| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Breast Imaging |
1 From the Depts of Medicine and Epidemiology and Biostatistics (K.K.) and Radiology (R.S.B., E.A.S.), Univ of California, San Francisco; General Internal Medicine Section, San Francisco Veterans Affairs Medical Ctr, 111A1, 4150 Clement St, San Francisco, CA 94121 (K.K.); Ctr for Health Studies, Group Health Cooperative, Seattle, Wash (L.A., C.L., W.E.B.); Dept of Radiology, Univ of Washington Medical Ctr, Seattle Cancer Care Alliance, Seattle, Wash (C.L.); Dept of Radiology, Univ of North Carolina, Chapel Hill (B.C.Y.); Applied Research Program, DCCPS, National Cancer Institute, Bethesda, Md (R.B.B.); Dept of Biostatistics, Univ of Washington, Seattle (W.E.B.); Cancer Research and Biostatistics, Seattle, Wash (W.E.B.); Ctr for Research Methods and Biometry, The Cooper Institute, Denver, Colo (J.H.V.); Health Promotion Research, Univ of Vermont, College of Medicine, Burlington (B.M.G.); and Norris Cotton Cancer Ctr/Dartmouth-Hitchcock Medical Ctr/Dept of Community and Family Medicine, Dartmouth Medical School, Lebanon, NH (P.A.C.). Received Dec 17, 2003; revision requested Feb 23, 2004; final revision received May 11; accepted Jun 18. Supported by a NCI-funded Breast Cancer Surveillance Consortium cooperative agreement (U01CA63740, U01CA86076, U01CA86082, U01CA63736, U01CA70013, U01CA63731, U01CA70040). Address correspondence to K.K. (e-mail: kerliko@itsa.ucsf.edu).
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
MATERIALS AND METHODS: From January 1996 to December 1999, Breast Imaging Reporting and Data System assessments and recommendations were collected prospectively for 1171792 screening examinations in 758 015 women aged 40–89 years at seven mammography registries in Breast Cancer Surveillance Consortium. Registries obtained waiver of signed consent or collected signed consent in accordance with institutional review boards at each location. Diagnosis of invasive cancer or ductal carcinoma in situ within 24 months of screening examination and tumor stage and size for invasive cancer were determined through linkage to pathology database or tumor registry. 
2 test was used to determine significant differences between groups.
RESULTS: Overall, 5.2% of first and 1.7% of subsequent screens included recommendation for short-interval follow-up, which was similar to likelihood of recommendation for diagnostic evaluation (first screens, 4.6%; subsequent, 2.6%). Most recommendations for short-interval follow-up were based on screening mammography alone (86.2% of first screens, 77.5% of subsequent). Yield of cancer for screening examinations with probably benign finding (PBF) and recommendation for short-interval follow-up based on screening mammography alone tended to be lower than in those with PBF and recommendation for short-interval follow-up after additional work-up (first screens: 0.54% vs 0.96%, P = .10; subsequent: 1.50% vs 1.73%, P = .26). Proportion of stage II and higher disease tended to be higher for examinations with PBF and recommendation for short-interval follow-up based on screening mammography alone compared with those recommended for short-interval follow-up after additional work-up (first screens: 34.7% vs 24.4%, P = .43; subsequent: 27.5% vs 19.2%, P = .13).
CONCLUSION: Many first screening examinations include recommendation for short-interval follow-up based on screening mammography alone. Cancer yield for these examinations is low and is lower than that with diagnostic work-up prior to short-interval follow-up recommendation. Absence of diagnostic work-up prior to short-interval follow-up recommendation may result in periodic surveillance of a high proportion of benign lesions.
© RSNA, 2005
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
The conventional approach has been for radiologists to recommend short-interval follow-up examinations to monitor PBFs only after a full diagnostic imaging work-up has been performed. This is done to establish that the risk of cancer for a PBF is sufficiently low that immediate biopsy is not warranted. The yield of cancer for these fully imaged PBFs is low (0.3%–1.8%) (4–12). Others may perform periodic short-interval follow-up examinations on the basis of screening mammography alone without further characterization of what appears to be a PBF (1). The yield of cancer for this latter group of PBFs is unknown.
The purpose of this study was to compare the cancer yield for screening examinations recommended for short-interval follow-up after a diagnostic imaging work-up versus after screening mammography only.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
These registries collect information on screening and diagnostic mammographic examinations performed in their defined catchment area. Each mammography registry links women in their registry to a state tumor registry or regional Surveillance, Epidemiology, and End Results program that collects population-based cancer data. Five of seven registries additionally link to pathology databases. Each registry obtains annual approval from its institutional review board to collect registry information for pooled analyses across the seven registries. Registries obtained a waiver of signed consent or collected signed consent per the institutional review board at each location. Linkage procedures are performed in accordance with human subject protocols to maintain participant confidentiality.
Subjects
The study sample included women aged 40–89 years who underwent screening mammography between January 1996 and December 1999. Screening examinations that occurred after December 1999 were excluded to ensure (a) at least 24 months for reporting of cancers to tumor registries after screening examinations and (b) 95% or higher completeness of cancer reporting. A screening mammographic examination was defined as a bilateral two-view craniocaudal and mediolateral oblique examination indicated by the radiologist as being performed for screening.
To minimize the inclusion of diagnostic examinations, examinations were excluded if they were obtained within 9 months of a screening examination. In other words, women that underwent more than one screening examination could have results of both examinations included in the analysis, as long as the examinations were more than 9 months apart.
From a total of 1 370 902 examinations among 881 640 women, examinations were excluded if a woman had self-reported breast augmentation (n = 11 233 [0.82%]) or prior diagnosis of breast cancer (n = 38 278 [2.8%]). Examination results were also excluded if the time interval since last mammographic examination was missing (n = 93 581 [6.8%]), if a Breast Imaging Reporting and Data System (BI-RADS) (14) assessment (n = 2482 [0.18%]) or management recommendation (n = 12 505 [0.91%]) was missing, or if the assessment and associated management recommendation could not be classified into one of the groups of interest (n = 37 150 [2.7%]). Among women with a final recommendation for short-interval follow-up, data from all subsequent examinations within 24 months (n = 3802 [0.28%]) were excluded, since these examinations were assumed to be part of periodic surveillance of the PBF.
Examinations with a final recommendation of short-interval follow-up within 24–36 months of a prior examination with the same recommendation were also excluded (n = 79 [0.006%]). This was based on the assumption that these examinations were part of a 3-year rather than a 2-year surveillance protocol for a PBF. The final study sample included 1 171 792 screening examinations in 758 015 women aged 40–89 years (mean age, 57 years).
Measurements and Definitions
At the time of each screening examination, demographic information and a self-reported breast health history that includes questions about previous mammographic examinations were obtained. Time between mammographic examinations was determined by using dates of prior mammographic examinations recorded in each mammography registry and self-reported information collected at the screening examination. When dates of prior mammographic examinations and self-reported information were available, actual dates were used whenever possible. Screening examinations were considered "first examinations" if a woman reported no history of prior mammography or if time between mammographic examinations was 5 years or more. Screening examinations were considered "subsequent examinations" if time between screening examinations was less than 5 years.
Assessments based on screening mammography only were defined as the initial assessment reported by the radiologist. Initial assessments at screening that were associated with additional imaging on the same date or within 3 months of the screening examination were classified as "additional imaging evaluation needed" (BI-RADS score 0) for analyses with cancer outcomes. For final assessments and recommendations, additional imaging performed within 3 months of the screening examination was taken into account. If additional imaging occurred within 3 months of the screening examination, the assessment and recommendation associated with the additional imaging became the final assessment and recommendation. If there was no additional imaging, only the initial assessment and recommendation were reported. When more than one follow-up test was recommended (18 829 of 1 171 792 [1.6%] assessments), we chose the most immediate or invasive follow-up test.
We stratified the data into seven groups according to BI-RADS (14) initial assessment, final assessment, and final management recommendation for follow-up, respectively: group 1, negative or benign findings, no additional imaging evaluation performed, and routine follow-up; group 2, PBF, no additional imaging evaluation performed, and routine follow-up; group 3, PBF, no additional imaging evaluation performed, and short-interval follow-up; group 4, PBF, no additional imaging evaluation per-formed, and diagnostic work-up; group 5, additional imaging evaluation needed, negative or benign findings at final assessment, and routine follow-up; group 6, additional imaging evaluation needed, PBF at final assessment, and short-interval follow-up (hereafter referred to as conventional PBF assessment); and group 7, additional imaging evaluation needed, and suspicious abnormality or abnormality highly suggestive of malignancy at final assessment and diagnostic work-up.
We combined examinations that had similar diagnostic evaluations and management recommendations with one of the groups noted above. (a) Examinations with an initial assessment of PBF, additional imaging performed on the same day or within 90 days, and final assessment of PBF and recommendation for routine follow-up were included in group 5. (b) Examinations with an initial assessment of PBF, additional imaging performed on the same day or within 90 days, and a final assessment of PBF and recommendation for short-interval follow-up were included in group 6 (conventional PBF assessment). (c) Examinations with an initial assessment of PBF, additional imaging performed on the same day or within 90 days, and a final assessment of additional imaging evaluation needed, suspicious abnormality, or abnormality highly suggestive of malignancy and final recommendation for diagnostic work-up were included in group 7.
Women were considered to have breast cancer if reports from a breast pathology database; Surveillance, Epidemiology, and End Results program; or state tumor registry showed any invasive carcinoma or ductal carcinoma in situ within 24 months of a screening examination. If a woman with a diagnosis of cancer underwent more than one screening examination during the study period, the diagnosis was associated with the most recent screening examination performed before the diagnosis was assigned. Women with only lobular carcinoma in situ were considered not to have cancer. All breast cancers were classified according to the American Joint Committee on Cancer staging system (15). Invasive cancers were categorized by tumor size, grade, and lymph node status.
Statistical Analysis
All analyses were performed by using SAS statistical software, version 8.0 (SAS Institute, Cary, NC). Data were analyzed for first and subsequent screening examinations separately because the rates of positive examination findings and cancer are lower among women who have undergone screening mammography previously compared with those who have never undergone mammography (16).
Frequency distributions of examinations according to assessment and recommendation were determined for seven analytic groups. The yield of cancer was calculated for all seven groups, while the distribution of tumor size (
10 mm and >10 mm) and stage (0 and I vs II, III, and IV) were determined for groups 2, 3, 4, and 6. The age-adjusted cancer yield, tumor size, and stage distribution for each assessment according to recommendation category were computed by using a weighted average, where the weights were calculated from the age distribution in a referent population. The referent population was determined on the basis of the type of screening examination performed and the outcome of interest. The referent populations for the age-adjusted analyses were screening examinations for calculation of cancer yield, invasive cancers for calculation of tumor size distribution, and breast cancers for calculation of stage distribution. In addition, the referent group was defined according to type of mammography (initial or subsequent). Cancer statistics were adjusted for age so that comparisons between groups were for a similar age distribution and differences seen were not confounded by age. Confidence intervals for the age-adjusted proportions and comparisons between these proportions were computed by using a normal approximation to the binomial distribution.
Frequency distributions of cancers according to time to detection (<12 months vs 12–24 months) and tumor stage (0 and I vs II, III, and IV), lymph node status (positive vs negative), and tumor grade (I and II vs III and IV) were determined for groups 2, 3, 4, and 6 for subsequent examinations only. 2 tests were computed by using the FREQ procedure in the software (SAS Institute). These tests provided P values that were used to determine whether differences between groups were statistically significant. A two-sided P value of less than .05 was considered to indicate a statistically significant difference for all analyses.
Age-adjusted Cox proportional hazards models were fit by using the PHREG procedure in the software (SAS Institute). The response of interest for these models was time to cancer diagnosis. The models provided survivor function estimates that were averaged for each assessment according to recommendation category and time to cancer diagnosis and were weighted according to age distribution. These weighted estimates were plotted to determine whether differences occurred in the time to cancer detection according to assessment and recommendation categories.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
Table 1 shows assessments reported by radiologists on the basis of screening mammography only. The proportion of screening examinations with an initial assessment of PBF decreased from 7.0% for first screens to 2.9% for subsequent screens (Table 1). Overall, 5.2% of first screens and 1.7% of subsequent screens included a recommendation for short-interval follow-up. This was similar to the percentage of screens with a recommendation for diagnostic evaluation (first screens, 4.6%; subsequent screens, 2.6%).
|
|
Yield of cancer varied for the three types of PBF assessment assigned on the basis of screening mammography only compared with the cancer yield for PBF assessments assigned after full imaging work-up (conventional group). Yield of cancer for PBF assessments that followed screening mammography only and included recommendation for routine follow-up was low and was much lower than PBF assessments assigned after a full imaging work-up (subsequent screens: age-adjusted, 0.49% vs 1.73%, P < .001) (Table 2). Yield of cancer for conventional PBF assessments tended to be higher than that for PBF assessments assigned after screening mammography only with a recommendation for short-interval follow-up (first screens: 0.96% vs 0.54%, P = .10; subsequent screens: 1.73% vs 1.50%, P = .26) (Table 2), but the difference was not statistically significant. Yield of cancer for PBF assessments assigned after screening mammography only that included recommendation for diagnostic evaluation was similar to that of conventional PBF assessments (first screens: 1.29% vs 0.96%, P = .39; subsequent screens: 2.0% vs 1.73%, P = .22) (Table 2).
Tumor Size and Breast Cancer Stage
The proportion of invasive cancers larger than 10 mm tended to be higher for PBF assessments assigned after screening mammography that included a recommendation for short-interval follow-up than for conventional PBF assessments (first screen: age-adjusted, 80.0% vs 63.2%, P = .14; subsequent screen: age-adjusted, 59.4% vs 54.5%, P = .52) (Table 3), but the difference was not statistically significant.
|
|
|
|
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
As is consistent with previous studies (4–12), we found that the yield of cancer for a conventional PBF assessment at first screening examination is low (0.96%). However, the yield of cancer is even lower for lesions recommended for short-interval follow-up on the basis of screening alone (0.54%). This suggests that a high proportion of benign lesions is being monitored with periodic surveillance when full diagnostic work-up is not performed and that full diagnostic work-up appears to lead to identification of benign lesions that do not need short-interval surveillance. We also found a nonsignificant trend toward a higher proportion of advanced cancer in PBFs recommended for short-interval follow-up on the basis of screening alone compared with conventional PBF assessment.
Studies in the literature report that about 1%–2% of first screening examinations are followed by conventional PBF assessments; the 1.6% reported in the present study is consistent with these results (4–7,9–11,17). Studies that do not distinguish PBF assessments recommended for short-interval follow-up after screening mammography only from those recommended after immediate diagnostic work-up have reported a much higher (4%–7%) prevalence of PBF assessments (1–3). To our knowledge, our study is the first to report that a high proportion of PBF assessments for first screening examinations is based on assessments conducted according to screening results alone and that the risk of cancer is low for these examinations.
BI-RADS quotes a performance benchmark for a recall rate of 5%–10% (14), with recall examinations defined as screening examinations with an initial BI-RADS assessment of additional imaging evaluation needed, suspicious abnormality, or abnormality highly suggestive of malignancy. In BI-RADS, PBF assessments are considered to be normal assessments and are therefore not counted in the calculation of recall rate. One explanation for radiologists who assign a PBF assessment on the basis of screening alone might be the goal of keeping the recall rate at or below the benchmark ceiling of 10% (14). For example, the recall rate for first screening examinations in our study, taking into account diagnostic imaging that occurred within 90 days of the initial screen, was 8.6%. If all PBFs had been assessed as requiring additional imaging evaluation with the intent of performing diagnostic work-up prior to deciding whether to recommend short-interval follow-up, the recall rate for first screening examinations would have been 14.0%, which is high, given that the sensitivity of mammography increases little beyond a recall rate of 4.8% (18).
For several years, radiologists have debated the wisdom of conducting PBF assessments on the basis of screening mammography alone (17,19–22). The preponderance of published opinion favors conducting these assessments only after a full diagnostic imaging work-up has been completed (17,20–22). This is also the approach recommended by BI-RADS (14). The results of our study support two of the arguments that have been made in support of restricting PBF assessments to diagnostic examinations.
First, it has been proposed that many benign lesions will appear to be PBFs at screening mammography but will be identified as benign after diagnostic imaging (eg, cysts at ultrasonography [US], intramammary lymph nodes at fine-detail mammography or US, skin calcifications at tangential-view mammography). In our study, the yield of cancer for first-screening PBF assessments is only about half as high (0.54%) for PBF assessments assigned at screening alone with a recommendation for short-interval follow-up compared with conventional PBF assessments (0.96%). This suggests that many lesions that would have been identified as benign after diagnostic imaging are being recommended for short-interval follow-up. Thus, prompt diagnostic imaging will (a) eliminate the need for any follow-up at all for about half of the PBF cases on the basis of screening alone, (b) allow a diagnosis to be assigned promptly, and (c) presumably reduce anxiety that may accompany a recommendation for short-interval follow-up.
Second, it has been proposed that some cancers will appear to be PBFs at screening mammography but will be identified accurately as suspicious for malignancy at diagnostic mammography or US (eg, by displaying subtle irregularity in contour, indistinctness, or spiculation in margins). PBF assessments based on screening mammography only lead to follow-up instead of immediate biopsy for some cancers that would have been given a final assessment of suspicious abnormality had full imaging work-up been done. This would have the effect of including some cancers in PBF assessments assigned at screening only that would not have been included with the conventional PBF assessments. This both inflates the cancer yield of PBF assessments assigned after screening only and could lead to tumors with a larger size and higher stage among this group. In other words, completion of a full diagnostic imaging work-up may reduce the number of cancers for which follow-up is recommended and permit earlier diagnosis in some cases. Results from our study indicate a trend toward more advanced cancers (larger size, higher stage) among patients with PBFs who undergo screening alone compared with the cohort involving conventional PBF assessment.
There are two other arguments in favor of assigning PBF assessments only after a full diagnostic imaging work-up has been completed, for which our study provides no guidance. One is that additional diagnostic images provide the baseline from which stability of the PBF is best assessed during subsequent follow-up imaging. Although conventional screening mammographic images are obtained routinely at follow-up, these images sometimes suggest but do not indicate clearly the presence of interval progression. In such borderline cases, fine-detail mammography or US may well permit a more definitive comparison, but only if similar images were obtained at the outset. It is difficult to assess stability versus interval change when current fine-detail mammograms and sonograms are compared with screening mammographic images obtained previously.
The second argument applies to mammography practices in which screening examinations are interpreted in batches after women have left a mammography facility. In this situation, the recommendation for short-interval follow-up is conveyed to the woman by a letter with boilerplate text sent by the mammography facility and by means of discussion with her referring clinician, who may not be sufficiently knowledgeable to fully explain the rationale for periodic surveillance rather than biopsy. It has been observed in clinical practice that such a woman is more likely to undergo immediate biopsy (rather than the recommended follow-up), when compared with a woman who undergoes a conventional PBF assessment that is fully explained in person at the completion of diagnostic imaging, confidently and competently, by a radiologist, technologist, or breast health educator in a mammography facility (17,21,22).
Traditionally, PBF assessments are assigned when interpreting results of a woman’s first screening mammographic examination or a subsequent examination when previous images are not available for comparison. Short-interval follow-up examinations for PBFs are not recommended if previous images are available, since comparison images can be used to determine whether a lesion is growing, stable, or decreased in size. Lesions that are growing require biopsy, whereas stable or decreasing lesions are followed annually rather than at shorter intervals (21,23). Thus, it is surprising that 1.7% of subsequent examinations included a final assessment of PBF and a recommendation for short-interval follow-up. This constituted 39% of all final recommendations exclusive of those assigned routine follow-up.
One explanation for this finding is that radiologists are not conducting PBF assessments as recommended in the literature (21,23). Also, BI-RADS has not provided guidance on the lack of need for short-interval follow-up for subsequent screening examinations. Finally, comparison images may not be readily available for some women who are undergoing screening mammography, and this may result in otherwise unnecessary surveillance of PBFs. Women and referring physicians should assist radiologists by ensuring that previous images are available for comparison whenever possible to minimize unnecessary surveillance of PBFs and to reduce the number of false-positive results (24–26).
Our study has several strengths and limitations. The accuracy of our data depends on completeness of cancer reporting to tumor registries and pathology laboratories at the mammography registries, which has been estimated to be more than 94.3% complete for the Breast Cancer Surveillance Consortium (27). Tumor size and stage were missing for 8.0%–13.6% of tumors, in part because of a change in coding of tumor size and stage by the Surveillance, Epidemiology, and End Results program between July 1998 and July 1999, which resulted in some invasive cancers with an in situ component being coded with an unknown size. However, it is unlikely that there is a tumor reporting bias to cancer registries related to mammography assessment or recommendation. Finally, we may not have acquired all of the additional imaging data that were obtained after a screening examination, which may have resulted in some misclassification of conventional PBF assessments and those based on screening mammography only. Any misclassification is likely to have led to underestimation of the differences in cancer yield between groups.
Use of short-interval follow-up to monitor PBFs should be limited to as small a number of women as possible because the cancer yield is low for these lesions. Performance of a full diagnostic imaging work-up prior to recommendation of short-interval follow-up limits the number of women with benign lesions that will undergo periodic surveillance and may also result in more timely detection of early-stage breast cancer. The recently published fourth edition of BI-RADS (28) provides extensive guidance on the use of PBF assessments and management recommendations for short-interval follow-up. Monitoring whether there are decreases in the prevalence of short-interval follow-up for surveillance of PBFs in the time periods before and after the publication of the fourth edition of BI-RADS will be important. Lack of a decline may suggest the need for more intensive interventions than just publication of guidelines to change the evaluation of these lesions.
|
FOOTNOTES |
|---|
Authors stated no financial relationship to disclose.
Author contributions: Guarantors of integrity of entire study, K.K., C.L., R.S.B., E.A.S.; study concepts, K.K., C.L., B.C.Y., R.S.B., R.B.B., W.E.B., P.A.C., E.A.S.; study design, K.K., C.L., B.C.Y., R.S.B., W.E.B., E.A.S.; literature research, K.K., E.A.S.; data acquisition, K.K., B.C.Y., W.E.B., J.H.V., B.M.G., P.A.C., E.A.S.; data analysis/interpretation, K.K., L.A., C.L., B.C.Y., R.S.B., R.B.B., W.E.B., B.M.G., P.A.C., E.A.S.; statistical analysis, L.A., W.E.B.; manuscript preparation, K.K., C.L., B.C.Y., R.S.B., R.B.B., W.E.B., E.A.S.; manuscript definition of intellectual content, K.K., C.L., B.C.Y., R.S.B., W.E.B., P.A.C., E.A.S.; manuscript editing, K.K., L.A., C.L., B.C.Y., R.S.B., R.B.B., W.E.B., J.H.V., P.A.C., E.A.S.; manuscript revision/review and final version approval, all authors.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODS RESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  E. J. A. Bowles, D. L. Miglioretti, E. A. Sickles, L. Abraham, P. A. Carney, B. C. Yankaskas, and J. G. Elmore Accuracy of Short-Interval Follow-Up Mammograms by Patient and Radiologist Characteristics Am. J. Roentgenol., May 1, 2008; 190(5): 1200 - 1208. [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]
|
|
|
|
 |
|
 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. L. Stefanick, G. L. Anderson, K. L. Margolis, S. L. Hendrix, R. J. Rodabough, E. D. Paskett, D. S. Lane, F. A. Hubbell, A. R. Assaf, G. E. Sarto, et al. Effects of Conjugated Equine Estrogens on Breast Cancer and Mammography Screening in Postmenopausal Women With Hysterectomy JAMA, April 12, 2006; 295(14): 1647 - 1657. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
F. M. Hall, E. A. Sickles, C. Lehman, R. Smith-Bindman, B. C. Yankaskas, B. M. Geller, P. A. Carney, and K. Kerlikowske Short-Interval Follow-up in Screening Mammography * Dr Sickles and colleagues respond: Radiology, October 1, 2005; 237(1): 371 - 373. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
