Pilot Study Evaluating the Efficacy of Breast Displacement During CT Coronary
Angiography for Dose Reduction and Image Quality Improvement Using the
Chrysalis Breast Displacement System.
Charles M. Swaney, MD,
The Chrysalis breast displacement system was developed in response to the need to reduce breast
radiation in those CT examinations where the breasts are in the plane of imaging and the plane of
imaging is relatively limited. It was postulated that displacing the breasts cephalad as much as possible
would reduce breast dose, particularly in the periareolar and upper –outer quadrants. It was also
thought that breast displacement might improve image quality in patients with large breasts. This
report reveals the findings of the pilot study of patients evaluating the use of the Chrysalis device in a
community hospital.
Methods:
CT Coronary Angiography(CTCA) was performed in 10 patients who provided informed consents and
were randomly selected to have the breasts displaced with the Chrysalis breast displacement system or
to be imaged in the conventional manner with the breasts non-displaced. All CT exams were performed
on the Toshiba Aquilion 64-slice CT scanner and included calculation of the Calcium score and CTCA.
Radiation exposure was measured with placement of OSL dosimeters (Landauer, Inc.) on the breasts at
multiple locations in order to accurately measure the dose to different parts of the breasts. On the first
2 patients, one non-displaced and the other displaced, a total of 7 OSL dosimetry devices were placed
on the breasts bilaterally. The devices were placed on the skin using the standard of breast ultrasound
localizations using clock measures. Dosimeters were placed at: 1) 6:00 at the inferior margin of the
breast where it meets the chest wall, but on the breast, 2) 4:30 halfway between the nipples and the
chest wall, 3) 7:30 halfway between the nipples and the chest wall, 4) Immediately lateral to the
nipples, 5) 10:30 halfway between the nipples and the chest wall, 6) 1:30 halfway between the nipples
and the chest wall, 7) at the axillae, bilaterally. After the first 2 patients, the OSL device locations were
adjusted to include only 4 devices bilaterally. In the control group, the OSL devices were located at the
6:00, inferior-medial quadrants (4:30 right, 7:30 left) halfway between the nipples and the chest wall,
immediately lateral to the nipples, and at the upper-inner quadrants (1:30 right, 10:30 left) halfway
between the nipples and the chest wall. In the displaced group, the OSL devices were located at the
6:00, inferior lateral quadrants (7:30 right, 4:30 left) halfway between the nipples and the chest wall,
immediately lateral to the nipples, and at the upper-outer quadrants (10:30 right, 1:30 left) halfway
between the nipples and the chest wall. Devices were placed with the patient in the supine position.
Although there was a difference in the medial and lateral positions of the dosimeters between the
displaced and non-displaced groups, the cephalo-caudal positions between the two groups was
equivalent by design. Differences in the medial and lateral positioning in the two groups were intended
to insure that the breasts didn’t cover the dosimeters because of positioning issues in patients with
large breasts while maintaining the critical cephalo-caudal relationships between the two groups.
Following placement of the dosimeters, the imaging study was performed in the conventional manner in
the non-displaced group. In the intervention group, the principal investigator manually displaced the
breasts upwards and fitted the Chrysalis device, taking care not to disturb the dosimeters. Imaging was
then performed in the conventional manner and after completion of the examination the dosimeters
were removed and sent for measurement. The patients in the intervention group filled out a
questionnaire regarding their experience wearing the Chrysalis device.
Patient demographics included weight and height for calculation of BMI, age, and bra cup size.
Evaluation of the images included objective measures of standard deviation at defined locations on 2
mm reconstructed axial images in order to obtain the most homogeneous site for measurement at each
location. Where possible, patients were matched using BMI and bra cup size, applying the measures of
standard deviation. Over-all, the groups of patients were compared for BMI and bra cup size.
Results:
Table 1 compares the radiation doses in mrad at each location measured in each group of patients. The
median dose is also included because of a single patient in the intervention group where the breasts
were not displaced effectively because of technical reasons. The mean doses in the intervention group
are lower than the control group at all locations, especially cephalad in the periareolar and upper outer
quadrants of the breasts. The dose at the inferior margin of the breasts was comparable between the
two groups indicating that the potential overall exposure to the patients was similar.
Table 1:
Dose (mrad) from dosimeters at quadrants of the breasts
Location
Inferior Margin
Inferior Quadrant
Nipple/ Periareolar
Upper Quadrant
Non-displaced Group
Left
Right
Mean
Median
Mean
Median
11625
11493
9105
12467
12187
13576
10420
13396
9347
10314
11575
10337
13334
13783
10245
10994
11862
6541
4445
1033
11994
11203
7313
8629
1141
3895
560
1053
10238
9176
774
610
Displaced Group
Left
Right
Mean
Median
Mean
Median
Doses in the upper quadrants of the breasts in the control group were greater than the lower margin of
the breasts in 60% of measurements and none were greater in the intervention group. This shows that
breast exposure is relatively equivalent over the breast in the control patients, whereas displacement
with the device effectively decreases breast dose, particularly in the periareolar and upper quadrants
where the largest concentration of breast parenchyma is located.
Radiation dose reduction in patients wearing the Chrysalis breast displacement system compared to
controls ranged from a median reduction of 1.6% at the inferior margin of the left breast to a median
95.2% reduction in the upper quadrant of the left breast, see Table 2. In the upper parts of the breasts
at the periareolar and upper quadrants, the median dose reduction ranged from 87.8-95.2%.
Table 2:
Dose Reduction in Patients with Chrysalis Displacement Compared to NonDisplaced Group
Location
Left
Right
Inferior Margin
Mean
Median
Mean
Median
Inferior Quadrant
Nipple/ Periareolar
Upper Quadrant
+2%
-43.1%
-51.2%
-91.7%
-1.6%
-17.5%
-30.8%
-35.6%
-87.8%
-62.2%
-95.2%
-89.8%
-23.2%
-33.4%
-92.4%
-94.5%
In Table 3 the relationship between the doses within the body of the breasts (upper 3 locations) to the
inferior margin of the breasts, in the same patients, is compared between the those fitted with Chrysalis
displacement and the control group in order to give an internal “control” for each patient. The wide
range in the displaced patients is directly attributable to a single patient in whom the breasts were
ineffectively displaced, whereas measurements in the patients with effective displacement were much
more closely aligned. Comparisons between the inferior quadrants of the breasts relative to upper
locations in the intervention group yield median dose measurements of 7.6 and 9.5% in the periareolar
location and 6.0 and 4.7% in the upper quadrant, compared with 76.8 and 76.0% in the periareolar
location and 82.5 and 95.0% in the upper quadrant in the control group at these levels, thereby
quantifying the degree of dose reduction with Chrysalis displacement. Median dose reduction at the
inferior quadrants of the breasts is ranges from 39.1 and 10.4%.
Table 3:
Percentage Dose in Quadrants Compared to Inferior Margin of the Breasts
Location
Inferior Quadrant
Nipple/ Periareolar
Upper Quadrant
Non-displaced Group
Left
Right
Mean
Range
Median
Mean
Range
Median
98.9%
78.3%
107.2%
85.5-110.0%
85.5%
98.7%
61.9-91.4%
76.7%
76.0%
37.1-128.8%
95.0%
76.1%
70.9-112.4%
103.4%
59.9-97.4%
76.8%
32.1-112.3%
82.5%
Displaced Group
Left
Right
Mean
Range
Median
Mean
Range
Median
55.1%
37.5%
8.7%
14.1-91.9%
60.9%
77.0%
4.2-130.0%
9.5%
34.8%
3.0-23.4%
4.7%
8.8%
34.1-121.5%
89.6%
4.8-109.7%
7.6%
3.5-16.1%
6.0%
Matching subjects proved difficult because there was a significant difference in body sizes in the
two groups, with controls having an average BMI of 29 and the intervention group having an average
BMI of 43. However, matching subjects 8 and 7 (Table 4), was possible and objective image quality was
compared. The mean ROI of the measurements in standardized locations was 31.56 in the control
subject and 19.47 in the displaced subject, showing an improvement in image quality of 38.3%. Image
quality improvement is expected to be greatest in patients with large breasts. Summation of all the ROI
values in both groups yielded almost exactly the same value, which is significant given that the average
BMI of the intervention group was much higher than the control group.
Conclusions:
Chrysalis displacement has been shown to be effective in the majority of patients. With median
radiation dose reductions of between 88% and 95% in the periareolar and upper quadrants of the
breasts, this improvement in radiation exposure is clinically significant and has the further benefit of
improving image quality. The benefits gained by using the Chrysalis are additive to other dose reduction
strategies at each institution and from each CT manufacturer because the breasts are displaced out of
the imaging plane.
Table 4: Objective Image Quality, Matched Subjects
BMI
Bra cup size
SD Pulmonary Artery, HU
SD Left Ventricle, HU
SD Right Ventricle, HU
SD Aorta, HU
SD Left Atrium, HU
SD Fat at Chest Wall, HU
SD Left Ventric. Wall, HU
SD Fat Adj. to Aorta, HU
SD Fat Adj. to Pulm. Artery,
HU
Mean SD ROI, HU
Percentage Improvement
Subject with Chrysalis
Subject 8, with
Chrysalis
43
DDD
24.8906
12.948
12.8628
22.0445
19.9608
16.094
13.0595
27.6986
25.6866
Subject 7,
non- displaced
42
D
24.0
41.3497
30.5553
24.6378
26.676
38.8654
30.9426
35.2194
32.1756
19.47
31.56
38.3%