| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Breast Imaging |
1 From the Departments of Radiology (S.P.P.), Community and Family Medicine (A.N.T., M.R.G., P.A.C.), Medicine (A.N.T.), and Pathology (W.A.W.), Dartmouth-Hitchcock Medical Center, 1 Medical Center Dr, HB 7999, Lebanon, NH 03756. From the 1999 RSNA scientific assembly. Received August 13, 1999; revision requested September 29; final revision received May 1, 2000; accepted May 22. Supported by grant DASD17-94-J-4109 from the United States Department of Defense and Cancer Center Support Grant 2P30 CA23108-22 from the National Cancer Institute. Address correspondence to S.P.P. (e-mail: steven.p.poplack@hitchcock.org).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: Mammographic data from 47,651 screening and 6,152 diagnostic examinations from November 1, 1996, to October 31, 1997, were linked to 1,572 pathologic results. Mammographic outcomes were based on BI-RADS assessments and recommendations reported by the interpreting radiologist. The consistency of BI-RADS recommendations was evaluated.
RESULTS: Screening mammography had a sensitivity of 72.4% (95% CI: 66.4%, 78.4%), specificity of 97.3% (95% CI: 97.25%, 97.4%), and positive predictive value of 10.6% (95% CI: 9.1%, 12.2%). Diagnostic mammography had higher sensitivity, 78.1% (95% CI: 71.9%, 84.3%); lower specificity, 89.3% (95% CI: 88.5%, 90.1%); and better positive predictive value, 17.1% (95% CI: 14.5%, 19.8%). The cancer detection rate with screening mammography was 3.3 per 1,000 women, with a biopsy yield of 22.4%, whereas the interval cancer rate was 1.2 per 1,000. Nearly 80% of screening-detected invasive malignancies were node negative. The recall rate for screening mammography was 8.3%. Ultrasonography was used in 3.5% of screening and 17.5% of diagnostic examinations. BI-RADS recommendations were generally consistent, except for probably benign assessments.
CONCLUSION: The sensitivity of screening mammography in this population-based sample is lower than expected, although other performance indicators are commendable. BI-RADS "probably benign" assessments are commonly misused.
Index terms: Breast neoplasms, radiography, 00.11, 00.30 • Cancer screening, 00.11, 00.30 • Diagnostic radiology, observer performance, 00.11, 00.30
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
The purpose of this study was to describe key performance measures of screening and diagnostic breast radiography in a geographically defined subject population and to evaluate the use of the American College of Radiology’s Breast Imaging Reporting and Data System (BI-RADS) (23) by interpreting radiologists. Our data are derived from a diverse group of mammography facilities, the majority of which are community based.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
All data contained in the NHMN database are scanned from standardized forms completed by women undergoing mammography, mammography technologists, and interpreting radiologists. The NHMN does not capture examinations limited to US. Women undergoing mammography provide demographic and some breast cancer risk factor information. Mammography technologists obtain additional risk and clinical information in a face-to-face interview with women undergoing mammography. During the pilot phase of NHMN development, test-retest reliability studies were conducted on all questions used in data collection for women, including information they provide to technologists during direct interviews. The test-retest results on final data collection forms were more than 90% reliable.
Radiologists record interpretive data by using BI-RADS terminology (23), including use of breast US, breast composition, assessment status, and recommendation for each breast. We created and distributed to participating radiologists a breast density atlas to assist and standardize coding of radiographic breast density. The atlas displays examples of borderline composition categories (fat vs scattered density vs heterogeneously dense vs extremely dense) and identifies correct density coding for each example. We also conducted quality assurance on interpretive data on 20 randomly selected cases from each facility by comparing data submitted to the NHMN by radiologists with the corresponding clinical text reports. Agreement between the NHMN project forms and the radiologists’ text reports was consistently greater than 96%.
Participating New Hampshire pathology laboratories send clinical pathology reports on all breast specimens, including fine-needle aspiration biopsy, core needle biopsy including the advance breast biopsy instrumentation, excisional biopsy, lumpectomy, and mastectomy, to the NHMN project office. These are abstracted and entered into a separate pathology database. The most serious pathologic outcome is applied when there are multiple pathologic results for the same breast, except when a suspicious-looking cytologic specimen precedes a benign histologic specimen. Linkages between the mammography and pathology databases are performed approximately every 6 months by using a probability-based matching program (INTEGRITY; Vality Technology, Boston, Mass) with demonstrated effectiveness.
Study Population
Mammographic examinations performed between November 1, 1996, and October 31, 1997, were eligible for inclusion in these analyses. During this period, 95 radiologists representing 20 radiology groups interpreted mammograms in 36 facilities (87.8%) in New Hampshire and contributed data. We excluded 5,482 women in whom mammography was performed in six of these mammography facilities because corresponding pathologic data were not available for these facilities. We also excluded 805 women who were missing interpretive assessments for both breasts. Mammographic data in 53,803 women were complete and met our inclusion criteria. These were linked with 1,572 benign and malignant pathologic results submitted by 82% (14 of 17) of the pathology laboratories in the state of New Hampshire. For 47,651 of these women, the initial indication for their examination was screening and for 6,152 women the initial indication was diagnostic.
We defined the nature of a mammographic examination on the basis of the presenting indication. We used a hierarchy of the following three independent data sources to identify screening indications: (a) technologist form—the woman undergoing mammography reported no current breast concerns (valid breast concerns were limited to lump, nipple discharge, and skin changes) and had no record in the NHMN database of a prior mammogram of any type within 270 days; (b) radiologist form—the type of examination was recorded as screening (asymptomatic) mammography by the interpreting radiologist; and (c) patient form—routine screening examination was selected as the indication for mammography. All other examinations not meeting these criteria were considered diagnostic. The evaluation of a valid clinical breast concern, as defined at the beginning of the paragraph, and short-term (<270 days) follow-up imaging were the primary diagnostic indications. Immediate supplementary imaging, defined as occurring within 45 days after the index screening examination, was not considered a diagnostic indication but was linked to the initial screening examination (Fig 1).
|
We analyzed the association of specific recommendations with final assessment categories for each woman. Multiple nonroutine recommendations were reported and may have included a less serious recommendation for the contralateral breast, since recommendations were not analyzed according to laterality. However, recommendations for routine follow-up, nonroutine follow-up, and the absence of a recommendation were considered mutually exclusive.
We linked indeterminate screening mammograms, defined as incomplete assessment and/or recommendation for or inclusion of immediate additional evaluation by means of mammography and/or US, with subsequent imaging that occurred within 45 days. All linked examinations were considered screening examinations because the initial indication was screening. The outcome of screening mammography reported here reflects the final assessment and recommendation status of associated examinations (Fig 1). An incomplete assessment status implies lack of resolution within 45 days after an indeterminate screening mammogram. We limited the time between imaging examinations to 45 days after an analysis of 338 examinations with initially incomplete results (category 0) revealed that more than 98% (172 of 175) of women who underwent supplementary imaging within 120 days underwent their examination within 45 days.
We defined the recall rate as the proportion of initial screening examinations assessed as incomplete and/or recommending or using additional imaging to arrive at a final assessment. This did not include definitively abnormal assessments (categories 4 and 5) or probably benign assessments (category 3) rendered solely on the basis of the initial screening examination.
We defined a positive cancer status as any tissue specimen, including malignant cytologic findings, revealing invasive carcinoma or ductal carcinoma in situ. We considered a malignant fine-needle aspiration biopsy outcome to reflect invasive carcinoma. We defined a negative cancer status as a benign result from tissue sampling and/or the absence of malignancy reported within the follow-up interval. Lobular carcinoma in situ, atypical epithelial proliferative disorders, and suspicious-looking cytologic findings without correlative histologic findings were considered benign in these analyses.
Statistical Methods
Summary statistics were used to describe patient and examination characteristics for screening and diagnostic mammograms separately. By using the mammographic outcome criteria and cancer status definitions described earlier, mammograms were linked with cancer outcomes to identify true-positive, true-negative, false-positive, and false-negative examinations. True-positive and false-positive status was defined as a positive mammographic interpretation with (true-positive result) or without (false-positive result) a cancer diagnosis reported within 365 days. A false-positive status was designated irrespective of whether a biopsy was performed. A true-negative result was a negative mammographic interpretation, including a probably benign assessment, with no report of cancer within the 365-day follow-up interval. Similarly, a false-negative result was defined as a negative mammographic interpretation with cancer diagnosed within the subsequent 365 days. On the basis of these classifications, sensitivity (true-positive/ [true-positive + false-negative]), specificity (true-negative/[false-positive + true-negative]), positive predictive value (true-positive/[true-positive + false-positive]), and negative predictive value (true-negative/[false-negative + true-negative]) were estimated.
Logistic regression, which was used to model the odds of a positive mammogram after controlling for cancer status, age less than 50 versus 50 or more years old, breast density (dense vs not dense), and history of a prior mammogram (yes vs no or unknown), was used to account for the influence of varying case mix on the operating characteristics of screening mammography. For purposes of analysis, "dense" breasts were categorized as those rated as heterogeneously dense or extremely dense, whereas breasts with fatty and scattered density were considered "not dense." To facilitate comparisons between sensitivity and specificity in the population in our study with those in other studies, sensitivity and specificity for women with particular characteristics were estimated by using the logistic regression model.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
|
Tables 2 and 3 list the frequency of final assessments with corresponding recommendations and cancer outcomes for both screening and diagnostic mammography. No recommendation accompanied the assessment in 0.5% (224 of 47,651) of women presenting for screening and in 0.8% (46 of 6,152) of women presenting for diagnostic mammography. The majority, 90.1% (n = 42,925), of screening mammograms were negative (categories 1 or 2), and 98.9% (n = 42,440) of negative screening mammograms had recommendations for routine follow-up. A smaller proportion, 68.7% (n = 4,227), of diagnostic mammograms were considered negative. Approximately 11.1% (n = 472) of negative diagnostic examinations had nonroutine recommendations. Assessments with results that were suspicious or highly suggestive of malignancy composed 1.8% (n = 842) of screening and 6.5% (n = 402) of diagnostic examinations. A recommendation for either biopsy or surgical consultation accompanied 78.6% (n = 602) of screening examinations with suspicious-looking results and 92.1% (n = 70) of those with results highly suggestive of malignancy. This pattern was also seen with diagnostic mammography.
|
|
Tables 2 and 3 also show the frequency of malignancy associated with specific assessment categories. As expected, the frequency of malignancy increases with the severity of the assessment code. Unresolved incomplete screening assessments had a malignancy rate similar to that of the probably benign category but were more highly associated with malignancy in women with diagnostic indications. Malignancy was present in less than 2% of the probably benign assessments, which is commensurate with published results (27,28).
Screening mammography helped detect malignancy in 3.3 per 1,000 women. Diagnostic mammography was used to identify cancer in 21.5 per 1,000 patients. Malignancy was diagnosed in 59 of 46,194 women after a negative screening examination and in 37 of 5,381 women with negative results at diagnostic mammography. The interval cancer rate was 1.2 per 1,000 women for screening and 6.0 per 1,000 for diagnostic mammography.
Table 4 outlines sensitivity, specificity, positive predictive value, and negative predictive value of screening and diagnostic mammography. Table 5 shows estimated sensitivity and specificity of screening mammography according to mammographic history, breast density, and age (less than 50 years vs 50 years or older). Similar results (not shown) were seen in the diagnostic mammography population.
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
We based mammographic outcome on the prospective report of the BI-RADS assessment and recommendation encoded by the interpreting radiologist. Although BI-RADS is useful for standardization, it is not always used correctly and does not always address complex imaging and clinical circumstances. Unlike the results of most other studies, the mammographic results in our study reflect the status of completed imaging evaluations. We classified a probably benign (category 3) assessment as a negative mammographic outcome. Almost half, 46% (27 of 59), of the interval cancers after screening mammography were assessed as probably benign. Had we classified probably benign mammograms as positive, our sensitivity and specificity for screening mammography would have been 84.6% and 90.4%, respectively.
We defined false-negative results (interval cancers) on the basis of the report of cancer outcomes in the 365 days after negative mammography. Thus, negative mammograms with no pathology report were considered as true-negative results, whereas positive mammograms with no pathology report were considered false-positive results. Because women with positive examinations are more likely to have aggressive follow-up, there is a potential for underestimating the false-negative rate and overestimating the true-negative rate. This phenomenon, known as verification bias (29), may lead to overestimation of sensitivity and underestimation of specificity. We have minimized this potential bias by obtaining cancer outcomes by using breast pathology reports from participating laboratories, as well as New Hampshire tumor registry data, and by excluding data from mammography centers associated with nonparticipating laboratories.
This study may provide a more comprehensive account of positive disease outcomes than have other studies (10, 13,15,16,21,30,31) that have estimated false-negative mammographic cancer outcomes or relied exclusively on tumor registry data. Our lower estimate of sensitivity compared with those of these other studies may have been the result of our more complete capture of cancer outcomes.
The timing of mammography with respect to clinical breast examination may also alter operating characteristics, especially sensitivity. Some studies (32–34) that reported lower interval cancer rates offered clinical breast examination at the time of screening mammography, which will decrease the interval cancer rate owing to coincidental detection by means of clinical breast examination of mammographically occult cancers.
These methodologic differences may help explain our sensitivity of 72.4% for screening mammography and corresponding interval cancer rate of 1.2 per 1,000 women. Although the sensitivity we report is within the range of sensitivities of 68%–88% (detection method) noted by Fletcher and colleagues (1) for seven randomized control trials, it is lower than those of 91%–93% in studies (10,11) from single expert centers. Our sensitivity estimate more closely approximates the rate of 79.9% for linked screening examinations noted with the New Mexico Mammography Project (21).
The interval cancer rate of 1.2 per 1,000 women in our study is also higher than that in other published studies (31,33–35). Interestingly, the tumor sizes and nodal status of the interval cancers in our study were relatively favorable, especially when compared with the staging characteristics of the malignancies identified at diagnostic mammography. We believe this may relate to the preponderance of prior mammography (92%) in the population undergoing screening mammography. We hypothesize that prior screening mammography may have been effective in leading to the extraction of larger tumors from that population, leaving smaller less detectable cancers available for discovery at the subsequent screening examination. This may also reflect a clinical decision to perform a biopsy in the setting of a probably benign mammogram or an initially indeterminate examination that resolved to a negative status (categories 1, 2, or 3) after supplementary imaging was completed.
The cancer detection rate of 3.3 cancers per 1,000 women who underwent screening mammography in our study is comparable with that in other studies (17,35–37), given the age distribution and history of prior mammography in the population in our study. One would expect to detect two to four cancers per 1,000 women at annual follow-up screening mammography and six to 10 cancers per 1,000 women at baseline screening mammography (38).
The characteristics of the screening-detected cancers in our study compare favorably with those of other studies (3,10,11,21,37–39). Roughly 21% of our screening-detected cancers were ductal carcinoma in situ, which is within the range of 19%–27% from prior North American studies (10,11,21,37–39). Mean and median tumor sizes of the invasive cancers in our study were equal or smaller (10,11,21,37–39). The rate of axillary nodal metastases of 20% for invasive malignancy is also comparable (11, 21,39), given that studies (10,38) reporting lower axillary node positivity rates have included ductal carcinoma in situ.
Other measures of screening mammographic performance, including specificity (97.3%), positive predictive value (10.6%), and recall rate (8.3%), meet the standards of the Agency of Health Care Policy and Research (14). We recognize that these estimates are somewhat inflated by our decision to base mammographic outcome on a completed imaging work-up, to consider probably benign assessment and short-term follow-up to reflect a negative interpretive outcome, and to define recall rate as we did. We defined recall rate according to the guidelines described by Linver et al (40), which differ from more inclusive abnormality rates reported by other investigators (10,19,21).
In addition to the traditional performance indicators described earlier, we also evaluated the use of BI-RADS by our interpreting physicians. It was reassuring to note that an appropriate recommendation followed the BI-RADS assessment most frequently, but there were a small number of inappropriate recommendations for all assessments for both screening and diagnostic mammography. Gross misapplications of BI-RADS, such as a recommendation for biopsy or surgical consultation in the setting of a negative or benign assessment or a recommendation for routine follow-up in the setting of an assessment that was suspicious or highly suggestive of malignancy, occurred rarely.
Although we suspect that some inappropriate recommendations represent coding errors, misclassification of screening and diagnostic indications, indecisiveness resulting in multiple recommendations, or additional nonroutine recommendations for the contralateral breast, a detailed analysis of these cases is beyond the scope of this article. Some discordance may reflect the difficulty of applying a rigid coding system to a complex and sometimes ambiguous set of clinical management alternatives. However, to some extent, the presence of discordant recommendations may indicate a lack of understanding of BI-RADS by some interpreting radiologists.
Discordant recommendations were especially evident for mammograms interpreted as probably benign, which were associated with a considerable number of routine follow-up recommendations (22%, screening; 21%, diagnostic) and a higher-than-expected rate of immediate additional imaging (14%, screening; 11%, diagnostic), predominantly US (11%, screening; 10%, diagnostic). In these instances, the interpreters appear to have misclassified benign and incomplete assessments as probably benign; this underscores the need for training mammographers in the use of BI-RADS, especially as it relates to the appropriate classification and corresponding recommendations of benign, probably benign, and incomplete assessments.
Although one of the strengths of this study includes the collection of standardized data from a diverse group of mammography facilities, we rely on many individuals to provide correctly coded data instruments, to oversee data entry, and to meticulously manage a large database. We acknowledge that despite extensive quality assurance measures, human error, including misclassification and incorrect or incomplete coding is a concern.
Another concern is the composition of our subject population. Approximately 98% of our study population is white, which is similar in ethnicity to other population-based mammography databases in the Northeast and the Northwest (41,42), but differs in ethnic distribution compared with study populations in mammography databases in other regions of the country (21,36).
In conclusion, our data suggest that the sensitivity of screening mammography (72.4%) is lower than generally believed, although other indicators of interpretive performance, including cancer detection rate, specificity, positive predictive value for a completed imaging work-up, recall rate, and the characteristics of screening-detected cancers, satisfy or exceed standards. Part, but not all, of the reduction in sensitivity can be explained by the preponderance of prior mammographic screening in the population in our study. We also learned that roughly 8% of women presenting for screening mammography had indeterminate examination results necessitating supplementary imaging, which included US 23% of the time. Approximately 90% of screening mammograms were considered negative or definitively benign; 7%, probably benign; 2%, suspicious or highly suggestive of malignancy; and 1%, indeterminate. Appropriate recommendations followed these assessment categories most of the time, although in the setting of a probably benign finding (category 3), recommendations frequently were misapplied. Further education of radiologists in the intended use of the BI-RADS lexicon may help address this problem.
|
FOOTNOTES |
|---|
Author contributions: Guarantor of integrity of entire study, S.P.P.; study concepts, S.P.P., A.N.T., P.A.C., W.A.W.; study design, S.P.P., A.N.T., P.A.C.; definition of intellectual content, S.P.P., A.N.T., P.A.C.; literature research, S.P.P., A.N.T., P.A.C.; clinical studies, S.P.P., P.A.C., W.A.W.; experimental studies, S.P.P., P.A.C., A.N.T., M.R.G.; data acquisition, P.A.C., M.R.G.; data analysis, M.R.G., A.N.T.; statistical analysis, M.R.G., A.N.T.; manuscript preparation and editing, S.P.P., A.N.T., M.R.G., P.A.C.; manuscript review, all authors.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  S. Taplin, L. Abraham, W. E. Barlow, J. J. Fenton, E. A. Berns, P. A. Carney, G. R. Cutter, E. A. Sickles, D. Carl, and J. G. Elmore Mammography Facility Characteristics Associated With Interpretive Accuracy of Screening Mammography J Natl Cancer Inst, June 18, 2008; 100(12): 876 - 887. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 W. F. Good, G. S. Abrams, V. J. Catullo, D. M. Chough, M. A. Ganott, C. M. Hakim, and D. Gur Digital Breast Tomosynthesis: A Pilot Observer Study Am. J. Roentgenol., April 1, 2008; 190(4): 865 - 869. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Gur Tomosynthesis: Potential Clinical Role in Breast Imaging Am. J. Roentgenol., September 1, 2007; 189(3): 614 - 615. [Full Text] [PDF]
|
|
|
|
|
|
S. P. Poplack, T. D. Tosteson, C. A. Kogel, and H. M. Nagy Digital Breast Tomosynthesis: Initial Experience in 98 Women with Abnormal Digital Screening Mammography Am. J. Roentgenol., September 1, 2007; 189(3): 616 - 623. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. P. Poplack, T. D. Tosteson, W. A. Wells, B. W. Pogue, P. M. Meaney, A. Hartov, C. A. Kogel, S. K. Soho, J. J. Gibson, and K. D. Paulsen Electromagnetic Breast Imaging: Results of a Pilot Study in Women with Abnormal Mammograms Radiology, May 1, 2007; 243(2): 350 - 359. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
P. A. Carney, L. A. Abraham, D. L. Miglioretti, K. R. Yabroff, E. A. Sickles, D. S. M. Buist, C. J. Kasales, B. M. Geller, R. D. Rosenberg, M. B. Dignan, et al. Factors Associated with Imaging and Procedural Events Used to Detect Breast Cancer After Screening Mammography Am. J. Roentgenol., February 1, 2007; 188(2): 385 - 392. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. A. Carney, E. Steiner, M. E. Goodrich, A. J. Dietrich, C. J. Kasales, J. E. Weiss, and T. MacKenzie Discovery of Breast Cancers Within 1 Year of a Normal Screening Mammogram: How Are They Found? Ann. Fam. Med, November 1, 2006; 4(6): 512 - 518. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. D. Rosenberg, B. C. Yankaskas, L. A. Abraham, E. A. Sickles, C. D. Lehman, B. M. Geller, P. A. Carney, K. Kerlikowske, D. S. M. Buist, D. L. Weaver, et al. Performance Benchmarks for Screening Mammography Radiology, October 1, 2006; 241(1): 55 - 66. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. M. Eberl, C. H. Fox, S. B. Edge, C. A. Carter, and M. C. Mahoney BI-RADS Classification for Management of Abnormal Mammograms J Am Board Fam Med, March 1, 2006; 19(2): 161 - 164. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 E. D. Pisano, C. Gatsonis, E. Hendrick, M. Yaffe, J. K. Baum, S. Acharyya, E. F. Conant, L. L. Fajardo, L. Bassett, C. D'Orsi, et al. Diagnostic Performance of Digital versus Film Mammography for Breast-Cancer Screening N. Engl. J. Med., October 27, 2005; 353(17): 1773 - 1783. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Q. Zhu, E. B. Cronin, A. A. Currier, H. S. Vine, M. Huang, N. Chen, and C. Xu Benign versus Malignant Breast Masses: Optical Differentiation with US-guided Optical Imaging Reconstruction Radiology, October 1, 2005; 237(1): 57 - 66. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 D. L. Weaver, P. M. Vacek, J. M. Skelly, and B. M. Geller Predicting Biopsy Outcome After Mammography: What Is the Likelihood the Patient Has Invasive or In Situ Breast Cancer? Ann. Surg. Oncol., August 1, 2005; 12(8): 660 - 673. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. E. Hendrick, G. R. Cutter, E. A. Berns, C. Nakano, J. Egger, P. A. Carney, L. Abraham, S. H. Taplin, C. J. D'Orsi, W. Barlow, et al. Community-Based Mammography Practice: Services, Charges, and Interpretation Methods Am. J. Roentgenol., February 1, 2005; 184(2): 433 - 438. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. P. Poplack, P. A. Carney, J. E. Weiss, L. Titus-Ernstoff, M. E. Goodrich, and A. N. A. Tosteson Screening Mammography: Costs and Use of Screening-related Services Radiology, January 1, 2005; 234(1): 79 - 85. [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]
|
|
|
|
|
|
S. P. Poplack, K. D. Paulsen, A. Hartov, P. M. Meaney, B. W. Pogue, T. D. Tosteson, M. R. Grove, S. K. Soho, and W. A. Wells Electromagnetic Breast Imaging: Average Tissue Property Values in Women with Negative Clinical Findings Radiology, May 1, 2004; 231(2): 571 - 580. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. A. Roubidoux, J. E. Bailey, L. A. Wray, and M. A. Helvie Invasive Cancers Detected after Breast Cancer Screening Yielded a Negative Result: Relationship of Mammographic Density to Tumor Prognostic Factors Radiology, January 1, 2004; 230(1): 42 - 48. [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]
|
|
|
|
|
|
J. M. Lewin, C. J. D'Orsi, R. E. Hendrick, L. J. Moss, P. K. Isaacs, A. Karellas, and G. R. Cutter Clinical Comparison of Full-Field Digital Mammography and Screen-Film Mammography for Detection of Breast Cancer Am. J. Roentgenol., September 1, 2002; 179(3): 671 - 677. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. E. Barlow, C. D. Lehman, Y. Zheng, R. Ballard-Barbash, B. C. Yankaskas, G. R. Cutter, P. A. Carney, B. M. Geller, R. Rosenberg, K. Kerlikowske, et al. Performance of Diagnostic Mammography for Women With Signs or Symptoms of Breast Cancer J Natl Cancer Inst, August 7, 2002; 94(15): 1151 - 1159. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. L. Rosen, J. A. Baker, and M. S. Soo Malignant Lesions Initially Subjected to Short-term Mammographic Follow-up Radiology, April 1, 2002; 223(1): 221 - 228. [Abstract] [Ful |
