- Email: [email protected]
Optimization of the Oblique Angles in the Treatment of Prostate Cancer During Six-Field Conformal Radiotherapy
~
Medical Dosimetry, Vol. 19, No.4, pp. 237-254, 1994 Copyright © 1994 American Association of Medical Dosimetrists Printed in the USA. All rights reserved 0958-3947/94 $6.00 + .00
Pergamon
0958-3947(94)00033-6
OPTIMIZATION OF THE OBLIQUE ANGLES IN THE TREATMENT OF PROSTATE CANCER DURING SIX-FIELD CONFORMAL RADIOTHERAPY BARBY PICKETT, MS, MACK ROACH,
III,
MD, PAUL HORINE, BS,
LYNN VERHEY, PH.D., and THEODORE L. PHILLIPS, MD University of California, San Francisco, CA 94143
Abstract-Historically, four perpendicular treatment fields or bilateral arcs have been used in the treatment of prostate cancer. As new techniques challenge the four-field box technique for their superiority of tumor coverage and adjacent critical structure sparing, oblique beam angles (in addition to right and left laterals) have been introduced as an alternative to anterior (AP) and posterior (PA) beams. Among the most popular of these alternative approaches is a six-field technique. Traditionally 45° angles have been used with this technique. In this study, opposed coplanar oblique beams angled 20, 25, 30, 35, 40, and 45° otT the lateral beam position, were compared for their ability to minimize adjacent critical structure doses, while maintaining maximum clinical target volume (CTV) coverage. This analysis compared rectum, bladder, and femoral head dose volume histograms (DVH) for each of these varying oblique gantry angles. As the angle of the posterior oblique beams became more horizontal, it is more difficult to encompass the apex of the prostate in the 95 % isodose value. On inferior CT slices near the apex of the prostate, the density of the pelvic bones in the path of the posterior oblique fields causes the beam to be slightly attenuated, thereby underdosing the CTV. The oblique angles most atTected by this bone heterogeneity are beams angled from 20 to 30° otT the lateral beam position. As the angles approach the vertical direction, rectal and bladder doses increase, while femoral head doses decrease. Oblique gantry angles approaching the horizontal direction result in a decrease in rectal and bladder doses, while femoral head doses increase. Of the oblique angles studied, 35° otT the lateral position provides lower rectum and bladder doses than 30, 40, and 45°; lower femoral head doses than 20, 25, and 30°, and the maximum CTV coverage on all CT slices studied. Key Words: Prostate cancer, Normalization, Three-dimensional treatment planning, Oblique angles.
treatment planning system was used to calculate doses to adjacent critical structures with varying opposed gantry angles, of 20, 25, 30, 35, 40, and 45° off the lateral beam direction. 8,9 Conformal treatment plans consisting of lateral (RTL and LTL), and oblique beams (RAO, LAO, LPO, and RPO) with similar techniques were analyzed, with and without heterogeneity corrections, to evaluate the impact of bone density on resulting dose distributions. Each patient was planned with 6-MV photons. A previous study showed little or no difference in dose distribution when photon energies from 6 MV to 18 MV and six or more conformal fields Were used. Blocks were designed to allow for daily treatment inaccuracies, internal organ movement, uncertainties associated with extracapsular extension, and homogeneous dose distributions. The block design has been published previously and is referred to as "nonuniform.,,7,10 The isodose distributions were evaluated on six planes of calculation including the most inferior axial, the central axis axial, and the most superior axial cuts, along with reconstructed midline sagittal, coronal, and oblique views to assess the homogeneity of isodose distributions and CTV coverage in all planes of calculation.
PURPOSE/OBJECTIVE This study was undertaken to determine and evaluate the optimal oblique gantry angles to be used for coplanar conformal treatment of the prostate by comparing the doses to adjacent critical structures and the isodose distributions relative to clinical target volume (CTV) coverage.
MATERIALS AND METHODS For this study, 10 patients were simulated in the supine position, and a urethrogram was obtained to assist in defining the apex of the prostate. 1. 4 Each patient had a treatment planning CT scan with 5-mm slice thicknesses. To minimize the risk of patient and organ movement, patients were simulated and treated with custom immobilization and scanned with their rectums empty and their bladders full. 5 . 7 After identifying the prostate, seminal vesicles, bladder, rectum, and femoral heads on axial CT slices, a U -M PLAN (University, of Michigan) three-dimensional (3-D)
Reprint requests to: Barby Pickett, MS, University of California, San Francisco, CA 94143.
237
238
Medical Dosimetry
3-D treatment plans were designed for each of the six gantry angles evaluated, with and without heterogeneity correction. In each of these cases, isodose distributions were renormalized until the apex of the prostate was encompassed within the 95% line. This renormalization increased doses to the adjacent critical structures relative to the prescription dose. Dose volume histograms (DVH) of the critical structures and CTV for the 10 patient series, with and without heterogeneity corrections, were generated, averaged, and compared for each gantry angle.
Volume 19, Number 4, 1994
RESULTS
To assess the impact of beam angle on the dose to the patients, rectum, bladder, and femoral head DVHs were generated using angles varying, every 5° from 20 to 45° off the lateral beam direction. The data for the 10 patients were averaged and graphed by critical structure site. Figures 1-6 show central plane axial isodose distributions by gantry angle with and without heterogeneity correction. Figures 7 -12 show apical axial isodose distributions for these same gantry
R
L
R
L
Fig. 1. Central axis axial isodose distribution for opposed oblique gantry angled 20° off the horizontal (70, 110, 250, and 290), with heterogeneity corrections on the top. and without heterogeneity corrections on the bottom, show the variation of dose attenuated by the pelvic bones in conformal treatment of the prostate, on the central axis axial slice.
Oblique angles in prostate cancer radiotberapy • B.
PrCKETI et
al.
239
R
L
R
L
Fig. 2. Central axis axial isodose distribution for opposed oblique gantry angled 25° off the horizontal (65, 115, 245, and 295), with heterogeneity corrections on the top, and without heterogeneity corrections on the bottom, show the variation of dose attenuated by the pelvic bones in conformal treatment of the prostate.
240
Medical Dosimetry
Volume 19, Number 4, 1994
R
L
R
L
Fig. 3. Central axis axial isodose distribution for opposed oblique gantry angled 30° off the horizontal (60, 120, 240, and 300), with heterogeneity corrections on the top, and without heterogeneity corrections on the bottom, show the variation of dose attenuated by the pelvic bones in conformal treatment of the prostate.
Oblique angles in prostate cancer radiotherapy. B. PICKETT et
at.
241
R
L
R
L
Fig. 4. Central axis axial isodose distribution for opposed oblique gantry angled 35" off the horizontal (55, 125, 235, and 305), with heterogeneity corrections on the top, and without heterogeneity corrections on the bottom, show the variation of dose attenuated by the pelvic bones in conformal treatment of the prostate.
242
Medical Dosimetry
Volume 19, Number 4, 1994
R
L
R
L
Fig. 5. Central axis axial isodose distribution for opposed oblique gantry angled 40° off the horizontal (50, 130, 230, and 310), with heterogeneity corrections on the top, and without heterogeneity corrections on the bottom, show the variation of dose attenuated by the pelvic bones in conformal treatment of the prostate.
Oblique angles in prostate cancer radiotherapy. B.
PICKETT
et al.
243
R
L
R
L
Fig. 6. Central axis axial isodose distribution for opposed oblique gantry angled 45° off the horizontal (45, 135, 225, and 315), with heterogeneity corrections on the top, and without heterogeneity corrections on the bottom, show the variation of dose attenuated by the pelvic bones in conformal treatment of the prostate.
244
Medical Dosimetry
Volume 19. Number 4. 1994
40
60~_ _ _ _ _,¥~
L
R
L
R
Fig. 7. Inferior axial isodose distribution for opposed oblique gantry angled 200 off the horizontal. with heterogeneity corrections on the top. and without heterogeneity correction on the bottom. compare the underdose of radiation near the apex of the prostate.
Oblique angles in prostate cancer radiotherapy. B.
PiCKETT
et al.
R
245
L
-------~\10
R
-15
-10
Fig. 8. Inferior axial isodose distribution for opposed oblique gantry angled 2SO off the horizontal, with heterogeneity corrections on the top, and without heterogeneity correction on the bottom, compare the underdose of radiation near the apex of the prostate.
L
246
Medical Dosimetry
Volume 19, Number 4, 1994
R
L
R
L
Fig. 9. Inferior axial isodose distribution for opposed oblique gantry angled 30° off the horizontal, with heterogeneity corrections on the top, and without heterogeneity correction on the bottom, compare the underdose of radiation near the apex of the prostate.
Oblique angles in prostate cancer radiotherapy. B.
PiCKETI et
al.
247
R
L
R
L
Fig. 10. Inferior axial isodose distribution for opposed oblique gantry angled 35° off the horizontal, with heterogeneity corrections on the top, and without heterogeneity correction on the bottom, compare the underdose of radiation near the apex of the prostate.
248
Medical Dosimetry
Volume 19, Number 4, 1994
R
L
R
L
Fig. 11. Inferior axial isodose distribution for opposed oblique gantry angled 40° off the horizontal, with heterogeneity corrections on the top, and without heterogeneity correction on the bottom, compare the underdose of radiation near the apex of the prostate.
Oblique angles in prostate cancer radiotherapy. B.
PICKETT et
al.
249
R
L
R
L
Fig. 12. Inferior axial isodose distribution for opposed oblique gantry angled 45° off the horizontal, with heterogeneity corrections on the top, and without heterogeneity correction on the bottom, compare the underdose of radiation near the apex of the prostate.
Medical Dosimetry
250
Volume 19, Number 4, 1994
PROSTATE DOSE WITHOUT HETEROGENEITY
RECTAL DOSE WITHOUT HETEROGENEITY
~,-----------------------~ 100 4---t--.--U1... 1Il,
lOIS
100
---
100
- - - 2DDtC
,
90
12~ -.-----------------------~
--0--
20DtC
---0---
2!5 DtC
----0-- --
30DtC
----0----
--f).--
SIS DtC
- t r - 3ISDrC
- - -19- - -
40DtC
---tB---
_._ .. .. . _.
4IS DtC
.....
_._ _.
2~
2!5DtC 30DrC
40DrC 4ISDtC
04----.----r---,---~r_~ Q Q Q .,.; o
110
g
....
Dose in%
Dose in %
Fig. 13. This graph compares averaged DVH of the CTV for 10 randomly selected patients, with varying opposed oblique gantry angles and without heterogeneity correction. Isodose distributions with conformal nonuniform margins Were used to generate the graph. When gantry angles varying from 20 to 30° off the horizontal were used, the apex of the prostate was underdosed due to bone attenuation, requiring a lower prescribed isodose value to cover of the prostate.
Fig. 15. This graph compares averaged DVH of the rectum for 10 randomly selected patients, with varying opposed oblique gantry angles and without heterogeneity correction. Isodose distributions with conformal nonuniform margins were used to generate the graph. Without regard to off-axis CTV coverage, as the gantry angles approach the horizontal beam position, the dose to the rectum is the lowest.
angles. Figures 13-16 show averaged DVH for the prostate, rectum, bladder, and femoral heads, respectively, without heterogeneity correction_ The technique used for these 10 patients weighted the individual beams to their own d m• xo with the isodose distribution
normalized at the combined isocenter of the plan. Figures 17 - 20 show averaged DVH by structure site with heterogeneity correction_ These figures reflect isocenter doses renormalized to achieve 95% isodose coverage at the apex of the prostate with varying oblique gantry angles_ 9 Note that the gantry angles from 20 to 30° off the lateral position show higher doses to the
FEMORAL HEADS WITHOUT HETEROGENEITY
BLADDER DOSES WITHOUT HETEROGENEITY
~,--------------------~
--
.•.•.•.....~
.!II "-a jo; ~tt
'..
7S
----0----
...
., ~;
z
30 IlI!lG
;
---&1---
40 IlI!lG
_._
4IS IlI!lG
.....-.
".
0
:15
110
;1 Jo;
7~
7S
--0--
20DtC
----0----
30DtC
2!5DtC
-I!r- SlSDrC
~o
~~
---tB--- 40DtC
_._....._.
25
'.
0
---
100
2!51l1!lG
- t r - SIS IlI!lG
••
110
12~
--0-- 201lI!lG
4ISDrC
100 0 Q
Doee In %
Fig. 14. This graph compares averaged DVH of the femoral head for 10 randomly selected patients, with varying opposed oblique gantry angles and without heterogeneity correction_ Isodose distributions with conformal nonuniform margins were used to generate the graph. Without regard to off-axis CTV coverage, as the gantry angles approach the horizontal beam position, the dose to the femoral heads is the highest; likewise, as the gantry approaches the vertical position, the femoral head doses are lower.
0
Q
.,.;
....
Q
0
on
Q
.,.;
....
Q
Q
~
~
0
.,.;
Doeein%
Fig_ 16. This graph compares averaged DVH of the bladder for 10 randomly selected patients, with varying opposed oblique gantry angles and without heterogeneity correction. Isodose distributions with conformal nonuniform margins were used to generate the graph. Without regard to off-axis CTV coverage, as the gantry angles approach the horizontal beam position, the dose to the bladder is the lowest.
Oblique angles in prostate cancer radiotherapy. B.
RENORMALIZED PROSTATE DOSES WITH HETEROGENEITY
..
~,-----------------------,
----
-0--
"'\
•••• (> •••
r
,
., ...
"...
....
--- --
~
251
et al.
RENORMALIZED RECTAL DOSES WITH HETEROGENEITY
125 , - - - - - - - - - - - - - - - , lOOt---~~~...,
PICKETT
_._
_.
100
---0-
20 Dee
-0--
~Dee
··· · 0· ···
30 Dee
:aDDI!:G
:III DI!:G 3ODI!:G lI5DI!:G 4ODI!:G
.sCI m
Cl8
9~
.... 0
~#.
4ISDI!:G
--A-
.
110
---.---
_._ -._.
~
0 0
.,
"
2IS
110
m
100
35 Dee 40 Dee
411 Dee
~
Dose in %
Dcneln%
Fig. 17. This graph compares averaged DVH of the CTV for 10 randomly selected patients, with varying opposed oblique gantry angles and with heterogeneity correction. Isodose distributions with conformal nonuniform margins were used to generate the graph. Renormalization of the rectal doses were needed to reflect the increase in prescription dose on apical slices with inadequate doses.
Fig. 19. This graph compares averaged DVH of the rectum for 10 randomly selected patients, with varying opposed oblique gantry angles and with heterogeneity correction. Isodose distributions with conformal nonuniform margins were used to generate the graph. Renormalization of doses were needed to reflect the increase in prescription dose on apical slices with inadequate doses.
rectum and bladder than the plans with 35° gantry angle due to heterogeneity correction needed to increase the CTV coverage at the apex of the prostate and subsequent renormalization of the isodose distributions. Traditionally, 45° angles have been used for the six-field technique. The results of this analysis show that 45° angles give the highest rectal and bladder doses of the angles studied; 35° angles provide 95% isodose
coverage to the entire volume of the CTV, lower rectal and bladder doses, and tolerable femoral head doses.
DISCUSSION The 10 patients in our series were simulated and CT scanned with an empty rectum and full bladder RENORMALIZED BLADDER DOSES WITH HETEROGENErI'Y
RENORMALIZED FEMORAL HEAD DOSES WITH HETEROGENEITY 12~
~,-----------------~ 100
#.
~! 9~
.... 0 0
--0--
m
- 0 --
~dr&
... . 0-.••
30 Dee
.
35~
--6-
110
---IB- -_._ -._.
:>
~
0 0
~
110
m
40~
411 Dee
--
100
20 Dee
~!
J~
~#.
75
~
Dose in %
Fig. 18. This graph compares averaged DVH of the femoral head for 10 randomly selected patients, with varying opposed oblique gantry angles and with heterogeneity correction. Isodose distributions with conformal nonuniform margins were used to generate the graph. Renormalization of doses were needed to reflect the increase in prescription dose on apical slices with inadequate doses.
--'-0- " -
30DrC
..
---.---
_._ _.
25
0
100
2IIDrC
-tr-
~o
0
on
N
0
on
...
on
~
20DrC
-0--
3ISDrC 40DrC ClSDrC
on
~
Dose 1D %
Fig. 20. This graph compares averaged DVH of the bladder for 10 randomly selected patients, with varying opposed oblique gantry angles and with heterogeneity correction. Isodose distributions with conformal nonuniform margins were used to generate the graph. Renormalization of doses were needed to reflect the increase in prescription dose on apical slices with inadequate doses.
252
Medical Dosimetry
Volume 19, Number 4, 1994
(a) Fig. 21. This is an apical axial CT slice of a patient displaying a 25° (a) and 35° (b) oblique gantry angle relative to the density of the pelvic bones along the central axis of the treatment beam.
allowing treatment planning with the prostate and seminal vesicles in the most posterior position possible. In so doing, the movement of the CTV is minimized in the posterior direction, and nonuniform field edges were used to accommodate CTV coverage. All prostate movement was found to be adequately covered within the prescribed 95% isodose value when the "nonuniform margins" were designed. 7•10 Our study compared the homogeneity of CTV isodose distributions with varying opposed oblique gantry angles, with and without heterogeneity corrections. To provide the most uniform coverage of the CTV, special
attention needs to be given to the apex of the prostate. There is a tendency for the dose near the apex of the prostate to be lower than the dose on the central axis. It can be seen that as the gantry is angled more laterally, a greater amount of bone intersects the center of the field, attenuating the beam slightly. When isodose distributions were compared with and without density corrections, it was determined that there was 2-5% attenuation of the beam due to high bone density in this area depending on oblique gantry angle. Refer to Fig. 3, when 30° oblique angles were used, and note that the 95% isodose curve encompasses the CTV on the
Oblique angles in prostate cancer radiotherapy. B. PICKETI el al.
253
(b)
Fig. 2l. (Continued) central axis both with and without heterogeneity correction. However, Fig. 9 shows that it requires the 90% isodose value to cover the apex of the prostate and, therefore, to treat adequately the CTV with 30° angles. Prescribing treatment to the 90% isodose line, rather than the 95% isodose line, increases the adjacent critical structure doses by the additional 5%. The difference in apical dose coverage, with and without heterogeneity correcti~n, was used to determine the magnitude of attenuation resulting from absorption of dose in the pelvic bones. Angling the gantry slightly more vertically, so that the central axis of the posterior oblique beam is outside the pelvic bones, decreases the differ-
ence in isodose value needed to cover the -apex of the CTV, and in effect minimizes the dose to adjacent structures. Treating the patient with the gantry angled 35° off the horizontal position is a large enough difference to move the central axis out of the attenuating bones, and to allow the apex of the prostate to be covered by the 95% isodose line. Figures 21A and 21B show an inferior slice of a CT scan near the apex of the prostate with 25° oblique and 35° oblique angles, respectively. Notice the different effect of pelvic bone densities at the apex of the prostate depending the oblique gantry angle used. Our study used DVH to evaluate the doses to the
254
Medical Dosimetry
prostate, rectum, bladder, and femoral heads relative to gantry angle. It should be noted that as the gantry angles approach the horizontal beam position, the doses to the rectum and bladder are the lowest and became higher as the gantry angles approached the lateral beam position. The doses to the femoral heads are the lowest for more vertical gantry positions and became higher as the gantry angles approached the horizontal beam position (AP or PA). Using 45° obliqued gantry angles would therefore result in the most homogeneous isodose distribution, with less apical dose discrepancy, but the doses to the rectum and bladder would be the highest. Using 20° obliqued gantry angles would result in the very high femoral head doses, with a suboptimal diamond-shaped isodose distribution represented in Fig. 1. Thirty-five degree oblique gantry angles result in an adequately homogeneous dose distribution, tolerable critical structure doses, and adequate apical prostate coverage. CONCLUSION With the availability of PSA screening, recent surgical data suggest that fewer men are likely to present with advanced disease, and more men will be candidates for local prostate irradiation. Using opposed oblique gantry angles less than 35° off the horizontal require prescribed isodose values lower than 90% to cover the apex of the prostate. Gantry angles from 35 to 45° encompass the apex of the prostate within the 95% isodose line; however, the more posterior is the gantry angle, the higher the rectum and bladder dose. The more lateral the gantry is angled, the higher the femoral dose. Treating adenocarcinoma of the prostate with 35° oblique angles off the lateral position delivers
Volume 19, Number 4, 1994
doses to the rectum and bladder that are within tolerance and maintains homogeneous isodose coverage of the CTV. REFERENCES I. Hunt, M.; Schultbeiss, T.; Hanks, G. A comparison of set-up
variations using EPID devices and film for patients treated for prostate cancer. Abstract at ASTRO;. 1993. 2. Roach, M.; Pickett, B.; Holland, J.; Zapotowski, K.; Marsh, D.; Tatera, B. Urethrogram during simulation for localized prostate cancer. Int. J. Radiat. BioI. Phys. 25:299-307; 1993. 3. Schild, S.; Buskirk, S.; Robinow, J. Prostate cancer: retrograde urethrography to improve treatment planning for radiation tberapy. Radiology. 181:885-887; 1991. 4. Sweeney, P.; Vijayakumar, S.; Sibley, G.; Salehpour, M.; Myriantbopoulos, L.; Rubin, S.; Sutton, H. Comparison of CT-based treatment planning and retrograde urethrography in determining the prostatic apex at simulation. Med. Dosim. 18:21-28; 1992. 5. Pilepich, M.; Prasad, S.; Perez, C. Computed tomography in definitive radiotberapy of prostatic carcinoma, part 2: definition of target volume. Int J Radiat. Oneol. Bioi. Phys. 8:235-240; 1982. 6. Rosenthal, S.; Roach, M.; Goldsmith, B.; Pickett, B.; Doggett, E.; Ryu, J. The accuracy of patient positioning during radiation for prostate cancer using a six field conformal technique with and without immobilization. Int. J. Radiat. BioI. Phys. 27(4):921-926; 1993. 7. Pickett, B.; Roach III, M.; Horine, P.; Akazawa, C.; Verhey, L.; Phillips, T.L. The value of "ideal margins" in tbe treatment of localized prostate cancer. Int. J. Radiat. Oneol. BioI. Phys. [In press]. 8. Pickett, 8.; Altieri, G. Normalization: what does it really mean? Med. Dosim. 17:15-27; 1992. 9. Roach, M.; Pickett, B.; Phillips, T. An analysis oftbe advantages as well as the physical and clinical limitations of three dimensionally (3-D) based co-planar conformal external beam irradiation (XRT) in tbe treatment of localized prostate cancer. In: Pierre Minet, M. ed. Three dimensional treatment planning. LaSalle, Belgium: Etienne RIGA; 1992:149-162. 10. Roach, M.; Pickett, 8.; Rosenthal, S.; Verhey, L.; Phillips, T. Defining treatment margins for six field conformal irradiation of localized prostate cancer. Int. J. Radiat. Oneol. BioI. Phys. 28(1):267-275; 1993.
Medical Dosimetry, Vol. 19, No.4, pp. 237-254, 1994 Copyright © 1994 American Association of Medical Dosimetrists Printed in the USA. All rights reserved 0958-3947/94 $6.00 + .00
Pergamon
0958-3947(94)00033-6
OPTIMIZATION OF THE OBLIQUE ANGLES IN THE TREATMENT OF PROSTATE CANCER DURING SIX-FIELD CONFORMAL RADIOTHERAPY BARBY PICKETT, MS, MACK ROACH,
III,
MD, PAUL HORINE, BS,
LYNN VERHEY, PH.D., and THEODORE L. PHILLIPS, MD University of California, San Francisco, CA 94143
Abstract-Historically, four perpendicular treatment fields or bilateral arcs have been used in the treatment of prostate cancer. As new techniques challenge the four-field box technique for their superiority of tumor coverage and adjacent critical structure sparing, oblique beam angles (in addition to right and left laterals) have been introduced as an alternative to anterior (AP) and posterior (PA) beams. Among the most popular of these alternative approaches is a six-field technique. Traditionally 45° angles have been used with this technique. In this study, opposed coplanar oblique beams angled 20, 25, 30, 35, 40, and 45° otT the lateral beam position, were compared for their ability to minimize adjacent critical structure doses, while maintaining maximum clinical target volume (CTV) coverage. This analysis compared rectum, bladder, and femoral head dose volume histograms (DVH) for each of these varying oblique gantry angles. As the angle of the posterior oblique beams became more horizontal, it is more difficult to encompass the apex of the prostate in the 95 % isodose value. On inferior CT slices near the apex of the prostate, the density of the pelvic bones in the path of the posterior oblique fields causes the beam to be slightly attenuated, thereby underdosing the CTV. The oblique angles most atTected by this bone heterogeneity are beams angled from 20 to 30° otT the lateral beam position. As the angles approach the vertical direction, rectal and bladder doses increase, while femoral head doses decrease. Oblique gantry angles approaching the horizontal direction result in a decrease in rectal and bladder doses, while femoral head doses increase. Of the oblique angles studied, 35° otT the lateral position provides lower rectum and bladder doses than 30, 40, and 45°; lower femoral head doses than 20, 25, and 30°, and the maximum CTV coverage on all CT slices studied. Key Words: Prostate cancer, Normalization, Three-dimensional treatment planning, Oblique angles.
treatment planning system was used to calculate doses to adjacent critical structures with varying opposed gantry angles, of 20, 25, 30, 35, 40, and 45° off the lateral beam direction. 8,9 Conformal treatment plans consisting of lateral (RTL and LTL), and oblique beams (RAO, LAO, LPO, and RPO) with similar techniques were analyzed, with and without heterogeneity corrections, to evaluate the impact of bone density on resulting dose distributions. Each patient was planned with 6-MV photons. A previous study showed little or no difference in dose distribution when photon energies from 6 MV to 18 MV and six or more conformal fields Were used. Blocks were designed to allow for daily treatment inaccuracies, internal organ movement, uncertainties associated with extracapsular extension, and homogeneous dose distributions. The block design has been published previously and is referred to as "nonuniform.,,7,10 The isodose distributions were evaluated on six planes of calculation including the most inferior axial, the central axis axial, and the most superior axial cuts, along with reconstructed midline sagittal, coronal, and oblique views to assess the homogeneity of isodose distributions and CTV coverage in all planes of calculation.
PURPOSE/OBJECTIVE This study was undertaken to determine and evaluate the optimal oblique gantry angles to be used for coplanar conformal treatment of the prostate by comparing the doses to adjacent critical structures and the isodose distributions relative to clinical target volume (CTV) coverage.
MATERIALS AND METHODS For this study, 10 patients were simulated in the supine position, and a urethrogram was obtained to assist in defining the apex of the prostate. 1. 4 Each patient had a treatment planning CT scan with 5-mm slice thicknesses. To minimize the risk of patient and organ movement, patients were simulated and treated with custom immobilization and scanned with their rectums empty and their bladders full. 5 . 7 After identifying the prostate, seminal vesicles, bladder, rectum, and femoral heads on axial CT slices, a U -M PLAN (University, of Michigan) three-dimensional (3-D)
Reprint requests to: Barby Pickett, MS, University of California, San Francisco, CA 94143.
237
238
Medical Dosimetry
3-D treatment plans were designed for each of the six gantry angles evaluated, with and without heterogeneity correction. In each of these cases, isodose distributions were renormalized until the apex of the prostate was encompassed within the 95% line. This renormalization increased doses to the adjacent critical structures relative to the prescription dose. Dose volume histograms (DVH) of the critical structures and CTV for the 10 patient series, with and without heterogeneity corrections, were generated, averaged, and compared for each gantry angle.
Volume 19, Number 4, 1994
RESULTS
To assess the impact of beam angle on the dose to the patients, rectum, bladder, and femoral head DVHs were generated using angles varying, every 5° from 20 to 45° off the lateral beam direction. The data for the 10 patients were averaged and graphed by critical structure site. Figures 1-6 show central plane axial isodose distributions by gantry angle with and without heterogeneity correction. Figures 7 -12 show apical axial isodose distributions for these same gantry
R
L
R
L
Fig. 1. Central axis axial isodose distribution for opposed oblique gantry angled 20° off the horizontal (70, 110, 250, and 290), with heterogeneity corrections on the top. and without heterogeneity corrections on the bottom, show the variation of dose attenuated by the pelvic bones in conformal treatment of the prostate, on the central axis axial slice.
Oblique angles in prostate cancer radiotberapy • B.
PrCKETI et
al.
239
R
L
R
L
Fig. 2. Central axis axial isodose distribution for opposed oblique gantry angled 25° off the horizontal (65, 115, 245, and 295), with heterogeneity corrections on the top, and without heterogeneity corrections on the bottom, show the variation of dose attenuated by the pelvic bones in conformal treatment of the prostate.
240
Medical Dosimetry
Volume 19, Number 4, 1994
R
L
R
L
Fig. 3. Central axis axial isodose distribution for opposed oblique gantry angled 30° off the horizontal (60, 120, 240, and 300), with heterogeneity corrections on the top, and without heterogeneity corrections on the bottom, show the variation of dose attenuated by the pelvic bones in conformal treatment of the prostate.
Oblique angles in prostate cancer radiotherapy. B. PICKETT et
at.
241
R
L
R
L
Fig. 4. Central axis axial isodose distribution for opposed oblique gantry angled 35" off the horizontal (55, 125, 235, and 305), with heterogeneity corrections on the top, and without heterogeneity corrections on the bottom, show the variation of dose attenuated by the pelvic bones in conformal treatment of the prostate.
242
Medical Dosimetry
Volume 19, Number 4, 1994
R
L
R
L
Fig. 5. Central axis axial isodose distribution for opposed oblique gantry angled 40° off the horizontal (50, 130, 230, and 310), with heterogeneity corrections on the top, and without heterogeneity corrections on the bottom, show the variation of dose attenuated by the pelvic bones in conformal treatment of the prostate.
Oblique angles in prostate cancer radiotherapy. B.
PICKETT
et al.
243
R
L
R
L
Fig. 6. Central axis axial isodose distribution for opposed oblique gantry angled 45° off the horizontal (45, 135, 225, and 315), with heterogeneity corrections on the top, and without heterogeneity corrections on the bottom, show the variation of dose attenuated by the pelvic bones in conformal treatment of the prostate.
244
Medical Dosimetry
Volume 19. Number 4. 1994
40
60~_ _ _ _ _,¥~
L
R
L
R
Fig. 7. Inferior axial isodose distribution for opposed oblique gantry angled 200 off the horizontal. with heterogeneity corrections on the top. and without heterogeneity correction on the bottom. compare the underdose of radiation near the apex of the prostate.
Oblique angles in prostate cancer radiotherapy. B.
PiCKETT
et al.
R
245
L
-------~\10
R
-15
-10
Fig. 8. Inferior axial isodose distribution for opposed oblique gantry angled 2SO off the horizontal, with heterogeneity corrections on the top, and without heterogeneity correction on the bottom, compare the underdose of radiation near the apex of the prostate.
L
246
Medical Dosimetry
Volume 19, Number 4, 1994
R
L
R
L
Fig. 9. Inferior axial isodose distribution for opposed oblique gantry angled 30° off the horizontal, with heterogeneity corrections on the top, and without heterogeneity correction on the bottom, compare the underdose of radiation near the apex of the prostate.
Oblique angles in prostate cancer radiotherapy. B.
PiCKETI et
al.
247
R
L
R
L
Fig. 10. Inferior axial isodose distribution for opposed oblique gantry angled 35° off the horizontal, with heterogeneity corrections on the top, and without heterogeneity correction on the bottom, compare the underdose of radiation near the apex of the prostate.
248
Medical Dosimetry
Volume 19, Number 4, 1994
R
L
R
L
Fig. 11. Inferior axial isodose distribution for opposed oblique gantry angled 40° off the horizontal, with heterogeneity corrections on the top, and without heterogeneity correction on the bottom, compare the underdose of radiation near the apex of the prostate.
Oblique angles in prostate cancer radiotherapy. B.
PICKETT et
al.
249
R
L
R
L
Fig. 12. Inferior axial isodose distribution for opposed oblique gantry angled 45° off the horizontal, with heterogeneity corrections on the top, and without heterogeneity correction on the bottom, compare the underdose of radiation near the apex of the prostate.
Medical Dosimetry
250
Volume 19, Number 4, 1994
PROSTATE DOSE WITHOUT HETEROGENEITY
RECTAL DOSE WITHOUT HETEROGENEITY
~,-----------------------~ 100 4---t--.--U1... 1Il,
lOIS
100
---
100
- - - 2DDtC
,
90
12~ -.-----------------------~
--0--
20DtC
---0---
2!5 DtC
----0-- --
30DtC
----0----
--f).--
SIS DtC
- t r - 3ISDrC
- - -19- - -
40DtC
---tB---
_._ .. .. . _.
4IS DtC
.....
_._ _.
2~
2!5DtC 30DrC
40DrC 4ISDtC
04----.----r---,---~r_~ Q Q Q .,.; o
110
g
....
Dose in%
Dose in %
Fig. 13. This graph compares averaged DVH of the CTV for 10 randomly selected patients, with varying opposed oblique gantry angles and without heterogeneity correction. Isodose distributions with conformal nonuniform margins Were used to generate the graph. When gantry angles varying from 20 to 30° off the horizontal were used, the apex of the prostate was underdosed due to bone attenuation, requiring a lower prescribed isodose value to cover of the prostate.
Fig. 15. This graph compares averaged DVH of the rectum for 10 randomly selected patients, with varying opposed oblique gantry angles and without heterogeneity correction. Isodose distributions with conformal nonuniform margins were used to generate the graph. Without regard to off-axis CTV coverage, as the gantry angles approach the horizontal beam position, the dose to the rectum is the lowest.
angles. Figures 13-16 show averaged DVH for the prostate, rectum, bladder, and femoral heads, respectively, without heterogeneity correction_ The technique used for these 10 patients weighted the individual beams to their own d m• xo with the isodose distribution
normalized at the combined isocenter of the plan. Figures 17 - 20 show averaged DVH by structure site with heterogeneity correction_ These figures reflect isocenter doses renormalized to achieve 95% isodose coverage at the apex of the prostate with varying oblique gantry angles_ 9 Note that the gantry angles from 20 to 30° off the lateral position show higher doses to the
FEMORAL HEADS WITHOUT HETEROGENEITY
BLADDER DOSES WITHOUT HETEROGENEITY
~,--------------------~
--
.•.•.•.....~
.!II "-a jo; ~tt
'..
7S
----0----
...
., ~;
z
30 IlI!lG
;
---&1---
40 IlI!lG
_._
4IS IlI!lG
.....-.
".
0
:15
110
;1 Jo;
7~
7S
--0--
20DtC
----0----
30DtC
2!5DtC
-I!r- SlSDrC
~o
~~
---tB--- 40DtC
_._....._.
25
'.
0
---
100
2!51l1!lG
- t r - SIS IlI!lG
••
110
12~
--0-- 201lI!lG
4ISDrC
100 0 Q
Doee In %
Fig. 14. This graph compares averaged DVH of the femoral head for 10 randomly selected patients, with varying opposed oblique gantry angles and without heterogeneity correction_ Isodose distributions with conformal nonuniform margins were used to generate the graph. Without regard to off-axis CTV coverage, as the gantry angles approach the horizontal beam position, the dose to the femoral heads is the highest; likewise, as the gantry approaches the vertical position, the femoral head doses are lower.
0
Q
.,.;
....
Q
0
on
Q
.,.;
....
Q
Q
~
~
0
.,.;
Doeein%
Fig_ 16. This graph compares averaged DVH of the bladder for 10 randomly selected patients, with varying opposed oblique gantry angles and without heterogeneity correction. Isodose distributions with conformal nonuniform margins were used to generate the graph. Without regard to off-axis CTV coverage, as the gantry angles approach the horizontal beam position, the dose to the bladder is the lowest.
Oblique angles in prostate cancer radiotherapy. B.
RENORMALIZED PROSTATE DOSES WITH HETEROGENEITY
..
~,-----------------------,
----
-0--
"'\
•••• (> •••
r
,
., ...
"...
....
--- --
~
251
et al.
RENORMALIZED RECTAL DOSES WITH HETEROGENEITY
125 , - - - - - - - - - - - - - - - , lOOt---~~~...,
PICKETT
_._
_.
100
---0-
20 Dee
-0--
~Dee
··· · 0· ···
30 Dee
:aDDI!:G
:III DI!:G 3ODI!:G lI5DI!:G 4ODI!:G
.sCI m
Cl8
9~
.... 0
~#.
4ISDI!:G
--A-
.
110
---.---
_._ -._.
~
0 0
.,
"
2IS
110
m
100
35 Dee 40 Dee
411 Dee
~
Dose in %
Dcneln%
Fig. 17. This graph compares averaged DVH of the CTV for 10 randomly selected patients, with varying opposed oblique gantry angles and with heterogeneity correction. Isodose distributions with conformal nonuniform margins were used to generate the graph. Renormalization of the rectal doses were needed to reflect the increase in prescription dose on apical slices with inadequate doses.
Fig. 19. This graph compares averaged DVH of the rectum for 10 randomly selected patients, with varying opposed oblique gantry angles and with heterogeneity correction. Isodose distributions with conformal nonuniform margins were used to generate the graph. Renormalization of doses were needed to reflect the increase in prescription dose on apical slices with inadequate doses.
rectum and bladder than the plans with 35° gantry angle due to heterogeneity correction needed to increase the CTV coverage at the apex of the prostate and subsequent renormalization of the isodose distributions. Traditionally, 45° angles have been used for the six-field technique. The results of this analysis show that 45° angles give the highest rectal and bladder doses of the angles studied; 35° angles provide 95% isodose
coverage to the entire volume of the CTV, lower rectal and bladder doses, and tolerable femoral head doses.
DISCUSSION The 10 patients in our series were simulated and CT scanned with an empty rectum and full bladder RENORMALIZED BLADDER DOSES WITH HETEROGENErI'Y
RENORMALIZED FEMORAL HEAD DOSES WITH HETEROGENEITY 12~
~,-----------------~ 100
#.
~! 9~
.... 0 0
--0--
m
- 0 --
~dr&
... . 0-.••
30 Dee
.
35~
--6-
110
---IB- -_._ -._.
:>
~
0 0
~
110
m
40~
411 Dee
--
100
20 Dee
~!
J~
~#.
75
~
Dose in %
Fig. 18. This graph compares averaged DVH of the femoral head for 10 randomly selected patients, with varying opposed oblique gantry angles and with heterogeneity correction. Isodose distributions with conformal nonuniform margins were used to generate the graph. Renormalization of doses were needed to reflect the increase in prescription dose on apical slices with inadequate doses.
--'-0- " -
30DrC
..
---.---
_._ _.
25
0
100
2IIDrC
-tr-
~o
0
on
N
0
on
...
on
~
20DrC
-0--
3ISDrC 40DrC ClSDrC
on
~
Dose 1D %
Fig. 20. This graph compares averaged DVH of the bladder for 10 randomly selected patients, with varying opposed oblique gantry angles and with heterogeneity correction. Isodose distributions with conformal nonuniform margins were used to generate the graph. Renormalization of doses were needed to reflect the increase in prescription dose on apical slices with inadequate doses.
252
Medical Dosimetry
Volume 19, Number 4, 1994
(a) Fig. 21. This is an apical axial CT slice of a patient displaying a 25° (a) and 35° (b) oblique gantry angle relative to the density of the pelvic bones along the central axis of the treatment beam.
allowing treatment planning with the prostate and seminal vesicles in the most posterior position possible. In so doing, the movement of the CTV is minimized in the posterior direction, and nonuniform field edges were used to accommodate CTV coverage. All prostate movement was found to be adequately covered within the prescribed 95% isodose value when the "nonuniform margins" were designed. 7•10 Our study compared the homogeneity of CTV isodose distributions with varying opposed oblique gantry angles, with and without heterogeneity corrections. To provide the most uniform coverage of the CTV, special
attention needs to be given to the apex of the prostate. There is a tendency for the dose near the apex of the prostate to be lower than the dose on the central axis. It can be seen that as the gantry is angled more laterally, a greater amount of bone intersects the center of the field, attenuating the beam slightly. When isodose distributions were compared with and without density corrections, it was determined that there was 2-5% attenuation of the beam due to high bone density in this area depending on oblique gantry angle. Refer to Fig. 3, when 30° oblique angles were used, and note that the 95% isodose curve encompasses the CTV on the
Oblique angles in prostate cancer radiotherapy. B. PICKETI el al.
253
(b)
Fig. 2l. (Continued) central axis both with and without heterogeneity correction. However, Fig. 9 shows that it requires the 90% isodose value to cover the apex of the prostate and, therefore, to treat adequately the CTV with 30° angles. Prescribing treatment to the 90% isodose line, rather than the 95% isodose line, increases the adjacent critical structure doses by the additional 5%. The difference in apical dose coverage, with and without heterogeneity correcti~n, was used to determine the magnitude of attenuation resulting from absorption of dose in the pelvic bones. Angling the gantry slightly more vertically, so that the central axis of the posterior oblique beam is outside the pelvic bones, decreases the differ-
ence in isodose value needed to cover the -apex of the CTV, and in effect minimizes the dose to adjacent structures. Treating the patient with the gantry angled 35° off the horizontal position is a large enough difference to move the central axis out of the attenuating bones, and to allow the apex of the prostate to be covered by the 95% isodose line. Figures 21A and 21B show an inferior slice of a CT scan near the apex of the prostate with 25° oblique and 35° oblique angles, respectively. Notice the different effect of pelvic bone densities at the apex of the prostate depending the oblique gantry angle used. Our study used DVH to evaluate the doses to the
254
Medical Dosimetry
prostate, rectum, bladder, and femoral heads relative to gantry angle. It should be noted that as the gantry angles approach the horizontal beam position, the doses to the rectum and bladder are the lowest and became higher as the gantry angles approached the lateral beam position. The doses to the femoral heads are the lowest for more vertical gantry positions and became higher as the gantry angles approached the horizontal beam position (AP or PA). Using 45° obliqued gantry angles would therefore result in the most homogeneous isodose distribution, with less apical dose discrepancy, but the doses to the rectum and bladder would be the highest. Using 20° obliqued gantry angles would result in the very high femoral head doses, with a suboptimal diamond-shaped isodose distribution represented in Fig. 1. Thirty-five degree oblique gantry angles result in an adequately homogeneous dose distribution, tolerable critical structure doses, and adequate apical prostate coverage. CONCLUSION With the availability of PSA screening, recent surgical data suggest that fewer men are likely to present with advanced disease, and more men will be candidates for local prostate irradiation. Using opposed oblique gantry angles less than 35° off the horizontal require prescribed isodose values lower than 90% to cover the apex of the prostate. Gantry angles from 35 to 45° encompass the apex of the prostate within the 95% isodose line; however, the more posterior is the gantry angle, the higher the rectum and bladder dose. The more lateral the gantry is angled, the higher the femoral dose. Treating adenocarcinoma of the prostate with 35° oblique angles off the lateral position delivers
Volume 19, Number 4, 1994
doses to the rectum and bladder that are within tolerance and maintains homogeneous isodose coverage of the CTV. REFERENCES I. Hunt, M.; Schultbeiss, T.; Hanks, G. A comparison of set-up
variations using EPID devices and film for patients treated for prostate cancer. Abstract at ASTRO;. 1993. 2. Roach, M.; Pickett, B.; Holland, J.; Zapotowski, K.; Marsh, D.; Tatera, B. Urethrogram during simulation for localized prostate cancer. Int. J. Radiat. BioI. Phys. 25:299-307; 1993. 3. Schild, S.; Buskirk, S.; Robinow, J. Prostate cancer: retrograde urethrography to improve treatment planning for radiation tberapy. Radiology. 181:885-887; 1991. 4. Sweeney, P.; Vijayakumar, S.; Sibley, G.; Salehpour, M.; Myriantbopoulos, L.; Rubin, S.; Sutton, H. Comparison of CT-based treatment planning and retrograde urethrography in determining the prostatic apex at simulation. Med. Dosim. 18:21-28; 1992. 5. Pilepich, M.; Prasad, S.; Perez, C. Computed tomography in definitive radiotberapy of prostatic carcinoma, part 2: definition of target volume. Int J Radiat. Oneol. Bioi. Phys. 8:235-240; 1982. 6. Rosenthal, S.; Roach, M.; Goldsmith, B.; Pickett, B.; Doggett, E.; Ryu, J. The accuracy of patient positioning during radiation for prostate cancer using a six field conformal technique with and without immobilization. Int. J. Radiat. BioI. Phys. 27(4):921-926; 1993. 7. Pickett, B.; Roach III, M.; Horine, P.; Akazawa, C.; Verhey, L.; Phillips, T.L. The value of "ideal margins" in tbe treatment of localized prostate cancer. Int. J. Radiat. Oneol. BioI. Phys. [In press]. 8. Pickett, 8.; Altieri, G. Normalization: what does it really mean? Med. Dosim. 17:15-27; 1992. 9. Roach, M.; Pickett, B.; Phillips, T. An analysis oftbe advantages as well as the physical and clinical limitations of three dimensionally (3-D) based co-planar conformal external beam irradiation (XRT) in tbe treatment of localized prostate cancer. In: Pierre Minet, M. ed. Three dimensional treatment planning. LaSalle, Belgium: Etienne RIGA; 1992:149-162. 10. Roach, M.; Pickett, 8.; Rosenthal, S.; Verhey, L.; Phillips, T. Defining treatment margins for six field conformal irradiation of localized prostate cancer. Int. J. Radiat. Oneol. BioI. Phys. 28(1):267-275; 1993.