• race;
  • ethnicity;
  • diagnostic mammography;
  • breast cancer


  1. Top of page
  2. Abstract


A previous study compared the performance (sensitivity, specificity, positive predictive value, and cancer detection rate) of screening mammography in Black and White women. No study, to the authors' knowledge, has evaluated the difference in the performance of diagnostic mammography between Black and White women.


Univariate analysis was used to evaluate differences in characteristics and cancers between Black and White women. Stratified and adjusted logistic regression analyses were used to test the association of Black and White race with performance measures of diagnostic mammography.


The sensitivity of diagnostic mammography was higher (91% vs. 84%) and specificity was lower (86% vs. 90%) among Black women compared with White women. After controlling for age, density, self-reported breast problems, and previous mammography, sensitivity was significantly higher (odds ratio [OR] = 1.82, 95% confidence interval [CI] = 1.22–2.80) and specificity was significantly lower (OR = 0.75, 95% CI = 0.70–0.81) among Black women. The crude cancer detection rate of mammography was higher for Black women (42.6/1000) than for White women (31.0/1000) and Black women had a higher proportion of cancers that were > 2.0 cm (57.4% vs. 46.2%) that were more often poorly differentiated (61.7% vs. 49.3%) and were more often estrogen-receptor and progesterone-receptor negative.


Black women have lower specificity of diagnostic mammography and, consequently, more unnecessary workups than White women. Black women have higher sensitivity of diagnostic mammography, with cancers that are larger and more advanced than White women. Delay in responding to signs and symptoms would explain the size and later stage. However, more research is needed to understand the biologic differences of breast cancer characteristics between Black and White women. Cancer 2005. © 2005 American Cancer Society.

Diagnostic mammography consists of extra views in addition to the standard 2-view (cranial–caudal and mediolateral view) screening study and is performed when women are seen for signs or symptoms or as continued workup after an indeterminate or positive screening study. Women may present with lump, discharge, pain, or skin changes. Because of increased scrutiny of diagnostic mammograms, and higher prevalence of signs and symptoms in the population, diagnostic mammography has been shown to have higher sensitivity and positive predictive value (PPV), and lower specificity, than screening mammography.1, 2

There have been only a few studies published on performance of diagnostic mammography.1–4 There was no published study found that compared diagnostic mammography in Black compared with White women. A previous population-based study on screening mammography performance from the Carolina Mammography Registry (CMR) found no difference in sensitivity, specificity, or PPV of screening mammography between Black and White women.5 However, the study found that Black women who reported symptoms before their screening mammograms had larger, higher grade tumors when compared with White women who reported symptoms. The authors suggest that Black women may have a longer delay in screening after the onset of symptoms than White women. We were interested in whether these findings comparing screening performance in Black and White women would persist for diagnostic mammography.

There is no reason to believe and no data to show that mammographic images of Black women differ from White women. There should be no difference in the sensitivity or specificity of diagnostic mammogram for Black women compared with White women, unless it is related to the risk of cancer or characteristics of the women themselves. In addition to examining the difference in performance of diagnostic mammograms between Black and White women, we were also interested in testing whether cancer outcomes from diagnostic mammography would differ for Black and White women, as seen previously in screening mammography.


  1. Top of page
  2. Abstract

Data for this study came from CMR, a population based mammography registry that links prospectively collected mammography data with population-based cancer data.6 The data were limited to all diagnostic mammograms in CMR performed for Black and White women between January 1, 1994 and December 31, 2000 and were further restricted to only women who were residents of North Carolina. A mammogram was considered to be diagnostic if it was classified as symptomatic or diagnostic by the radiologist or technologist at the time of the study. Mammograms were excluded if women were younger than 25 years of age, had a history of implants, or a personal history of breast cancer. They were also excluded if a valid ACR/BI-RADS™ (American College of Radiology Breast Imaging Reporting and Data System, Reston, VA) assessment code was missing.7 The patient provided demographic information and history through a self-administered questionnaire at the beginning of the imaging visit. The technologist and radiologist, at each visit, recorded the reason for the patient visit, assessment of the diagnostic mammogram, and recommendations for future evaluations.

Patient's age was categorized into five mutually exclusive categories, 25–39 years, 40–49 years, 50–59 years, 60–69 years, and ≥ 70 years. A missing response was coded as a negative response to questions about self-reported history, including personal history of breast cancer, first-degree family history of breast cancer, history of breast problems, breast implants, or use of hormones. Self-reported breast problems included presence of lump, discharge, and other (excluding pain). When symptoms were evaluated specifically, the categories were mutually exclusive: 'lump only', 'discharge only' and 'other', which included women who had both a lump and discharge or had some other symptom. We understand that women report symptoms and signs; we use the term “symptoms” to refer to both.

Women chose one unique race from a list of racial groups. A woman's history of previous mammography was based on prior mammography present in the database. The woman's self-report was used if no record of a prior mammogram existed in the database. If there were no data present indicating previous mammography and the date of previous mammography could not be determined by self-report, it was assumed that the woman did not have previous mammography. Each woman was assigned rural or urban status based on her residential zip code, by using the Metropolitan Statistical Area score determined by the Federal Office of Management and Budget, following a set of official standards published by the interagency Federal Executive Committee on Metropolitan Areas.8 A woman's self-reported education attainment level was categorized into three groups: 1) less than high school, 2) high school graduate and/or some college, and 3) college graduate.

Breast density and assessment of diagnostic mammography were coded according to the BI-RADS assessments. Breast density was categorized into four categories: 1) extremely dense, 2) heterogeneously dense, 3) scattered fibroglandular densities, and 4) almost entirely fat. Assessments of diagnostic mammograms were categorized as follows: 0, needs further evaluation; 1, normal; 2, benign finding; 3, probably benign; 4, suspicious abnormality; and 5, suspicious for cancer. The radiologist made recommendations for additional workup based on evaluation of the diagnostic mammogram. Recommendations included ultrasound, magnetic resonance imaging (MRI), computed tomography (CT), biopsy, and surgical consult. The radiologist also made recommendations for the timing of the additional workup, either immediate, short-term (< 6 mos), or routine (1 yr, 2 yrs, or age 40 yrs).

Diagnostic mammography data were linked on a unique identifier to a database containing information about the breast cancer status of women. The assignment of identification (ID) to women in both databases was accomplished by computer software package, Matchware (MatchWare, Inc., Tampa, FL). The breast cancer database contained data compiled from three different sources, the North Carolina Central Cancer Registry, pathology labs, and from the mammography facilities. The breast cancer database contained extensive information on the pathologic characteristics of the tumors, including tumor size, grade, progesterone receptor status, estrogen receptor status, and type of cancer (invasive or ductal carcinoma in situ). Both the mammography data and the pathology data go through extensive quality control at various stages of data collection and data management. All identifying information for patients, radiologists, and facilities had been removed from the data set before any analyses were performed. Protection of identifying information was strictly adhered to, as outlined by Carney.9 CMR has been approved annually by the University of North Carolina at Chapel Hill School of Medicine Institutional Review Board (IRB) and by the IRBs of other participating hospitals. CMR also holds a Public Health Service (PHS) certificate of confidentiality.

To determine the performance of diagnostic mammography, each diagnostic mammogram was initially categorized as true positive (TP), false negative (FN), true negative (TN), and false positive (FP) according to the BI-RADS code assigned by the radiologist and the cancer outcome in the follow-up period. The follow-up period was defined as one calendar year after the date of the diagnostic mammogram. This definition corresponds to the follow-up period that is commonly used for screening mammography. A positive diagnostic mammogram was one in which the BI-RADS code assigned by the radiologist was a 0, 4, or 5, or a BI-RADS of 3, when associated with a recommendation of immediate follow-up. All other diagnostic mammograms were considered to be negative at the time of assessment. A positive diagnostic mammogram with a cancer diagnosis of ductal carcinoma in situ (DCIS) or invasive cancer in the follow-up period was considered a true positive. All other positive cancers were false positive. A negative diagnostic mammogram with no cancer in the follow-up period was categorized as a true negative; if cancer was detected in the follow-up period, the negative diagnostic mammogram was considered a false negative.

Initially, a descriptive analysis was performed of characteristics of the women associated with the diagnostic mammograms and of signs and symptoms that these women reported. Performance measures were calculated including sensitivity, specificity, cancer detection rate, and PPV. Sensitivity was calculated as the proportion of positive diagnostic mammograms among women who had a cancer diagnosis in the follow-up period [TP/(TP + FN)]. Specificity was calculated as the proportion of negative diagnostic mammograms among women who did not have a cancer diagnosis in the follow up period [TN/(TN + FP)]. The cancer detection rate was calculated as the number of true positive diagnostic mammograms divided by the total number of diagnostic mammograms [TP/(TP + FN + TN + FP)]. PPV was calculated as the proportion of diagnostic mammograms with a diagnosis of breast cancer in the follow-up period compared with all positive diagnostic mammograms [TP/(TP + FP)].

Sensitivity, specificity, and PPV were evaluated separately in logistic regression models, using the covariates of age, breast density, any self-reported breast symptoms, previous mammography, and racial group. These covariates (excluding race) have been shown to influence performance of screening mammography in previous studies.9–15 These covariates are likely to influence diagnostic mammography in a similar manner. All models were run using the PROC GENMOD procedure in SAS (SAS Institute, Cary, NC).16 For sensitivity, we restricted the data to diagnostic mammograms among women with a cancer diagnosis in the follow-up period and estimated the odds ratio (OR) for a TP diagnostic assessment versus a FN diagnostic assessment for each level of a covariate. For specificity, the data were restricted to all negative diagnostic mammograms and the OR was estimated for a TN diagnostic assessment versus a FP diagnostic assessment for each level of a covariate. For PPV, the data were restricted to all positive diagnostic mammograms, and the OR was estimated for a TP diagnostic assessment versus a FP diagnostic mammogram for each level of a covariate. An OR > 1.00 indicated better performance for any one stratum compared with the reference group. The ratio of the odds for a correct interpretation in a given variable category relative to a reference category was calculated after controlling for all other covariates in the model. Results from all logistic models were tested for significance at the 0.05 significance level.

For those women who developed cancer in the follow-up period, we described the cancer characteristics by racial group. We evaluated whether there were any differences between Black and White women in terms of type of breast cancer (DCIS or Invasive), tumor grade, estrogen receptor status, progesterone receptor status, and tumor size, by using univariate chi-square statistics.


  1. Top of page
  2. Abstract

Demographics and Health History

For the years 1994–2000, there were a total of 9488 diagnostic mammograms for 8245 Black women and 47,182 diagnostic mammograms for 37,292 White women in CMR. Restricting the data to mammograms of Black and White women 25 years of age or older with no personal history of breast cancer and no history of breast implants yielded 46,984 diagnostic mammograms from 40,751 women. The results cited for Black and White women heretofore refer to the mammograms among Black and White women. Comparing mammograms among Black to those of White women (Table 1), Black women were younger (67.8% of Black women were < 50 yrs of age compared with 52.7% of White women); they were less educated (15.6% of Black women received less than a high school education compared with 7.3% of White); they had a higher proportion living in rural locations (41.3%) compared with White women (38.2%); they had a lower proportion with a history of breast procedures (26.9% vs. 34.7%); and Black women were less likely than White women to have had a previous mammography examination (70.8% vs. 83.5%). There was little difference in reported family history of breast cancer. The numbers in this study are very large; consequently, most differences between Black and White women were statistically significant. The absolute differences should be considered.

Table 1. Characteristics Associated with Diagnostic Mammograms of Black and White Women
CharacteristicsBlack womenWhite womenTotal
  • All differences between black and white women are statistically significant (P < 0.0001).

  • a

    Column percentages for each characteristic group.

  • b

    Includes missing values.

  • c

    Lump, discharge are mutually exclusive.

  • d

    Cases where more than one symptom was reported.

Total Age8610100.038,374100.046,984
 < 40275232.0854722.311,299
 ≥ 706067.0448011.75086
 < HS134415.628177.34161
 HS graduate329038.214,17136.917,461
 > College127014.8761219.88882
Rural/urban status     
Family Hx breast cancer     
Biopsy/surgical history     
Breast density     
 Extremely dense99011.5456811.95558
 Heterogeneously dense356341.416,75343.720,316
 Scat. fibrogland. dens.317036.814,30237.317,472
 Entirely fat4355.113373.51772
Previous mammography     
Self-reported breast problems    
 Any problem735685.430,93180.638,287
  Other problemsd316936.813,48235.116,651
 No problemb125414.6744319.48697

The distribution of breast density among the mammograms showed only a slightly lower proportion of dense breasts for Black women (52.9%) compared with White women (55.6%) (Table 1). Symptoms were reported for 85% of Black women and 80% of White women. The proportion reporting a lump was very close for Black women (43.5%) and White women (42.2%). Reporting of discharge alone was infrequent overall (3.6%) and higher in Black women than White women (5.1%, vs. 3.3%). Rates for lump and discharge were higher for Black women in every age group.

Cancer Detection Rates

Cancer detection rates overall were 42.6/1000 for Black women and 31.0/1000 for White women. The rates were higher in Black compared with White in every subgroup evaluated (Table 2). The largest differences were in the youngest and oldest age groups: Black women younger than 40 years of age had cancer detection rates that were 1.8 times higher than White women in the same age group; Black women in the age categories 60–69 and ≥ 70 had cancer detections rates that were 1.9 and 1.8 times higher, respectively, when compared with White women in the same age categories.

Table 2. Cancer Detection Rates by Characteristics Associated with Diagnostic Mammograms of Black and White Women
 Black womenWhite womenRatio B/W
No. cancersCancer detection rate/1000 mammogramsaNo. cancersCancer detection rate/1000 mammogramsa
  • a

    Includes missing values.

  • b

    Cancer detection rate = TP/all mammograms.

  • c

    Lump, discharge are mutually exclusive.

  • d

    Cases where more than one symptom was reported.

Total Age40342.6140731.01.4
 < 407323.612813.21.8
 ≥ 7090141.939379.21.8
 < HS9664.715448.61.3
 HS graduate12732.843824.81.3
 > College3626.020720.11.3
Rural/urban status     
Family Hx breast cancer     
Biopsy/surgical history     
Breast density     
 Extremely dense2721.212320.61.0
 Heterogeneously dense16341.356727.71.5
 Scat. fibrogland. dens.15847.961437.81.3
 Entirely fat1632.23421.71.5
Previous mammography     
Self-reported breast problems     
 Any problem38147.3127134.91.4
  Other problemsd32721.828517.51.2
 No problema2215.213614.51.0

Reporting any symptom at presentation for diagnostic mammography was associated with a significantly higher cancer detection rate for Black (47.3/1000) compared with White women (34.9/1000). The majority of cancers detected are related to presence of a lump: the cancer detection rate among the women who were positive for lump was 72.4/1000 for Black women and 51.2 for White women. Women reporting only discharge had cancer detection rates, for Black and White women of 18.1/1000 and 12.7/1000, respectively, very close to the rate in women without symptoms, 15.2/1000 and 14.5/1000. Women with no problems or discharge alone had the lowest cancer detection rates.

When there was a positive history of breast procedures associated with the mammograms, detection rates were similar (27.7/1000 for Black women and 24.2/1,000 for White women). When the history was negative for history of breast procedures, Black women had significantly higher cancer detection rates (48.1/1000) than White women (34.5/1,000). Black women in urban areas had cancer detection rates (45.0/1000) that were 5% higher than cancer detection rates for Black women in rural areas (39.3/1000). For White women, the difference in cancer detection rates between urban (31.1/1000) and rural women (30.6/1000) was only 1%.

Performance Measures

Overall, sensitivity among diagnostic mammograms was higher for Black women (91%) than for White women (84%), specificity was lower in Black women (86%) than in White women (90%), and PPV was 25% for both Black and White women (Table 3)

Table 3. Performance Measures by Characteristics Associated with Diagnostic Mammograms of Black and White Women
Patient characteristicSensitivity (95% CI)Specificity (95% CI)PPV (95% CI)
  • a

    Includes missing values.

  • b

    Lump, discharge are mutually exclusive.

  • c

    Cases where more than one symptom was reported.

Total Age0.91 (0.88–0.94)0.84 (0.82–0.86)0.86 (0.85–0.87)0.90 (0.89–0.90)0.25 (0.23–0.27)0.25 (0.24–0.26)
 < 400.89 (0.82–0.96)0.88 (0.83–0.94)0.85 (0.84–0.86)0.88 (0.88–0.89)0.14 (0.11–0.17)0.10 (0.09–0.12)
 40–490.90 (0.84–0.95)0.82 (0.78–0.86)0.86 (0.85–0.87)0.90 (0.89–0.90)0.19 (0.16–0.23)0.17 (0.15–0.19)
 50–590.91 (0.85–0.98)0.83 (0.79–0.87)0.89 (0.87–0.91)0.91 (0.91–0.92)0.30 (0.24–0.37)0.27 (0.24–0.30)
 60–690.89 (0.81–0.97)0.79 (0.74–0.84)0.89 (0.87–0.91)0.92 (0.92–0.93)0.43 (0.34–0.52)0.35 (0.32–0.39)
 ≥ 700.96 (0.89–0.99)0.90 (0.87–0.93)0.85 (0.82–0.88)0.91 (0.90–0.92)0.53 (0.45–0.61)0.50 (0.46–0.54)
Family Hx breast cancer     
 Yes0.97 (0.85–1.00)0.77 (0.71–0.84)0.88 (0.86–0.90)0.90 (0.89–0.91)0.27 (0.19–0.36)0.28 (0.24–0.32)
 Noa0.91 (0.88–0.94)0.85 (0.83–0.87)0.86 (0.86–0.87)0.90 (0.90–0.91)0.25 (0.22–0.27)0.24 (0.24–0.26)
Biopsy/surgical history     
 Yes0.84 (0.76–0.92)0.77 (0.73–0.81)0.86 (0.86–0.87)0.90 (0.89–0.90)0.17 (0.14–0.22)0.20 (0.18–0.22)
 Noa0.93 (0.90–0.95)0.87 (0.85–0.89)0.86 (0.85–0.88)0.90 (0.90–0.91)0.27 (0.25–0.30)0.27 (0.26–0.29)
Previous mammography     
 Yes0.90 (0.87–0.94)0.82 (0.80–0.84)0.88 (0.88–0.89)0.91 (0.91–0.92)0.24 (0.21–0.27)0.25 (0.24–0.26)
 Noa0.92 (0.88–0.96)0.93 (0.90–0.95)0.81 (0.80–0.83)0.85 (0.84–0.86)0.26 (0.22–0.30)0.24 (0.22–0.27)
Breast density      
 Extremely dense0.78 (0.58–0.91)0.76 (0.69–0.84)0.83 (0.81–0.85)0.88 (0.87–0.89)0.11 (0.07–0.17)0.15 (0.12–0.18)
 Heterogeneously dense0.90 (0.86–0.95)0.82 (0.79–0.85)0.86 (0.85–0.87)0.90 (0.90–0.90)0.24 (0.20–0.27)0.22 (0.20–0.24)
 Scat. fibrogland. dens.0.96 (0.93–0.99)0.88 (0.86–0.91)0.88 (0.87–0.89)0.92 (0.91–0.92)0.30 (0.26–0.34)0.32 (0.30–0.34)
 Entirely fat0.88 (0.62–0.98)0.85 (0.69–0.95)0.91 (0.89–0.94)0.94 (0.93–0.96)0.28 (0.16–0.42)0.28 (0.20–0.38)
 Missing0.85 (0.73–0.96)0.87 (0.79–0.95)0.79 (0.75–0.83)0.84 (0.82–0.86)0.28 (0.20–0.37)0.22 (0.17–0.27)
Self–reported breast problems     
 Any problem0.91 (0.89–0.94)0.85 (0.83–0.87)0.86 (0.85–0.87)0.89 (0.89–0.90)0.26 (0.24–0.28)0.26 (0.24–0.27)
  Lumpb0.92 (0.89–0.96)0.87 (0.84–0.89)0.82 (0.81–0.83)0.87 (0.86–0.87)0.31 (0.28–0.34)0.29 (0.27–0.30)
  Dischargeb0.67 (0.35–0.90)0.53 (0.35–0.71)0.87 (0.84–0.91)0.91 (0.90–0.93)0.13 (0.06–0.24)0.13 (0.08–0.20)
  Other problemsc0.91 (0.84–0.97)0.83 (0.78–0.87)0.89 (0.88–0.91)0.93 (0.92–0.93)0.18 (0.14–0.22)0.20 (0.17–0.22)
 No problema0.86 (0.65–0.97)0.79 (0.73–0.86)0.91 (0.89–0.92)0.94 (0.93–0.94)0.14 (0.09–0.21)0.19 (0.16–0.23)

Sensitivity generally increased with age for Black women. In contrast, for White women, sensitivity was highest in the youngest (< 40) and oldest (≥ 70) age groups and lower in intermediate age groups (40–69 yrs). For both Black and White women, the sensitivity was higher in the presence of breast symptoms compared with no breast symptoms and was very low when the only symptom was discharge (not significantly lower in Black women, most likely because of the small number of cancers, but significantly lower in White women.) Sensitivity was higher with a negative history of previous breast procedures compared with a positive history of breast procedures.

Overall specificity of the mammograms was lower for Black women compared with White women, and this relation persisted across strata of all covariates. Overall PPV of the mammograms was similar for Black and White women and tended to be similar across strata of covariates. The PPV was lowest when the symptom was discharge, and the PPV in presence of lump was significantly higher than in presence of discharge for Black and White women.

Comparisons of sensitivity, specificity, and PPV, using logistic regression models, controlling for the covariates age, density, any self-reported breast problems, and previous mammography revealed that sensitivity was significantly higher (OR = 1.82, 95% CI = 1.22–2.80), specificity was significantly lower (OR = 0.75, 95% CI = 0.70–0.81), and PPV was marginally higher (OR = 1.12, 95% CI = 0.96–1.31) when comparing mammograms of Black women with White women (Table 4)

Table 4. Logistic Regression Results by Characteristics Associated with Diagnostic Mammograms of Black and White Womenab
 Probability of (+) mammogram given diseaseProbability of (−) mammogram given no diseaseProbability of disease, given (+) mammogram
No.OR95% CINo.OR95% CINo.OR95% CI
  • CI: confidence interval; OR: odds ratio; PPV: positive predictive value.

  • a

    Records missing density in the BI-RADS classification system were excluded from the analysis.

  • b

    Each variable is adjusted for all other variables in the table.

  • c

    Includes missing or unknown values.

All women         
 Black women3641.821.22–2.807,7940.750.70–0.8113591.120.96–1.31
 White women13381.00(ref. group)35,6221.00(ref. group)45021.00(ref. group)
Black women         
 Age < 40670.210.03–0.9425551.330.99–1.784300.130.08–0.20
 ≥ 70771.00(ref. group)4751.00(ref. group)1401.00(ref. group)
Previous mammography         
 Yes2161.00(ref. group)55761.00(ref. group)8221.00(ref. group)
 Extremely dense270.770.10–4.369630.470.32–0.691840.600.27–1.39
 Heterogeneously dense1631.930.27–8.8134000.570.39–0.806211.270.64–2.62
 Scat. fibrogland. dens.1585.080.67–26.6730120.670.46–0.955041.590.81–3.28
 Entirely fat161.00(ref. group)4191.00(ref. group)501.00(ref. group)
Any self-reported breast probs.        
 Yes3521.00(ref. group)66291.00(ref. group)12481.00(ref. group)
White women         
 Age < 401190.840.42–1.7481731.010.87–1.1610570.100.08–0.13
 ≥ 703771.00(ref. group)39191.00(ref. group)6621.00(ref. group)
Previous mammography         
 Yes10541.00(ref. group)29,8281.00(ref. group)34021.00(ref. group)
 Extremely dense1230.670.21–1.8344450.470.36–0.606280.930.56–1.56
 Heterogeneously dense5670.970.32–2.4316,1860.540.42–0.6820881.200.76–1.95
 Scat. fibrogland. dens.6141.530.50–3.8413,6880.630.49–0.7916821.500.95–2.42
 Entirely fat341.00(ref. group)13031.00(ref. group)1041.00(ref. group)
Any self-reported breast probs.        
 Yes12101.00(ref. group)28,5721.00(ref. group)39831.00(ref. group)

Cancer Characteristics

The proportion of DCIS among all cancers was only slightly lower for Black women compared with White women, with DCIS detected in 5.7% of Black women and 6.8% of White women (P = 0.43) (Table 5). Among invasive cancers however, Black women had a significantly higher proportion of higher grade cancers than White women; 61.7% had poorly differentiated tumors compared with 49.3% of White women (P < 0.001). Black women also had a higher proportion of cancers that were > 2.0 cm compared with White women, 57.4% vs.46.2%, (P = 0.015). In addition, among Black women, a significantly higher proportion of cancers were ER negative and PR negative when compared with tumors of White women.

Table 5. Characteristic of Cancers Following Diagnostic Mammograms for Black And White Women
Tumor characteristicsBlackWhiteChi-square
No.% of nonmissingaNo.% of nonmissingaP value
  • a

    All percentages are column percentages, based on the nonmissing data within an analysis subgroup.

  • b

    P value for Mantel–Haenszel statistic, based on all nonmissing categories.

  • c

    P value for Fisher exact test, based on positive and negative results only.

Total Grade3801001311100 
 Well differentiated167.212114.2< 0.001b
 Moderately differentiated6931.131036.5 
 Poorly differentiated13761.741949.3 
 Missing (% missing)158(41.6)461(35.2) 
Estrogen receptor status     
 Positive11157.848974.0< 0.001c
Not done/missing (% missing)188(49.5)650(49.6) 
Progesteron receptor status     
 Positive9650.843166.3< 0.001c
 Not done/missing (% missing)191(50.3)661(50.4) 
Tumor size     
 ≤ 10 mm5119.417118.80.015b
 11–20 mm6123.231835.0 
 21–30 mm7227.421423.5 
 > 30 mm7930.020622.7 
 Missing (% missing)117(30.8)402(30.7) 


  1. Top of page
  2. Abstract

The objective of this study was to compare performance of diagnostic mammography in Black women with White women from a large population-based mammography registry. At the time of diagnostic mammography exam, information collected included personal information about the women's medical and family histories, the results and recommendations of the imaging examinations performed, and cancer and benign outcomes. In comparing performance of diagnostic mammography, we found that mammograms of Black women, compared with White women, had higher sensitivity, lower specificity, equal PPV, and a greater proportion of larger, higher stage cancers. In our study, after controlling for previous mammography, age, breast density, and self-reported breast symptoms in the logistic models, there continued to be differences in performance of diagnostic mammography between Black and White women.

Black women also had a larger proportion of larger, higher grade, ER- and PR-tumors. There was a large amount of missing size and grade data for mammograms in both Black and White women; however, the proportion of missing data was similar for Black and White women. Other studies conducted in different populations have found similar associations of cancer characteristics comparing Black and White women. These studies also reported that cancers in Black women are larger, higher grade and have a higher proportion of ER and PR negativity when compared with White women.10, 17, 18

In our previous study of screening mammography, we found no differences in screening performance when comparing mammography between Black and White women. We did find similar but smaller differences between Black and White women in tumor characteristics such as size, stage, ER status, and PR status.5 A few studies have reported performance of diagnostic mammography but not for Black and White women separately.1, 2 These studies were undertaken in predominantly White or mixed non-Black populations. Performance estimates of diagnostic mammography in these studies are similar to performance measures of diagnostic mammography in White women in our study.

Specificity was lower for mammography in Black women compared with White women. The lack of a previous mammogram, (29.2% for Black women, 16.5% for White women) results in more false positive results. Radiologists are more likely to work up a suspicious finding when no previous films are present.19–23

Sensitivity is higher in diagnostic mammography than in screening mammography.2, 24, 25 These differences are most likely related to greater presence of symptoms and larger size of tumors that are found in diagnostic mammography.2, 26 Presence of a lump raises the probability of a cancer diagnosis27 and is one explanation for the higher sensitivity in diagnostic mammography. Another explanation for the difference in sensitivity between screening and diagnostic mammography is the difference in absolute size of the tumor. Although there were differences in tumor size in Black and White women in our previous study on screening mammography, these differences did not result in an appreciable difference in the sensitivity of screening mammography of Black and White women.5 In diagnostic mammography, where the tumors that are found are larger in size (and are larger in Black women than in White women), the sensitivity of diagnostic mammography is greater for Black women than for White women.1

Whereas Black women had slightly larger proportions of all symptoms (lump, discharge, and symptoms overall), we believe it is tumor size that drives the sensitivity and cancer detection rates in the diagnostic setting. Fifty-seven percent of Black women had invasive tumors > 2 cm, compared with 46.2% of White women. The proportion who reported lump was very similar; however, the cancer detection rate among Black women with a lump was 72.4/1000 compared with 51.2/1000 for White women. Other research in a variety of populations has also shown the relation between tumor size, grade, higher sensitivity, and cancer detection rates.28–34 Although more Black women reported discharge (the values for Black women and White women were 5.1% and 3.3%, respectively), the absolute proportions are too small to account for the difference in performance measures. Cancer detection rates in the presence of discharge were much lower than for lump and similar to detection rates that were found when no specific problem was recorded for Black and White women. The few published studies on discharge and mammography performance were conducted on case series only.35, 36

We are left with the question of why Black women present with larger tumors. Is it because they have faster growing tumors or do Black women as a group wait longer to get evaluated for their symptoms? The literature suggests that there are several factors operating. In a few studies, it has been shown that larger proportions of Black women, compared with White women, do not have a previous screening mammogram at the time of their diagnostic mammogram or are less likely to have had a screening mammogram within the previous 3 years.30, 32, 33 In addition, Black women wait longer to be seen for symptoms.28, 29, 37 As the ER and PR have a higher proportion of negativity, it is possible that there is something different about the cancers at a genetic or molecular level, which has yet to be identified. The differential roles of biologic, cultural, and other factors in the risk for and detection of breast tumors need to be further investigated.

In conclusion, in studying the performance of diagnostic mammography needs when symptoms are present, we found differences in results for mammograms of Black and White women in a community-based mammography registry. Black women have larger, higher grade tumors that lead to higher sensitivity, higher cancer detection rates, and lower specificity. Biologic contributions cannot be ruled out. In the short term, Black women need to be encouraged to respond more quickly when discovering breast signs or symptoms that need to be evaluated. Screening has been increasing among Black women, and this trend may reduce the disparities seen in this study.


  1. Top of page
  2. Abstract
  • 1
    Barlow WE, Lehman CD, Zheng Y, Ballard-Barbash R, Yankaskas BC, Cutter GR, et al. Performance of diagnostic mammography for women with signs or symptoms of breast cancer. J Natl Cancer Inst. 2002; 94: 11511159.
  • 2
    Poplack SP, Tosteson AN, Grove MR, Wells WA, Carney PA. Mammography in 53,803 women from the New Hampshire mammography network. Radiology. 2000; 217: 832840.
  • 3
    Robertson CL. A private breast imaging practice: medical audit of 25,788 screening and 1,077 diagnostic examinations. Radiology. 1993; 187: 7579.
  • 4
    Dee KE, Sickles EA. Medical audit of diagnostic mammography examinations: comparison with screening outcomes obtained concurrently. AJR Am J Roentgenol. 2001; 176: 729733.
  • 5
    Gill KS, Yankaskas BC. Screening mammography performance and cancer detection among black women and white women in community practice. Cancer. 2004; 100: 139148.
  • 6
    Yankaskas BC, Jones MB, Aldrich TE. The Carolina Mammography Registry: A population-based mammography and cancer surveillance project. J Registry Manag. 1996; 23: 173178.
  • 7
    D'Orsi CJ, Bassett L, Feig SA, et al. editors. Breast Imaging Reporting and Database System. Reston: American College of Radiology, 1998.
  • 8
    Kwok RK, Yankaskas B. The use of census data for determining race and education as SES indicators: a validation study. AEP. 2001; 11: 171177.
  • 9
    Carney PA, Miglioretti DL, Yankaskas BC, Kerlikowske K, Rosenberg R, Rutter CM, et al. Individual and combined effects of age, breast density, and hormone replacement therapy use on the accuracy of screening mammography. Ann Intern Med. 2003; 138: 168175.
  • 10
    Elmore JG, Wells CK, Howard DH, Feinstein AR. The impact of clinical history on mammographic interpretations. JAMA. 1997; 277: 4952.
  • 11
    Kerlikowske K, Grady D, Barclay J, Sickles EA, Ernster V. Effect of age, breast density, and family history on the sensitivity of first screening mammography. JAMA. 1996; 276: 3338.
  • 12
    Kerlikowske K, Grady D, Barclay J, Sickles EA, Eaton A, Ernster V. Positive predictive value of screening mammography by age and family history of breast cancer. JAMA. 1993; 270: 24442450.
  • 13
    Rosenberg RD, Hunt WC, Williamson MR, Gilliland FD, Wiest PW, Kelsey CA, et al. Effects of age, breast density, ethnicity, and estrogen replacement therapy on screening mammographic sensitivity and cancer stage at diagnosis: review of 183,134 screening mammograms in Albuquerque, New Mexico. Radiology. 1998; 209: 511518.
  • 14
    Slanetz PJ, Giardino AA, McCarthy KA, Hall DA, Halpern EF, Moore RH, et al. Previous breast biopsy for benign disease rarely complicates or alters interpretation on screening mammography. AJR Am J Roentgenol. 1998; 170: 15391541.
  • 15
    Brenner RJ, Pfaff JM. Mammographic changes after excisional breast biopsy for benign disease. AJR Am J Roentgenol. 1996; 167: 10471052.
  • 16
    Stokes ME, Davis CS, Koch GS. Categorical data analysis using the SAS system. 2nd ed. Cary, NC: SAS Institute Inc., 2000.
  • 17
    Furberg H, Millikan R, Dressler L, Newman B, Geradts J. Tumor characteristics in African American and white women. Breast Cancer Res Treat. 2001; 68: 3343.
  • 18
    Elledge RM, Clark GM, Chamness GC, Osborne CK. Tumor biologic factors and breast cancer prognosis among white, Hispanic, and black women in the United States. J Natl Cancer Inst. 1994; 86: 705712.
  • 19
    Kavanagh AM, Giles GG, Mitchell H, Cawson JN. The sensitivity, specificity, and positive predictive value of screening mammography and symptomatic status. J Med Screen. 2000; 7: 105110.
  • 20
    Kleit AN, Ruiz JF. False positive mammograms and detection controlled estimation. Health Serv Res. 2003; 38: 12071228.
  • 21
    Callaway MP, Boggis CRM, Astley SA, Hunt I. The influence of previous films on screening mammographic interpretation and detection of breast carcinoma. Clin Radiol. 1997; 52: 527529.
  • 22
    Thurfjell MG, Vitak B, Azavedo E, Svane G, Thurfjell E. Effect on sensitivity and specificity of mammography screening with or without comparison of old mammograms. Acta Radiologica. 2000; 41: 5256.
  • 23
    Burnside ES, Sickles EA, Sohlich RE, Dee KE. Differential value of comparison with previous examinations in diagnostic versus screening mammography. AJR Am J Roentgenol. 2002; 179: 11731177.
  • 24
    Duijm LE, Guit GL, Zaat JO, Koomen AR, Willebrand D. Sensitivity, specificity and predictive values of breast imaging in the detection of cancer. Br J Cancer. 1997; 76: 377381.
  • 25
    Eltahir A, Jibril JA, Squair J, Heys SD, Ah-See AK, Needham G, et al. The accuracy of “one-stop” diagnosis for 1,110 patients presenting to a symptomatic breast clinic. J R Coll Surg Edinb. 1999; 44: 226230.
  • 26
    Libstug AR, Moravan V, Aitken SE. Results from the Ontario breast screening program, 1990–1995. J Med Screen. 1998; 5: 7380.
  • 27
    Aiello EJ, Buist DS, White E, Seger D, Taplin SH. Rate of breast cancer diagnoses among postmenopausal women with self-reported breast symptoms. J Am Board Fam Pract. 2004; 17: 408415.
  • 28
    Coates RJ, Bransfield DD, Wesley M, Hankey B, Eley JW, Greenberg RS, et al. Differences between black and white women with breast cancer in time from symptom recognition to medical consultation. Black/White Cancer Survival Study Group. J Natl Cancer Inst. 1992; 84: 938950.
  • 29
    Vernon SW, Tilley BC, Neale AV, Steinfeldt L. Ethnicity, survival, and delay in seeking treatment for symptoms of breast cancer. Cancer. 1985; 55: 15631571.
  • 30
    Jones BA, Patterson EA, Calvocoressi L. Mammography screening in African American women: evaluating the research. Cancer. 2003; 91(Suppl 1): 258272.
  • 31
    Hunter CP, Redmond CK, Chen VW, Austin DF, Greenberg RS, Correa P, et al. Breast cancer: factors associated with stage at diagnosis in black and white women. Black/White Cancer Survival Study Group. J Natl Cancer Inst. 1993; 85: 11291137.
  • 32
    Jones BA, Kasl SV, Curnen MG, Owens PH, Dubrow R. Can mammography screening explain the race difference in stage at diagnosis of breast cancer? Cancer. 1995; 75: 21032113.
  • 33
    McCarthy EP, Burns RB, Coughlin SS, Freund KM, Rice J, Marwill SL, et al. Mammography use helps to explain differences in breast cancer stage at diagnosis between older black and white women[comment]. Ann Intern Med. 1998; 128: 729736.
  • 34
    Gwyn K, Bondy ML, Cohen DS, Lund MJ, Liff JM, Flagg EW, et al. Racial differences in diagnosis, treatment, and clinical delays in a population-based study of patients with newly diagnosed breast carcinoma. Cancer. 2004; 100: 15951604.
  • 35
    Simmons R, Adamovich T, Brennan M, Christos P, Schultz M, Eisen C, et al. Nonsurgical evaluation of pathologic nipple discharge. Ann Surg Oncol. 2003; 10: 113116.
  • 36
    Nakahara H, Namba K, Watanabe R, Furusawa H, Matsu T, Akiyama F, et al. A comparison of MR imaging, galactography and ultrasonography in patients with nipple discharge. Breast Cancer. 2003; 10: 320329.
  • 37
    Olivotto IA, Gomi A, Bancej C, Brisson J, Tonita J, Kan L, et al. Influence of delay to diagnosis on prognostic indicators of screen-detected breast carcinoma. Cancer. 2002; 94: 21432150.