| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1
Department of Radiology, UCSF Medical Center, Box 1667, San Francisco, CA
94143-1667.
2
Present address: Department of Radiology, University of Michigan Health
System, 1500 E. Medical Center Dr., TC 2910F, Ann Arbor, MI 48109-0326.
Received July 28, 1999;
accepted after revision October 12, 1999.
Presented in part at the annual meeting of the American Roentgen Ray
Society, San Francisco, May 1998.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. Eighty-eight thousand three hundred forty-six consecutive screening mammography examinations were performed from April 1985 to August 1997. Patient weights were normalized to ideal weight correcting for height and were subdivided into adiposity cohorts including underweight by greater than 10%; ideal weight ± 10%; overweight by 11-24%; overweight by 25-39%; and overweight by greater than 40%. The rates of recall, biopsy, cancer detection, and cancer stage were calculated and were analyzed using the chi-square test for trend. Cancer size was analyzed using linear regression analysis.
RESULTS. Reliable (p < 0.05) and meaningful differences were seen between cohorts of increasing adiposity for rates of recall, biopsy, and cancer detection. An increase in recall rate occurred with progressively increasing adiposity (3.88%, 4.89%, 5.11%, 5.47%, 5.55% [p < 0.0001]). The rate of biopsy also increased with increasing adiposity (0.98%, 1.31%, 1.35%, 1.59%, 1.65% [p < 0.0002]), as did the rate of screening-detected cancer (number of cases of cancer per 1000 women screened) (3.74, 4.29, 5.34, 4.70, 6.04 [p < 0.015]). Finally, increased adiposity also correlated with increased median cancer size (p < 0.02) and with more advanced stage at diagnosis (p = 0.046).
CONCLUSION. Increasing adiposity correlates with progressive increases in the rates of recall, biopsy, and cancer detection for women undergoing screening mammography. Increasing adiposity also correlates with increased cancer size and stage, providing further support for obesity as a risk factor for breast cancer.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
A paucity of data exists that assess obesity as a risk factor for breast cancer detected at screening mammography. A retrospective study among postmenopausal women undergoing screening mammography showed that women with breast cancer had a significantly greater Quetelet index for obesity (weight / height2) than those who did not have breast cancer [1], suggesting that screening-detected breast cancer occurs more frequently in obese women. Reeves et al. [27] showed a strong relationship between increased body mass index and the likelihood of nonlocalized disease in women with breast cancer presenting as a palpable mass. However, they found no association between body mass index and extent of disease in asymptomatic women with breast cancer detected on screening mammography. Although that study suggests that no differences exist in mammographically detected cancer as a function of obesity, the descriptors of tumor extent were limited to local versus regional versus distant disease and did not use more detailed descriptors such as tumor size and stage [27].
We conducted a retrospective review of prospectively collected data on a large number of consecutive screening mammography examinations, in which we assessed for differences in outcome measures between adiposity cohorts of women defined by their weights corrected for height, compared with ideal weight. The clinical outcomes evaluated include rates of recall, biopsy, and cancer detection, in addition to tumor size and cancer stage.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We conducted a retrospective review of prospectively collected data on 88,346 consecutive screening mammography examinations. Self-reported weights were correlated with self-reported heights and normalized to ideal weight using a standard height-weight table [31]. Prospectively collected outcome measures were compared retrospectively among adiposity cohorts defined as follows: under-weight by greater than 10%; ideal weight ± 10%; overweight by 11-24%; overweight by 25-39%; and overweight by greater than or equal to 40%. Of the 88,346 examinations, 3475 patients were in the underweight cohort, 39,656 patients were in the ideal weight cohort, 20,789 patients were in the overweight by 11-24% cohort, 12,345 patients were in the overweight by 25-39% cohort, and 12,081 patients were in the overweight by greater than or equal to 40% cohort. Modest age differences were noted among these adiposity cohorts, with median ages of 48, 48, 52, 52, and 52 years, respectively. Mean ages for these cohorts were 51.2, 50.3, 53.5, 54.0 and 53.2 years, respectively.
Mammography Procedure
The screening examination involved mediolateral oblique and craniocaudal
mammographic views of each breast using a dedicated American College of
Radiology-accredited mammography unit (Mamex DC, Soredex, Conroe, TX; or Alpha
III, Instrumentarium Imaging, Milwaukee, WI). These examinations were
interpreted by board-certified staff radiologists: two radiologists who
specialize in breast imaging and four general diagnostic radiologists. Each
examination was interpreted by one of these staff radiologists, a breast
imaging fellow, and from zero to three radiology residents, with the staff
radiologist giving the final interpretation. Mammographic findings were
reported as normal or abnormal (recall), with specific recommendations for
prompt further evaluation of each abnormal case.
For 28,323 consecutive cases, beginning in July 1993, the interpreting radiologist subjectively determined breast density as seen on mammography according to the four categories of density described in the American College of Radiology Breast Imaging Reporting and Data System lexicon [32]. These categories are almost entirely fat (density 1), scattered fibroglandular densities (density 2), heterogeneously dense (density 3), and extremely dense (density 4).
Follow-Up
Clinical outcomes for all women with screening examinations interpreted as
having abnormal findings were determined by contacting each woman's personal
physician and by searching our institution's radiology and pathology
databases, as described previously
[29]. This search enabled us
to determine for each weight cohort the rates of recall, biopsy, and cancer
detection, as well as cancer size and stage of all screening-detected cancer.
"Breast cancer" was defined as a histopathologic diagnosis of
ductal carcinoma in situ or any type of invasive carcinoma. Tumor size was
recorded as size based on pathologic examination, unless that was unavailable,
in which case mammographic size was used. In addition, data for all women who
underwent screening mammography were linked by computer to our regional
Surveillance, Epidemiology, and End Results tumor registry, which collects
population-based cancer data from nine contiguous counties in our state,
including the six counties in our service area
[30]. Thus, we could identify
additional cases of interval cancer, defined as cancer detected not at
screening but with tissue diagnosis within 1 year of screening. Mammographic
data were linked to the tumor registry at least 2 years after screening to
facilitate more complete reporting of cancer cases to the tumor registry.
Data Analysis
The chi-square test for trend, a test of linear association, was used to
compare the adiposity cohorts for various screening mammography outcomes.
Outcomes evaluated include rate of recall, biopsy recommended,
screening-detected and interval cancer, cancer stage, breast density
(comparing proportions of breast densities 1 and 2 with breast densities 3 and
4), and frequency of initial versus subsequent screening mammographic
examinations. Linear regression analysis was used to assess differences in
size of cancerous tumors in the adiposity cohorts. Statistical analysis was
performed using Prism statistical software (GraphPad, San Diego, CA).
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
The recall rate increased progressively with increasing adiposity (Table 1 [p < 0.0001]). A 14% increase in recall rate was found in the heaviest weight cohort compared with the ideal weight cohort. Combining all overweight groups, a 9% increase in recall rate was observed compared with the ideal weight cohort. A 21% decrease in recall rate was found in the underweight cohort compared with the ideal weight cohort.
|
The biopsy rate also increased with increasing patient adiposity (Table 1 [p < 0.0002]). Twenty-six percent more biopsies were performed in the most obese cohort than in the ideal weight cohort and 14% more biopsies in the combined overweight groups than in the ideal weight cohort. Twenty-five percent fewer biopsies were performed in the underweight cohort than in the ideal weight cohort.
The cancer detection rate also increased with increasing adiposity (Table 1 [p < 0.015]). The cancer detection rate was 44% higher for the most obese cohort than for the ideal weight cohort, and the combined overweight groups had a 25% higher cancer detection rate than the ideal weight cohort. The underweight cohort had a 13% decrease in cancer detection compared with the ideal weight cohort. The detected cases of invasive breast cancer had progressively increased size with increasing adiposity (p < 0.02), from a mean tumor diameter of 11 mm in the underweight cohort to 19 mm in the heaviest overweight cohort (Table 2). Furthermore, the rate of stage II and higher cancer increased with increasing adiposity (Table 3 [p = 0.046]). Thus, the cancerous tumors detected on screening mammography were larger and of more advanced stage in the obese cohorts.
|
|
To determine whether improved mammographic visualization in breasts with greater fat replacement could at least partially account for the observed increased recalls, biopsies, and cases of cancer detected in obese patients, we assessed the mammographic breast density for each adiposity cohort. Comparing less dense breasts (densities 1 and 2) with denser breasts (densities 3 and 4), we found that increasing adiposity does result in decreased mammographic breast density (Table 4 [p < 0.0001]). This correlation was also observed when the mammograms that detected breast cancer within each adiposity cohort were compared independently (p < 0.0001). The mean breast densities of the women who had screening-detected cancer were similar to those of the entire population within each adiposity cohort (Table 5).
|
|
The increased rates of recall, biopsy, and cancer detection in obese patients cannot be explained by improved mammographic visualization because we did not find fewer cases of interval cancer in the heavier weight cohorts. Fifty cases of interval cancer were identified by linkage of 63,329 women to our regional tumor registry. A small increase, instead of a decrease, occurred in cases of interval cancer in the overweight cohorts compared with the ideal weight cohort. A larger increase in interval cancer rate was seen in the underweight cohort than in the ideal weight cohort, although the very small number of interval cancer cases in the underweight cohort limits the reliability of this observation. None of the observed differences in interval cancer rates was statistically significant (p = 0.98), possibly because of small numbers of cancer cases (Table 6). Furthermore, if improved cancer detection in obese patients was caused by improved mammographic visualization of breast cancer, one would expect to find smaller and less advanced cancer in obese women. We found the opposite. The mean and median sizes of screening-detected invasive cancer increased with increasing adiposity (Table 2 [p < 0.02]). Also, the rate of stage II and higher cancer increased with increasing adiposity (Table 3 [p = 0.046]).
|
We compared the frequency of initial screening examinations versus subsequent screening examinations as a function of obesity. This comparison was done to determine if difference existed in use of screening mammography among the adiposity cohorts. We found that as adiposity increased, the percentage of examinations that were initial screening examinations increased (Table 7 [p < 0.0001]). The percentage of initial screening examinations increased steadily from 45.1% in underweight women to 53.5% in the most obese cohort. This figure suggests that obese women were less likely to undergo screening on a regular basis, and hence that obese women were more likely to have prevalent (as opposed to incident) cancer detected at screening. To further assess this issue, we analyzed initial screening mammographic examinations for screening-detected cancer as a function of adiposity. We observed a trend of increase in screening-detected cancer with increasing adiposity that was of only borderline significance (Table 8 [p = 0.05]). If the increased initial screening examinations explain the increase in screening-detected cancer in obese women, one would expect a less significant trend when the rate of screening-detected cancer as a function of adiposity is analyzed for the combination of initial and subsequent examinations. In fact, we observed the opposite. When all screening examinations were analyzed together, the increase in screening-detected cancer as a function of increasing adiposity is statistically significant (Table 1 [p < 0.015]). Therefore, the trend of increased screening-detected cancer in obese women is not explained by the increased initial screening examinations in obese women.
|
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Previous studies have shown that for all women diagnosed with breast cancer, whether detected at screening mammography or at palpation, obesity is associated with tumors that are larger at presentation [15,16,17,18,19,20] and poorer in prognosis [13, 15, 21,22]. A study by Reeves et al. [27] assessing breast cancer detected at screening mammography did not show differences in extent of cancer (local versus regional versus distant disease) as a function of body mass index. Our study expands on these findings because we have more detailed information about screening outcomes, including recall and biopsy rates, tumor size and stage, and interval cancer. Our observation that screening-detected cancerous tumors are significantly larger (p < 0.02) and more advanced in stage (p = 0.046) in heavier women suggests that mode of cancer detection does not affect the association of obesity with the discovery of larger tumors with poorer prognosis.
We have discussed several factors that could confound interpretation of our results. Fatty breast density has been shown to be associated with increased mammographic sensitivity for detection of malignancy [30]. Fatty breast density is also believed to permit improved mammographic visualization of malignancy because of less obscuration by adjacent dense tissue. We indeed observed a progressive increase in fatty density as a function of increasing obesity. However, improved mammographic visualization does not appear to explain the increase in frequency of screening-detected cancer in obese patients because we observed a slight increase, not a decrease, in the frequency of interval cancer in our obese cohorts and because our screening-detected cancerous tumors were larger, not smaller, in obese women. These findings are most consistent with the conclusion that more cases of cancer are detected on screening mammography in obese women simply because breast cancer develops more frequently and perhaps more rapidly in obese women.
Patient age is another possible confounding factor. However, the age differences we observed between adiposity cohorts were small and did not show a consistent trend. Mean ages were 3 years 6 months greater for the overweight cohorts than for the ideal weight cohort, but no age differences existed between the individual overweight cohorts. Furthermore, the mean age of the underweight cohort was 1 year more, not less, than that of the ideal weight cohort. The small magnitude and inconsistency of these age differences do not appear to explain the statistically significant and progressive increases we observed for heavier women in rates of recall, biopsy, cancer detection, tumor size, and stage.
Our results are not explained by differences in the use of mammography as a function of adiposity. We observed an increase in frequency of initial compared with subsequent screening mammographic examinations in obese women (p < 0.0001). These findings suggest that obesity can adversely affect a woman's likelihood of undergoing screening. Because initial mammographic examinations are associated with higher rates of recall, biopsy, cancer detection, and larger tumor size than are subsequent screening examinations, our results could hypothetically be explained by the increased frequency of initial mammographic examinations in obese women. However, progressive increases in the rate of cancer detection as a function of adiposity were maintained when the initial and subsequent mammography subgroups were analyzed independently (Table 8). Although differences in the use of mammography cannot explain our results, we cannot exclude this as a contributing factor to the magnitude of the trends we observe.
Insofar as none of the possible confounding factors appears to explain our observed results, the observed increased rate of screening-detected cancer in heavier women is likely the result of a direct effect of obesity in increasing the risk for development of breast cancer. The current hypothesis to explain increased breast cancer risk in obese postmenopausal women is that they have higher circulating levels of estrogen, which may support breast cancer growth and metastatic potential [27]. This hypothesis is based on the observation that the main source of endogenous estrogen after menopause is conversion in adipose tissue of the androgen precursor androstenedione to estrone [3]. Obesity also is associated with decreased production of sex-hormone-binding globulin [33], which results in an increase in the biologically active, unbound form of estradiol [27, 34]. Further evidence to support the role of estrogen in breast cancer development is the increased frequency of breast cancer observed in postmenopausal women who take exogenous estrogen [13, 31].
Previous epidemiologic studies have indicated the presence of an obesity-related increase in risk for the development of breast cancer in postmenopausal women, but this trend has not been observed in premenopausal women [1,2,3,4,5,6,7,8,9,10,11,12,13,14]. We observe an increase in screening-detected cancer as a function of obesity in our patient population, which includes both premenopausal and postmenopausal women. We cannot address the issue of a difference in risk for premenopausal versus postmenopausal women because we do not have menopause information in our database.
We also observed increased size and stage of breast cancer detected on screening mammography with increasing obesity. Broader studies assessing breast cancer detected by any means have observed poorer prognosis in obese patients, with higher recurrence rates [21, 23, 24] and higher mortality from breast cancer [21, 25, 26]. These observations may be explained by faster tumor growth and enhanced metastatic potential in obese women caused by estrogen hormonal enhancement [16, 18, 21, 27].
In summary, increasing obesity correlates with an increase in recall rate, biopsy rate, and cancer detection rate for women undergoing screening mammography. We observed larger and more advanced-stage screening-detected cancerous tumors and a slight increase (rather than decrease) in interval cancer in obese women. Our results provide further support for obesity as a risk factor [1] for the development of breast cancer and [2] for a poorer prognosis for obese patients with breast cancer.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  F. Sassi, H. S. Luft, and E. Guadagnoli Reducing Racial/Ethnic Disparities in Female Breast Cancer: Screening Rates and Stage at Diagnosis Am J Public Health, December 1, 2006; 96(12): 2165 - 2172. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 G. A. Porter, K. M. Inglis, L. A. Wood, and P. J. Veugelers Effect of Obesity on Presentation of Breast Cancer Ann. Surg. Oncol., March 1, 2006; 13(3): 327 - 332. [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]
|
|
|
|
 |
|
 M. A. Roubidoux, J. S. Kaur, K. A. Griffith, J. Sloan, C. Wilson, P. Novotny, and M. Lobell Correlates of Mammogram Density in Southwestern Native-American Women Cancer Epidemiol. Biomarkers Prev., June 1, 2003; 12(6): 552 - 558. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. Demark-Wahnefried, B. L. Peterson, E. P. Winer, L. Marks, N. Aziz, P. K. Marcom, K. Blackwell, and B. K. Rimer Changes in Weight, Body Composition, and Factors Influencing Energy Balance Among Premenopausal Breast Cancer Patients Receiving Adjuvant Chemotherapy J. Clin. Oncol., May 1, 2001; 19(9): 2381 - 2389. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
