IMRT plans, especially in view of our efforts to deliver higher curative radiation doses while minimizing radiation reactions.

In this report, mean target doses remained escalated above 70 Gy, despite excessive dosimetric variations introduced by study design. Higher escalated average mean doses above 74 Gy were observed for the prostate, while average mean seminal vesicle doses also remained escalated above 73.6 Gy, with the exception of one plan (71.8 Gy).

However, setup uncertainties increased target dose inhomogeneity, as evidenced by the significant changes in the target minimum doses and EUDs. Specifically, mean EUD decreases of 1% for prostate and 4% for seminal vesicles were observed as a result of a 5-m shift in the superior direction only. A similar shift in the inferior direction did not decrease target EUD values. Correlations between target minimum doses and EUDs revealed that minimum doses of at least 61 Gy for the prostate and 6 Gy for the seminal vesicles are necessary to achieve EUDs higher than 72.5 Gy, which have been reported to be associated with better outcomes.24

Table 3. Percent volumes of normal tissues receiving doses greater than prescription limit* of all 10 studied patients Rectum Colon Bladder

Patient T1 H1 H2 T1 H1 H2 T1 H1 H2

*Prescription limit: 15% rectum and colon volume to exceed 68 Gy; 3% bladder volume to exceed 65 Gy.

Table 4. The equivalent uniform dose (EUD) for prostate and seminal vesicles for each of the 10 studied patients using all dosimetric plans

(Parte 1 de 2)