Medical Dosimetry, Vol. 23, No. 2, pp.

89 –94, 1998
1998 American Association of Medical Dosimetrists
Printed in the USA. All rights reserved
0958-3947/98 $19.00 1 .00

PII S0958-3947(98)00009-0

COMPARISON OF THE HOMOGENEITY OF BREAST DOSE

DISTRIBUTIONS WITH AND WITHOUT THE MEDIAL WEDGE
CYNTHIA L. IKNER, M.S., D.O., RICHARD RUSSO, R.T.T., C.M.D.,
MATTHEW B. PODGORSAK, PH.D., GARY M. PROULX, M.D., and R. JEFFREY LEE, M.D.
Department of Radiation Medicine, Roswell Park Cancer Institute, Buffalo, NY 14263-0001

Abstract—Radiation of the intact breast often requires medial and lateral wedges to improve dose homogeneity
of its pyramidal shape and to achieve acceptable cosmesis. There is some concern that radiation scatter from the
medial wedge may contribute to cancer in the uninvolved breast, yet treatment without the medial wedge is
associated with inhomogeneity of magnitudes that affect cosmesis. These homogeneities are identified on
treatment plans generated at the central axis (CAX). It is not known if comparing isodose curves at the central
axis reflect homogeneity in superior and inferior planes. A study was undertaken to both examine inhomogeneity
with and without the medial wedge, and to determine if plan selection at the CAX was representative of
homogeneity above and below the CAX. Ten consecutive patients with early breast cancers had cranial, CAX,
and caudal CT images of each breast compared with two wedging conditions, lateral only (LW) and medial and
lateral wedged conditions (dual wedges 5 DW). Dosimetry was optimized at the CAX for DW and LW
conditions. Dose distributions and hot spots relative to prescribed dose were compared for cranial, CAX, and
caudal images. Mean chest wall separations were measured. Six of ten patients had equivalent LW and DW
distributions at the levels examined. Only one of these patients had a single off-axis hot spot > 20%. Six patients
had comparable LW and DW dosimetry and acceptable hot spots at the central axis, as well as chest wall
separations < 22 cm. In conclusion, if isodose configurations are commensurate at the CAX, these patients will
have homogeneity above and below the CAX. In patients with chest wall separations < 22 cm, treatment without
the medial wedge is feasible, sparing the contralateral breast dose with little compromise to inhomogeneity in the
treated breast. © 1998 American Association of Medical Dosimetrists.

Key Words: Breast, Inhomogeneity, Dosimetry, Wedges, Off-axis.

INTRODUCTION mesis, omitting the medial wedge is currently suggested

Breast conserving surgery and adjuvant radiation therapy
avoid mutilating surgery without compromising survival.
Radiation, a known carcinogen, is a required component
to conserve the breast. Its application requires careful
consideration to prevent causing the very problem it may
solve. Carcinogenesis of the chest wall and the contralat-
eral breast have been reported with increased incidence
following irradiation, particularly in patients less than 45
years old when treated.1–3 The etiology, however, is not
known.
It has been suggested that the radiation scatter from
the medial wedge increases the contralateral breast
dose.4 – 8 The wedges are used to compensate for the
more narrow nipple/areolar apex which extends outward
from its widened base at the chest wall. With wedge
heels toward the apex, dose homogeneity as measured in
the axial midplane can be achieved.
Acceptable dosimetry with minimum to maximum
dose is measured at the central axis (CAX).9 Increased
telangectasia and fibrosis of the breast has been reported
when the inhomogeneity reaches 20%.8 Despite its asso-
ciation with inhomogeneity and reported effects on cos-
Reprint requests to: Dr. Cynthia Ikner, Department of Radiation
Medicine, Roswell Park Cancer Institute, Elm and Carlton Street,
Buffalo, NY 14263-0001.
89
for women # 45 years old based on data from the
Connecticut Breast Registry.7 In this study of 41,109
patients receiving breast and chest wall irradiation, an
increased relative risk of contralateral breast cancers was
seen in this group compared to women $ 45 years.
Inhomogeneity seen at the central axis, measured by
3D and pseudo-3D algorithms, is compounded cranial
and caudal to the CAX in large breasts (chest wall
separation $ 28 cm).9 The extent of inhomogeneity
above and below the CAX has been shown to be 6 20%
when optimization and lung corrections are used.8 In the
same study it was concluded that wedges only achieve
tissue compensation in a single plane and this may have
important clinical implications for both local control and
normal tissue damage. These findings have been corrob-
orated in 3D treatment planning where it is possible to
provide necessary tissue compensation to fit breast ge-
ometry and correct inhomogeneity.10,11 In pseudo-3D it
is possible to display multiple CT slices, but the dose
distribution is calculated by area on any one and then
added to give a “volume”. This requires 10 mm images
throughout the breast, exposing the patient to additional
radiation for a less than accurate result. Treatment plan-
ning for pseudo-3D systems frequently rely on dosimetry at
the central axis. It is not known if comparison of plans at the
central axis is reflective of the dosimetry above and below.
90 Medical Dosimetry Volume 23, Number 2, 1998

Table 1. Chest wall separation (s), field sizes and wedges at Planning images. The length of the treatment field
the CA was determined by a superior point at the uppermost
Chest Wall breast tissue and a lower margin 2 cm below the breast
Separation Field sizes tissue. A central mean was computed and the superior
Patient (s in cm) (cm) DW°/LW° image was taken at the centimeter measure equidistant
A 18 6.1 3 16.8 6 45/2 45 from the center of the field to the superior field margin.
B 25 11 3 23.4 6 30/1 45 Likewise, an inferior image was taken at a point equi-
C 21 10.8 3 19 6 30/1 30 distant from the field center to the inferior field edge.
D 22 8 3 18 1 16/1 30
E 20 9 3 19.8 6 20/1 45 Dosimetry. The ROCS treatment planning system
F 24.5 11.5 3 18 6 30/2 45
G 22 11.5 3 23 6 30/1 45 was used to calculate the isodose distribution produced
H 19 7.4 3 18.3 6 30/2 45 by tangential fields to the breast using 6 MV X rays and
I 21 10.3 3 18 6 30/2 45 using an asymmetric jaw from a Varian 600 C linear
J 21 10 3 19 6 30/2 45
accelerator to minimize lung irradiation from diverging
beams. The technique used was the standard lateral
shift.8 Three transaxial slices and two treatment plans per
A study was undertaken to determine if the tech- slice from each patient were developed using the beam
nique of comparing CAX scans adequately reflects ho- parameters as previously determined on a Varian Ximet-
mogeneity above and below the CAX in consecutive ron simulator. The dose normalization point was chosen
patients with early breast cancers. Likewise, can exam- at a point which is at the lung-chest wall interface, in the
ination of the CAX reflect homogeneity above and below plane of the perpendicular bisector of the posterior field
the central axis so as to determine which patients can be edges of the two beams at isocenter. This point is nor-
treated without the medial wedge? malized to the dose prescribed by the physician (200
cGy). The 200 cGy volume was duplicated from the CA
METHODS slices in the DW and LW conditions for each patient.
Patient selection Lung corrections were used. Optimization was attempted
Eleven consecutive patients with early stage breast to keep the tumor volume determined by the treating
cancer and an intact breast after breast conserving sur- physician to within 110% (220 cGy). This isodose plan
gery were selected. was duplicated by adjusting the field weightings and
using the appropriate wedge(s) as needed. Plans were
CT scanning evaluated by a physicist to determine comparability.

Positioning. Using a Radiation Oncology Comput-

ers Systems, Inc. (ROCS) Density Phantom,12 patients
were placed in the treatment position, which included the
ipsilateral arm above the head and chin turned to the
contralateral shoulder, tilted out of the beam edge. The
contralateral arm was placed at the patient’s side and a
triangular sponge was placed beneath the patient’s knees.
RESULTS
Data were obtained on 11 patients and 11 breasts.
The median age of the patients was 54 years, with a
range of 39 –79 years. One patient had only the dual
wedged condition completed and the data were not sig-
nificantly different with inclusion; therefore, the data

Fig. 1. Superior image for patient D using dual wedges.

 Homogeneity of breast dose distribution ● C. IKNER et al. 91

Fig. 2. CAX of patient D using dual wedges.

included only the 10 patients with all three measures in
dual wedged and single wedged conditions. Table 1
shows a description of field size, chest wall separation
and wedges used at the CAX for both dual wedged and
single wedged conditions. Eight patients had s 5 20.3 6
1.2 cm and two had s 5 24 6 1.8 (patients B and F),
consistent with medium and large breasts, respective-
ly.13
Patients A, C, D, E, H, and I were able to achieve
equivalent wedged and unwedged volumes on CT plan-
ning. In these patients, if the dual and single wedged
CAXs showed similar isodose curves, the superior and
inferior slices were comparable in the presence or ab-
sence of the medial wedge. Figures 1, 2, and 3 show dose
distributions for superior, CAX, and inferior slices for
the DW condition for patient D. Figures 4, 5, and 6 show
the LW conditions. Note the comparability of the isodose
patterns at the CAX regardless of the medial wedge.
Table 2 (DW) shows the absolute values and per-
cent isodose for the hot spots in each axis cut. The mean
hot spot for breasts in the DW condition is 114 6 4.0,
112 6 2.74, and 115 6 4.0% for superior, CAX, and
inferior plans. Three of ten patients with dual wedges
showed inhomogeneity greater than 20%, patient F infe-
riorally and patients B and J superiorally. The unaccept-
able values were in off-axis cuts for these patients treated
with dual wedges; therefore, plan selection based on
homogeneity at the CAX would not have been represen-
tative of superior and inferior plans. Patients B, F, and J
also had chest wall separation values of 25, 24.5, and 21
cm, respectively.
Single wedge hot spots in the central, upper, and
lower planes with the lateral wedge only are seen in
Table 3. Inhomogeneities were most common in the
superior plane (B 5 123%, F 5 126%, and J 5 126%),
and inferiorally (C 5 122%, F 5 130%), and of
greater magnitude than at the CA (F 5 126%, J 5
122%). A summary table comparing mean hot spot
percentages of the prescribed doses can be seen in
Table 4.

Fig. 3. Inferior image of patient D using dual wedges, 1 16 medially and 1 30 laterally.
92 Medical Dosimetry
Fig. 4. Superior image of patient D using a single lateral wedge, 1 30°.
DISCUSSION
In this study, representative isodoses and hot spots
were examined throughout the breast to determine if
patients could achieve acceptable distributions using
only the lateral wedge and whether or not this was
reflected on examination of the CAX views alone. Six of
ten consecutive patients were able to show comparable
CA isodose distributions with the superior and inferior
distribution patterns remaining comparable without a
medial wedge as evaluated by a three slice pseudo-3D
algorithm. The isodose patterns in these patients were
similar at the CAX and the four patients requiring a
medial wedge for homogeneity were not comparable at
the CAX. Two of the patients showing comparable pat-
terns had hot spots .20%, one at the CAX (J) and one in
the inferior axis (C). It is not known if these are clinically
relevant volumes.
In the early 1980’s Frass, Robertson, and Lichter at
the University of Michigan described three contributing
Fig. 5. CAX image of patient D using lateral wedge only.
Volume 23, Number 2, 1998
factors to opposite breast dose (OBD) based on patient
and phantom data using the NCI (h/3) prescription point.
The OBD is comprised of (1) photon scatter from the
wedge which increases OBD at depth in the presence of
a medial wedge, (2) increased dose per monitor unit to
compensate for wedge attenuation, increasing the OBD
with the thin edge of the beam pointing toward the
opposite breast allowing increasing transmission com-
pared to the central ray, and (3) non-alignment of deep
beam edges increasing OBD. OBD of 50 –200 cGy was
measured for a typical 50 Gy treatment course using 200
cGy fractions. The authors demonstrated that with lung
density corrections the need for wedges was reduced and
recommended optimization with the lateral wedge only
and adding the medial wedge sparingly.
In the mid 1980’s Webb et al. examined isodose
patterns from CT scans at mid-plane and 1 cm inside the
upper- and lower-field borders on 30 patients. Dosimetry
was calculated first with dual wedges, then dual wedges
 Homogeneity of breast dose distribution ● C. IKNER et al. 93

Fig. 6. Inferior image of patient D using a single lateral wedge.

and lung corrections, and then reoptimization altering
wedge angles and field weightings to reduce the dose
gradient to 5–10% in the medial plane. Despite the
reoptimization, inhomogeneity remained 20% in the bor-
der planes with no change in the magnitude of hot and
cold spots.8 In 6 of 10 patients this criteria was met.
In the late 1980’s and early 1990’s 3D planning
allowed defined volumes of over- and underdosed ar-
eas.9 –11,13 A study to determine the number of calcula-
tion planes necessary for 3D determination of homoge-
neity in order to compare pseudo-3D with 3D was done;
pseudo-3D are those planning systems capable of multi-
level dosimetry planning using a planar algorithm. Three
breast sizes in three patients comprised the study pool.
Three slices 6 6 cm from the central axis were sufficient
for treatment planning and the pseudo-3D and full CT
agreed well for small and medium breasted patients,
chest wall separations of 18 and 21 cm, respectively. In
large breasted patients (chest wall separation 5 28) 3D
compensators were required for homogeneity. No at-
tempt was made to evaluate the contribution of single vs.
dual wedge homogeneity.
In a study of 100 sequential patients Das et al. used
wedging to “minimize the size and magnitude of hot
spots” and showed that hot spots for medium breasts and
large breasts were 108.5 6 2.7 and 112.6 6 3.1% of the
prescribed dose, respectively. There was no description
of the wedging used. It was concluded that the hot spot
magnitude was related to chest wall separation, the only
correlate to inhomogeneity when field size, breast size
and breast height were considered.13
The hot spots for medium and large breasts after
optimization in the Das study compared to the current
dual wedged data of 112.8 6 2.7 (mean 6 SD) and
116 6 4.4, the lateral wedge hot spot mean 6 SD. As
100% of the current study sample had medium or medi-
um/large breasts, more inhomogeneity would be ex-
pected off-axis, yet, only one patient had hot spots
greater than 20% in all three planes, and 6 other patients
were able to demonstrate isodose patterns that would
allow treatment without a medial wedge.
In the current study, patients with chest wall sepa-
rations $ 24 cm required dual wedges. This was in
agreement with Buchholz et al., who measured volumes
using a pseudo-3D algorithm and a single lateral wedge.
Reported isodose volumes $ 115% occurred only in
patients with chest wall separations of $ 27 cm. Isodose

Table 2. Hot spot with dual wedges Table 3. Hot spot with single lateral wedge

Patient Superior Central Axis Inferior Patient Superior Central Axis Inferior

isodose: cGy/% cGy/% cGy/% isodose: cGy/% cGy/% cGy/%

A 221/111 220/110 227/114 A 234/117 233/116 235/118

B 243/122 228/114 233/116 B 246/123 235/118 239/120
C 218/109 224/112 232/116 C 238/119 238/119 245/122
D 230/115 222/111 227/114 D 229/114 221/110 226/113
E 235/118 224/112 240/120 E 234/117 225/112 237/118
F 237/118 235/118 246/123 F 252/126 251/126 261/130
G 222/111 223/112 235/118 G 234/117 231/116 241/120
H 228/114 220/110 228/114 H 234/117 227/113 236/118
I 230/115 224/112 222/111 I 237/118 233/116 224/112
J 224/122 234/117 220/110 J 253/126 232/122 238/119
94 Medical Dosimetry Volume 23, Number 2, 1998

Table 4. Summary Table: Hot spot as percent 4. Fraass, M.A.; Roberson, P.L.; Lichter, A.S. Dose to the contralat-
prescribed dose eral breast due to primary breast irradiation. Int. J. Radiat. Oncol.
Biol. Phys. 11:485– 497; 1985.
N Mean 6 SD Minimum Maximum 5. Kelly, C.A.; Wang, X.G.; Chu, J.C.; Hartsell, W.F. Dose to con-
tralateral breast: a comparison of four primary breast irradiation
DW Superior 10 114 3.98 109 122 techniques. Int. J. Radiat. Oncol. Biol. Phys. 34(3):727–732; 1996.
LW Superior 10 119 4.14 114 126 6. Kurtz, J.M.; Amalric, R.; Brandone, J.; Ayme, Y.; Spitalier, J.M.
DW Central 10 115 3.95 110 123 Contralateral breast cancer and other second malignancies in pa-
LW Central 10 119 4.14 114 126 tients treated by breast-conserving therapy with radiation. Int. J.
DW Inferior 10 116 4.39 110 126 Radiat. Oncol. Biol. Phys. 15:227–284; 1988.
LW Inferior 10 119 4.94 112 130 7. Harris, J.R.; Morrow, M. Treatment of early-stage breast cancer.
In: Harris, J.R.; Lippman, M.E.; Morrow, M.; Hellman, S., editors.
Diseases of the breast. Philadelphia: Lippincott-Raven; 1996:530.
8. Webb, S.; Leach, M.O.; Bentley, R.E.; Maureemootoo, K.; Kar-
volumes between 110 –115% occurred in 1–3% of nold, J.R.; Toms, M.A.; Gardiner, J.; Parton, D. Clinical dosimetry
for radiotherapy to the breast based on imaging with the prototype
breasts with chest wall separations , 25 cm.14 Royal Marsden Hospital CT simulator. Phys. Med. Biol. 32(2):835–
In patients with chest wall separations # 24 cm, hot 845; 1987.
spot inhomogeneities of , 20% can be expected; there- 9. Cheng, C.; Das, I.; Stea, B.; The effect of the number of computed
tomographic slices on dose distributions and evaluation of treat-
fore, sparing the medial wedge can occur in a majority of
ment planning systems for radiation therapy of intact breast. Int. J.
patients with little compromise in inhomogeneity when Radiat. Oncol. Biol. Phys. 30:183–195; 1994.
isodose patterns are compared on the central axis. To 10. Chin, L.M.; Cheng, C.W.; Siddon, R.L.; Rice, R.K.; Mijnheer,
determine the positive predictive value of CAX compar- B.J.; Harris, J.R. Three-dimensional photon dose distributions with
and without lung corrections for tangential breast intact treatments.
isons in defining breast dose homogeneity, an examina- Int. J. Radiat. Oncol. Biol. Phys. 17:1327–1335; 1989.
tion of pseudo-3D vs. 3D plans is a logical next step. 11. Solin, L.J.; Chu, J.C.H.; Sontag, M.R.; Brewster, L.; Cheng, E.;
Dopple, K.; Drzymala, R.E.; Hunt, M.; Kuske, R.; Manolis, J.M.;
REFERENCES McCormick, B.; Munzenrider, J.E. Three-dimensional photon treat-
ment planning of the intact breast. Int. J. Radiat. Oncol. Biol. Phys.
1. Boice, J.D.; Harvey, E.G.; Blettner, M.; Stovall, M.; Flannery, J.T. 21:193–203; 1991.
Cancer in the contralateral breast after radiotherapy for breast 12. ROCS Treatment Planning System Users Manual. Using the
cancer. N. Engl. J. Med. 326(12):781–5; 1992. ROCS density phantom. Revision 2.B. Carlsbad, CA: Radiation
2. Choy, A.; Barr, L.C.; Serpell, J.W.; Baum, M. Radiation-induced Oncology Computer Systems, Inc.; 1996 4-4.
sarcoma of the retained breast after conservative surgery and ra- 13. Das, I.; Cheng, C.; Fein, D.; Fowble, B. Patterns of dose variability
diotherapy for early breast cancer. Eur. J. Surg. Oncol. 19:376 – in radiation prescription of breast cancer. Radiother. Oncol. 44:83–
393; 1993. 89; 1997.
3. Early Breast Cancer Trialists’ Collaborative Group. Effects of 14. Buchholz, T.; Gurgoze, E.; Bice, W.; Prestidge, B. Dosimetric
radiotherapy and surgery in early breast cancer—an overview of the analysis of intact breast irradiation in off-axis planes. Int. J. Radiat.
randomized trials. N. Engl. J. Med. 333(22):1444 –1445; 1995. Oncol. Biol. Phys. 39:261–267; 1997.