- Email: [email protected]
Prostate motion during standard radiotherapy as assessed by fiducial markers
R
ADIOTHERAPY
a8~c0~0G~ ELSEVIER
Radiotherapy and Oncology 37 (1995) 35-42
Prostate motion during standard radiotherapy as assessedby fiducial markers J.M. Crook*a, Y. Raymond, D. Salhanib, H. Yangb, B. Eschea aDepartment of Radiation Oncology, Ottawa Regional Cancer Centre. Ottawa, Canada bDepartment of Clinical Physics. Ottawa Regional Cancer Centre. Ottawa. Cana& Received 24 January 1995; revision received 13 July 1995;accepted 25 July 1995
Abstract
From November 1993to August 1994,55 patients with localized prostate carcinoma had three gold seedsplaced in the prostate under transrectal ultrasound guidance prior to the start of radiotherapy in order to track prostate motion. Patients had a planning CT scan before initial simulation and again at about 40 Gy, just prior to simulation of a field reduction. Seedposition relative to fixed bony landmarks (pubic symphysis and both ischial tuberosities) was digitized from each pair of orthogonal films from the initial and boost simulation using the Nucletron brachytherapy planning system. Vector analysis was performed to rule out the possibility of independent seedmigration within the prostate between the time of initial and boost simulation. Prostate motion was seenin the posterior (mean: 0.56 cm; SD: 0.41 cm) and inferior directions (mean:0.59 cm; SD: 0.45 cm). The baseof the prostate was displaced more than 1 cm posteriorly in 30% of patients and in 11% in the inferior direction. Prostate position is related to rectal and bladder tilling. Distension of these organs displaces the prostate in an anterosuperior direction, with lesser degreesof tilling allowing the prostate to move posteriorly and inferiorly. Conformal therapy planning must take this motion into consideration. Changes in prostate position of this magnitude preclude the use of standard margins. Keywordr: Prostate cancer; Radiotherapy; Prostate motion
1. Introduction
Advances in radiological imaging techniques have improved the ability to delineate tumor volumes, while the development of 3-dimensional treatment planning and conformal radiation therapy permits the shaping of treatment volumes with an accuracy never before achievable. The treatment of prostate cancer provides an ideal proving ground for this exciting technology. On the one hand, there is a clear dose-responserelationship indicating that higher doses may cure more patients [7,10,13] while on the other hand the prostate is in close
tumor dose [5,8,11]. Rigid patient immobilization
[6,16]
may be comforting to the oncologist, but ignores target motion within the patient. Variation in bladder and rectal tilling have been shown to affect prostate position within the pelvis [2,9,15,17], to an extent which may require field adjustments during the course of radiotherapy. This study was undertaken to assessthe usefulnessof radio-opaque prostate markers in tracking prostate motion through a course of radiotherapy, to quantitate the movement observed, and to determine optimal margins for conformal radiotherapy.
proximity to critical radiosensitive structures, the bladder and rectum. Sparing of the bladder and rectum with
conformal therapy has improved the toxicity profile [6,14,16] and has led to experimentation with increasing l Corresponding author, 501 Smyth Rd., Ottawa, Ontario, Canada, KIH 8L6.
2. Material and methods
SinceNovember 1993,55 patients with localized prostate cancer have had placement of three gold seeds(Rest Industries) in the prostate under transrectal ultrasound (TRUS) guidance, prior to initial simulation for pelvic
0167-8140/95BO9.50 0 1995Elsevier Science Ireland Ltd. All rights reserved SSDI 0167-8140(95)01613-L
36
J.M. Crook et al. /Radiotherapy
radiotherapy. The seedsare cylindrical, measuring 0.08 cm in diameter and 3 cm in length. Antibiotic coverage with ciprofloxacin (500 mg q8h x 3 doses)was provided. The seedswere positioned at the baseof the prostate near the seminal vesicles (seed l), the posterior aspect (seed2), and the apex of the prostate (seed 3). A planning pelvic CT scan was obtained for each patient, and an isodose distribution was produced for a four-field box technique. At the time of first simulation, a urethrogram was performed for the AP film and lo- 15 cc of barium was used to opacify the rectum for the lateral film. Patients were scanned, simulated and treated supine, with a full bladder, and with their lower legs supported in a Styrofoam immobilization device from knees to ankles which ensures a reproducible set up with respectto the angle of hip flexion and degreeof pelvic tilt. The dose prescribed was 46-50 Gy to the initial volume, followed by 16-22 Gy to a reduced volume. Fraction sizewas 2 Gy per day and the overall treatment time was 6.5-7 weeks. Both initial and boost volumes were treated using a four-field box technique. The stagedistribution for the 55 patients was Tlc:14, T2a: 11, T2b: 13, T2c:3, T3: 13, T4: 1. Pre-radiotherapy hormonal cytoreduction was used in 15 patients with T2b to T4 tumors. Three patients lost a seed prior to boost simulation and were excluded. Two of the three lost seedswere apical and one was lost from the base. Weekly port films were obtained to track prostate motion during treatment. A detailed analysis of prostate motion based on port films was not undertaken. Seed position relative to the field edge was checked in order to avoid geographic miss, but as the port films do not include the entire bony pelvis, seedposition relative to bony landmarks could not readily be determined. Films were taken using 6 or 18 MV photons at 0 and either 90 or 270”, using Kodak XRP fast film without filtration. Seedswere readily visible on all AP port films and usually visible on the lateral films. Based on early observations on prostate motion in the first 10 patients, a second planning CT scan was introduced between 36 and 40 Gy, to plan the reduced volume for the final 16-22 Gy. This second planning CT was performed with a full bladder and no gastro-intestinal contrast. Imageswere taken every 5 mm from the acetabulae to the bottom of the ischial tuberosities. The second simulation, for the reduced volume, was also performed with the patient having a full bladder. No gastro-intestinal contrast was used. Patients were set up in the immobilization device and flouroscoped on their initial marks in the presenceof a radiation oncologist (JC) to ensure accuracy of set-up before proceeding with simulation of the boost volume. The coordinates of the fiducial markers and selected landmarks were obtained from the reconstruction of the anteroposterior (AP) and lateral radiographs taken at the time of simulation for the initial and boost stagesof
and Oncology 37 (1995) 35-42
treatment. Selected landmarks included the anterior aspectof the pubis symphysisand the inferior aspectsof the right and left ischial tuberosities. Localization was performed using the NucletronTM Brachytherapy Planning System [3]. Digitization of the fiducial markers was done at the midpoint along the major axis of the seedsas seen on the simulator films. For this study, the seedlocalization error was determined to be typically
(Ai “%)
(1) where Ai” represents the vectors of the boost markers relative to the fixed coordinate system. The displacement of the boost frame relative to the initial frame is determined from vectors defining the selected bony landmarks. These landmarks form a rigid body. By ‘rigid’ body is meant a set of points whose mutual distances are invariable. Hence, the vectors (bt-b) and (RI+) are different representations of the same vector; their components are different only becausethey are measuredrelative to different coordinate systems.Thus the displacement of the rigid body is a measureof displacement of the boost coordinate system relative to the fixed coordinate system. The displacement of a rigid body can be reduced to a translation, followed by a rotation about some base point. If bi, say, is chosen as that base point then the vectors in the boost frame can be transformed, i.e., X X’, by the translation vector, @l-W followed by a rotation about bi:
(2)
X” = RX’
(3)
31
J.M. Crook et al. /&diotherapy and Oncology 37 (1995) 35-42
where X’ represents a vector defined in the translated boost coordinate system, and R the transformation matrix that operates on X ’ to produce X ” in the fixed frame. The operator R is the product of two rotations, one in the AP plane and one in the left-right plane. Their product is given by the following transformation matrix:
R=
cos(u)
cos(b)sin(a)
sin(b)
-sin(a)
cos(b)cos(a)
sin(b)cos(a)
-sin(b)
cos(b)
0
The anglesa and b represent the angles between the vectors (hi-b) and (Bi I-4’) projected onto xy (AP) and yz (right-left) planes in the fixed coordinate system, respectively. This study is concerned with prostatic movement with time which is inferred from the displacement of the markers. To determine this it is necessaryto separateactual prostate motion from motion reported due to seed migration and/or topological changes of the prostate itself. Movement due to seedmigration and/or prostate distortion can be determined by considering the relative vectors between the seedsin the initial and boost systems. In the absenceof migration and/or distortion the relative vectors will remain unchanged. Therefore, if the difference in magnitude of the corresponding relative vectors exceeded 0.3 cm, the patient was eliminated from further analysis on the assumption that migration and/distortion had occurred (n = 9). At present, there is insufficient data to determine topological changesof the gland. However, if one considers plausible forces that could produce distortion, it is unlikely to be a major factor in the absenceof a rectal mass or an unusual degree of pelvic muscle tension. Nonetheless, a second seriesof patients who have complete CT data with 5-mm cuts through the prostate on both the initial and boost planning CT scansis currently being analyzed. It is expected that topological changes can be determined from this type of analysis. For the remaining patients (n = 37), the vectors were transformed from the boost coordinate system to the fixed systemusing Eqns. (2) and (3). In this analysis the pubic symphysisand the left ischial tuberosity were used to define the transformation. The pubic symphysis was chosenas the basepoint. Rotation of the boost frame relative to the fixed frame about the base point is readily determined, as is the displacement of each marker (using Ew. (1)). The stage distribution of the 37 patients included for detailed analysis was Tlc:ll, T2a:7, T2b:lO, T2c:2, T3:6 and T4:l.
Table 1 Summary of patient rotation Component of rotation
Average (degrees)
SD (degrees)
AP Right-left
1.63 2.09
1.55 1.81
3. Results The tip of the urethrogram cone varied in position from 0 to 2.8 cm above the most inferior aspect of the ischial tuberosities (mean 1.5 cm) and was < 1 cm above in 24% of patients. The apparent thickness of the urogenital diaphragm was determined by measuring the distance from the tip of the urethrogram cone to the apical
0.6
I
!
-0.4 .-------
0.6
Fig. 1. Histograms showing change in seedco-ordinates in the x-axis (lateral) at boost simulation relative to initial reference frame. Each vertical bar representsdata for one patient. Maximum displacement: 0.5 cm, seed 1: base, seed 2: posterior, seed 3: apex.
38
J.M. Crook et al. /Radiotherapy
seedin the prostate and ranged from 0.3 to 2.8 cm (mean 1.5cm). At the time of initial simulation, the apex of the prostate, as indicated by the apical seed, was <2 cm above the ischial tuberosities in 42% of patients (n = 22), < 1.5cm in 19%and < 1 cm in 8%. Becauseof variability in the thickness of the urogenital diaphragm, only 56% (12/22) of these low-lying prostates would have been detected by urethrogram. Table 1 shows the average rotations of the boost frame relative to the fixed frame about the base point. This confirms that reproducibility of set up between the initial and boost simulations was excellent. There was very little prostate motion in the lateral (x) direction. For seeds1, 2 and 3 the average motion was 0.04,0.09 and 0.1 cm with a range of O-O.5cm (standard
and Oncology 37 (1995) 35-42
deviation: 0.2 cm). Motion occurred with equal frequency to the right and to the left (see Fig. 1). Movement in the craniocaudal Q) axis was usually in the inferior or caudal direction. For seeds1,2 and 3 the average motions were 0.54, 0.53 and 0.49 cm with a maximum of 1.6 cm (standard deviation: 0.5 cm) (see Fig. 2). Forty-three percent of patients showed more than 0.5 cm inferior displacement of the prostate and 11%more than 1 cm (seeTable 2). Ten patients showed prostate displacement in the superior or cranial direction (maximum 0.7 cm). In the anteroposterior (z) axis, movement was usually seenin the posterior or dorsal direction. For seeds1, 2 and 3 the averagedisplacementswere 0.72,0.62 and 0.46 cm with a maximum of 1.8 cm (standard deviation: 0.45
ISeed 1
-2
~-.--_L-----
Fig. 2. Histograms showing change in seed co-ordinates in y-axis (supero-inferior) at boost simulation relative to initial referenceframe. Each vertical bar representsdata for one patient. Negative direction indicates superior displacement. Maximum displacement: 1.8cm, seed I: base, seed 2: posterior, seed 3: apex.
Fig. 3. Histograms showing change in seed coordinates in z-axis (antero-posterior) at boost simulation relative to initial reference frame. Each vertical bar represents data for one patient. Negative direction indicates posterior displacement. Maximum displacement: 1.7 cm, seed I: base, seed 2: posterior, seed 3: apex.
J.M. Crook et al. /Radiotherapy and Oncology 37 (1995) 35-42
39
Table 2 Summary of prostate motion (cm) Z-axis
Y/-axis
X-axis Range
Mean
SD
Range
Mean
SD
Range
Mean
SD
Seed#I (base) Seed#2 (posterior) Seed63 (apex)
0.36 to -0.37 0.53 to -0.24 0.55 to -0.24
0.03
0.18 0.19
-0.34 -0.25
0.54 0.53 0.49
0.36 to -1.79 0.22 to -1.5 0.16 to -1.67
-0.72 -0.62
0.53 0.41
0.17
0.69 to -1.63 0.74 to -1.64 0.73 to -1.31
-0.31
0.09 0.1
-0.46
0.44
Centre of mass of A defined by three seeds
0.48 to -0.22
0.05
0.15
0.66 to -1.37
-0.59
0.5
0.19 to -1.69
-0.60
0.41
X-axis = lateral; Y-axis = supero-inferior; o= superior, 0 = inferior (caudal); Z-axis = antero-posterior; o= anterior, @ = posterior (dorsal)
,
I
cm) (seeFig. 3). Sixty per cent of patients showed more than 0.5 cm posterior displacement of the base prostate and 30% more than 1.0 cm (seeTable 2). Two prostates moved in the anterior or ventral direction (maximum 0.3 cm). The coordinates of the centre of mass of the triangle defined by the three seedswere determined for each patient (Fig. 4) and prostate motion was summarized by the length of the displacement vector joining the centre of massat the time of initial simulation to that at boost simulation. Prostate motion was then examined with respect to stage of disease.For Tic and T2a carcinomas, the averagemagnitude of the shift of the centre of mass was 0.68 cm, while for T2b, T2c and T3 tumors it was 0.55 cm (p = 0.34). Rectal diameter was measuredon the initial and boost CT scans at the level of the base of the prostate. The mean decreasein diameter was 1.5 cm (range: 0.5 cm larger to 5 cm smaller). No correlation could be detected between change in rectal diameter determined by CT and prostate displacement determined by fiducial markers. The effect of bladder filling was examined in six patients for whom double setsof simulator films were obtained, first with a full bladder and then with an empty bladder. The direction and magnitude of prostate displacement was similar to that seen between initial and boost simulations (see Fig. 5).
--
0.6
Average
- 0.12
1
~
I
4(a)
-
-___
4(b) / I
0.5
/
I
A”&T,e
= 0.562
j
j
4. Discussion There have been several reports on the effect of bladder and rectal filling on prostate position [2,9,15,17]. Accurate target volume definition is crucial in the delivery of precision radiotherapy. Since the prostate is not visible on portal imaging, margins must be adequate to encompassthis movement if geographic miss is to be avoided. Fig. 4. Displacement of centre of mass of triangle defined by three seeds.(a) Lateral direction (x-axis); (b) supero-inferior direction (JJaxis) (negative indicates superior displacement); (c) antero-posterior direction (z-axis) (negative indicates posterior displacement).
J.M. Crook ei al. /Radiotherapy
.-.--0
-1 Y :
,2 l-2
-3
5 -4
i
-5 cm
a: a
b
Fig. 5. (a) Displacement of the triangle formed by the three seeds,relative to fixed bony reference point, for one patient. Top right: initial simulation. Bottom left: boost simulation. (b) Displacement seen following emptying of bladder for one patient. Top right: full bladder. Bottom left: post void.
This study documents an alarming and consistent displacement of the prostate in the posterior and inferior direction relative to bony landmarks during pelvic radiotherapy. Eleven percent of patients showed an inferior shift of the prostate of more than 1 cm and 30% showed a posterior shift of more than 1 cm. Although the average displacement is only approximately 0.6 cm in both the posterior and inferior directions, and is encompassed within currently acceptable margins, the range of motion is large. Alteration of margins to include extremesof motion would mean treating a larger than necessaryvolume of rectum in the majority of patients. We did not see a significant difference in the mean prostate movement between early stage (Tic, T2a) and locally advanced(T2b, T2c, T3) tumors. The magnitude of the difference in mobility may have been reduced by the use of pre-radiotherapy hormonal cytoreduction of bulkier tumors. Downstaging and increase in prostate
and Oncology 37 (1995) 35-42
mobility prior to radiotherapy was not systematically recorded. The conditions under which the patients were scanned, simulated and treated were in no way unusual. Patients were instructed to have a full bladder “in order to displace normal structures out of the way of the beam“. Compliance with these instructions was not routinely checked,although periodic reminders were given. There was no intervention in the state of rectal filling. Rectal filling is certainly one of the important factors determining prostate position. We found a mean decrease in rectal diameter of 1.5 cm between the pretreatment planning CT scan and the scan repeatedat 40 Gy. There was not, however, a direct correlation between larger decreasesin rectal diameter and prostate displacement. A partial explanation may be that the measuresof prostate displacement and rectal diameter were taken on different days, the former using seedposition at boost simulation, the latter using the boost planning CT scan. This study was not designed to use CT data from the initial and boost planning CT scans to confirm prostate motion. The gold seedsare readily visible on the CT scans, but early in the study, many patients did not have them inserted until after their initial planning CT. In addition, since the first planning scan usually scanned the whole pelvis, cuts were often done at l-cm intervals and therefore yield minimal prostate information. These cuts are not readily comparable to cuts at 5-mm intervals obtained for the boost volume. Analysis of a second series of patients using serial CT data is currently underway. Bladder distension also has an important influence on prostate position. All patients in this study were asked to have a full bladder for all scans, simulations and treatment. It has been suggestedthat patients should be treated with an empty bladder to eliminate uncertainty in prostate position caused by variation in bladder filling. This is despite the obvious advantage of using a full bladder to reduce normal tissue toxicity by displacing the anterior bladder wall and small bowel from the high dose volume. Schild [ 151observed that bladder distension displaced the posterior border of the prostate up to 0.8 cm (median 0.2 cm) posteriorly. Our data show that post void simulation was associatedwith a posterior and inferior displacement of the prostate. Contraction of pelvic floor musculature to maintain voluntary continence pulls the prostate superiorly. With seedsin the prostate, this can easily be demonstrated under fluoroscopy. Relaxation of the pelvic floor after voiding may allow the prostate to sag posteriorly and inferiorly. Ten Haken [ 171has demonstrated that rectal distension with as little as 60 cc of contrast shifted the prostate in the anterior/superior direction a mean distance of 0.5 cm (range: O-2.0 cm). Measurementswere based on the shift in position of a Foley balloon which was pulled down snugly on the prostate. This is in agreementwith
J.M. Crook et al. /Radiotherapy
the present series.Since the rectum tends to becomeprogressively less distended during a course of pelvic radiotherapy, the predominant motion is in the posterior and inferior direction. Ten Haken’s data have been interpreted by some authors as indicating that larger margins may be required anteriorly and superiorly [11,12], but this is only true if the patient is somehow scanned and simulated with the prostate at its most posterior and inferior position. Balter et al. [l] recently published a detailed analysis of prostate motion in 10 patients using implanted markers and serial port films. They report a maximum expected prostate movement of 0.2-1.0 cm. However, by not using the initial simulator film as a baseline, they do not have the same ‘time-zero’ reference as was chosen for the present study. Analysis of only the simulation films provides better anatomic detail, and allows fluoroscopic verification of the set-up prior to boost simulation. Foman [4] used weekly CT scans on 10 patients undergoing conformal radiotherapy and found movement of the prostate and seminal vesiclesof up to 3.5 cm (60% anteriorly and 40% posteriorly). The degree of bladder filling was not stated. Perhaps the conformal techniquesdecreasedthe incidence and severity of radiation proctitis and cystitis, so that there was no consistent effect on rectal and bladder filling and the direction of prostate movement was random. The magnitude of prostate motion remained highly significant and is in agreement with the present series and with previously published data. Melian [9] reported on four serial CT scansperformed in each of 12 patients with and without instillation of 30 cc of rectal air. Shifts in target volumes of up to 3.0 cm were observed in the anterior-posterior direction and up to 1.5cm in the lateral direction. Such a degreeof lateral displacement was not observed in this study.
5. conclusions There is significant prostate motion during radiotherapy. Variation in the magnitude of prostate motion does not allow the use of standard margins. Boost volumes can not be determined adequately from a planning CT scan performed prior to initial treatment. Indeed, the magnitude of prostate motion detected in this study may have implications on the margins chosen for the initial volume. In order to tailor radiation fields optimally, we recommend that radio-opaque markers be placed in the prostate prior to the start of radiotherapy. If on-line portal imaging is available, field position should be verified relative to the prostate markers rather than surrounding skeletal anatomy. If a field reduction is planned, a second planning CT scan should be performed just before the boost simulation.
and Oncology 37 (1995) 35-42
41
Acknowledgements Y. Raymond was supported by an Ivan Smith Summer Studentship. The authors wish to thank Dr. V. Zaleski for ultrasonographic expertise, and Carolle Brazeau and Jacquelyn Cafferty for secretarial assistance. References PI Baher, J.M., Sandler, H.M., Lam, K., Bree, R.L., Lichter, AS. and Ten Haken, R. Measurement of prostate motion over the courseof routine radiotherapy using implanted markers. Int. J. Radiat. Oncol. Biol. Phys. 31: 113-118, 1995. PI Beard, C.J., Bussiere, M.R.., Plunkett, M.E., Coleman, C.N. and Kijewski, P.K. Analysis of prostate and seminal vesicle motion. Int. J. Radiat. Oncol. Biol. Phys. 27: 136, 1993. I31 Brachytherapy Treatment Planning, UPS Module, User Manual 090.403, 1993. I41 Forman, J.D., Mesina, C.F., He, T., Devi, S.B., Ben-Josef,E., Pehzzari, C., Vijayakumar and Chen, G.T. Evaluation of changesin the location and shape of the prostate and rectum during a sevenweek course of conformal radiotherapy. Int. J. Radiat. Oncol. Biol. Phys. 27: 222, 1993(abstr.). 151Forman, J.D., Orton, C., Ezzell, G. and Porter A.T. Prehminary results of a hyperfractionated dose escalation study for locally advanced adenocarcinoma of the prostate. Radiother. Oncol. 27: 203-208, 1993. 161Hanks, G.E. Conformal radiation in prostate cancer: Reduced morbidity with hope of increased local control. Int. J. Radiat. Oncol. Biol. Phys. 25: 377-378, 1993. I71 Hanks, G.E., Martz,K.L. and Diamond, J.J. Effect of dose on local control of prostate cancer. Int. J. Radiat. Oncol. Biol. Phys. 15: 1299-1306, 1988. 181Leibel, S.A., Heimann, R., Kutcher, G.J., Zelefsky, M.J., Burman, CM., Melian E., Orazem, J.P., Mohan, R., LoSasso, T.J., Lo,Y.-C., Wiseberg,J.A., Chapman, D.S., Ling,C.C. and Fuks, Z. Three-dimensional conformal radiation therapy in locally advancedcarcinoma of the prostate: Preliminary results of a phase I dose-escalationstudy. Int. J. Radiat. Oncol. Biol. Phys. 28: 55-65, 1994. r91 Melian, E., Kutcher, G., Leibel, S., Zelefsky, M., Baldwin, B. and Fuks, Z. Variation in prostate position: Quantitation and implications for three-dimensional conformal radiation therapy. Int. J. Radiat. Oncol. Biol. Phys. 27: 137, 1993(abstr.). 1101Perez, C.A., Lee, H.K., Georgiou, A., Logsdon, M.D., Lai, P.P. and Lockett, M.A. Technical and tumor-related factors affecting outcome of definitive irradiation for localized carcinoma of the prostate. Int. J. Radiat. Oncol. Biol. Phys. 26: 581-591, 1993. llil Phillips, T.J. Three-Dimensional conformal radiation therapy in locally advanced carcinoma of the prostate: Preliminary results of a phase I dose escalation study. Int. J. Radiat. Oncol. Biol. Phys. 28: 325-326, 1994. WI Roach, M. III, Pickett, B., Rosenthal, S.A., Verhey, L. and Phillips, T.L. Defining treatment margins for six field conforma1 irradiation of localized prostate cancer. Int. J. Radiat. Oncol. Biol. Phys. 28: 267-275, 1994. I131 Sandler, H.M., McShan, D.L. and Lichter, A.S. Potential improvement in the results of irradiation for prostate carcinoma using improved dose distribution. Int. J. Radiat. Oncol. Biol. Phys. 22: 361-367, 1992. I141 Sandler, H.M., Perez-Tamayo, C., Ten Haken, R.K. and Lichter, A.S. Dose escalation for Stage C (T3) prostate cancer:
42
J.M. Crook et al. /Radiotherapy
Minimal rectal toxicity observed using conformal therapy. Radiother. Oncol. 23: 53-54, 1992. 1151 Schild, SE., Casale, H.E. and Bellefontaine, L.P. Movements of the prostate due to rectal and bladder distension: Implications for radiotherapy. Med. Dosim. 18: 13-15, 1993. [ 161 Soffen, E.M., Hanks, G.E., Hunt, M.A. and Epstein, B.E. Conformal static field radiation therapy treatment of early prostate
and Oncology 37 (1995) 35-42
cancer versus non-confonnal techniques: A reduction in acute morbidity. Int. J. Radiat. Oncol. Biol. Phys. 24: 485-488, 1992. 1171 Ten Haken, R.K., Forman, J.D., Heimburger, D.K., Gerhardsson, A., McShan, D.L., Perez-Tamayo, C., Schoeppel, S.L. and Lichter AS. Treatment planning issuesrelated to prostate movement in responseto differential filling of the rectum and bladder. Int. J. Radiat. Oncol. Biol. Phys. 20: 1317-1324, 1991.
ADIOTHERAPY
a8~c0~0G~ ELSEVIER
Radiotherapy and Oncology 37 (1995) 35-42
Prostate motion during standard radiotherapy as assessedby fiducial markers J.M. Crook*a, Y. Raymond, D. Salhanib, H. Yangb, B. Eschea aDepartment of Radiation Oncology, Ottawa Regional Cancer Centre. Ottawa, Canada bDepartment of Clinical Physics. Ottawa Regional Cancer Centre. Ottawa. Cana& Received 24 January 1995; revision received 13 July 1995;accepted 25 July 1995
Abstract
From November 1993to August 1994,55 patients with localized prostate carcinoma had three gold seedsplaced in the prostate under transrectal ultrasound guidance prior to the start of radiotherapy in order to track prostate motion. Patients had a planning CT scan before initial simulation and again at about 40 Gy, just prior to simulation of a field reduction. Seedposition relative to fixed bony landmarks (pubic symphysis and both ischial tuberosities) was digitized from each pair of orthogonal films from the initial and boost simulation using the Nucletron brachytherapy planning system. Vector analysis was performed to rule out the possibility of independent seedmigration within the prostate between the time of initial and boost simulation. Prostate motion was seenin the posterior (mean: 0.56 cm; SD: 0.41 cm) and inferior directions (mean:0.59 cm; SD: 0.45 cm). The baseof the prostate was displaced more than 1 cm posteriorly in 30% of patients and in 11% in the inferior direction. Prostate position is related to rectal and bladder tilling. Distension of these organs displaces the prostate in an anterosuperior direction, with lesser degreesof tilling allowing the prostate to move posteriorly and inferiorly. Conformal therapy planning must take this motion into consideration. Changes in prostate position of this magnitude preclude the use of standard margins. Keywordr: Prostate cancer; Radiotherapy; Prostate motion
1. Introduction
Advances in radiological imaging techniques have improved the ability to delineate tumor volumes, while the development of 3-dimensional treatment planning and conformal radiation therapy permits the shaping of treatment volumes with an accuracy never before achievable. The treatment of prostate cancer provides an ideal proving ground for this exciting technology. On the one hand, there is a clear dose-responserelationship indicating that higher doses may cure more patients [7,10,13] while on the other hand the prostate is in close
tumor dose [5,8,11]. Rigid patient immobilization
[6,16]
may be comforting to the oncologist, but ignores target motion within the patient. Variation in bladder and rectal tilling have been shown to affect prostate position within the pelvis [2,9,15,17], to an extent which may require field adjustments during the course of radiotherapy. This study was undertaken to assessthe usefulnessof radio-opaque prostate markers in tracking prostate motion through a course of radiotherapy, to quantitate the movement observed, and to determine optimal margins for conformal radiotherapy.
proximity to critical radiosensitive structures, the bladder and rectum. Sparing of the bladder and rectum with
conformal therapy has improved the toxicity profile [6,14,16] and has led to experimentation with increasing l Corresponding author, 501 Smyth Rd., Ottawa, Ontario, Canada, KIH 8L6.
2. Material and methods
SinceNovember 1993,55 patients with localized prostate cancer have had placement of three gold seeds(Rest Industries) in the prostate under transrectal ultrasound (TRUS) guidance, prior to initial simulation for pelvic
0167-8140/95BO9.50 0 1995Elsevier Science Ireland Ltd. All rights reserved SSDI 0167-8140(95)01613-L
36
J.M. Crook et al. /Radiotherapy
radiotherapy. The seedsare cylindrical, measuring 0.08 cm in diameter and 3 cm in length. Antibiotic coverage with ciprofloxacin (500 mg q8h x 3 doses)was provided. The seedswere positioned at the baseof the prostate near the seminal vesicles (seed l), the posterior aspect (seed2), and the apex of the prostate (seed 3). A planning pelvic CT scan was obtained for each patient, and an isodose distribution was produced for a four-field box technique. At the time of first simulation, a urethrogram was performed for the AP film and lo- 15 cc of barium was used to opacify the rectum for the lateral film. Patients were scanned, simulated and treated supine, with a full bladder, and with their lower legs supported in a Styrofoam immobilization device from knees to ankles which ensures a reproducible set up with respectto the angle of hip flexion and degreeof pelvic tilt. The dose prescribed was 46-50 Gy to the initial volume, followed by 16-22 Gy to a reduced volume. Fraction sizewas 2 Gy per day and the overall treatment time was 6.5-7 weeks. Both initial and boost volumes were treated using a four-field box technique. The stagedistribution for the 55 patients was Tlc:14, T2a: 11, T2b: 13, T2c:3, T3: 13, T4: 1. Pre-radiotherapy hormonal cytoreduction was used in 15 patients with T2b to T4 tumors. Three patients lost a seed prior to boost simulation and were excluded. Two of the three lost seedswere apical and one was lost from the base. Weekly port films were obtained to track prostate motion during treatment. A detailed analysis of prostate motion based on port films was not undertaken. Seed position relative to the field edge was checked in order to avoid geographic miss, but as the port films do not include the entire bony pelvis, seedposition relative to bony landmarks could not readily be determined. Films were taken using 6 or 18 MV photons at 0 and either 90 or 270”, using Kodak XRP fast film without filtration. Seedswere readily visible on all AP port films and usually visible on the lateral films. Based on early observations on prostate motion in the first 10 patients, a second planning CT scan was introduced between 36 and 40 Gy, to plan the reduced volume for the final 16-22 Gy. This second planning CT was performed with a full bladder and no gastro-intestinal contrast. Imageswere taken every 5 mm from the acetabulae to the bottom of the ischial tuberosities. The second simulation, for the reduced volume, was also performed with the patient having a full bladder. No gastro-intestinal contrast was used. Patients were set up in the immobilization device and flouroscoped on their initial marks in the presenceof a radiation oncologist (JC) to ensure accuracy of set-up before proceeding with simulation of the boost volume. The coordinates of the fiducial markers and selected landmarks were obtained from the reconstruction of the anteroposterior (AP) and lateral radiographs taken at the time of simulation for the initial and boost stagesof
and Oncology 37 (1995) 35-42
treatment. Selected landmarks included the anterior aspectof the pubis symphysisand the inferior aspectsof the right and left ischial tuberosities. Localization was performed using the NucletronTM Brachytherapy Planning System [3]. Digitization of the fiducial markers was done at the midpoint along the major axis of the seedsas seen on the simulator films. For this study, the seedlocalization error was determined to be typically
(Ai “%)
(1) where Ai” represents the vectors of the boost markers relative to the fixed coordinate system. The displacement of the boost frame relative to the initial frame is determined from vectors defining the selected bony landmarks. These landmarks form a rigid body. By ‘rigid’ body is meant a set of points whose mutual distances are invariable. Hence, the vectors (bt-b) and (RI+) are different representations of the same vector; their components are different only becausethey are measuredrelative to different coordinate systems.Thus the displacement of the rigid body is a measureof displacement of the boost coordinate system relative to the fixed coordinate system. The displacement of a rigid body can be reduced to a translation, followed by a rotation about some base point. If bi, say, is chosen as that base point then the vectors in the boost frame can be transformed, i.e., X X’, by the translation vector, @l-W followed by a rotation about bi:
(2)
X” = RX’
(3)
31
J.M. Crook et al. /&diotherapy and Oncology 37 (1995) 35-42
where X’ represents a vector defined in the translated boost coordinate system, and R the transformation matrix that operates on X ’ to produce X ” in the fixed frame. The operator R is the product of two rotations, one in the AP plane and one in the left-right plane. Their product is given by the following transformation matrix:
R=
cos(u)
cos(b)sin(a)
sin(b)
-sin(a)
cos(b)cos(a)
sin(b)cos(a)
-sin(b)
cos(b)
0
The anglesa and b represent the angles between the vectors (hi-b) and (Bi I-4’) projected onto xy (AP) and yz (right-left) planes in the fixed coordinate system, respectively. This study is concerned with prostatic movement with time which is inferred from the displacement of the markers. To determine this it is necessaryto separateactual prostate motion from motion reported due to seed migration and/or topological changes of the prostate itself. Movement due to seedmigration and/or prostate distortion can be determined by considering the relative vectors between the seedsin the initial and boost systems. In the absenceof migration and/or distortion the relative vectors will remain unchanged. Therefore, if the difference in magnitude of the corresponding relative vectors exceeded 0.3 cm, the patient was eliminated from further analysis on the assumption that migration and/distortion had occurred (n = 9). At present, there is insufficient data to determine topological changesof the gland. However, if one considers plausible forces that could produce distortion, it is unlikely to be a major factor in the absenceof a rectal mass or an unusual degree of pelvic muscle tension. Nonetheless, a second seriesof patients who have complete CT data with 5-mm cuts through the prostate on both the initial and boost planning CT scansis currently being analyzed. It is expected that topological changes can be determined from this type of analysis. For the remaining patients (n = 37), the vectors were transformed from the boost coordinate system to the fixed systemusing Eqns. (2) and (3). In this analysis the pubic symphysisand the left ischial tuberosity were used to define the transformation. The pubic symphysis was chosenas the basepoint. Rotation of the boost frame relative to the fixed frame about the base point is readily determined, as is the displacement of each marker (using Ew. (1)). The stage distribution of the 37 patients included for detailed analysis was Tlc:ll, T2a:7, T2b:lO, T2c:2, T3:6 and T4:l.
Table 1 Summary of patient rotation Component of rotation
Average (degrees)
SD (degrees)
AP Right-left
1.63 2.09
1.55 1.81
3. Results The tip of the urethrogram cone varied in position from 0 to 2.8 cm above the most inferior aspect of the ischial tuberosities (mean 1.5 cm) and was < 1 cm above in 24% of patients. The apparent thickness of the urogenital diaphragm was determined by measuring the distance from the tip of the urethrogram cone to the apical
0.6
I
!
-0.4 .-------
0.6
Fig. 1. Histograms showing change in seedco-ordinates in the x-axis (lateral) at boost simulation relative to initial reference frame. Each vertical bar representsdata for one patient. Maximum displacement: 0.5 cm, seed 1: base, seed 2: posterior, seed 3: apex.
38
J.M. Crook et al. /Radiotherapy
seedin the prostate and ranged from 0.3 to 2.8 cm (mean 1.5cm). At the time of initial simulation, the apex of the prostate, as indicated by the apical seed, was <2 cm above the ischial tuberosities in 42% of patients (n = 22), < 1.5cm in 19%and < 1 cm in 8%. Becauseof variability in the thickness of the urogenital diaphragm, only 56% (12/22) of these low-lying prostates would have been detected by urethrogram. Table 1 shows the average rotations of the boost frame relative to the fixed frame about the base point. This confirms that reproducibility of set up between the initial and boost simulations was excellent. There was very little prostate motion in the lateral (x) direction. For seeds1, 2 and 3 the average motion was 0.04,0.09 and 0.1 cm with a range of O-O.5cm (standard
and Oncology 37 (1995) 35-42
deviation: 0.2 cm). Motion occurred with equal frequency to the right and to the left (see Fig. 1). Movement in the craniocaudal Q) axis was usually in the inferior or caudal direction. For seeds1,2 and 3 the average motions were 0.54, 0.53 and 0.49 cm with a maximum of 1.6 cm (standard deviation: 0.5 cm) (see Fig. 2). Forty-three percent of patients showed more than 0.5 cm inferior displacement of the prostate and 11%more than 1 cm (seeTable 2). Ten patients showed prostate displacement in the superior or cranial direction (maximum 0.7 cm). In the anteroposterior (z) axis, movement was usually seenin the posterior or dorsal direction. For seeds1, 2 and 3 the averagedisplacementswere 0.72,0.62 and 0.46 cm with a maximum of 1.8 cm (standard deviation: 0.45
ISeed 1
-2
~-.--_L-----
Fig. 2. Histograms showing change in seed co-ordinates in y-axis (supero-inferior) at boost simulation relative to initial referenceframe. Each vertical bar representsdata for one patient. Negative direction indicates superior displacement. Maximum displacement: 1.8cm, seed I: base, seed 2: posterior, seed 3: apex.
Fig. 3. Histograms showing change in seed coordinates in z-axis (antero-posterior) at boost simulation relative to initial reference frame. Each vertical bar represents data for one patient. Negative direction indicates posterior displacement. Maximum displacement: 1.7 cm, seed I: base, seed 2: posterior, seed 3: apex.
J.M. Crook et al. /Radiotherapy and Oncology 37 (1995) 35-42
39
Table 2 Summary of prostate motion (cm) Z-axis
Y/-axis
X-axis Range
Mean
SD
Range
Mean
SD
Range
Mean
SD
Seed#I (base) Seed#2 (posterior) Seed63 (apex)
0.36 to -0.37 0.53 to -0.24 0.55 to -0.24
0.03
0.18 0.19
-0.34 -0.25
0.54 0.53 0.49
0.36 to -1.79 0.22 to -1.5 0.16 to -1.67
-0.72 -0.62
0.53 0.41
0.17
0.69 to -1.63 0.74 to -1.64 0.73 to -1.31
-0.31
0.09 0.1
-0.46
0.44
Centre of mass of A defined by three seeds
0.48 to -0.22
0.05
0.15
0.66 to -1.37
-0.59
0.5
0.19 to -1.69
-0.60
0.41
X-axis = lateral; Y-axis = supero-inferior; o= superior, 0 = inferior (caudal); Z-axis = antero-posterior; o= anterior, @ = posterior (dorsal)
,
I
cm) (seeFig. 3). Sixty per cent of patients showed more than 0.5 cm posterior displacement of the base prostate and 30% more than 1.0 cm (seeTable 2). Two prostates moved in the anterior or ventral direction (maximum 0.3 cm). The coordinates of the centre of mass of the triangle defined by the three seedswere determined for each patient (Fig. 4) and prostate motion was summarized by the length of the displacement vector joining the centre of massat the time of initial simulation to that at boost simulation. Prostate motion was then examined with respect to stage of disease.For Tic and T2a carcinomas, the averagemagnitude of the shift of the centre of mass was 0.68 cm, while for T2b, T2c and T3 tumors it was 0.55 cm (p = 0.34). Rectal diameter was measuredon the initial and boost CT scans at the level of the base of the prostate. The mean decreasein diameter was 1.5 cm (range: 0.5 cm larger to 5 cm smaller). No correlation could be detected between change in rectal diameter determined by CT and prostate displacement determined by fiducial markers. The effect of bladder filling was examined in six patients for whom double setsof simulator films were obtained, first with a full bladder and then with an empty bladder. The direction and magnitude of prostate displacement was similar to that seen between initial and boost simulations (see Fig. 5).
--
0.6
Average
- 0.12
1
~
I
4(a)
-
-___
4(b) / I
0.5
/
I
A”&T,e
= 0.562
j
j
4. Discussion There have been several reports on the effect of bladder and rectal filling on prostate position [2,9,15,17]. Accurate target volume definition is crucial in the delivery of precision radiotherapy. Since the prostate is not visible on portal imaging, margins must be adequate to encompassthis movement if geographic miss is to be avoided. Fig. 4. Displacement of centre of mass of triangle defined by three seeds.(a) Lateral direction (x-axis); (b) supero-inferior direction (JJaxis) (negative indicates superior displacement); (c) antero-posterior direction (z-axis) (negative indicates posterior displacement).
J.M. Crook ei al. /Radiotherapy
.-.--0
-1 Y :
,2 l-2
-3
5 -4
i
-5 cm
a: a
b
Fig. 5. (a) Displacement of the triangle formed by the three seeds,relative to fixed bony reference point, for one patient. Top right: initial simulation. Bottom left: boost simulation. (b) Displacement seen following emptying of bladder for one patient. Top right: full bladder. Bottom left: post void.
This study documents an alarming and consistent displacement of the prostate in the posterior and inferior direction relative to bony landmarks during pelvic radiotherapy. Eleven percent of patients showed an inferior shift of the prostate of more than 1 cm and 30% showed a posterior shift of more than 1 cm. Although the average displacement is only approximately 0.6 cm in both the posterior and inferior directions, and is encompassed within currently acceptable margins, the range of motion is large. Alteration of margins to include extremesof motion would mean treating a larger than necessaryvolume of rectum in the majority of patients. We did not see a significant difference in the mean prostate movement between early stage (Tic, T2a) and locally advanced(T2b, T2c, T3) tumors. The magnitude of the difference in mobility may have been reduced by the use of pre-radiotherapy hormonal cytoreduction of bulkier tumors. Downstaging and increase in prostate
and Oncology 37 (1995) 35-42
mobility prior to radiotherapy was not systematically recorded. The conditions under which the patients were scanned, simulated and treated were in no way unusual. Patients were instructed to have a full bladder “in order to displace normal structures out of the way of the beam“. Compliance with these instructions was not routinely checked,although periodic reminders were given. There was no intervention in the state of rectal filling. Rectal filling is certainly one of the important factors determining prostate position. We found a mean decrease in rectal diameter of 1.5 cm between the pretreatment planning CT scan and the scan repeatedat 40 Gy. There was not, however, a direct correlation between larger decreasesin rectal diameter and prostate displacement. A partial explanation may be that the measuresof prostate displacement and rectal diameter were taken on different days, the former using seedposition at boost simulation, the latter using the boost planning CT scan. This study was not designed to use CT data from the initial and boost planning CT scans to confirm prostate motion. The gold seedsare readily visible on the CT scans, but early in the study, many patients did not have them inserted until after their initial planning CT. In addition, since the first planning scan usually scanned the whole pelvis, cuts were often done at l-cm intervals and therefore yield minimal prostate information. These cuts are not readily comparable to cuts at 5-mm intervals obtained for the boost volume. Analysis of a second series of patients using serial CT data is currently underway. Bladder distension also has an important influence on prostate position. All patients in this study were asked to have a full bladder for all scans, simulations and treatment. It has been suggestedthat patients should be treated with an empty bladder to eliminate uncertainty in prostate position caused by variation in bladder filling. This is despite the obvious advantage of using a full bladder to reduce normal tissue toxicity by displacing the anterior bladder wall and small bowel from the high dose volume. Schild [ 151observed that bladder distension displaced the posterior border of the prostate up to 0.8 cm (median 0.2 cm) posteriorly. Our data show that post void simulation was associatedwith a posterior and inferior displacement of the prostate. Contraction of pelvic floor musculature to maintain voluntary continence pulls the prostate superiorly. With seedsin the prostate, this can easily be demonstrated under fluoroscopy. Relaxation of the pelvic floor after voiding may allow the prostate to sag posteriorly and inferiorly. Ten Haken [ 171has demonstrated that rectal distension with as little as 60 cc of contrast shifted the prostate in the anterior/superior direction a mean distance of 0.5 cm (range: O-2.0 cm). Measurementswere based on the shift in position of a Foley balloon which was pulled down snugly on the prostate. This is in agreementwith
J.M. Crook et al. /Radiotherapy
the present series.Since the rectum tends to becomeprogressively less distended during a course of pelvic radiotherapy, the predominant motion is in the posterior and inferior direction. Ten Haken’s data have been interpreted by some authors as indicating that larger margins may be required anteriorly and superiorly [11,12], but this is only true if the patient is somehow scanned and simulated with the prostate at its most posterior and inferior position. Balter et al. [l] recently published a detailed analysis of prostate motion in 10 patients using implanted markers and serial port films. They report a maximum expected prostate movement of 0.2-1.0 cm. However, by not using the initial simulator film as a baseline, they do not have the same ‘time-zero’ reference as was chosen for the present study. Analysis of only the simulation films provides better anatomic detail, and allows fluoroscopic verification of the set-up prior to boost simulation. Foman [4] used weekly CT scans on 10 patients undergoing conformal radiotherapy and found movement of the prostate and seminal vesiclesof up to 3.5 cm (60% anteriorly and 40% posteriorly). The degree of bladder filling was not stated. Perhaps the conformal techniquesdecreasedthe incidence and severity of radiation proctitis and cystitis, so that there was no consistent effect on rectal and bladder filling and the direction of prostate movement was random. The magnitude of prostate motion remained highly significant and is in agreement with the present series and with previously published data. Melian [9] reported on four serial CT scansperformed in each of 12 patients with and without instillation of 30 cc of rectal air. Shifts in target volumes of up to 3.0 cm were observed in the anterior-posterior direction and up to 1.5cm in the lateral direction. Such a degreeof lateral displacement was not observed in this study.
5. conclusions There is significant prostate motion during radiotherapy. Variation in the magnitude of prostate motion does not allow the use of standard margins. Boost volumes can not be determined adequately from a planning CT scan performed prior to initial treatment. Indeed, the magnitude of prostate motion detected in this study may have implications on the margins chosen for the initial volume. In order to tailor radiation fields optimally, we recommend that radio-opaque markers be placed in the prostate prior to the start of radiotherapy. If on-line portal imaging is available, field position should be verified relative to the prostate markers rather than surrounding skeletal anatomy. If a field reduction is planned, a second planning CT scan should be performed just before the boost simulation.
and Oncology 37 (1995) 35-42
41
Acknowledgements Y. Raymond was supported by an Ivan Smith Summer Studentship. The authors wish to thank Dr. V. Zaleski for ultrasonographic expertise, and Carolle Brazeau and Jacquelyn Cafferty for secretarial assistance. References PI Baher, J.M., Sandler, H.M., Lam, K., Bree, R.L., Lichter, AS. and Ten Haken, R. Measurement of prostate motion over the courseof routine radiotherapy using implanted markers. Int. J. Radiat. Oncol. Biol. Phys. 31: 113-118, 1995. PI Beard, C.J., Bussiere, M.R.., Plunkett, M.E., Coleman, C.N. and Kijewski, P.K. Analysis of prostate and seminal vesicle motion. Int. J. Radiat. Oncol. Biol. Phys. 27: 136, 1993. I31 Brachytherapy Treatment Planning, UPS Module, User Manual 090.403, 1993. I41 Forman, J.D., Mesina, C.F., He, T., Devi, S.B., Ben-Josef,E., Pehzzari, C., Vijayakumar and Chen, G.T. Evaluation of changesin the location and shape of the prostate and rectum during a sevenweek course of conformal radiotherapy. Int. J. Radiat. Oncol. Biol. Phys. 27: 222, 1993(abstr.). 151Forman, J.D., Orton, C., Ezzell, G. and Porter A.T. Prehminary results of a hyperfractionated dose escalation study for locally advanced adenocarcinoma of the prostate. Radiother. Oncol. 27: 203-208, 1993. 161Hanks, G.E. Conformal radiation in prostate cancer: Reduced morbidity with hope of increased local control. Int. J. Radiat. Oncol. Biol. Phys. 25: 377-378, 1993. I71 Hanks, G.E., Martz,K.L. and Diamond, J.J. Effect of dose on local control of prostate cancer. Int. J. Radiat. Oncol. Biol. Phys. 15: 1299-1306, 1988. 181Leibel, S.A., Heimann, R., Kutcher, G.J., Zelefsky, M.J., Burman, CM., Melian E., Orazem, J.P., Mohan, R., LoSasso, T.J., Lo,Y.-C., Wiseberg,J.A., Chapman, D.S., Ling,C.C. and Fuks, Z. Three-dimensional conformal radiation therapy in locally advancedcarcinoma of the prostate: Preliminary results of a phase I dose-escalationstudy. Int. J. Radiat. Oncol. Biol. Phys. 28: 55-65, 1994. r91 Melian, E., Kutcher, G., Leibel, S., Zelefsky, M., Baldwin, B. and Fuks, Z. Variation in prostate position: Quantitation and implications for three-dimensional conformal radiation therapy. Int. J. Radiat. Oncol. Biol. Phys. 27: 137, 1993(abstr.). 1101Perez, C.A., Lee, H.K., Georgiou, A., Logsdon, M.D., Lai, P.P. and Lockett, M.A. Technical and tumor-related factors affecting outcome of definitive irradiation for localized carcinoma of the prostate. Int. J. Radiat. Oncol. Biol. Phys. 26: 581-591, 1993. llil Phillips, T.J. Three-Dimensional conformal radiation therapy in locally advanced carcinoma of the prostate: Preliminary results of a phase I dose escalation study. Int. J. Radiat. Oncol. Biol. Phys. 28: 325-326, 1994. WI Roach, M. III, Pickett, B., Rosenthal, S.A., Verhey, L. and Phillips, T.L. Defining treatment margins for six field conforma1 irradiation of localized prostate cancer. Int. J. Radiat. Oncol. Biol. Phys. 28: 267-275, 1994. I131 Sandler, H.M., McShan, D.L. and Lichter, A.S. Potential improvement in the results of irradiation for prostate carcinoma using improved dose distribution. Int. J. Radiat. Oncol. Biol. Phys. 22: 361-367, 1992. I141 Sandler, H.M., Perez-Tamayo, C., Ten Haken, R.K. and Lichter, A.S. Dose escalation for Stage C (T3) prostate cancer:
42
J.M. Crook et al. /Radiotherapy
Minimal rectal toxicity observed using conformal therapy. Radiother. Oncol. 23: 53-54, 1992. 1151 Schild, SE., Casale, H.E. and Bellefontaine, L.P. Movements of the prostate due to rectal and bladder distension: Implications for radiotherapy. Med. Dosim. 18: 13-15, 1993. [ 161 Soffen, E.M., Hanks, G.E., Hunt, M.A. and Epstein, B.E. Conformal static field radiation therapy treatment of early prostate
and Oncology 37 (1995) 35-42
cancer versus non-confonnal techniques: A reduction in acute morbidity. Int. J. Radiat. Oncol. Biol. Phys. 24: 485-488, 1992. 1171 Ten Haken, R.K., Forman, J.D., Heimburger, D.K., Gerhardsson, A., McShan, D.L., Perez-Tamayo, C., Schoeppel, S.L. and Lichter AS. Treatment planning issuesrelated to prostate movement in responseto differential filling of the rectum and bladder. Int. J. Radiat. Oncol. Biol. Phys. 20: 1317-1324, 1991.