UK Breast Screening Programme Research Articles

Mammography Breast Cancer Screening Triage Using Deep Learning: A UK Retrospective Study.

Background Breast screening enables early detection of cancers; however, most women have normal mammograms, resulting in repetitive and resource-intensive reading tasks. Purpose To investigate if deep learning (DL) algorithms can be used to triage mammograms by identifying normal results to reduce workload or flag cancers that may be overlooked. Materials and Methods In this retrospective study, three commercial DL algorithms were investigated using consecutive mammograms from two UK Breast Screening Program sites from January 2015 to December 2017 and January 2017 to December 2018 on devices from two mammography vendors. Normal mammograms with a 3-year follow-up and histopathologically proven cancer detected at screening, the subsequent round, or in the 3-year interval were included. Two algorithm thresholds were set: in scenario A, 99.0% sensitivity for rule-out triage to a lone reader, and in scenario B, approximately 1.0% additional recall providing a rule-in triage for further assessment. Both thresholds were then applied to the screening workflow in scenario C. The sensitivity and specificity were used to assess the overall predictive performance of each DL algorithm. Results The data set comprised 78 849 patients (median age, 59 years [IQR, 53-63 years]) and 887 screening-detected, 439 interval, and 688 subsequent screening round-detected cancers. In scenario A (rule-out triage), models DL-1, DL-2, and DL-3 triaged 35.0% (27 565 of 78 849), 53.2% (41 937 of 78 849), and 55.6% (43 869 of 78 849) of mammograms, respectively, with 0.0% (0 of 887) to 0.1% (one of 887) of screening-detected cancers undetected. In scenario B, DL algorithms triaged in 4.6% (20 of 439) to 8.2% (36 of 439) of interval and 5.2% (36 of 688) to 6.1% (42 of 688) of subsequent-round cancers when applied after the routine double-reading workflow. Combining both approaches in scenario C resulted in an overall noninferior specificity (difference, -0.9%; P < .001) and superior sensitivity (difference, 2.7%; P < .001) for the adaptive workflow compared with routine double reading for all three algorithms. Conclusion Rule-out and rule-in DL-adapted triage workflows can improve the efficiency and efficacy of mammography breast cancer screening. © RSNA, 2023 Supplemental material is available for this article. See also the editorial by Nishikawa and Lu in this issue.

Left-right breast asymmetry and risk of screen-detected and interval cancers in a large population-based screening population.

To assess the associations between automated volumetric estimates of mammographic asymmetry and breast cancers detected at the same ("contemporaneous") screen, at subsequent screens, or in between (interval cancers). Automated measurements from mammographic images (N = 79,731) were used to estimate absolute asymmetry in breast volume (BV) and dense volume (DV) in a large ethnically diverse population of attendees of a UK breast screening programme. Logistic regression models were fitted to assess asymmetry associations with the odds of a breast cancer detected at contemporaneous screen (767 cases), adjusted for relevant confounders.Nested case-control investigations were designed to examine associations between asymmetry and the odds of: (a) interval cancer (numbers of cases/age-matched controls: 153/646) and (b) subsequent screen-detected cancer (345/1438), via conditional logistic regression. DV, but not BV, asymmetry was positively associated with the odds of contemporaneous breast cancer (P-for-linear-trend (Pt) = 0.018). This association was stronger for first (prevalent) screens (Pt = 0.012). Both DV and BV asymmetry were positively associated with the odds of an interval cancer diagnosis (Pt = 0.060 and 0.030, respectively). Neither BV nor DV asymmetry were associated with the odds of having a subsequent screen-detected cancer. Increased DV asymmetry was associated with the risk of a breast cancer diagnosis at a contemporaneous screen or as an interval cancer. BV asymmetry was positively associated with the risk of an interval cancer diagnosis. The findings suggest that DV and BV asymmetry may provide additional signals for detecting contemporaneous cancers and assessing the likelihood of interval cancers in population-based screening programmes.

Abstract P4-07-04: Genomic analysis of paired DCIS and subsequent recurrence to assess clonal relatedness in screen detected DCIS

Abstract In order to understand the clonal relatedness of DCIS and subsequent recurrence, 40 patients with primary DCIS that developed a subsequent ipsilateral event (28 invasive and 12 pure DCIS) and 17 with a subsequent contralateral event, were identified within the Sloane project, a prospective national cohort study of DCIS embedded in the UK Breast Screening Programme. Median follow up was 5.4 years. All cases were reviewed by a breast histopathologist and macro-dissected to separate the DCIS or invasive disease from the normal tissue. DNA was extracted using the AllPrep DNA/RNA FFPE Kit (Qiagen). Somatic copy number aberrations were assessed using the HumanCytoSNP-12 BeadChip Kit (Illumina) and ASCAT. Targeted sequencing of a custom panel of 120 breast cancer-associated genes was performed using the SureSelect XT low input Target Enrichment System (Agilent Technologies). Prior to undertaking clonal related analysis, the relatedness of the samples was confirmed by cryptic relatedness analysis in PLINK v1.07. Clonal relatedness was assessed using two methods. The first used copy number data, available for all samples, where each breakpoint was compared between the pairs of tumours from the same individual and a concordance score calculated based on the presence of shared and unique breakpoints. The second used targeted sequencing, available on 23 paired ipsilateral cases and 7 contralateral paired recurrences, that was run through the previously described Clonality package (Biometrics. 2018). The results of the clonal relatedness analysis are summarized in below. Evidence of clonal relatedness on copy number and / or mutation analysisEquivocal on mutation analysisNot clonal on mutation or copy number analysisNot clonal by copy numberPrimary DCIS &amp; Ipsilateral invasive recurrence (28 pairs)1648Primary DCIS &amp; Ipsilateral DCIS recurrence (12 pairs)1101Primary DCIS &amp; Contralateral recurrence (16 pairs)1213Recurrent synchronous DCIS and invasive disease (12 pairs)1110 There were 2 cases where mutation data showed clear evidence of clonal relatedness but copy number did not, suggesting that the copy number approach may underestimate relatedness. There was also a proportion of cases where the pairs shared only one mutation and it was not possible to ascertain conclusively whether they were related, particularly when the shared mutation was a common driver mutation in PIK3CA that could have arisen by chance in unrelated samples. Of the 28 primary DCIS that developed ipsilateral invasive recurrence, 8 cases (29%) showed no evidence of clonal relatedness by one or other method. Of these 3 recurred in the original tumour bed, 3 in a different quadrant and in 2 the site of recurrence was unknown. Of the 12 primary DCIS that developed an ipsilateral DCIS recurrence 92% showed evidence of clonal relatedness and also recurred earlier than the invasive recurrences, mean time to recurrence 2.9 years compared to 4.15 years for recurrent invasive disease (P=0.04, Student’s t-test). The mutations that were most commonly shared between the primary and recurrence were PIK3CA and TP53, followed by KMT2C, CAD and DNAH2. As expected, the majority of the contralateral cases were unrelated. Of the 28 cases of DCIS that recurred ipsilaterally with invasive disease 16 had synchronous DCIS at the time of detection of the invasive disease and in 12 cases we were able to analyse the two synchronous components separately. In all cases the synchronous recurrent DCIS and invasive components showed evidence of clonal relatedness. This study shows synchronous DCIS and IDC are clonally related as is DCIS that recurs as DCIS in the majority of cases. However, 29% of DCIS that recurred as invasive disease showed no evidence of clonal relatedness. It remains to be seen whether exome sequencing of the unrelated lesions would reveal evidence of relatedness or whether they are truly new primaries. Citation Format: Vandna Shah, Anargyros Megalios, Rana Shami, Mathini Sridharan, Carolina Salinas de Souza, Tapsi Kumar, Karen Clements, Andrew Futreal, Sarah Pinder, Alastair Thompson, Elinor Jane Sawyer, Sloane Steering Committee, CRUK Grand Challenge PRECISION Team. Genomic analysis of paired DCIS and subsequent recurrence to assess clonal relatedness in screen detected DCIS [abstract]. In: Proceedings of the 2019 San Antonio Breast Cancer Symposium; 2019 Dec 10-14; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2020;80(4 Suppl):Abstract nr P4-07-04.

Ethnic and age differences in right-left breast asymmetry in a large population-based screening population.

Exposure to sex hormones is important in the pathogenesis of breast cancer and inability to tolerate such exposure may be reflected in increased asymmetrical growth of the breasts. This study aims to characterize, for the first time, asymmetry in breast volume (BV) and radiodense volume (DV) in a large ethnically diverse population. Automated measurements from digital raw mammographic images of 54,591 cancer-free participants (aged 47-73) in a UK breast screening programme were used to calculate absolute (cm3) and relative asymmetry in BV and DV. Logistic regression models were fitted to assess asymmetry associations with age and ethnicity. BV and DV absolute asymmetry were positively correlated with the corresponding volumetric dimension (BV or DV). BV absolute asymmetry increased, whilst DV absolute asymmetry decreased, with increasing age (P-for-linear-trend <0.001 for both). Relative to Whites, Blacks had statistically significantly higher, and Chinese lower, BV and DV absolute asymmetries. However, after adjustment for the corresponding underlying volumetric dimension the age and ethnic differences were greatly attenuated. Median relative (fluctuating) BV and DV asymmetry were 2.34 and 3.28% respectively. After adjusting for the relevant volumetric dimension (BV or DV), age and ethnic differences in absolute breast asymmetry were largely resolved. Previous small studies have reported breast asymmetry-breast cancer associations. Automated measurements of asymmetry allow the conduct of large-scale studies to further investigate these associations.

Breast thickness based DRLs in screening mammography

The motivation for use of dose audit and DRLs is to promote continuing optimisation. General radiology practice quotes a DRL for an average patient and has been widely adopted for standard imaging examinations. However, for mammography, the range of breast sizes encountered in clinical practice limits the utility of DRLs. For a number of years, we have been conducting an annual clinical dose survey based on 100 consecutive client examinations collected over a contemporaneous time period. A MATLAB programme was developed to extract demographic and exposure data from the DICOM image headers. This was subsequently compiled in a database for calculation of mean glandular dose (MGD) developed for the UK breast screening programme. In order to improve the utility of DRLs for routine audit without adding significant complexity, we have determined three distinct DRLs based on the 95th percentile of the distribution of MGD values within three bands of compressed breast thickness; Average breast (50 mm–70 mm), Small breast (<40 mm) and Large breast (>80 mm). Completing the dose audit cycle requires dissemination of results in a relatable way and in order to achieve this, we have produced a poster for display in each mammography room which radiographers can use to reference against any examination performed. This provides a more meaningful comparison of their practice with the norm. We anticipate that as radiographers become accustomed to viewing and evaluating MGDs, they will become more familiar with the expected behaviour of a system and differences between systems.

The Potential Impact of Digital Breast Tomosynthesis on the Benign Biopsy Rate in Women Recalled within the UK Breast Screening Programme.

Background Multiple studies showed digital breast tomosynthesis (DBT) has a higher cancer detection rate and a lower recall rate than full-field digital mammography (FFDM). However, there is a paucity of studies on whether the benign biopsy rate will be lower with DBT. Purpose To evaluate the benign biopsy rate of DBT versus that of FFDM in women recalled after breast screening within the National Health Service Breast Screening Programme. Materials and Methods This prospective single institution study included women who underwent screening FFDM or screening breast MRI between November 13, 2015, and July 29, 2016, and were recalled. Women who were recalled underwent a breast examination, additional imaging (mammography, US), and biopsy, if required. The number of interventions and the outcome of assessment were recorded. FFDM was performed at the screening appointment, and DBT images were acquired when participants attended a screening assessment within 3 weeks after screening FFDM. Two of four readers blinded to biopsy results analyzed DBT images in isolation and then in conjunction with FFDM images and any additional two-dimensional mammographic or US images. Two-tailed McNemar tests were used to test differences in sensitivity and specificity of the two conditions at 5% significance level. Results During the study period, 30 933 women underwent FFDM. A total of 1470 women were recalled (recall rate, 4.8%), and 827 were included after exclusion criteria were applied. Their mean age was 56.7 years ± 7.7 (standard deviation). A total of 145 breast cancers were detected (142 with FFDM, two with DBT only, one with surveillance MRI). Triple assessment without DBT resulted in 571 breast biopsies and enabled detection of 142 cancers. The addition of DBT would have resulted in 298 biopsies and detection of 142 cancers, reducing the number of biopsies from 571 of 827 (69.0%) to 298 of 827 (36.0%). Conclusion When compared with full-field digital mammography, digital breast tomosynthesis can reduce the benign biopsy rate while maintaining the cancer detection rate. © RSNA, 2019 Online supplemental material is available for this article. See also the editorial by Skaane in this issue.

Blurred digital mammography images: an analysis of technical recall and observer detection performance.

Blurred images in full-field digital mammography are a problem in the UK Breast Screening Programme. Technical recalls may be due to blurring not being seen on lower resolution monitors used for review. This study assesses the visual detection of blurring on a 2.3-MP monitor and a 5-MP report grade monitor and proposes an observer standard for the visual detection of blurring on a 5-MP reporting grade monitor. 28 observers assessed 120 images for blurring; 20 images had no blurring present, whereas 100 images had blurring imposed through mathematical simulation at 0.2, 0.4, 0.6, 0.8 and 1.0 mm levels of motion. Technical recall rate for both monitors and angular size at each level of motion were calculated. χ2 tests were used to test whether significant differences in blurring detection existed between 2.3- and 5-MP monitors. The technical recall rate for 2.3- and 5-MP monitors are 20.3% and 9.1%, respectively. The angular size for 0.2- to 1-mm motion varied from 55 to 275 arc s. The minimum amount of motion for visual detection of blurring in this study is 0.4 mm. For 0.2-mm simulated motion, there was no significant difference [χ2 (1, N = 1095) = 1.61, p = 0.20] in blurring detection between the 2.3- and 5-MP monitors. According to this study, monitors ≤2.3 MP are not suitable for technical review of full-field digital mammography images for the detection of blur. Advances in knowledge: This research proposes the first observer standard for the visual detection of blurring.

Radiation doses received in the United Kingdom breast screening programme in 2010 to 2012.

To review the radiation doses received by women attending the UK breast-screening programme between 2010 and 2012. To compare doses with previous years and to quantify the impact on dose of changing from analogue to digital imaging and to analyse doses by type of imaging system. Measurements of doses to samples of about 50-100 women attending for screening were collected across the whole of the UK breast-screening programme. Data were collected for 87,122 exposures, using 449 X-ray sets, for 25,408 women. The average mean glandular dose (MGD) was 1.79 mGy for mediolateral oblique images and 1.58 mGy for craniocaudal images. The average MGD per two-view examination was 4.01 mGy for film-screen imaging and 3.03 mGy for direct digital radiography (DR) and 4.69 mGy for computed radiography. The MGD to women attending breast screening has been reduced on average by about 25% where DR systems have replaced film-screen systems. The dose reduction was greatest for breasts with the largest compressed thickness. There are large variations in dose between the different models of DR system provided by different manufacturers. There should be further work to ensure that all DR systems are operated at the optimal dose level to ensure the best cancer detection while balancing the detriment caused by using radiation. Changes in the radiation dose in breast screening over time have been determined. Specifically, the impact on radiation dose of introducing different types of DR and computed radiography system into breast screening has been quantified.

Reduction in interval cancer rates following the introduction of two-view mammography in the UK breast screening programme

Background:The introduction of two-view mammography at incident (subsequent) screens in the National Health Service Breast Screening Programme (NHSBSP) has led to an increased number of cancers detected at screen. However, the effect of two-view mammography on interval cancer rates has yet to be assessed.Methods:Routine screening and interval cancer data were collated from all screening programmes in the United Kingdom for women aged 50–64, screened between 1 April 2003 and 31 March 2005. Interval cancer rates were compared based on whether two-view mammography was in use at the last routine screen.Results:The reduction in interval cancers following screening using two-view mammography compared with one view was 0.68 per 1 000 women screened. Overall, this suggests the introduction of two-view mammography at incident screen was accompanied by a 15–20% reduction in interval cancer rates in the NHSBSP.Conclusion:The introduction of two-view mammography at incident screens is associated with a reduction in incidence of interval cancers. This is consistent with previous publications on a contemporaneous increase in screen-detected cancers. The results provide further evidence of the benefit of the use of two-view mammography at incident screens.

O-81 The impact of caseload on practice patterns in breast cancer: Evidence from the UK Breast Screening Programme

O-81 The impact of caseload on practice patterns in breast cancer: Evidence from the UK Breast Screening Programme

Absolute Numbers of Lives Saved and Overdiagnosis in Breast Cancer Screeninq, from a Randomized Trial and from the Breast Screening Programme in England

ObjectivesTo estimate the absolute numbers of breast cancer deaths prevented and the absolute numbers of tumours overdiagnosed in mammographic screening for breast cancer at ages 50–69 years.SettingThe Swedish Two-County randomized trial of mammographic screening for breast cancer, and the UK Breast Screening Programme in England, ages 50–69 years.MethodsWe estimated the absolute numbers of deaths avoided and additional cases diagnosed in the study group (active study population) of the Swedish Two-County Trial, by comparison with the control group (passive study population). We estimated the same quantities for the mortality and incidence rates in England (1974–2004 and 1974–2003, respectively). We used Poisson regression for statistical inference.ResultsA substantial and significant reduction in breast cancer mortality was associated with screening in both the Two-County Trial (P < 0.001) and the screening programme in England (P < 0.001). The absolute benefits were estimated as 8.8 and 5.7 breast cancer deaths prevented per 1000 women screened for 20 years starting at age 50 from the Two-County Trial and screening programme in England, respectively. The corresponding estimated numbers of cases overdiagnosed per 1000 women screened for 20 years were, respectively, 4.3 and 2.3 per 1000.ConclusionsThe benefit of mammographic screening in terms of lives saved is greater in absolute terms than the harm in terms of overdiagnosis. Between 2 and 2.5 lives are saved for every overdiagnosed case.

A Rapid-Response Economic Evaluation of the UK NHS Cancer Reform Strategy Breast Cancer Screening Program Extension via a Plausible Bounds Approach

The 2007 National Health Service Cancer Reform Strategy includes a proposed extension of the UK breast screening program to women aged 47 to 49 years. The aim of this study is to undertake a preliminary assessment of this proposal to identify the key factors determining its cost-effectiveness and to determine whether there is sufficient uncertainty that requires more thorough analyses. An economic model was constructed. For simplicity, the health impact of screening was estimated by calculating the lives saved through shifts in prognostic group. A "plausible bounds" approach was used to derive distributions for model parameters for probabilistic sensitivity analysis. UK data were used to populate the model. The cost-effectiveness of the extension is estimated to be pound27,400 per quality-adjusted life-year (QALY) with a 29% probability of cost-effectiveness at a threshold of pound20,000 per QALY. The deterministic estimate of benefit becomes negative if the anxiety due to a false-positive result is set at 0.028 QALYs. Including a small positive benefit from a negative screen has a dramatic impact on the cost-effectiveness of screening. The impact of other factors appears less marked. Because the vast majority of women aged 47 to 49 years will test negative when screened for breast cancer and most of those who test positive will actually be free of the disease, the psychological impacts of screening are likely to drive cost-effectiveness for this age group. Therefore, a research priority should be to better understand and quantify these effects.

Modelling the Impact of Changes in Sensitivity on the Outcomes of the Uk Breast Screening Programme

To assess the impact on breast cancer mortality of improving the sensitivity of breast screening programmes. A Markov model was populated with data obtained from published statistics describing the UK National Health Service Breast Screening Programme and the incidence and mortality of breast cancer in the UK. The model was used to study the impact of changes to the sensitivity of screening. The effects on cancer detection rates and breast cancers and total mortality was studied for a cohort of women followed from age 45 to age 89. Running the model from age 45 to 89, with sensitivity set at the baseline value of 75%, predicts the detection at screening of 44 cancers per thousand of population and the detection outside screening of 82 cancers per thousand of population. Running the model with values of sensitivity from 75% to 95% shows the proportion of cancers detected at screening increasing as screening improves, and deaths from breast cancer falling. The drop in breast cancer deaths is however modest. Increasing sensitivity from 75% to 85% reduces the number of breast cancer deaths from 28 to 27 per thousand. Likely achievable improvements in the sensitivity of screening do not have a marked effect on breast cancer mortality.

Increasing participant recruitment into large-scale screening trials: experience from the CADET II study

Multicentre randomized trials frequently encounter difficulties in meeting their recruitment targets, resulting in extension of the trial and delays in implementation of the findings. We report on recruitment strategies implemented in a randomized evaluation of computer-aided detection in women attending routine screening in the UK Breast Screening Programme. The target population for the trial was identified from an existing NHS database of women aged 50-70 invited for routine mammography in Coventry, Manchester and Nottingham, UK. Women were asked to consent to their mammograms being randomly allocated (in a ratio of 28:1:1) to one of three film-reading protocols. Trial information was mailed to women, along with the invitation to attend screening, and informed consent was obtained at the mammography appointment. Several strategies were implemented to increase recruitment rates. Recruitment rate increased significantly over time in the study (P < 0.0010 in all centres) with an overall acceptance rate of 46% of those attending screening. Mailing of the trial information sheet separate from the screening invitation in Coventry and Nottingham increased the recruitment rate, even after adjustment for the trend over time and for socioeconomic status of the attendees (P < 0.001). Extension of recruitment to mobile screening units in Nottingham, and the presence of an additional member of staff also improved recruitment (P < 0.001). Simplification of the trial information sheet had little effect. Increases in recruitment rate were not attributable to socioeconomic status of the attendees. In multicentre trials, monitoring of local recruitment protocols is required to ensure that each centre can maximize accrual targets.

Breast screening unit size and performance on self-assessment (PERFORMS)

The UK Breast Screening Programme (UKBSP) is comprised of approximately 100 individual Breast Screening Units (BSUs) that vary in size (measured by number of women screened). Previous research using UKBSP real life data (Blanks, Bennett, Wallis and Moss, 2002), attributed differences in performance, related to BSU size, to smaller units' slightly lowered cancer detection rates and positive predictive value scores (compared to large/medium units). All BSUs on the UKBSP annually take part in the PERFORMS scheme as way of self-assessing their film-reading skills. We looked at the performance of all film-readers who had completed the last PERFORMS round (SA08) by BSU size in order to explore any group differences mediated by unit capacity.

Mammography—oncogenecity at low doses

Controversy exists regarding the biological effectiveness of low energyx-rays used for mammography breast screening. Recent radiobiology studieshave provided compelling evidence that these low energy x-rays may be4.42 ± 2.02 times more effective in causing mutational damage than higher energy x-rays.These data include a study involving in vitro irradiation of a human cell line using amammography x-ray source and a high energy source which matches the spectrum ofradiation observed in survivors from the Hiroshima atomic bomb. Current radiation riskestimates rely heavily on data from the atomic bomb survivors, and a direct comparisonbetween the diagnostic energies used in the UK breast screening programme and those usedfor risk estimates can now be made. Evidence highlighting the increase in relative biologicaleffectiveness (RBE) of mammography x-rays to a range of x-ray energies implies that therisks of radiation-induced breast cancers for mammography x-rays are potentiallyunderestimated by a factor of four. A pooled analysis of three measurementsgives a maximal RBE (for malignant transformation of human cells in vitro) of4.02 ± 0.72 for 29 kVp (peak accelerating voltage) x-rays compared to high energy electrons and higherenergy x-rays.For the majority of women in the UK NHS breast screening programme, it is shown thatthe benefit safely exceeds the risk of possible cancer induction even when this higherbiological effectiveness factor is applied. The risk/benefit analysis, however, implies theneed for caution for women screened under the age of 50, and particularly for those with afamily history (and therefore a likely genetic susceptibility) of breast cancer.In vitro radiobiological data are generally acquired at high doses, and there are differentextrapolation mechanisms to the low doses seen clinically. Recent low dose in vitro datahave indicated a potential suppressive effect at very low dose rates and doses. Whilstmammography is a low dose exposure, it is not a low dose rate examination, andprotraction of dose should not be confused with fractionation. Although there is potentialfor a suppressive effect at low doses, recent epidemiological data, and several internationalradiation risk assessments, continue to promote the linear no-threshold (LNT) model.Finally, recent studies have shown that magnetic resonance imaging (MRI) is moresensitive than mammography in detecting invasive breast cancer in women with a geneticsensitivity. Since an increase in the risk associated with mammographic screening wouldblur the justification of exposure for this high risk subgroup, the use of other (non-ionising)screening modalities is preferable.

Full-Field Digital versus Screen-Film Mammography: Comparison within the UK Breast Screening Program and Systematic Review of Published Data

To (a) compare the performance of full-field digital mammography (FFDM), using hard-copy image reading, with that of screen-film mammography (SFM) within a UK screening program (screening once every 3 years) for women aged 50 years or older and (b) conduct a meta-analysis of published findings along with the UK data. The study complied with the UK National Health Service Central Office for Research Ethics Committee guidelines; informed patient consent was not required, since analysis was carried out retrospectively after data anonymization. Between January 2006 and June 2007, a London population-based screening center performed 8478 FFDM and 31 720 SFM screening examinations, with modality determined by the type of machine available at the screening site. Logistic regression was used to assess whether breast cancer detection rates and recall rates differed between screening modalities. For the meta-analysis, random-effects models were used to combine study-specific estimates, if appropriate. A total of 263 breast cancers were detected. After adjustment for age, ethnicity, area of residence, and type of referral, there was no evidence of differences between FFDM and SFM in terms of detection rates (0.68 [95% confidence interval {CI}: 0.47, 0.89] vs 0.72 [95% CI: 0.58, 0.85], respectively, per 100 screening mammograms; P = .74), recall rates (3.2% [95% CI: 2.8, 3.6] vs 3.4% [95% CI: 3.1, 3.6]; P = .44), positive predictive value (PPV) of an abnormal mammogram, or characteristics of detected tumors. Meta-analysis of data from eight studies showed a slightly higher detection rate for FFDM, particularly at 60 years of age or younger (pooled FFDM-SFM difference: 0.11 [95% CI: 0.04, 0.18] per 100 screening mammograms), but no clear modality differences in recall rates or PPVs. Within a routine screening program, FFDM with hard-copy image reading performed as well as SFM in terms of process indicators; the meta-analysis was consistent with FFDM yielding detection rates at least as high as those for SFM.

Vacuum-assisted percutaneous removal of breast papillomas and radial scars

Introduction: Image-guided removal of breast lesions has been described. The UK breast screening program pathology sub-group have defined a group of lesions with uncertain malignant potential (B3) when surgical excision is recommended. However large-core vacuum-assisted devices provide an alternative means of removal. We review the Maidstone experience in the vacuum-assisted removal of papillomas and radial scars.

Non-operative diagnosis—effect on repeat-operation rates in the UK breast screening programme

Non-operative diagnosis rates in the UK breast screening programme have improved dramatically from 48.8% in 1994/95 (only nine units achieved the then minimum standard of 70%) to 94% in 2005/06 (only seven units failed to achieve the target of 90%). Preoperative and operative history of all 120,550 women diagnosed with screen-detected breast cancer in the UK between April 1994 and March 2006 was derived from different national databases. In 2005/06, 2,790 (17.8%) of the 15,688 women having surgery needed two or more operations. In 2001/02 (non-operative diagnosis rate 87%), the re-operation rate was 23.8% (2,377 of 9,969). Extrapolation backwards to 1994/95 (non-operative diagnosis rate 48.8%) suggests a re-operation rate of 62%. Analysis over the 4 years from April 2002 (n=34,198) demonstrates that 4,089 (12%) women with a correct non-operative diagnosis of invasive disease required additional surgery compared to 1,166 (48%) of women who were under-staged (diagnosed as non-invasive based on core biopsy, but actually suffering from invasive disease). Failure to achieve a non-operative diagnosis of invasive disease (n=1,542) or non-invasive disease (n=2,247) resulted in re-operation rates of 65 and 43% respectively. Given the impact of not having a diagnosis pre-operatively, or of under-staging invasive carcinoma, it seems timely to introduce more sophisticated standards.

CADET II: A prospective trial of computer-aided detection (CAD) in the UK Breast Screening Programme

508 Background: Computer-aided detection (CAD) systems that attract the reader's attention to potentially suspicious features on a mammogram are now in clinical use in screening mammography in the USA. Double reading increases small cancer detection by up to 10%, and the aim of this trial was to evaluate whether the performance of a single reader using CAD can match that of standard double reading in the UK Breast Screening Programme. Methods: Information about the study was posted to women invited for routine screening mammography at three UK screening centers. Informed consent was obtained at the time of the mammography appointment. Film batches were randomly assigned in a ratio of 28:1:1 to one of three film reading regimes: mammograms double read and single read with CAD, mammograms double read only, mammograms single read with CAD only. Double reading was conducted according to standard protocol. For single reading with CAD, mammograms were digitized and analyzed by an ImageChecker CAD system V8.1 and readers viewed the CAD prompted images on a low resolution computer display screen or hard copy printouts. The study was designed as a prospective multicenter randomized trial with matched comparison of single reading with CAD and double reading in terms of the number of women recalled for assessment, cancer detection rate and staging of the two reading regimes. Equivalence was defined as a 95% confidence interval on the difference between cancer detection rates of no more than 10% in either direction and the estimated sample size was 30,000 women. Results: 31,293 women were recruited into the trial between September 2006 and August 2007. The cancer detection rates of the two reading regimes were comparable (86.5% for single reading with CAD and 87.8% for double reading). This equates to a difference of −1.3% (95%CI −8.4%, +5.7%; p=0.7). Recall rates were 3.9% for single reading with CAD and 3.3% for double reading, a small but significant, absolute difference of 0.6% (95%CI 0.3%, 0.8%; p<0.007), a relative difference of 18%. Conclusions: The cancer detection rate of single reading with CAD was comparable to that achieved by double reading with only a small increase in recall rate. The results of this study should inform policy on the potential of CAD to improve efficiency in the UK Breast Screening Programme. No significant financial relationships to disclose.