| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Original Research |
1 Norris Cotton Cancer Center, Dartmouth Medical School, Lebanon, NH
03756.
2 Department of Family Medicine, Oregon Health and Science University, 3181 SW
Sam Jackson Park Rd., Portland, OR 97239.
3 Group Health Cooperative, Center for Health Studies, Seattle, WA 98101.
4 Applied Research Program, Division of Cancer Control and Population Sciences,
National Cancer Institute, Bethesda, MD 20892.
5 Department of Radiology, University of California, San Francisco, San
Francisco, CA 94115.
6 Health Promotions Research, University of Vermont, Burlington, VT 05401.
7 Department of Radiology, University of New Mexico, Albuquerque, NM
87131.
8 Cooper Institute, Lakewood, CO 80401.
9 Department of Pathology, University of Vermont School of Medicine, Burlington,
VT 05405.
10 Department of Medicine, University of California, San Francisco, CA
94115.
Received September 28, 2005;
accepted after revision May 25, 2006.
Supported by cooperative agreements UO1CA63731, UO1CA63736, UO1CA63740,
UO1CA69976, UO1CA70013, UO1CA70040, UO1CA86076, and UO1CA86082 from the
National Cancer Institute as part of the NCI Breast Cancer Surveillance
Consortium.
Abstract
 Top Abstract IntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
MATERIALS AND METHODS. The data source was 584,470 women with no previous breast cancer from six states in the Breast Cancer Surveillance Consortium. In this observational study, we linked data from 1,207,631 routine screening mammograms performed between January 1, 1996, and December 31, 2002, to data on additional imaging, interventional procedures, and biopsy outcome (benign or malignant). Additional examinations were categorized into diagnostic mammography, sonography, or both. Events were further subdivided by whether they were performed on the same day as the screening examination and whether patients reported breast symptoms. Logistic regression analysis was used to examine the association between additional evaluation performed and the likelihood of biopsy and the likelihood of subsequent breast cancer diagnosis after adjustment for patient and screening mammographic characteristics.
RESULTS. Most (92%) of the screening examinations did not include additional imaging. The probability of biopsy ranged from 0.4% for examinations with no follow-up to 20.1% for those with diagnostic mammography and sonography on the same day as screening among women without symptoms and from 2.1% for those with no follow-up to 18.9% for those with diagnostic mammography and sonography on a day different from screening among women with symptoms. Thirty percent of women without symptoms who underwent biopsy had cancer, whereas 27.1% of women with symptoms who underwent biopsy had cancer. Women who underwent biopsy after screening mammography with diagnostic mammography and sonography on the same day had the highest probability of breast cancer (37.6% among women without symptoms, 36.4% among women with symptoms), whereas those who underwent only sonography performed at a later date had the lowest probability of breast cancer (11.9% among women without symptoms, 17.1% among women with symptoms).
CONCLUSION. Women who undergo screening mammography followed by diagnostic mammography and sonography have a high probability of undergoing biopsy and having the biopsy result of breast cancer when follow-up imaging is performed on the same day as screening mammography whether or not breast symptoms are present. Biopsy performed after sonography in the absence of diagnostic mammography had a low yield of breast cancer.
Keywords: breast biopsy • breast cancer screening • breast neoplasms • mammography • sonography
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Kerlikowske et al. [17] reported that an average of two procedures are performed to evaluate each abnormal mammogram, regardless of age, except that on average women 60 years and older need 30% fewer procedures than women younger than 60 years. The most frequent procedure for evaluating an abnormal mammogram was additional mammographic views. Yoshihara et al. [18] found that certain characteristics on the images, including lobulated shape, irregular contour, and nonuniform echotexture on breast sonography and spiculated mass on mammography differentiated breast cancer from benign breast disease, indicating that both types of imaging are needed to detect breast cancer. In another study [15], investigators examined factors that influence the likelihood that different imaging and procedural events would follow screening mammography, but the small number of cases of cancer detected did not allow the authors to assess combinations of events that led to a cancer diagnosis. In a synthesis of studies on abnormal mammographic results and palpable breast masses, Kerlikowske et al. [19] found that findings on diagnostic mammography did not influence whether biopsy should be performed on a palpable mass.
We previously found that recommendations used in practice often do not correspond to the BI-RADS assessment [20, 21]. Radiologists undertake different approaches based on their own thresholds of concern, which may explain the variability that exists in interpretive practice. In the current study, we used patient characteristics to assess the imaging events and biopsy procedures after routine screening mammography that lead to a diagnosis of either breast cancer or benign breast disease.
None of the aforementioned studies [15-19] addressed differences based on whether all or part of the evaluation was done on the same day as screening mammography, a service mammographic facilities sometimes offer. Many questions exist about evaluation after breast cancer screening, including the relation between additional evaluation and patient self-report of symptoms and whether the evaluation is initiated or completed on the same day as the screening examination or on a different day. A better understanding of the potential added value of additional evaluation after screening mammography may help health care providers prepare their patients for the likelihood of various follow-up procedures after screening, which can cause marked anxiety [22], even when the findings are benign.
In the current study, we assessed imaging events and biopsy procedures leading to a diagnosis of either breast cancer or benign breast disease after routine screening mammography. We further examined whether additional imaging was performed on the same day as the screening examination, whether the decision for same-day follow-up was influenced by patient report of symptoms, and whether breast biopsy findings were more likely to be positive or negative in these situations. Because previous findings [23] suggested that biopsy findings in patients with self-reported symptoms are not associated with higher detection rates, we hypothesized that additional evaluation performed on the same day as screening would result in a higher rate of biopsy and a lower rate of detection of breast cancer compared with performance of screening and additional evaluation on different days.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
We included bilateral mammograms designated as screening by the interpreting radiologist. To ensure that mammograms included in the analysis were screening examinations, we excluded mammograms if the woman had undergone mammography or sonography within the preceding 9 months. We also excluded women with a history of breast cancer. We included women with at least two screening mammograms (evident through either self-report or existence in the Breast Cancer Surveillance Consortium database) because mammographic performance and outcome are different for first versus subsequent mammographic examinations [27] and because most mammographic examinations currently performed in the United States are subsequent mammograms (not a woman's first screening mammogram). Women could contribute more than one round of screening in the analysis.
We began with 1,416,091 screening examinations that initially met criteria. We excluded women with a history of breast reduction, breast augmentation, or breast reconstruction (combined n = 14,558, or 1%). In a stepwise manner we then excluded examinations that were missing key variables necessary for our adjusted analyses: 125,663 (9.0%) examinations that were missing the BI-RADS breast density assessment categories (1 = fatty, 2 = scattered densities, 3 = heterogeneously dense, 4 = extremely dense); 44,960 (3.2%) examinations in which time since last mammogram was undocumented; 10,561 (< 1%) examinations in which hormone replacement therapy use or menopausal status was undocumented; 10,742 (< 1%) examinations missing information related to family history of breast cancer; and 1,971 (< 1%) examinations in which type of additional imaging was unknown. We excluded five examinations missing date of cancer diagnosis. After taking these steps, we identified 1,207,631 routine screening mammographic examinations of 584,470 eligible women during the study period. A mean of 2.1 screening examinations per woman were included.
Relevant Definitions and Linkages for Analyses
For each screening mammogram, we looked ahead in 60-day increments from the
date of the initial screening examination to define in the data set radiologic
events, interventional procedures (biopsies), and biopsy outcome (benign or
malignant). We considered diagnostic mammography and sonography to be part of
the series if they were performed within 60 days of the initial screening
mammogram in the series. The series was extended if additional imaging was
performed within 60 days of the previous imaging. The series continued until
no further imaging was performed within 60 days after the last imaging
procedure, a biopsy was performed, or breast cancer was diagnosed. A biopsy or
cancer diagnosis was considered part of the series if it was performed or was
made within 60 days of the last imaging event included in the series. We chose
60-day increments because an assessment of the Breast Cancer Surveillance
Consortium data revealed that 85% of imaging events resolved within this time
period. The mean number of days between the date of the initial screening
examination and last imaging event in the series was 10.1 (SD, 11.1).
We classified the series of events into the following groups: screening mammography only, screening and diagnostic mammography, screening mammography and sonography, and screening and diagnostic mammography plus sonography. An important factor under study was whether the additional imaging occurred on the same day as the screening examination that initiated the series. We therefore further divided follow-up events on the basis of whether screening mammography and at least one part of a diagnostic evaluation were performed on the same day or on different days. Last, we examined the influence of women's reports of breast symptoms, including lump, nipple discharge, and other symptoms, even though radiologists indicated the examination was for screening purposes. We used evidence of benign or malignant breast pathologic results as an indication that women underwent interventional procedures to establish a breast cancer diagnosis.
Data Analyses
The series of events and patient characteristics were summarized with
frequency distributions. We used logistic regression analysis to examine the
association between type of evaluation and probability of biopsy after
inclusion of potential confounders in the models. It was also of interest to
investigate the series of events that may have led to a positive biopsy result
(diagnosis of ductal carcinoma in situ or invasive breast cancer). Therefore,
we chose screening examinations that were followed by biopsy and fit logistic
regression models in which the outcome was the probability of breast cancer
diagnosis. All models were fit by use of the Genmod procedure in SAS (SAS
Institute) and adjusted for patient and mammographic characteristics. In our
multivariate analyses we also used generalized estimating equations with an
exchangeable correlation structure to account for correlation among women
within the same mammographic facilities. Two-sided p < 0.05 was
considered statistically significant for all analyses.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
|
Overall, most (92%) of the screenings did not include additional imaging (Fig. 1). Most of the additional imaging that was done involved diagnostic views. Examinations followed by both diagnostic mammograms and sonograms were more likely to be followed by biopsy (16.7%) compared with screening examinations followed by diagnostic mammograms or sonograms only (7.6% and 12.3%, respectively). Twenty percent to 32% of the examinations followed by biopsy resulted in a diagnosis of breast cancer, the highest proportion (31.8%) of breast cancers being found in the women who went straight to biopsy after undergoing screening mammography only. This rate, however, was similar to that among women who underwent both diagnostic mammography and sonography after screening mammography (30.5%). These results did not take into account the report of breast symptoms or the date of additional imaging.
|
Although radiologists' indications for all mammograms included in this study were routine screening, almost 3% of the women had a self-report of breast symptoms other than pain (Table 2). Sonography was more likely to follow screening mammography for women with self-reported symptoms, especially when the examinations were performed on the same day. Screening examinations with additional imaging also appeared more likely to be followed by biopsy compared with examinations without additional imaging. In our unadjusted analysis, the likelihood of biopsy in most procedure groups was higher if symptoms were reported.
|
Overall, 1.1% of women without symptoms underwent biopsy compared with 3.6% of women with symptoms (Table 2). Women with and those without breast symptoms who underwent biopsy had similar rates of breast cancer (29.9% vs 27.1%). Twenty-seven percent of women who reported symptoms at the screening examination also had reported symptoms at a previous screening examination; thus, 73% of women reporting symptoms were reporting new symptoms.
Screening examinations followed by diagnostic mammography with or without sonography on the same date yielded the highest percentage of cases of cancer. The probability of biopsy was lower among women undergoing additional evaluation on the same day as screening mammography (8.1%; total number of women [i.e., those with and those without symptoms] undergoing additional imaging on same day and biopsy divided by total number of women undergoing imaging on same day) than among women undergoing additional evaluation later (10.5%; total number of women [with or without symptoms] undergoing imaging on a different day and biopsy divided by total number of women undergoing imaging on a different day).
Among women with no symptoms, mean invasive tumor diameter was 14.0 mm among women undergoing follow-up on a different day, 14.6 mm among women undergoing follow-up on the same day, and 15.7 mm among women undergoing no follow-up (data not shown). Among women with symptoms, mean tumor diameter was 17.3 mm for cancers among women who underwent follow-up on a day different from that of the screening examination, 22.4 mm for women undergoing evaluation on the same day, and 24.3 mm for those who received no follow-up (data not shown). These findings suggest that larger, more obvious cancers were evaluated on the same day as screening, women with the largest tumors going straight to biopsy.
The probability of biopsy increased with the amount of evaluation performed in all procedure groups (Table 3). The increased odds of breast biopsy among women with breast symptoms at screening mammography (compared with women who did not undergo follow-up imaging) ranged from 3.76 (95% CI, 2.85-4.97) to 9.51 (95% CI, 6.61-13.69) adjusted for mammographic registry, age at screening examination, time since last mammogram, family history of breast cancer, menopausal status, hormone use, and breast density and accounting for correlation within facility. The relation increased among women without symptoms, the increased odds of biopsy ranging from 10.12 (95% CI, 7.14-14.34) to 46.25 (95% CI, 34.1-62.73).
|
After mammographic registry, age at screening examination, time since last mammogram, family history of breast cancer, menopausal status, hormone therapy use, breast density, and correlation within facility were accounted for, the relative odds of presence of breast cancer after biopsy were essentially the same for women with and those without breast symptoms at screening mammography (Table 4). Screening mammography with sonography performed later among women with no symptoms was associated with the lowest odds of a breast cancer diagnosis (odds ratio, 0.29; 95% CI, 0.23-0.38) relative to screening examinations with no follow-up. Women with symptoms who underwent screening mammography, diagnostic mammography, and sonography all on the same day had higher odds of a diagnosis of breast cancer after biopsy compared with women who did not undergo follow-up (odds ratio, 2.17; 95% CI, 1.20-3.92). All other imaging groups had lower odds of a breast cancer diagnosis.
|
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
Reports [20, 21] of screening mammography in the United States show the percentage of cases in which patients report symptoms is small. These examinations usually are scheduled as screenings because the patient reports no symptoms when scheduling but then reports a symptom (e.g., pain, lump) when filling out the questionnaire on the day of service. Although most practices have policies for converting these mammograms to diagnostic examinations, this change is not always made, especially in practices that batch-interpret images after the patients have left the facility.
Overall, women with symptoms were more likely to undergo biopsy after screening mammography. Women most likely to undergo biopsy after screening mammography were those whose screening mammograms were followed by diagnostic mammography and sonography. Both women without symptoms and those with symptoms had higher odds of undergoing breast biopsy after additional imaging procedures than women undergoing screening mammography only. Women without symptoms, however, were less likely than women with symptoms to go straight to biopsy without additional evaluation. Alternatively, comparison of results for women without symptoms who underwent breast biopsy after additional imaging procedures with results for women who underwent screening mammography only showed that having breast symptoms was not associated with a greater likelihood of having breast cancer. This finding differs from those of Aiello et al. [28], who found that women reporting a breast lump had increased odds of having breast cancer compared with women without symptoms. Those authors did not find other breast symptoms, such as nipple discharge, pain, and other symptoms, associated with a breast cancer diagnosis.
We did not conduct our analysis at the level of the specific symptom but rather grouped all women with symptoms into one group, which may explain why the probability of cancer was not higher among women with breast symptoms in our study. We also conducted our analysis at the level of the type of diagnostic imaging evaluation performed immediately or soon after screening mammography, which may also explain the differences in our findings. Aiello et al. [28] compared screening views with diagnostic views separately rather than diagnostic views associated with initial screening views. Other research [7, 29] focused on diagnostic mammography has shown at least twofold greater breast cancer yield among women reporting symptoms compared with those not reporting symptoms. In one study [7], in which Breast Cancer Surveillance Consortium data also were used, 72% of women with a self-reported lump had breast cancer, as did 84% of women with other symptoms. In our study, 27% of women with self-reported symptoms other than pain had cancer. Although the previous study focused specifically on diagnostic mammography rather than on examinations that follow routine screening, women's self-reports of symptoms are important.
Biopsy after screening mammography accompanied by sonography but not diagnostic mammography had the lowest yield of breast cancer among women with and those without symptoms and the lowest yield for sonography performed on the same day as screening (within symptom category). This finding suggests that sonography in the absence of diagnostic mammography was not helpful in differentiating benign and malignant lesions and may have to be used more selectively, given the costs of screening associated with sonography. We found that of the 1,705 sonograms that led to breast biopsy, 332 were associated with a breast cancer diagnosis, 1,373 being benign. According to the national global allowable reimbursement amount for Medicare [30], these numbers account for $89,959 in costs associated with benign biopsy findings. Further research is needed to determine the type of benign lesion on a sonogram that leads to the high frequency of benign biopsy findings.
Regardless of breast symptoms, biopsy, diagnostic mammography, and sonography after screening mammography on the same day as the screening had a high probability of having positive results for breast cancer. Biopsy among women in all other imaging groups had a low probability of resulting in a breast cancer diagnosis, especially when the procedures were performed on another day. We were unable to determine the extent to which our same-day evaluations were the result of standing policies at the mammographic facilities of providing online interpretation of screening mammograms (i.e., as the mammograms are obtained, rather than batch interpretation performed later). We believe that online review and batch review of mammograms are very different interpretive experiences and that the difference appears to affect performance. At least one study [31] has shown that online (same day) interpretation results in more follow-up imaging than does batch interpretation. This finding indicates that the frequency of additional evaluation and the rate of benign biopsy findings may be affected by how services are delivered. Although it is true that patient compliance and the skill of the facility staff are important factors, we did not control for these variables in our study.
We also did not determine the extent of possible selection bias resulting from mammographic technologists recognizing abnormal findings (with or without approval of a radiologist) and then performing immediate additional imaging to avoid the inconvenience of certain recall. We did, however, determine the extent to which this bias affected the analysis. It is unlikely that a policy of providing online interpretation of screening examinations itself explains our findings, because the inherently shorter time to diagnosis in these cases should have resulted in the finding of smaller tumors. Our observation that tumor size was larger for women who underwent same-day procedures may indicate that women who received additional imaging on the same day as screening had more obvious (larger) tumors, which received prompt attention.
Our study focused specifically on imaging events that led to biopsy and the proportion of those biopsies that had positive results. We did not assess mammographic interpretation linked to recommendations, as is often done for imaging assessments, because we were more interested in the occurrence of events than in the specific behaviors of radiologists. For this reason, our findings are not directly comparable with those of other studies in the literature. Our approach, however, is useful in addressing the likelihood that certain events would occur under different conditions (e.g., the effect of symptoms and imaging evaluation on the same as opposed to different days) and, especially important, whether these events would lead to a diagnosis of breast cancer.
As with any large national study, our findings may have been affected by missing data, especially those related to patient characteristics, although we did our best to ensure data were complete. Another limitation was that unmeasured confounders may have affected when and what type of follow-up was done and the detection of cancer. For example, women with larger lesions may not have gone on to additional imaging because the tumor was obvious on the screening views. Thus, these women were more likely to go straight to biopsy, and it was more likely for the biopsy to show cancer.
Another limitation was that we used dates of events as reported by facilities to determine when imaging was done, which may have led to misclassification bias. We have no reason to believe, however, that the reported dates were inaccurate. We also assumed that procedures were performed on the same day as interpreted online while women waited, although we do not know this scenario to be true in all cases. Last, we did not include MRI or type of biopsy (imaging-guided or not) in this analysis because these data were not accurately recorded by all registries during the study period. These procedures are important for research and should be included in future studies.
In conclusion, women who undergo screening mammography followed by diagnostic mammography and sonography have the highest probability of undergoing biopsy and having the biopsy result be positive, whether or not the follow-up imaging is performed on the same day as screening and whether or not breast symptoms are present. Biopsy after sonography performed in the absence of diagnostic mammography has a low yield of breast cancer. Women should be informed about the possibility of follow-up evaluation so that they are adequately prepared for screening mammography.
|
TopAbstractIntroductionMaterials and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  H. P. Forman Back to the Beginning Am. J. Roentgenol., February 1, 2007; 188(2): 295 - 296. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
