.l&Yka/ Pnnted
Dosfmerr~. Vol. 17. pp. l-9 I” the U S.A 411 nghts reserved
Copyright
0
1992 American
0739421 l/92 Association of Medical
$5.00 + .W Doslmetnsts
TREATMENT PLANNING FOR THE BREAST PATIENT: WITH OR WITHOUT LUNG CORRECTION? LOUISE A. FARRADAY, DIP. (TH. RAD.) C.M.D. and GARY T. DOSWELL, Ontario Cancer Foundation, Toronto-Bayview Regional Cancer Centre, 2075 Bayview Avenue, Toronto, Ontario, Canada, M4N 3M5
M.A.Sc.
Abstract-Patients who have undergone a surgical lumpectomy for breast cancer are routinely referred to our clinic for a course of radiation therapy. This treatment consists of a pair of obliquely opposing radiation fields that encompass the breast and underlying structures. In order to treat the chest wall, a portion of lung must be included in this target volume, as breast tumours can invade the underlying chest wall due to direct invasion, blood invasion, or lymphatic permeation. At our clinic, we have always included this volume as normal tissue (with normal tissue density); thus, an inhomogeneity correction has never been generated on any breast plan. The aim of our study was to accurately measure the volume of lung within the treatment fields at time of simulation, using the simulator films and patient parameters, and to determine the point at which this volume of lung significantly altered the computer plan. This would necessitate the generation of a new distribution in order to produce a homogeneous dose to within +5%. A table was constructed to identify those patients who had excessive lung volumes. At time of simulation it would be possible, using the table, to determine the type of treatment plan required for each patient. Those patients demonstrating an excessive lung volume would require a more extensive planning procedure utilizing a (computerized tomography). An inhomogeneity correction could then be applied to the computerized distribution. Key Words:
Surgical lumpectomy,Normal tissue density, Inhomogeneity,Breast.
INTRODUCI’ION
Patients who have undergone a surgical lumpectomy for breast cancer are routinely referred to our clinic for a course of radiation therapy to the breast and surrounding areas. This treatment consists of a pair of obliquely opposing radiation fields that encompass the breast and underlying structures. In order to treat the chest wall, a portion of lung must be included in this target volume. This volume of lung is normally included, as breast tumours can invade the underlying chest wall due to direct invasion, blood invasion, or lymphatic permeation. At our clinic, routine planning of breast tangents have always included this lung volume as normal tissue (with normal tissue density), thus an inhomogeneity correction has never been generated on any breast plan. The aim of this study was to accurately measure the volume of lung within the treatment fields at the time of simulation using the simulator films and patient parameters and to determine the point at which this volume of lung significantly altered the computer plan. This would necessitate a new distribution to be generated in order to produce a homogeneous dose to within *Sk. A table was constructed in order to identify those patients who had an excessive lung volume (estimated from simulation). In our opinion, these pa* Theratronics is a trademark of Theratronics International Limited, Canada. * MEVATRON is a registered trademark of Siemens Medical Systems, Inc.
tients needed a more extensive planning procedure, perhaps utilizing CT (computerised tomography). This would enable an inhomogeneity correction to be applied to the computer plan. (Due to patient load and CT availability it was imperative that we identify only the patients who have an unacceptable amount of lung within the treatment volume.) At time of simulation it would be possible, using the table, to determine the type of treatment plan required for each patient. TREATMENT
TECHNIQUES
Overview For radiation treatment of the breast, the treatment unit chosen depends directly on the size of the target volume, i.e., the medial and lateral separation of the breast volume at the posterior field border. A small target volume with the total separation of 18 cm. or less is treated on our TheratronicTM* 780 cobalt unit. Larger breast volumes are treated on our medium range MevatrorPt or Theratronics 6MV units, and in rare cases, patients are treated on our Theratronics 25MV unit if the total separation is over 25 cm. T780 cobalt technique This technique takes advantage of the “head swivel” capability of this unit. The patient is positioned supine, resting on a sloped back support so that
Volume 17, Number
Medical Dosimetry
(
Fig. 1. Treatment
1,
1992
medial and lateral setup points
of breast tangents utilizing the “head swivel” capability of our cobalt unit.
the patient’s sternum is approximately parallel to the treatment couch. The contralateral arm is resting beside the patient and the wrist of the ipsilateral arm rests on the forehead. On each treatment day, a threefield laser setup ensures that the patient’s positioning is accurate. The posterior borders of the medial and lateral fields are the setup points for treatment. The center of the field is positioned on the posterior aspect of the medial field at 80 cm SSD, the correct gantry angle is selected, and the field is “swivelled” into position by rotating the head of the treatment unit by two or three degrees, depending on the divergence of the field (Fig. 1). Head swivel on this unit is about an axis, which includes the source. For the lateral field, this process is repeated by rotating the gantry through 180” and positioning the center of the beam on the posterior edge of the lateral border at 80
cm SSD. Again, the head of the treatment unit is swivelled until the posterior edge of the light field matches the border as marked on the patient’s skin (2” or 3”). We have found this a quick and accurate treatment technique, as the field borders are not on mobile breast tissue. Also, the posterior edges of the beams are colinear, thus reducing the dose penumbra in the region of the lung. M6/T6 breast technique This is an isocentric technique-the patient is again positioned on a sloped back support ensuring the sternum is parallel to the treatment couch (Fig. 2). The patient’s arm position differs such that the ipsilatera1 arm is supported at 90” to the body by an arm rest. Positioning is again verified by a three-point laser setup. To locate the isocenter, a midstemum point is
T6 isocentric technique
couch translation
Fig. 2. Treatment
of breast tangents utilizing the 6 MV SAD technique.
Treatment planning for the breast patient 0 L. A. FARRADAY and G. T.DOSWELL
T25SSD technique
Fig. 3. Treatment
to breast bridge
of breast tangents utilizing the 25 MV SSD technique using a breast bridge.
selected (as this is on stable tissue) and the depth of isocenter is set. The treatment couch is then translated laterally by a predetermined distance. (The isocenter is determined from tables that calculate the isocenter position, depending upon gantry angle, patient separation, and initial collimator setting). When the isocenter has been determined, the gantry is rotated and the medial field is verified on the patient’s skin. The lateral field is verified by rotating the gantry 180” and again viewing the field on the patient’s skin before initiating treatment. T25 Breast Technique For this technique we position the patient in the same way as in the cobalt technique. This is a nonisocentric setup at 100 cm SSD (Fig. 3). Again, a threepoint laser setup ensures that the patient’s positioning is accurate. The gantry is rotated and the medial field is positioned at 100 cm SSD to a breast bridge set at the patient’s breast separation. After the medial field has been treated, the lateral field is positioned and treated in the same way as the medial tangent. To treat the breast tissue in its entirety, 1 to 2 cm bolus is required due to the skin-sparing effect of the unit (D,,, is at 4 cm). This bolus is placed over the breast before initiating treatment. It is noted that the patient’s skin reaction will be intense and special skin care is given.
their arms supported by an arm support could not have moved through the aperture of the CT scanner. In those instances the arm position was altered so that the wrist of the ipsilateral arm rested on the forehead, as with the cobalt and T25 treatment techniques. The medial and lateral field borders were marked with aluminium markers, which can be easily identified on the scan. Each scan was 1 cm in thickness and taken at 2 cm intervals along the entire length of the treatment volume. The CT data recorded on the tapes was then transferred to our computer planning system. When we reviewed the scans, a number of patient parameters were readily determined:
STUDY PROCEDURE
Patients who were having routine breast fields simulated were chosen at random to have a CT scan of the breast and chest wall area (Fig. 4). During the CT session, the patients were positioned in their treatment position with one exception, patients who had
Fig. 4. CT slice through the central plane of a breast tangent treatment. Arrow indicates aluminum markers to indicate the posterior field borders.
2.8
2.4
3.5
2.3
2.1
2.0
2
3
4
5
6
7
4.5
5.6
2.7
6.2
3.8
5.3
4.3
B (4
,$z
$
+7 +40 -7 +20 +27
1.6 2.1 1.4 1.8 1.9
(measured from the CT s’,ans).
AVERAGE 220%
-7
1.4
AVERAGE 1.6
-27
1.1
cwMtio=1.5)
ERROR 96 (ReIative to r(~)
14.1
18.5
9.8
34.0
14.3
23.3
17.6
2
compared to calculated values from Theraplan.
iTAB = A manual calculation to obtain the lung area on the central slice. 2 Lung area = Theraplan calculated lung areas (on the central slice).
Chest wall ratio = zzz
) CHEST FYALL RATIO
Where: A = Deuth of lung. as measured from the Simulator film B = Patient’s separation - total chest wall thickness = SEPN - 2.5Y
2.6
A (4
1
PATIENT NUMBER
lung areas 2 (
TAB
+9 +ll +24 -3 +6 +37
21.4 12.9 27.5 10.1 17.4 10.3
AYERQGE + 15%
-17
21.3
(cm*)
ERROR % (MANUAL VS. CT) (CU
LUNG AREA
Table 1. Depth in lung as measured from the simulator films; comparison between the medial and lateral chest wall thicknesses as measured from the CT films (Chest Wall Ratio); manually calculated
P
Treatment planning for the breast patient 0 L. A. FARRADAY and G. T. DOSWELL
Fig. 5. Simulator film taken (gantry at 90”) to show the chest wall thickness in the vertical plane. Arrow indicates a wire on the posterior border of the medial tangent.
1. The medial and lateral chest wall thicknesses on
each patient were measured. 2. By utilizing the CT diagnostic package on our Theratronics planning system, the lung area within the treatment fields (on the central slice) was determined. 3. The average relative electron density of lung volume within the treatment fields was determined.
(Fig. 5). We selected a patient having a routine simulation of breast tangents and placed a wire on the post border of the medial field. Under fluoroscopy (with the simulator head at 90’7, the distance from the wire to the edge of the lung was measured. The following formula was used to obtain the actual chest wall thickness at the oblique angle of the medial field (Fig. 6): Y=-
The initial step in the study was to record the chest wall thickness of each patient. The actual thickness was measured utilizing the CT scans and it was found that the chest wall thickness varied greatly from patient to patient. We were, however, able to determine that on average, for a range of patients, the lateral chest wall was one and one-half times thicker than the medial chest wall (Table 1). We then attempted to see whether or not we would be able to measure the chest wall thickness at the time of initial simulation using fluoroscopy only
1
where
x
cos0
X = The vertical measurement from the skin edge to the limit of the chest wall. 0 = The angle between the tangent baseline and the vertical.
By multiplying the medial thickness by 1; we obtained an estimate of the lateral chest wall thickness. To verify these simulator measurements the patient was sent for a CT scan. Results showed that the estimated value of 3.3 cm, calculated on simulation,
wire on posterior border of medial tangent,
thickness thickness
Fig. 6. Geometry of the actual chest wall thickness at the oblique angle of the medial field.
Medical Dosimetry
Fig. 7. Mathematical
equation to calculate the area of lung within the target volume on the central slice.
compared favourably with the more accurate CT value of 3.6 cm for the medial tangent chest wall thickness. To determine manually the amount of lung within the treatment fields at the level of the central axis, we used only the simulator film and patient separation, and applied the following formula (Fig. 7): TAB Lung Area = cm2 2
where
Volume 17, Number 1, 1992
A = The depth of lung from the chest wall to the field edge at the central axis. (Measured directly from the simulator film in cm.) B= Patient’s separation - the total chest wall thickness (2.5 Y) 2 where 2.5 Y = medial chest wall thickness + lateral chest wall thickness (measured in the transverse plane using calipers).
This formula calculates the area of a hemi-ellipse that approximates the area of the lung within the target volume at the central axis of the beam (This has invariably been the area of greatest lung volume). A comparison was made between the estimated lung areas and the Theraplan*-calculated lung areas. The manually calculated values compared favourably with the Theraplan values (Table 1). With the help of an oncologist specializing in the treatment of breast carcinomas, a set of parameters was established to determine at what point the amount of lung within the treatment volume produced an unacceptable distribution. * Theraplan is a trademark ofTheratronics International, Limited, Canada.
These parameters were as follows: The maximum dose value on the plan was not taken as the “point maximum” but an isodose measuring an estimated 2 square cm. (Fig. 8). The covering isodose was a point measured f anteriorly from the posterior field border at the mid-separation of the contour’ (Fig. 8). If the range between these two values was 10% or less, the plan was acceptable. Anything greater than this value was deemed unacceptable. Approximately 50 distributions were planned using our Theraplan software on each treatment technique for small and large separations. For the cobalt and 6MV distributions the depth of the lung within the fields was increased by 5 mm until a measurement of 4 cm was reached on the final plan (as measured anteriorly at the centre of the lung volume). For the T25 MV distributions, up to 8 cm of lung was entered. The isodose distributions were analyzed and tabulated according to acceptability values of the plans for each treatment technique (Table 2). We then produced a series of plans with the same depth of lung but with an increase in the width of lung to see whether this had an even greater effect on the distribution (the effect was small but noticeable). In all cases, a bulk dSAR inhomogeneity calculation technique was used, with a lung density of 0.25 (the average value determined from our CT review). SUMMARY The aim of this study was to identify breast patients who had an unacceptable amount of lung within their treatment volume. In calculating this lung volume it was imperative that this extra step in the simulation be as simple as possible, so as not to
4.5, 6.2
~8.0, >8.0
25
12.5
2.0, 4.0
44
> 100
> 0.0
10.8
7.7
1.5, 3.3 1.5, 4.6
5.5
11.4
2.0, 3.6 1.0, 3.5
10.4
Cl’ LUNG AREA (cm*)
1.5, 4.4
LUNG DEPTHS A,B (cm)
>o.o
30
correction CT).
STANDARD
STANDARD
STANDARD OR ELONGATED’
STANDARD
ELONGATED
STANDARD
ELONGATED
STANDARD
ELONGATED
LUNG SHAPE’
POINTAT WH.ICH LUNG CORRECTION IS REQUIRED ie., CENTRAL SUCE DKlRlBuTHIN BECOMES BCMDERLINE UNACCElT”~
those patients requiring an inhomogeneity distribution (either manually or utilizing
25
21
20
16
Notes: a “Borderline Unacceptability”: an isodose area of 2 cm’ or more on the central slice distribution is ~8% higher than the prescription isodose on the ‘no lung’ distribution. * Lung Contour Shape: ELONGATED A/B 5 0.35 for A I 2 cm. STANDARD: 0.36 I A/B 5 0.60 for A 5 2 cm. A6MV central slice distributions with separation = 25 cm. No distribution is acceptable, with orwithout lung.
25MV
6MV
*co
BEAM ENERGY
BREAST SEPARATION (an)
Table 2. Table identifying to their computerized
Volume 17, Number 1, 1992
Medical Dosimetry anterior
area maximum
point dose maximum 110%
(2 cm>) 115%
2oow
2O”W .
0;
Fig.
8.
Breast
distribution showing the point maximum, volume maximum, prescription point (2/3 anteriorly from the posterior field border), relevant isodoses and lung volume.
lengthen the planning procedure. We designed an accurate but workable selection process utilizing patient parameters and data in Table 2. The dosimetrist can quickly decide at the time of simulation whether the lung area calculated would directly influence the distribution and whether the patient required an inhomogeneity correction applied to the distribution. Upon discussion with the oncologist, it would be decided whether the lung volume should be manually corrected for in the usual three-contour distribution, or whether the patient would require a planning CT. It is estimated that approximately 25% to 30% of breast patients would require this extra step in their planning procedure. CONCLUSIONS Our data indicate that the point at which an isodose distribution becomes unacceptable depends on many factors such as beam energy, breast separation, lung depth within the treatment volume and, to some extent, the shape of the lung volume. A summary of Table 2 follows: 1. For cobalt treatments the smaller the breast separation, the greater the amount of lung allowable within the target volume, i.e., 1.5 to 2 cm in depth as measured on the central axis. For larger separations a smaller amount of lung is allowable to produce the same effect, due primarily to the relatively low beam energy, i.e., 1 cm to 1.5 cm depending on the amount of lung volume elongation. 2. For 6MV treatments the same effect occurs. For an average breast separation of 2 1 cm, 1.5 cm to 2 cm of lung is allowable, depending on the total lung volume (on the central axis) or amount of lung elongation.
For separations 25 cm or greater, due primarily to the beam energy of approximately 1.8 MeV, any amount of lung within the treatment volume will have a negative effect on the distribution, i.e., it will cause the maximum to increase, yet the isodose prescription point will remain the same. 3. For 25 MV treatments the beam energy of approximately 8 MeV has a positive effect on the distribution. It is shown that for separations up to 25 cm, a depth in lung greater than 8 cm is needed to produce an unacceptable distribution (this is a highly unlikely situation). For separations of 30 cm or greater allowable lung depth reduces to 4.5 cm. We can generally say that all T25 distributions are acceptable regardless of lung volume. N.B. All patient separations for 25 MV must include a 1 cm to 2 cm layer of bolus, needed to counteract the skin sparing effect (d,,, = 4 cm).
Patient example Patient parameters: 1. Beam energy = 6 MV, 2. Breast separation at the central axis = 21 cm, 3. Depth of lung (measured from the simulator film on the central axis) = 2 cm, 4. Medial chest wall thickness in the vertical plane = 3 cm, 5. Gantry angle (measured from the vertical) = 43”, therefore, oblique medial chest wall thickness: y=x y=3
cos 0 cos 43”
Treatment planning for the breast patient 0 L. A. FARRADAY
Medial chest wall thickness at the oblique angle of the field Y = 4.1 cm. Lateral chest wall thickness (1: r) = 6.2 cm. Lung area (central slice): TAB = -_-cm2 2 = 7r2.0 x 5.38 2 = 16.9 cm2
Since $ = 0.37, the shape of the lung volume is standard.
and G. T. DOSWELL
9
Table 2 recommends that a lung correction be calculated on this distribution. (either by utilizing CT or by manually adding an internal lung contour).
would like to thank Dr. Robert MacKenzie of our institute for his helpful suggestions in the preparation of this manuscript.
Acknowledgment--We
REFERENCE 1. National Surgical Adjuvant Project for Breast and Bowel Project. NSAPB Protocol No. B-21, Sec. I 1.1.5, p. 17. NSAPB Operations Office, Pittsburgh, PA.