Radiologists' perceptions of computer aided detection versus double reading for mammography interpretation

RATIONALE AND OBJECTIVES

The aim of this study was to examine radiologists' use and perceptions of computer-aided detection (CAD) and double reading for screening mammography interpretation.

MATERIALS AND METHODS

A mailed survey of 257 community radiologists participating in the national Breast Cancer Surveillance Consortium was used to assess perceptions and practices related to CAD and double reading. Latent class analysis was used to classify radiologists' overall perceptions of CAD and double reading on the basis of their agreement or disagreement with specific statements about CAD and double reading.

RESULTS

Most radiologists (64%) reported using CAD for more than half the screening mammograms they interpreted, but only <5% reported double reading that much. More radiologists perceived that double reading improved cancer detection rates compared to CAD (74% vs 55% reported), whereas fewer radiologists thought that double reading decreased recall rates compared to CAD (50% vs 65% reported). Radiologists with the most favorable perceptions of CAD were more likely to think that CAD improved cancer detection rates without taking too much time compared to radiologists with the most unfavorable overall perceptions. In latent class analysis, an overall favorable perception of CAD was significantly associated with the use of CAD (81%), a higher percentage of workload in screening mammography (80%), academic affiliation (71%), and fellowship training (58%). Perceptions of double reading that were most favorable were associated with academic affiliation (98%).

CONCLUSIONS

Radiologists' perceptions were more favorable toward double reading by a second clinician than by a computer, although fewer used double reading in their own practice. The majority of radiologists perceived both CAD and double reading at least somewhat favorably, although for largely different reasons.

Performance of diagnostic mammography differs in the United States and Denmark

Diagnostic mammography is the primary imaging modality to diagnose breast cancer. However, few studies have evaluated variability in diagnostic mammography performance in communities, and none has done so between countries. We compared diagnostic mammography performance in community-based settings in the United States and Denmark. The performance of 93,585 diagnostic mammograms from 180 facilities contributing data to the US Breast Cancer Surveillance Consortium (BCSC) from 1999 to 2001 was compared to that of all 51,313 diagnostic mammograms performed at Danish clinics in 2000. We used the imaging workup's final assessment to determine sensitivity, specificity and an estimate of accuracy: area under the receiver-operating characteristics (ROCs) curve (AUC). Diagnostic mammography had slightly higher sensitivity in the United States (85%) than in Denmark (82%). In contrast, it had higher specificity in Denmark (99%) than in the United States (93%). The AUC was high in both countries: 0.91 in United States and 0.95 in Denmark. Denmark's higher accuracy may result from supplementary ultrasound examinations, which are provided to 74% of Danish women but only 37% to 52% of US women. In addition, Danish mammography facilities specialize in either diagnosis or screening, possibly leading to greater diagnostic mammography expertise in facilities dedicated to symptomatic patients. Performance of community-based diagnostic mammography settings varied markedly between the 2 countries, indicating that it can be further optimized.

Diagnosis of second breast cancer events after initial diagnosis of early stage breast cancer

To examine whether there are any characteristics of women or their initial tumors that might be useful for tailoring surveillance recommendations to optimize outcomes. We followed 17,286 women for up to 5 years after an initial diagnosis of ductal carcinoma in situ (DCIS) or early stage (I/II) invasive breast cancer diagnosed between 1996 and 2006. We calculated rates per 1,000 women years of recurrences and second breast primaries relative to demographics, risk factors, and characteristics of initial diagnosis: stage, treatment, mode of initial diagnosis. Nearly 4% had a second breast cancer event (314 recurrences and 344 second breast primaries). Women who used adjuvant hormonal therapy or were ≥ 80 years had the lowest rates of second events. Factors associated with higher recurrence and second primary rates included: initial DCIS or stage IIB, estrogen/progesterone receptor-negative, younger women (<50 years). Women with a family history or greater breast density had higher second primary rates, and women who received breast conserving surgery without radiation had higher recurrence rates. Roughly one-third of recurrences (37.6%) and second primaries (36.3%) were not screen-detected. Initial mode of diagnosis was a predictor of second events after adjusting for age, stage, primary treatment, and breast density. A recent negative mammogram should not falsely reassure physicians or women with new breast symptoms or changes because one-third of second cancers were interval cancers. This study does not provide any evidence in support of changing surveillance intervals for different subgroups.

Effect of previous benign breast biopsy on the interpretive performance of subsequent screening mammography

BACKGROUND

Most breast biopsies will be negative for cancer. Benign breast biopsy can cause changes in the breast tissue, but whether such changes affect the interpretive performance of future screening mammography is not known.

METHODS

We prospectively evaluated whether self-reported benign breast biopsy was associated with reduced subsequent screening mammography performance using examination data from the mammography registries of the Breast Cancer Surveillance Consortium from January 2, 1996, through December 31, 2005. A positive interpretation was defined as a recommendation for any additional evaluation. Cancer was defined as any invasive breast cancer or ductal carcinoma in situ diagnosed within 1 year of mammography screening. Measures of mammography performance (sensitivity, specificity, and positive predictive value 1 [PPV1]) were compared both at woman level and breast level in the presence and absence of self-reported benign biopsy history. Referral to biopsy was considered a positive interpretation to calculate positive predictive value 2 (PPV2). Multivariable analysis of a correct interpretation on each performance measure was conducted after adjusting for registry, year of examination, patient characteristics, months since last mammogram, and availability of comparison film. Accuracy of the mammogram interpretation was measured using area under the receiver operating characteristic curve (AUC). All statistical tests were two-sided.

RESULTS

A total of 2,007,381 screening mammograms were identified among 799,613 women, of which 14.6% mammograms were associated with self-reported previous breast biopsy. Multivariable adjusted models for mammography performance showed reduced specificity (odds ratio [OR] = 0.74, 95% confidence interval [CI] = 0.73 to 0.75, P < .001), PPV2 (OR = 0.85, 95% CI = 0.79 to 0.92, P < .001), and AUC (AUC 0.892 vs 0.925, P < .001) among women with self-reported benign biopsy. There was no difference in sensitivity or PPV1 in the same adjusted models, although unadjusted differences in both were found. Specificity was lowest among women with documented fine needle aspiration-the least invasive biopsy technique (OR = 0.58, 95% CI = 0.55 to 0.61, P < .001). Repeating the analysis among women with documented biopsy history, unilateral biopsy history, or restricted to invasive cancers did not change the results.

CONCLUSIONS

Self-reported benign breast biopsy history was associated with statistically significantly reduced mammography performance. The difference in performance was likely because of tissue characteristics rather than the biopsy itself.

Time trends in radiologists' interpretive performance at screening mammography from the community-based Breast Cancer Surveillance Consortium, 1996-2004

PURPOSE

To examine time trends in radiologists' interpretive performance at screening mammography between 1996 and 2004.

MATERIALS AND METHODS

All study procedures were institutional review board approved and HIPAA compliant. Data were collected on subsequent screening mammograms obtained from 1996 to 2004 in women aged 40-79 years who were followed up for 1 year for breast cancer. Recall rate, sensitivity, and specificity were examined annually. Generalized estimating equation (GEE) and random-effects models were used to test for linear trend. The area under the receiver operating characteristic curve (AUC), tumor histologic findings, and size of the largest dimension or diameter of the tumor were also examined.

RESULTS

Data on 2,542,049 subsequent screening mammograms and 12,498 cancers diagnosed in the follow-up period were included in this study. Recall rate increased from 6.7% to 8.6%, sensitivity increased from 71.4% to 83.8%, and specificity decreased from 93.6% to 91.7%. In GEE models, adjusted odds ratios per calendar year were 1.04 (95% confidence interval [CI]: 1.02, 1.05) for recall rate, 1.09 (95% CI: 1.07. 1.12) for sensitivity, and 0.96 (95% CI: 0.95, 0.98) for specificity (P < .001 for all). Random-effects model results were similar. The AUC increased over time: 0.869 (95% CI: 0.861, 0.877) for 1996-1998, 0.884 (95% CI: 0.879, 0.890) for 1999-2001, and 0.891 (95% CI: 0.885, 0.896) for 2002-2004 (P < .001). Tumor histologic findings and size remained constant.

CONCLUSION

Recall rate and sensitivity for screening mammograms increased, whereas specificity decreased from 1996 to 2004 among women with a prior mammogram. This trend remained after accounting for risk factors. The net effect was an improvement in overall discrimination, a measure of the probability that a mammogram with cancer in the follow-up period has a higher Breast Imaging Reporting and Data System assessment category than does a mammogram without cancer in the follow-up period.

When radiologists perform best: the learning curve in screening mammogram interpretation

PURPOSE

To examine changes in screening mammogram interpretation as radiologists with and radiologists without fellowship training in breast imaging gain clinical experience.

MATERIALS AND METHODS

In an institutional review board-approved HIPAA-compliant study, the performance of 231 radiologists who interpreted screen-film screening mammograms from 1996 to 2005 at 280 facilities that contribute data to the Breast Cancer Surveillance Consortium was examined. Radiologists' demographic data and clinical experience levels were collected by means of a mailed survey. Mammograms were grouped on the basis of how many years the interpreting radiologist had been practicing mammography, and the influence of increasing experience on performance was examined separately for radiologists with and those without fellowship training in breast imaging, taking into account case-mix and radiologist-level differences.

RESULTS

A total of 1 599 610 mammograms were interpreted during the study period. Performance for radiologists without fellowship training improved most during their 1st 3 years of clinical practice, when the odds of a false-positive reading dropped 11%-15% per year (P < .015) with no associated decrease in sensitivity (P > .89). The number of women recalled per breast cancer detected decreased from 33 for radiologists in their 1st year of practice to 24 for radiologists with 3 years of experience to 19 for radiologists with 20 years of experience. Radiologists with fellowship training in breast imaging experienced no learning curve and reached desirable goals during their 1st year of practice.

CONCLUSION

Radiologists' interpretations of screening mammograms improve during their first few years of practice and continue to improve throughout much of their careers. Additional residency training and targeted continuing medical education may help reduce the number of work-ups of benign lesions while maintaining high cancer detection rates.

Variability in interpretive performance at screening mammography and radiologists' characteristics associated with accuracy

PURPOSE

To identify radiologists' characteristics associated with interpretive performance in screening mammography.

MATERIALS AND METHODS

The study was approved by institutional review boards of University of Washington (Seattle, Wash) and institutions at seven Breast Cancer Surveillance Consortium sites, informed consent was obtained, and procedures were HIPAA compliant. Radiologists who interpreted mammograms in seven U.S. regions completed a self-administered mailed survey; information on demographics, practice type, and experience in and perceptions of general radiology and breast imaging was collected. Survey data were linked to data on screening mammograms the radiologists interpreted between January 1, 1998, and December 31, 2005, and included patient risk factors, Breast Imaging Reporting and Data System assessment, and follow-up breast cancer data. The survey was returned by 71% (257 of 364) of radiologists; in 56% (205 of 364) of the eligible radiologists, complete data on screening mammograms during the study period were provided; these data were used in the final analysis. An evaluation of whether the radiologists' characteristics were associated with recall rate, false-positive rate, sensitivity, or positive predictive value of recall (PPV(1)) of the screening examinations was performed with logistic regression models that were adjusted for patients' characteristics and radiologist-specific random effects.

RESULTS

Study radiologists interpreted 1 036 155 screening mammograms; 4961 breast cancers were detected. Median percentages and interquartile ranges, respectively, were as follows: recall rate, 9.3% and 6.3%-13.2%; false-positive rate, 8.9% and 5.9%-12.8%; sensitivity, 83.8% and 74.5%-92.3%; and PPV(1), 4.0% and 2.6%-5.9%. Wide variability in sensitivity was noted, even among radiologists with similar false-positive rates. In adjusted regression models, female radiologists or fellowship-trained radiologists had significantly higher recall and false-positive rates (P < .05, all). Fellowship training in breast imaging was the only characteristic significantly associated with improved sensitivity (odds ratio, 2.32; 95% confidence interval: 1.42, 3.80; P < .001) and the overall accuracy parameter (odds ratio, 1.61; 95% confidence interval: 1.05, 2.45; P = .028).

CONCLUSION

Fellowship training in breast imaging may lead to improved cancer detection, but it is associated with higher false-positive rates.

Screening-detected breast cancers: discordant independent double reading in a population-based screening program

PURPOSE

To analyze discordant and concordant screening-detected breast cancers in a nationwide population-based screening program by using independent double reading with consensus.

MATERIALS AND METHODS

The study is a part of the evaluation of the Norwegian Breast Cancer Screening Program and is covered by the Cancer Registry regulation. Analyses were based on prospective initial interpretation scores of 1 033 870 screenings that included 5611 breast cancers. A five-point scale for probability of cancer was used in the initial interpretation. Screening mammograms with a score of 2 or higher by either radiologist were discussed at consensus meetings where the decision whether to recall was made. A score of 1 by one reader and 2 or higher by the other was defined as a discordant interpretation and discordant cancer, whereas a score of 2 or higher by both readers was defined as a concordant recall and cancer.

RESULTS

Discordant interpretation was present in 5.3% (54 447 of 1 033 870) of the screenings, whereas 2.1% (21 928 of 1 033 870) were concordant positive interpretations. Of the screening-detected cancers, 23.6% (1326 of 5611) were diagnosed in women who were recalled because of screenings with discordant interpretation. One hundred seventeen interval breast cancers were diagnosed among the 40 312 screenings that were dismissed at consensus; these were 6.5% of all interval cancers. A significantly higher proportion of microcalcifications alone was present in discordant cancers (24.9% [304 of 1219]) compared with concordant cancers (17.7% [704 of 3972]) (P < .001).

CONCLUSION

Independent double reading with consensus at mammography screening has the potential to increase the cancer detection rate compared with single reading. Mammograms with microcalcifications alone are significantly more common among discordant cancers.

Bias associated with self-report of prior screening mammography

BACKGROUND

Self-reported screening behaviors from national surveys often overestimate screening use, and the amount of overestimation may vary by demographic characteristics. We examine self-report bias in mammography screening rates overall, by age, and by race/ethnicity.

METHODS

We use mammography registry data (1999-2000) from the Breast Cancer Surveillance Consortium to estimate the validity of self-reported mammography screening collected by two national surveys. First, we compare mammography use from 1999 to 2000 for a geographically defined population (Vermont) with self-reported rates in the prior two years from the 2000 Vermont Behavioral Risk Factor Surveillance System. We then use a screening dissemination simulation model to assess estimates of mammography screening from the 2000 National Health Interview Survey.

RESULTS

Self-report estimates of mammography use in the prior 2 years from the Vermont Behavioral Risk Factor Surveillance System are 15 to 25 percentage points higher than actual screening rates across age groups. The differences in National Health Interview Survey screening estimates from models are similar for women 40 to 49 and 50 to 59 years and greater than for those 60 to 69, or 70 to 79 (27 and 26 percentage points versus 14, and 14, respectively). Overreporting is highest among African American women (24.4 percentage points) and lowest among Hispanic women (17.9) with non-Hispanic White women in between (19.3). Values of sensitivity and specificity consistent with our results are similar to previous validation studies of mammography.

CONCLUSION

Overestimation of self-reported mammography usage from national surveys varies by age and race/ethnicity. A more nuanced approach that accounts for demographic differences is needed when adjusting for overestimation or assessing disparities between populations.

An assessment of the likelihood, frequency, and content of verbal communication between radiologists and women receiving screening and diagnostic mammography

RATIONALE AND OBJECTIVES

Research on communication between radiologists and women undergoing screening and diagnostic mammography is limited. We describe community radiologists' communication practices with patients regarding screening and diagnostic mammogram results and factors associated with frequency of communication.

MATERIALS AND METHODS

We received surveys from 257 radiologists (70% of those eligible) about the extent to which they talk to women as part of their health care visit for either screening or diagnostic mammograms, whether this occurs if the exam assessment is positive or negative, and how they use estimates of patient risk to convey information about an abnormal exam where the specific finding of cancer is not yet known. We also assessed characteristics of the radiologists to identify associations with more or less frequent communication at the time of the mammogram.

RESULTS

Two hundred and forty-three radiologists provided complete data (95%). Very few (<6%) reported routinely communicating with women when screening mammograms were either normal or abnormal. Fewer than half (47%) routinely communicated with women when their diagnostic mammograms were normal, whereas 77% often or always communicated with women when their diagnostic exams were abnormal. For positive diagnostic exams, female radiologists were more likely to be frequent communicators compared to males (87.1%-72.8%; P=.02) and those who spend 40%-79% of their time in breast imaging (94.6%) were more likely to be frequent communicators compared to those who spend less time (67.2%-78.9%; P=.02). Most radiologists convey risk information using general rather than numeric statements (57.7% vs. 28.5%).

CONCLUSIONS

Radiologists are most likely to convey information about diagnostic mammographic findings when results are abnormal. Most radiologists convey risk information using general rather than numeric statements.

Disclosing harmful mammography errors to patients

PURPOSE

To assess radiologists' attitudes about disclosing errors to patients by using a survey with a vignette involving an error interpreting a patient's mammogram, leading to a delayed cancer diagnosis.

MATERIALS AND METHODS

We conducted an institutional review board-approved survey of 364 radiologists at seven geographically distinct Breast Cancer Surveillance Consortium sites that interpreted mammograms from 2005 to 2006. Radiologists received a vignette in which comparison screening mammograms were placed in the wrong order, leading a radiologist to conclude calcifications were decreasing in number when they were actually increasing, delaying a cancer diagnosis. Radiologists were asked (a) how likely they would be to disclose this error, (b) what information they would share, and (c) their malpractice attitudes and experiences.

RESULTS

Two hundred forty-three (67%) of 364 radiologists responded to the disclosure vignette questions. Radiologists' responses to whether they would disclose the error included "definitely not" (9%), "only if asked by the patient" (51%), "probably" (26%), and "definitely" (14%). Regarding information they would disclose, 24% would "not say anything further to the patient," 31% would tell the patient that "the calcifications are larger and are now suspicious for cancer," 30% would state "the calcifications may have increased on your last mammogram, but their appearance was not as worrisome as it is now," and 15% would tell the patient "an error occurred during the interpretation of your last mammogram, and the calcifications had actually increased in number, not decreased." Radiologists' malpractice experiences were not consistently associated with their disclosure responses.

CONCLUSION

Many radiologists report reluctance to disclose a hypothetical mammography error that delayed a cancer diagnosis. Strategies should be developed to increase radiologists' comfort communicating with patients.

Probabilistic computer model developed from clinical data in national mammography database format to classify mammographic findings

PURPOSE

To determine whether a Bayesian network trained on a large database of patient demographic risk factors and radiologist-observed findings from consecutive clinical mammography examinations can exceed radiologist performance in the classification of mammographic findings as benign or malignant.

MATERIALS AND METHODS

The institutional review board exempted this HIPAA-compliant retrospective study from requiring informed consent. Structured reports from 48 744 consecutive pooled screening and diagnostic mammography examinations in 18 269 patients from April 5, 1999 to February 9, 2004 were collected. Mammographic findings were matched with a state cancer registry, which served as the reference standard. By using 10-fold cross validation, the Bayesian network was tested and trained to estimate breast cancer risk by using demographic risk factors (age, family and personal history of breast cancer, and use of hormone replacement therapy) and mammographic findings recorded in the Breast Imaging Reporting and Data System lexicon. The performance of radiologists compared with the Bayesian network was evaluated by using area under the receiver operating characteristic curve (AUC), sensitivity, and specificity.

RESULTS

The Bayesian network significantly exceeded the performance of interpreting radiologists in terms of AUC (0.960 vs 0.939, P = .002), sensitivity (90.0% vs 85.3%, P < .001), and specificity (93.0% vs 88.1%, P < .001).

CONCLUSION

On the basis of prospectively collected variables, the evaluated Bayesian network can predict the probability of breast cancer and exceed interpreting radiologist performance. Bayesian networks may help radiologists improve mammographic interpretation.

Comparing screening mammography for early breast cancer detection in Vermont and Norway

BACKGROUND

Most screening mammography in the United States differs from that in countries with formal screening programs by having a shorter screening interval and interpretation by a single reader vs independent double reading. We examined how these differences affect early detection of breast cancer by comparing performance measures and histopathologic outcomes in women undergoing opportunistic screening in Vermont and organized screening in Norway.

METHODS

We evaluated recall, screen detection, and interval cancer rates and prognostic tumor characteristics for women aged 50-69 years who underwent screening mammography in Vermont (n = 45 050) and in Norway (n = 194 430) from 1997 through 2003. Rates were directly adjusted for age by weighting the rates within 5-year age intervals to reflect the age distribution in the combined data and were compared using two-sided Z tests.

RESULTS

The age-adjusted recall rate was 9.8% in Vermont and 2.7% in Norway (P < .001). The age-adjusted screen detection rate per 1000 woman-years after 2 years of follow-up was 2.77 in Vermont and 2.57 in Norway (P = .12), whereas the interval cancer rate per 1000 woman-years was 1.24 and 0.86, respectively (P < .001). Larger proportions of invasive interval cancers in Vermont than in Norway were 15 mm or smaller (55.9% vs 38.2%, P < .001) and had no lymph node involvement (67.5% vs 57%, P = .01). The prognostic characteristics of all invasive cancers (screen-detected and interval cancer) were similar in Vermont and Norway.

CONCLUSION

Screening mammography detected cancer at about the same rate and at the same prognostic stage in Norway and Vermont, with a statistically significantly lower recall rate in Norway. The interval cancer rate was higher in Vermont than in Norway, but tumors that were diagnosed in the Vermont women tended to be at an earlier stage than those diagnosed in the Norwegian women.

Factors associated with imaging and procedural events used to detect breast cancer after screening mammography

OBJECTIVE

The purpose of this study was to characterize the type and frequency of diagnostic evaluations after screening mammography and to summarize their association with the likelihood of biopsy and subsequent breast cancer diagnosis.

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.

Communicating with women about mammography

BACKGROUND/METHODS

We report survey results of the types of tools used to communicate with women about breast cancer screening and the content areas included in each tool for member countries of the International Breast Cancer Screening Network (IBSN).

RESULTS

In addition to using pamphlets and invitation letters, new technologies are being used such as the Internet which allows for easy updating of information and can provide interactive modules. Several countries have addressed the needs of specific populations such as indigenous populations or blind women. All countries provide basic information, although they do not provide all the same information.

CONCLUSION

More research is needed to understand what women need to make an informed decision about mammography and to learn what the best modalities are to provide this information.

Risk factors for breast carcinoma in situ versus invasive breast cancer in a prospective study of pre- and post-menopausal women

Risk factors for breast carcinoma in situ and invasive breast cancer were compared using data from 61,844 women (61% post-menopausal) with no prior breast cancer and at least one screening mammogram between April 1, 1996 and June 30, 2001. The women were followed until a subsequent mammogram before July 1, 2001, or a benign biopsy or breast cancer diagnosis before June 30, 2002. A total of 1,191 breast cancers (300 in situ and 891 invasive) were diagnosed during an average follow-up of 3.1 years. Multivariate Cox regression models were used to estimate the relative risks (RR) of in situ and invasive cancer associated with family history of breast cancer, age at first childbirth or nulliparity, post-menopausal hormone use, body mass index (BMI), and mammographic breast density. Separate analyses were done for pre- and post-menopausal women. BMI was unrelated to risk of in situ cancer regardless of menopausal status, but was associated with an increased risk of invasive cancer in post-menopausal women (RR = 1.9 for BMI > or = 30 vs. BMI < 22, 95% confidence interval 1.4-2.5). Later childbearing and nulliparity were more strongly associated with in situ than invasive cancer in pre-menopausal women. Post-menopausal hormone use was more strongly associated with invasive disease. RR associated with family history and breast density were similar for in situ cancer and invasive cancer. Results indicating that BMI is related to post-menopausal invasive cancers but unrelated to in situ cancers are consistent with the hypothesis that concomitants of obesity activate proliferation.

Performance benchmarks for screening mammography

PURPOSE

To retrospectively evaluate the range of performance outcomes of the radiologist in an audit of screening mammography by using a representative sample of U.S. radiologists to allow development of performance benchmarks for screening mammography.

MATERIALS AND METHODS

Institutional review board approval was obtained, and study was HIPAA compliant. Informed consent was or was not obtained according to institutional review board guidelines. Data from 188 mammographic facilities and 807 radiologists obtained between 1996 and 2002 were analyzed from six registries from the Breast Cancer Surveillance Consortium (BCSC). Contributed data included demographic information, clinical findings, mammographic interpretation, and biopsy results. Measurements calculated were positive predictive values (PPVs) from screening mammography (PPV(1)), biopsy recommendation (PPV(2)), biopsy performed (PPV(3)), recall rate, cancer detection rate, mean cancer size, and cancer stage. Radiologist performance data are presented as 50th (median), 10th, 25th, 75th, and 90th percentiles and as graphic presentations by using smoothed curves.

RESULTS

There were 2 580 151 screening mammographic studies from 1 117 390 women (age range, <30 to >/=80 years). The respective means and ranges of performance outcomes for the middle 50% of radiologists were as follows: recall rate, 9.8% and 6.4%-13.3%; PPV(1), 4.8% and 3.4%-6.2%; and PPV(2), 24.6% and 18.8%-32.0%. Mean cancer detection rate was 4.7 per 1000, and the median [corrected] mean size of invasive cancers was 13 mm. The range of performance outcomes for the middle 80% of radiologists also was presented.

CONCLUSION

Community screening mammographic performance measurements of cancer outcomes for the majority of radiologists in the BCSC surpass performance recommendations. Recall rate for almost half of radiologists, however, is higher than the recommended rate.

Improving the concordance of mammography assessment and management recommendations

PURPOSE

To retrospectively compare the concordance of initial and final assessment categories for mammograms with management recommendations made before and after the final rules of the Mammography Quality Standards Act (MQSA) were in effect for screening and diagnostic mammography.

MATERIALS AND METHODS

The study included mammograms from 1996 to 2001 from the seven mammography registries of the Breast Cancer Surveillance Consortium (BCSC). The authors defined the pre-MQSA period as January 1, 1996-April 27, 1999, and the post-MQSA period as April 28, 1999-December 31, 2001 (2470151 screening and 194199 diagnostic mammograms). Assessment was cross-classified according to management recommendation. Changes in concordance between assessment and recommendation were evaluated by year and by period (before and after MQSA) for computer-linked data and for all data by using Pearson chi(2) test to evaluate differences. Mantel-Haenszel chi(2) test was used to measure change in concordance over time. Each registry and the BCSC Statistical Coordinating Center had a Federal Certificate of Confidentiality and approval from each institution's review board for protection of human subjects to collect and send data to coordinating center and conduct research with these data. Active consent was required at only one site in this HIPAA-compliant study.

RESULTS

Concordance increased significantly in the post-MQSA period for Breast Imaging Reporting and Data System categories 3-5 assessments at both screening and diagnostic mammography. The most substantial improvements were in the use of the management recommendation for "additional imaging," which decreased from 41% in 1996 to 15% in 2001 for screening mammograms with an initial assessment of category 4 (P < .001). Recommendation for short-interval follow-up in women with screening mammograms with a category 3 final assessment increased from 51% in 1996 to 76% in 2001 (P < .001). Concordance for diagnostic mammograms assigned category 0 improved from 65% in the pre-MQSA period to 81% in the post-MQSA period (P < .001).

CONCLUSION

This analysis demonstrates that over a relatively short period of time, major improvement in radiology reporting has occurred.

Print information to inform decisions about mammography screening participation in 16 countries with population-based programs

OBJECTIVE

To profile and compare the content and presentation of written communications related to informed decision-making about mammography.

METHODS

Materials from 16 screening programs organized at the national or regional level were analyzed according to five major information domains suggested by the international literature.

RESULTS

A majority of countries provided information on the program (interval, cost and quality). There was considerable variability in comprehensiveness of elements in the domains, e.g., test characteristics (false positive/negative) and pros and cons of screening. The majority noted the likelihood of recall for further tests, few commented on the risks of additional tests or finding unimportant tumors. The audit also found variation in presentation (words and pictures).

CONCLUSIONS

Presentation of comprehensive, but balanced information on screening benefits and risks is complex and daunting. Issues such as framing effects, coupled with debate about screening efficacy are challenging to the design of effective information tools. The objective of increasing screening prevalence at the population level must be balanced with objectively presenting complete and clear information. Additional research is needed on how information (and mode of presentation) impact screening decisions.

PRACTICE IMPLICATIONS

Public health officials need to articulate their objectives and review written communication according to important decision-making domains.

Prospective breast cancer risk prediction model for women undergoing screening mammography

BACKGROUND

Risk prediction models for breast cancer can be improved by the addition of recently identified risk factors, including breast density and use of hormone therapy. We used prospective risk information to predict a diagnosis of breast cancer in a cohort of 1 million women undergoing screening mammography.

METHODS

There were 2,392,998 eligible screening mammograms from women without previously diagnosed breast cancer who had had a prior mammogram in the preceding 5 years. Within 1 year of the screening mammogram, 11,638 women were diagnosed with breast cancer. Separate logistic regression risk models were constructed for premenopausal and postmenopausal examinations by use of a stringent (P<.0001) criterion for the inclusion of risk factors. Risk models were constructed with 75% of the data and validated with the remaining 25%. Concordance of the predicted with the observed outcomes was assessed by a concordance (c) statistic after logistic regression model fit. All statistical tests were two-sided.

RESULTS

Statistically significant risk factors for breast cancer diagnosis among premenopausal women included age, breast density, family history of breast cancer, and a prior breast procedure. For postmenopausal women, the statistically significant factors included age, breast density, race, ethnicity, family history of breast cancer, a prior breast procedure, body mass index, natural menopause, hormone therapy, and a prior false-positive mammogram. The model may identify high-risk women better than the Gail model, although predictive accuracy was only moderate. The c statistics were 0.631 (95% confidence interval [CI] = 0.618 to 0.644) for premenopausal women and 0.624 (95% CI = 0.619 to 0.630) for postmenopausal women.

CONCLUSION

Breast density is a strong additional risk factor for breast cancer, although it is unknown whether reduction in breast density would reduce risk. Our risk model may be able to identify women at high risk for breast cancer for preventive interventions or more intensive surveillance.

Pathologic findings from the Breast Cancer Surveillance Consortium: population-based outcomes in women undergoing biopsy after screening mammography

BACKGROUND

To the authors' knowledge, a comprehensive analysis of pathology outcomes after screening mammography, as it is practiced clinically in the U.S. general population, has not been performed.

METHODS

Breast Cancer Surveillance Consortium data from 1996-2001 were used to identify pathology specimens that were obtained within 1 year of screening mammograms performed on 786,846 women ages 40-89 years. The pathology results were classified as invasive carcinoma, ductal carcinoma in situ (DCIS), or benign. The associations between overall pathology outcomes and invasive tumor size and lymph node status were analyzed by age and mammography assessment category.

RESULTS

The rates of both recommending and performing a biopsy varied little with age. The 1,664,032 screening mammograms were followed by 26,748 known biopsies (1.6%) and 8815 diagnoses of breast carcinoma (0.53%). Overall, 81% of carcinomas were invasive, and 78% of those were pathologically lymph node-negative tumors, in contrast to the 66% prevalence observed in the Surveillance, Epidemiology, and End Results (SEER) data during the same period. Most invasive tumors measured between 0 mm and 10 mm (35%) or between 11 mm and 20 mm (36%) in greatest dimension, and 92% and 78% were lymph node-negative tumors, respectively: Biopsy results that were classified as malignant increased with age (P < 0.0001) and were most likely to follow Breast Imaging, Reporting, and Diagnosis System Category 5 and 4 assessments, respectively. Ductal hyperplasia (19.6%), fibroadenoma (18.5%), and other benign findings (56.1%) were the most common benign diagnoses.

CONCLUSIONS

Pathologically negative lymph nodes were more prevalent in this mammographically screened population than in the overall SEER population. The prevalence of invasive carcinoma, DCIS, and benign findings presented herein establish a range of expected biopsy outcomes for women after screening mammography in the general U.S. population.

Modeling the dissemination of mammography in the United States

OBJECTIVE

This paper presents a methodology for piecing together disparate data sources to obtain a comprehensive model for the use of mammography screening in the US population for the years 1975-2000.

METHODS

Two aspects of mammography usage, the age that a woman receives her first mammography and the interval between subsequent mammograms, are modeled separately. The initial dissemination of mammography is based on cross-sectional self report data from national surveys and the interval length between screening exams is fit using longitudinal mammography registry data.

RESULTS

The two aspects of mammography usage are combined to simulate screening histories for individual women that are representative of the US population. Simulated mammography patterns for the years 1994-2000 were found to be similar to observed screening patterns from the state level mammography registry for Vermont.

CONCLUSIONS

The model presented gives insight into screening practices over time and provides an alternative public health measure for screening usage in the US population. The comprehensive description of mammography use from its introduction represents an important first step to understanding the impact of mammography on breast cancer incidence and mortality.

The use of additional imaging increased specificity and decreased sensitivity in screening mammography

BACKGROUND AND OBJECTIVES

To examine the use of additional imaging after standard mammographic screening views to better understand the value of these additional testing in improving accuracy.

MATERIALS AND METHODS

Statewide population data on screening mammography were used to report accuracy measures at screening and after additional imaging tests. Pathology data from biopsies performed within 1 year following the screening mammogram were used to determine cancer outcome (489 in situ and invasive cancers). Pathologic and population characteristics of women receiving different types of imaging were assessed by chi-square or t-tests. Similar tests compared women with the same imaging and differing outcomes.

RESULTS

Of 77,799 women with screening mammograms 9.9% had additional imaging. Additional imaging reduced false positives from 7,765 (100/1,000 mammograms) to 1,112 (14/1,000 mammograms). The majority of false negatives (82%) occurred in women receiving only screening views, and additional imaging increased the number of false negatives from 82 (1/1,000 mammograms) to 115 (1.5/1,000 mammograms).

CONCLUSION

Additional imaging can reduce unnecessary biopsy but at the cost of some additional false negatives. Additional imaging's potential for improving the sensitivity of screening is limited because most missed cancers occur in women who do not have additional imaging.

Performance benchmarks for diagnostic mammography

PURPOSE

To evaluate a range of performance parameters pertinent to the comprehensive auditing of diagnostic mammography examinations, and to derive performance benchmarks therefrom, by pooling data collected from large numbers of patients and radiologists that are likely to be representative of mammography practice in the United States.

MATERIALS AND METHODS

Institutional review board approval was met, informed consent was not required, and this study was Health Insurance Portability and Accountability Act compliant. Six mammography registries contributed data to the Breast Cancer Surveillance Consortium (BCSC), providing patient demographic and clinical information, mammogram interpretation data, and biopsy results from defined population-based catchment areas. The study involved 151 mammography facilities and 646 interpreting radiologists. The study population included women 18 years of age or older who underwent at least one diagnostic mammography examination between 1996 and 2001. Collected data were used to derive mean performance parameter values, including abnormal interpretation rate, positive predictive value (for abnormal interpretation, biopsy recommended, and biopsy performed), cancer diagnosis rate, invasive cancer size, and the percentages of minimal cancers, axillary node-negative invasive cancers, and stage 0 and I cancers. Additional benchmarks were derived for these performance parameters, including 10th, 25th, 50th (median), 75th, and 90th percentile values.

RESULTS

The study involved 332,926 diagnostic mammography examinations. Mean performance parameter values were abnormal interpretation rate, 8.0%; positive predictive value for abnormal interpretation, 31.4%; positive predictive value for biopsy recommended, 31.5%; positive predictive value for biopsy performed, 39.5%; cancer diagnosis rate, 25.3 per 1000 examinations; invasive cancer size, 20.2 mm; percentage of minimal cancers, 42.0%; percentage of axillary node-negative invasive cancers, 73.6%; and percentage of stage 0 and I cancers, 62.4%.

CONCLUSION

The presented BCSC outcomes data and performance benchmarks may be used by mammography facilities and individual radiologists to evaluate their own performance for diagnostic mammography as determined by means of periodic comprehensive audits.