| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 All authors: Department of Radiology, University of Michigan Health System, 1500 E. Medical Center Dr., Ann Arbor, MI 48109-0326.
Received September 27, 1999;
accepted after revision February 8, 2000.
Address correspondence to M. A. Helvie.
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. Three hundred consecutive routine mammograms were reviewed. Patients were categorized according to body mass index. Compression thickness, compressive force, kilovoltage, and milliampere-seconds were recorded. Geometric unsharpness and contrast degradation were calculated for each body mass index category.
RESULTS. Body mass index categories were lean (3%), normal (36%), overweight (36%), and obese (25%). Body mass index was directly correlated with compressed thickness. In the mediolateral oblique view, the mean thickness of the obese category exceeded normal thickness by 18 mm (p < 0.01), corresponding to a 32% increase in geometric unsharpness. Mean obese thickness exceeded lean thickness by 33 mm (p < 0.01), corresponding to a 79% increase in unsharpness. Similar trends were observed for the craniocaudal view. In the mediolateral oblique projection, there was an increase of 1.0 kVp (p < 0.01) for obese compared with normal and 1.7 kVp (p < 0.01) between lean and obese, corresponding, respectively, to a 16% and a 25% decrease in image contrast because of scatter and kilovoltage changes. Milliampere-seconds increased by 47% on the mediolateral oblique images in the obese category compared with normal body mass index.
CONCLUSION. An increased body mass index was associated with greater compressed breast thickness, resulting in increased geometric unsharpness, decreased image contrast, and greater potential for motion unsharpness.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Recent studies suggest mammographically screened obese women have larger and higher stage tumors than normal-weight women [12, 13]. Bailey et al. [12] showed that mean tumor diameter significantly increased from 11.7 to 13.1 to 17.1 mm when comparing normal, overweight, and obese women undergoing annual mammographic screening. Likewise, Hunt and Sickles [13] found larger tumors and higher stage tumors in screened obese women than in their normal-weight counterparts. The mean tumor diameter of women exceeding 40% of their ideal body weight was 19 mm compared with 15 mm for women of ideal body weight. The cause of these differences in size at screening detection has not been established. It may relate to difficulties in mammographic detection, or it may reflect biologic differences.
We explored possible adverse effects of increased body weight on mammogram quality as a potential cause for the observed differences of tumor size in screening-detected cancer. In particular, we studied the hypothesis that the thickness of compressed breasts during mammography in overweight and obese women exceeds that of normal-weight women, causing image degradation. Increased compressed breast thickness contributes to image degradation through decreased geometric sharpness, decreased image contrast, and motion unsharpness [14, 15]. Such degradation of image quality could limit mammographic detection of early breast cancer and contribute to the worsened outcome for overweight and obese women with breast cancer. Additionally, we assessed measurements of mammographic positioning [16] to determine if the mammograms of overweight and obese women are limited by difficulties in breast positioning.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
As part of their examination, all patients completed a questionnaire that obtained information on height and weight. Body mass index was calculated for each patient using Quetelet's index (weight [kg] / height [m]2), a method for estimating total body fat that accounts for height [5, 17, 18]. Each patient was then assigned a body habitus classification on the basis of her body mass index, similar to those described in standards recently published by the National Institutes of Health [1]. Patients with a body mass index of less than 20 were categorized as lean; with a body mass index of 20-24.9, as normal weight; with a body mass index of 25-29.9, as overweight; and with a body mass index greater than 30, as obese.
Standard two-view mammographic examinations were performed for each patient on two identical mammography units (DMR; General Electric Medical Systems, Milwaukee, WI). The quantitative digital readout of compression thickness, applied compressive force, kilovoltage, milliampere-seconds, target, and filter (molybdenum or rhodium) were recorded for each examination. Each machine functioned in an automatic mode. Kilovoltage, filter, and target source were chosen by machine algorithms. Before beginning the study, measured compression thickness from the mammography systems was compared with the actual thickness of acrylic breast phantoms compressed at 8 dkN from 2-8 cm at 1 cm increments. Breast thickness readings were standardized to the measured norms, which did not vary more than 2 mm from the actual thicknesses of the phantoms. The examination of each breast was considered an independent event. Craniocaudal and mediolateral oblique images were reviewed from each examination. Film size was noted.
Mammographic positioning was assessed on the basis of previously described methods [16]. Posterior nipple line distance was measured as a perpendicular line from the nipple to the pectoral muscle or its expected position on the mediolateral oblique images, and as a perpendicular line from the nipple to the posterior film edge or pectoral muscle on craniocaudal images. On the mediolateral oblique images, we assessed the presence of the inferior extent of the pectoral muscle below the posterior nipple line and the presence of the inframammary fold. On craniocaudal images, the posterior nipple line distance was measured and compared with the posterior nipple line distance in the mediolateral oblique image. A posterior nipple line distance in the craniocaudal image no shorter than 1 cm less than the posterior nipple line distance in the mediolateral oblique image was considered adequate. The inclusion of all fibroglandular breast tissue at the medial aspect of the image was assessed on the craniocaudal image. Breast density was evaluated and categorized into one of four categories, category 1 indicating an almost entirely fatty breast, and category 4 indicating an extremely dense breast based on the Breast Imaging and Reporting Data System criteria [19].
All data were entered into a database program (Excel; Microsoft, Redmond, WA). Statistical analysis was done with the Sigma Plot statistical package for PCs (Statistical Package for the Social Sciences, Chicago, IL). A standard correlation was calculated between numeric data points of body mass index categories and each of the following: mean compression, thickness, kilovoltage, milliampere-seconds, and compressed force.
Geometric unsharpness (f x [d2 / d1], where f is the focal spot size, d2 is the object-to-film distance, and d1 is the source-to-object distance) was determined for surface and mid breast lesions in both projections. Image contrast degradation caused by scatter was estimated using the scatter data in mammography estimated by Monte Carlo methods [20]. Mean glandular dose in milligray per exposure was calculated on the basis of dosage tables [21]; and the mean kilovoltage, milliampere-seconds, and compression thicknesses, assuming a 50:50 fatty to glandular tissue ratio and a molybdenum—molybdenum target—filter, were also calculated.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Compression thickness correlated directly with body mass index for both the mediolateral oblique and craniocaudal images (Table 1). On the mediolateral oblique image, the difference in mean compression thickness between normal and obese women was 18 mm (p < 0.01), corresponding to a calculated increase in geometric unsharpness of 32% for a surface lesion and 24% for a lesion in the mid breast. A 33-mm difference (p < 0.01) in mean thickness between lean and obese women corresponded to an increase in unsharpness of 79% and 57% for surface and mid breast lesions, respectively. In the craniocaudal projection, the difference in mean compression thickness between normal-weight and obese women was 10 mm (p < 0.01), corresponding to an increase in geometric unsharpness of 19% and 14% for the two locations. The difference in thickness between lean and obese women was 24 mm (p < 0.01), which corresponded to an increase in geometric unsharpness of 56% and 41%.
|
Kilovoltage potentials directly correlated with body mass index (Table 2). In the mediolateral oblique projection, there was an increase of 1.0 (p < 0.01) in the mean kilovoltage used for obese women compared with that used for normal-weight counter-parts. Mean kilovoltage increased by 1.7 kVp between lean and obese patients (p < 0.01). These differences corresponded to 16% and 25% calculated decreases in image contrast because of scatter and kilovoltage changes, respectively, when molybdenum spectrum was assumed for all cases. A small (0.3 kVp) but statistically significant (p < 0.04) increase was also seen between the normal-weight and overweight participants on the craniocaudal images. The mean kilovoltage did not increase as much as would be expected because the mammography system switched to a rhodium filter to improve penetration when body mass indexes increased. In the mediolateral oblique projection, 6% of normal-weight and 57% of obese patients used rhodium filtration (p < 0.01).
|
The milliampere-seconds used in both mediolateral oblique and craniocaudal images also correlated directly with the body mass index (Table 2). The mean milliampere-seconds for the mammograms of obese women was 47% higher in the mediolateral oblique projection and 38% higher in the craniocaudal projection than for the mammograms of normal-weight subjects. Compared with lean women, the mean milliampere-seconds for obese women increased 220% in the mediolateral oblique projection and 185% in the craniocaudal projection.
Compression force used for the mammogram increased with increasing body mass index and increasing compressed thickness. Obese women tolerated, on average, a 2.7 dkN (20%) increase in force on the breast during examination in the mediolateral oblique projection and a 1.9 dkN (15%) increase in force in the craniocaudal projection when compared with normal-weight subjects (Table 1).
Positioning assessments are presented in Table 3. No adverse relationship was noted between obesity and the total number of image quality parameters fulfilled. The overweight and obese patients were more likely to fulfill image quality parameters (posterior nipple line distance parameters and inclusion of medial glandular tissue) assessed on craniocaudal images and were less likely to fulfill criteria on mediolateral oblique images than normal-weight women. Lean women fulfilled the fewest image quality positioning parameters. Obese women accounted for most of the large film studies (24 x 30 cm) performed. Forty-seven percent of the mammograms of obese women were imaged on large films, accounting for 81% of all large size films used in the study.
|
Breast density was inversely proportional to body mass index. The mean numeric categorization for breast density in lean women was 3.2, in normal-weight women was 2.9, in overweight women was 2.3, and in obese women was 2.2. The differences between normal and overweight and between normal and obese were significant (p < 0.05).
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Our study was performed to determine if image quality differences between the obese and the normal-weight woman might contribute to the observed differences in size of screening-detected tumors. In particular, we tested the hypothesis that compressed breast thickness increased with body mass index, resulting in degradation of image quality, which could contribute to the larger tumors found at screening in obese patients. We found a significant direct relationship between body mass index and compressed breast thickness. The increase in compression thickness causes degradation of quantifiable image quality factors in overweight and obese women. In particular, increased geometric unsharpness, decreased image contrast, and increased motion unsharpness all contribute to this degradation. These factors worsen image quality, which may cause interpretive difficulties and contribute to the larger tumors found in overweight and obese women.
The examination of image quality in over-weight and obese women is not an insignificant issue for the mammographer. Most American women older than 40 years are considered over-weight or obese by National Institutes of Health criteria [1]. Since 1980, the percentage of overweight and obese women in the United States has increased by 10%. Currently, 64% of women 50-59 years old (mean patient age in our study was 53 years) are classified as overweight or obese, similar to the 60% of our total population that is classified as overweight or obese. Hence, the mammographic image quality findings in overweight and obese women represent those of the typical mammography patient examined by a breast imager in the United States.
A key component of mammographic image quality is adequate compression thickness [14, 16, 22, 23]. Increased geometric unsharpness and decreased image contrast have been shown to occur in the imaging of thicker breasts [14, 15, 23,24,25]. Additionally, thicker compressed breasts have increased overlap of structures, decreased uniformity of breast tissue displayed, and increased beam hardening [22, 23]. The obese women in our study averaged a compression thickness that was 18 mm thicker than that of their normal-weight counterparts on the mediolateral oblique images, causing an increase in calculated geometric unsharpness of 32% for a surface lesion. Geometric unsharpness increased 79% between lean and obese women. On the craniocaudal images, the differences were less but still accounted for an increase in geometric unsharpness of 24%. This geometric unsharpness is caused by an increasing size of penumbrablur at the edge of an imaged object as it moves farther away from the X-ray film. With increasing penumbra size is a concomitant decrease in image contrast. This unsharpness has the potential to "blur out" subtle masses, spicules, or small microcalcifications. This blur is greater in obese women with thicker breasts, which would imply more limited visualization and detection. Women with thick breasts, even if not obese, would have similar image degradation.
Previous studies have shown that compression thicknesses on mediolateral oblique images are greater than those on craniocaudal images [14]. We showed mediolateral oblique compression thicknesses are also differentially affected by patient weight. Both analysis of variance and standard correlation showed a greater increase in compression thickness with increasing weight in the mediolateral oblique projection than that occurring in the craniocaudal view. Mean compression thicknesses differed between the mediolateral oblique and the craniocaudal images primarily in overweight and obese patients. The mediolateral oblique image best displays the upper outer quadrant and the axillary tail, the most common site of breast malignancy. Degradation in the quality of this image could significantly affect the detection of malignancy in overweight and obese patients.
Mammographic image quality is affected by the kilovoltage at which the study is performed. Lower kilovoltage allows absorption in the photoelectric range, creating maximum image contrast. Although higher kilovoltages allow greater penetration of thick tissue, they increase Compton scattering and decrease image contrast. Obese women had a significant increase in mean kilovoltage in the mediolateral oblique image when compared with their lean and normal-weight counterparts. Kilovoltage changes and increase in scatter resulted in decreases in image contrast of 25% and 16%, respectively. These numbers underestimate the loss of contrast because rhodium filtration was often (57%) used by the machine algorithms to reduce dose and duration of exposure in the obese patients. Rhodium filtration decreases contrast compared with molybdenum filtration.
We illustrate image quality differences associated with breast thickness in Figures 1A,1B and 2A,2B. Visible changes in the geometric sharpness and contrast associated with thinner thickness is evident in Figure 1A,1B. Anterior compression of the breast allowed thinner compressed thickness and confident visualization of an invasive tumor that was poorly seen in the routine image obtained with greater thickness 20 min previously. Figure 2A,2B shows the American College of Radiology breast phantom imaged at the usual thickness of 4.2 cm and at 7.2 cm, similar to the difference between thickness in a lean and an obese patient. Added thickness was achieved by adding 3 cm of BR12 phantom material to the American College of Radiology phantom. Note that the speck groups 1 and 2 are less visible because of decreased contrast and decreased geometric sharpness with the thick phantom. To maintain an exposure of less than 2 sec, the automatic exposure algorithm chose a rhodium target, causing further loss of contrast for the thick phantom.
|
|
|
|
Overweight and obese women had a significant increase in the milliampere-seconds used to produce their mammograms. Although mammography time remained less than 3 sec in all groups, it can be theorized that the increased exposure time could contribute to motion unsharpness in overweight and obese patients. Motion blur may further limit the capability of mammography to evaluate early or subtle lesions in the over-weight individual. The longer exposure increases the average glandular dose for overweight and obese patients. In the mediolateral oblique images, we estimated an increase in mean glandular dose of 136% (72 mGy) between normal-weight and obese women and an increase of 233% (156 mGy) between lean and obese women. These findings are similar to those of previous studies, which found a greater than twofold dose increase with 1.5-2 cm change in compressed thickness [15, 24, 25]. Current mammographic equipment is designed with multiple parameters and limits in mind. Kilovoltage, milliampere-seconds, and glandular dose are kept to a minimum while providing adequate tissue penetration for proper interpretation. In overweight and obese patients, one or more of these parameters must be more compromised to provide the best image.
Overweight and obese women did not show significant differences in positioning standards compared with normal-weight participants. Therefore, positioning differences do not explain the larger tumors found in screened women. Our study did not investigate difficulties experienced by our technologists in positioning the breasts of women of different weights. However, they report the most difficulty with positioning both extremes of weight, the very lean and the very obese.
Breast density declined with increasing body mass index, reflecting increasing proportions of fat in the breast. This would be expected to improve subject contrast if the decline in density was uniform, which could make interpretation better. Yet, paradoxically, larger tumors are found in obese screened women than in normal-weight screened women, which suggests that any potential benefit of lower density is negated by other factors. These factors include image quality degradation caused by thickness and the potential role of biologic differences between normal and overweight women.
Limitations of Study
Weight and height in this study were not independently measured, and
inaccuracies in weight reporting could limit the accuracy of our data.
However, most studies investigating breast cancer risks and body mass also use
self-reported weights, as does the National Institutes of Health in
calculating its statistics on national weight trends
[1,
6,
7,
9,
11]. A study comparing
self-reported and measured weight and height found an average underestimation
of weight by 3.1% and an overestimation of height by 6%
[26]. Any difference would not
be expected to significantly change our results. All cases were drawn from the
local community and therefore reflect the racial, ethnic, and weight makeup of
our area. The division of our patient population into body mass index
categories mirrored that described for the country at large
[1]. For women 50-59 years old
(mean age in our study, 53 years), 29% are classified as overweight, and 35%
are considered obese. This is similar to the body mass indexes measured in our
study.
Conclusion
Compressed breast thickness during mammography is increased in overweight
and obese women and results in degradation of image quality. This would be
expected to adversely affect mammography interpretaion, especially of small or
subtle lesions, and may contribute to the differences in the size of
screening-detected tumors. The challenge for manufacturers and mammographers
is to maximize image quality in these women, who represent most women at risk
for breast cancer in the United States. Higher output per unit area of target
could reduce potential for motion unsharpness. Improved detector efficiency
could also reduce dose and motion unsharpness. Increased target-to-breast
distance for thicker breasts could reduce geometric unsharpness.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  M ROBINSON and C J KOTRE Trends in compressed breast thickness and radiation dose in breast screening mammography Br. J. Radiol., March 1, 2008; 81(963): 214 - 218. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. N. Uppot, D. V. Sahani, P. F. Hahn, D. Gervais, and P. R. Mueller Impact of Obesity on Medical Imaging and Image-Guided Intervention Am. J. Roentgenol., February 1, 2007; 188(2): 433 - 440. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. G. Elmore, P. A. Carney, L. A. Abraham, W. E. Barlow, J. R. Egger, J. S. Fosse, G. R. Cutter, R. E. Hendrick, C. J. D'Orsi, P. Paliwal, et al. The Association Between Obesity and Screening Mammography Accuracy Arch Intern Med, May 24, 2004; 164(10): 1140 - 1147. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
