Usefulness of magnetic resonance imaging in benign and malignant breast lesions: a pictorial review

Magnetic resonance imaging (MRI) can play a role in the early detection of breast cancer or recurrent disease in cases when mammograms are not helpful. It is reported that malignant lesions enhance significantly following contrast injection while benign lesions show little or no enhancement. MRI has been shown to detect small breast lesions with sensitivity greater than mammography and can successfully image the dense young breast. 
 
We would like to highlight a number of interesting cases of benign and malignant breast lesions detected with MRI in our tertiary referral centre, including intracapsular and extracapsular implant rupture, silicon leak, fibroadenoma versus fibrosarcoma and a case of extensive fat necrosis. We would also like to demonstrate the usefulness of MRI in the preoperative assessment of patients with multifocal lobular and ductal carcinoma as well as postoperative surgical change and axillary node involvement. In all our cases, MRI was a valuable adjunct in the diagnoses and subsequent management of patients with breast disease.

(page number not for citation purposes) showed an increased rate of benign biopsy. This may be related to the increasing rate of wide bore needle (WBN) biopsies graded as B3 (indeterminate). Common B3 pathologies include atypical ductal hyperplasia (ADH), columnar cell change with hyperplasia or atypia (CCC) and intraduct papilloma (IP). Previous studies have shown an association of these lesions with malignancy [1,2]. Our practise is to recommend excision biopsy of these B3 lesions. We retrospectively audited surgical excision biopsies of B3 lesions between April 2004 and April 2005, recording mammogram findings, patient demographics, WBN and surgical excision pathological diagnoses. Twenty-five women age 50-70 (mean age 58) had excision biopsy of their B3 lesions; 64% were microcalcifications, 28% masses and the remainder distortions. The 14G core biopsy pathology included 38% ADH, 16% atypical lobular hyperplasia, 16% CCC and 12% IP. The surgical excision pathology available in 14 of these women showed ductal carcinoma in situ in seven and invasive ductal carcinoma in situ in three, justifying our practise. We discuss how the surgical pathology correlates with that of the WBN. This quantitative study retrospectively reviewed consecutive Mammotome procedures from January 2003 to July 2005. B3 outcomes were analysed by category combined with follow-up for any evidence of histological upgrade to carcinoma. Out of a total of 120 consecutive Mammotome procedures, 61 (58%) had a B3 outcome. The B3 category subdivided into: 37% (n = 23) atypical ductal hyperplasia, 37% (n = 23) as radial scars, with the remaining 26% (n = 15) in a mixed category containing mucoceles, lobular carcinoma in situ, and papillomata. A total of 42.6% (n = 26) of the B3 category underwent annual mammographic follow-up with no signs of recurrence, 41% (n = 25) proceeded to 3-yearly NHSBSP routine recall follow-up, and 9.8% (n = 6) proceeded to surgical followup with two patients being up-graded to carcinoma. Four patients were lost to follow-up. The incidence of carcinoma in the B3 category ranged between 3.6% and 6.3%.
Trends demonstrated that Mammotome excision for B3 lesions combined with annual mammographic follow-up can be safe practice providing each case is discussed within a multidisciplinary setting with regard to atypia, past history and concordance of imaging and results.

11
Radiological predictors of successful therapeutic wide local excision of ductal carcinoma in situ: findings from the Sloane project A Evans 1  The aim of this analysis was to ascertain whether mammographic unidimensional measurement (UDM), bidimensional product (BDP) measurement and pathological grade are helpful in predicting which patients could be offered a successful single therapeutic wide local excision (WLE) for ductal carcinoma in situ (DCIS). The study group was 505 patients with DCIS whose mammograms showed calcification, and in whom a nonoperative diagnosis had been obtained and a WLE attempted. Mammographic calcifications were measured in two planes at 90° on the oblique view and were classified pathologically as high, low or intermediate nuclear grade. In the sample, 342 patients had a successful first WLE and 163 patients had further surgery. A UDM <35 mm and a BDP <800 mm were associated with successful excision at first operation (69% vs 54%, P = 0.02 and 70% vs 27%, P = 0.0001, respectively). If the BBP cut-off had been applied to these cases, 16 unsuccessful WLEs would have been prevented but six successful WLEs may have been replaced by mastectomies. The histological nuclear grade did not influence the chance of a successful first WLE (66%, 69% and 80% for low, intermediate and high nuclear grade, respectively). The BDP maintained significance in subgroups based on nuclear grade more frequently than UDM. The BDP of mammographic calcification is a better predictor of successful WLE than UDM.

Clinical cases covering management of borderline lesions A Evans
Breast Institute, City Hospital, Nottingham, UK Breast Cancer Research 2006, 8(Suppl 1):P12 (doi: 10.1186/bcr1427) Lesions of uncertain malignant potential include radial scars, papillary lesions and mucoceles. Lobular neoplasia and atypical ductal hyperplasia (ADH) are often associated with such abnormalities and present similar problems. Columnar cell atypia and apocrine atypia, once their natural history has been elucidated, may join this group of lesions. The management of lesions of uncertain malignant potential has become a more common and complex problem in recent years. The introduction of first core biopsy and then vacuum-assisted biopsy devices has led to an increase in the nonoperative diagnosis of such lesions. These lesions may be incidental findings that do not represent the clinical or radiological abnormality. In the past, such lesions were managed by surgical excision (radial scar, papillary lesion and ADH) or by mammographic follow-up (lobular neoplasia). It is now recognised that the upgrade rates to ductal carcinoma in situ or invasive cancer vary in proportion to the degree of cellular atypia present and by the amount of tissue removed at percutaneous biopsy. Vacuum biopsy excision is also an option for some of these lesions. In this session we shall discuss a number of such cases to highlight the difficulties and dilemmas found when managing these lesions. No statistically significant differences were found concerning tumour size or involvement of axillary lymph nodes. Conclusions FFDM had a higher detection rate for DCIS but no difference was observed for invasive tumours. Recall rates were lower due to fewer technically inadequate examinations. FFDM performs well in high-volume population-based screening. In the United Kingdom, screening personnel (radiologists, advanced practitioners, breast physicians/clinicians and registrars) read breast screening cases and symptomatic radiologists read cases that have been referred to them. Our previous PERFORMS research has suggested that there may be differences in reading styles between these two groups owing to such differences in their real-life practice. We set out to investigate whether such previously noted trends in reading style were predictive of current performance in 2006. Consequently, we examined the proficiency of the two groups on the recent PERFORMS set of mammograms. The performance for a group of 15 symptomatic readers was examined as compared with 15 screening personnel over a set of 60 difficult mammographic cases that contained a range of features and mammographic classification types. Both groups were matched, as far as possible, on real-life factors that may affect reporting skill -such as case volume and reallife reading experience. Concentrating on the groups' specificity and sensitivity measures identified whether current tendencies in radiological reading style were comparable with those previously noted. Results indicate that, on this scheme, the symptomatic readers' tendency to 'over-read' comparative with screeners may still be evident.

15
Promoting early breast cancer presentation in women after their final routine breast screening mammogram A  encourage early help seeking among older women with breast symptoms. It will be delivered at the point when the women leave the routine protection afforded by the National Health Service Breast Screening Programme. It is in line with government-recommended practice and is complementary to the breast screening programme. The intervention is designed to increase women's knowledge about breast symptoms and risk, to promote disclosure of symptoms to someone, to reduce perceptions of barriers and to increase intentions to seek help. The rationale and evidence base for the intervention will be presented. We have shown that delayed presentation of breast cancer (≥3 months) is associated with poorer survival at all ages [1]. The intervention builds on evidence about risk factors for delayed presentation of breast cancer [2][3][4][5][6] and is informed by a theoretical framework about help-seeking for breast symptoms [7]. The ultimate aim of the intervention is to reduce the proportion of older women with breast cancer who delay their presentation, and thereby save lives. Results Of 300 biopsies, 264 were performed by radiographers. All cases had specimen radiography. The calcium retrieval rate was 80% in early 2004 and 95% by the end of July 2005. The calcifications were classified into cluster, tiny cluster and scattered cluster, according to the mammographic appearances. In the negative biopsies they were 32%, 43% and 25%, respectively. The comparison between the radiologist opinion against the final core biopsy results shows 90% concordance with pathology for opinions 2 and 5, and 75% for opinions 3 and 4 Conclusion Calcium retrieval rates for radiographers initially matched, then exceeded, those for radiologists. They also showed a steady improvement during the study period, rising to 95%. The cluster morphology did influence the calcium retrieval rate. There was a good concordance of radiology and pathology opinion. BUPA's mammography quality assurance programme comprises an annual audit visit to 31 hospital-based screening centres by one of a team of three quality assurance mammographers. Clinical competence and mammographic film quality are assessed and technical quality control data are also collected. Monitoring of overall image quality is also addressed by a 'quarterly' test object film. A patient dose survey has been undertaken with exposure and breast thickness data collected from all sites. The standard of mammography across screening centres is good and well within NHSBSP standards. Reject rates are less than 3% and are personalised to each radiographer, enabling feedback and learning. Continuing professional development is in place and sites have evidence of a system of peer review. Quality control checks are undertaken according to recommended standards and records are kept. The mean glandular dose for a lateral oblique film across all centres was 1.7 mGy, well below the recommended national reference dose level. Mean doses at individual hospitals ranged from approximately 1.1 mGy to 2.3 mGy. Image quality scores using the Leeds TORMAM are generally satisfactory and typical of elsewhere, with only a few films having scores less than optimum and driving a move to higher contrast film/screens at these sites. In The Netherlands a number of (screening) trials with digital mammography have started. The first trial began in 1999, in which the applicability of digital mammography was tested in a clinical environment. For this purpose a GE Senographe 2000D was installed in the Radboud University Nijmegen Medical Centre. The outcome of this trial was positive. Therefore in 2002 a second trial started at a static screening site in Utrecht with a Lorad Selenia system. In this trial, digital mammography was evaluated in a screening environment with its specific demands regarding workflow. In 2004 two more trials were started with mobile digital screening units. In these trials, a Fuji FCR Profect and an Agfa DM 1000 system were installed in the screening units. This summer a new trial will start at a static screening unit in Nijmegen in which digital mammography equipment from different vendors (General Electric, IMS, Sectra, Planmed) will be installed to test connectivity. Results of all trials will be presented with emphasis on physical and technical aspects and workflow issues. Problems with the mammography systems in the trials will be discussed. Besides this, some experiences with digital mammography equipment in Dutch hospitals will be discussed with emphasis on possible pitfalls.

19
Ultrasound and fine needle aspiration assessment of the axilla in patients with operable invasive breast cancer Methods The SPI is calculated from the minimum excision margin being divided by the total specimen weight (sw) to tumour diameter (td) ratio (sw/td). The standard of surgical performance increases with an increase in SPI. A review of histology reports provided SPIs for 97 WLE specimens in 96 patients treated by one surgeon. The mean SPIs for palpable tumours, ultrasound-guided tumours and wire-guided tumours were calculated. Results There was a significant difference between palpable tumours and impalpable tumours (Kruskal-Wallis test P = 0.007). There was no significant difference between ultrasound-guided and wire-guided WLEs (Mann-Whitney test P = 0.153). In the clinic, surgeons have to decide whether a patient has a true lump or not. This is often difficult, and particularly so for trainees. Nevertheless, the surgeon or breast physician must discipline themselves to formally characterise a symptomatic breast abnormality. This should be done by using the standard breast industry classification of 1-5. In this situation, P for palpability precedes the number. Surgeons will be assisted by their radiologists, who are able to offer a very high level of imaging support. The problem that arises is when a surgeon or clinician identifies a lump and the radiologist is unable to identify the lump at imaging. It should be remembered that approximately 15% of breast cancers are mammographically occult. In a recent case, a woman had an 8 cm lump in her breast. The radiologist reports a normal mammogram. Over the next 2 years five different junior doctors observe this lump. Finally, the patient sought a second opinion for her 8 cm Grade I cancer. Why did this occur?
The radiologist was a consultant, the requesting clinician was a (locum) junior. There is a natural tendency to rely on technology, which is usually better than a clinical examination -but not always. Consultant breast radiologists can appear intimidating to junior doctors. If that junior doctor is working in a poorly organised surgical breast clinic, then there is a potential for mistakes to occur. The solution is to have the diagnostic process in your breast unit so organised that risks such as these are reduced to a minimum. generalized scheme for quality evaluation will be presented. The most complex part of this scheme is the comparison of tissue samples obtained by image-guided biopsy with tissue samples subsequently obtained during surgery. Because image-guided biopsy programs retrieve histology specimens that are microscopically as valid as histology obtained from open surgery, comparing the histology from an image-guided breast biopsy with the histology from an open surgical biopsy is complex. One cannot use the well-known method of determining false-negative and false-positive rates. In addition, breast histology, itself, is quite complex. Some benign breast disease is quite focal and specific, such as fibroadenomas. Other benign breast disease is diffuse and not very specific, such as fibrocystic abnormalities. Furthermore, malignant breast disease is part of a histology spectrum starting with normal-looking breast tissue with atypical features, progressing to carcinoma in situ, and ending, finally, with infiltrating breast cancer. To illustrate how histological comparisons should be made for breast tissue, published results from a large, nationally funded study will be re-examined using the proposed scheme. Although the breast biopsy, itself, may seem like the hard work of a new breast biopsy program, it is not. After the first year of the program, follow-up of women who have been biopsied is the true, back-breaking, hard work. How a breast center should perform air-tight follow-up will be described. An update on the findings of this survey will be presented to the conference. Several recommendations were also made about the provision of regular mammographic services for women under 50 with a family history of breast cancer. It has become clear to the team developing and leading FH01 that service provision across the United Kingdom is very patchy and there appears to be significant geographical variation in the level of service the public can expect. We will examine some of that variation and consider whether mechanisms should be put in place to monitor service provision more closely. is a single-arm study of annual mammography in women aged 40-49 with a moderate family history of breast cancer. Originally, it was planned to recruit 10,000 women but this has been revised to 6,000. In this presentation we summarise the considerations that informed the study and its design, the problems encountered and the progress of the study so far. Implications for future studies in specific risk groups are discussed. Information gathered from the questionnaires, together with routine data from the KC62s, will allow us to compare the performance of units using radiographer-only double reading with that of other units. We will also be able to compare the performance before and after the change in reading protocol for units moving to radiographer-only double reading. The main outcome measures of performance will be cancer detection rates, standardised detection ratios and recall rates. In my experience the consultant radiographer within the breast services provides a similar role and clinical level of responsibility as that of the medical breast specialist. All professionals within such a role require the knowledge, skills and expertise to provide a high quality of service for all its users. The position has provided a key person who is used as a supportive resource for all the members of the multidisciplinary team. The consultant radiographer role has helped to enhance communication not only within the team, but also for the clients and patients and across other service boundaries. It has helped to promote the service and the sharing of good practice within the service, which in turn will enrich the service as a whole. Recruitment and retention of the staff is essential to the continued success of the service. The new training and educational pathways and recognition of the staff can only be of benefit to the service.

33
Workforce issues in breast imaging: the consultant radiologist's perspective Advanced practice means more than just a means to enable second reading of films. Without continuous feedback, involvement in assessment and support, this can become a boring task, with skill levels and motivation difficult to sustain. Increasingly, in many units practitioners are replacing other aspects of the traditional role of the consultant radiologist. Almost all advanced practitioners are film readers, but many have extended their roles further and now also do stereo tactic biopsy, ultrasound and ultrasound-guided biopsy, clinical examination, and localisation of impalpable tumours. We need to encourage the many eligible units who could participate in the current trial of radiographeronly screen film reading to join, in order to provide concrete evidence that radiographers are as good as radiologists in real-life practice.
With consultant radiologist posts becoming a little easier to fill than previously, we need to examine this role more carefully and decide what benefits practitioners can bring to a unit, and how to make this role a fulfilling and secure one for our radiographers in the future. This poster is a case presentation of a woman who was recalled from breast screening with suspicious microcalcifications. Follow-up mammograms revealed that it had been an artefact on the skin surface. The cause of this artefact was unknown and this poster attempts to identify a product that may have caused this calcific-like artefact. Previous research [1,2] highlights the potential problems associated with the use of deodorants and antiperspirants on mammographic interpretation. These studies demonstrate the association of the use of underarm deodorants with calcific-like artefacts. This poster not only examines deodorant as a possible source of the artefact, but also identifies other products as probable causes. Five products in total were used: stick deodorant, talc, body scrub, paint and body glitter. Samples of each of these were individually X-rayed using an anthropomorphic breast phantom. The subsequent films were compared with the original X-ray and one of the products was identified as the most probable cause of the artefact. As a result of this research, products not previously considered as potential causes of artefacts were identified. Our results are presented in a light-hearted format. We have a total of 10 film readers and have always double-reported with arbitration (by a single radiologist). Our sensitivity has been good with an SDR and a small cancer detection rate within the target QA standard.

Hormones and radiation as risk factors for breast cancer
Factors perceived as problematic were the increasing number of readers, a high proportion of inexperienced readers, combinations of cautious readers and lack of discussion. At the beginning of 2005 we adopted a new arbitration system. All prevalent round patients recalled by at least one film reader had their films reviewed by a minimum of three film readers, including at least one radiologist. This system was extended to include all incident round patients referred to arbitration. Recall rates and cancer detection rates were audited for 12 months.
We have now achieved a recall rate of 7.1/1,000 with no reduction in the cancer detection rate. Other benefits include an increased capacity in assessment clinics to accommodate age extension, reduced patient anxiety, better team working and a complimentary reduction in recall rates for incident round women.

42
A radiographer-delivered intervention to promote early presentation of breast cancer among women as they leave the routine protection of the NHS breast screening programme Our aim was to develop a radiographer-delivered psycho-educational intervention to be delivered at the point when the women leave the routine protection of the NHSBSP in order to: increase women's knowledge about breast symptoms and risk, promote disclosure of symptoms to someone close, reduce perceptions of barriers and increase intentions to seek help, and counter the interpretation by older women that their risk of breast cancer diminishes once routine screening ends. The ultimate aim of the intervention is to reduce the proportion of older women with breast cancer who delay their presentation, and to thereby save lives. We report the development of two variants of the intervention: a booklet and a 10-minute, radiographer-delivered, interview plus the booklet. Both variants of the intervention have been piloted within the South East London Breast Screening Programme and been shown to be acceptable and feasible to women and the Programme. A randomised controlled trial is planned to assess the effect of the intervention on delayed presentation and survival.

43
Using an aubergine as a phantom for practicing stereotactic guided core biopsy Acceptable variation is known as common cause variation and is deemed to be due to natural chance. The only way to reduce this type of variation is to remodel the whole process. Special cause variation is an exceptional variation that needs to be investigated to identify the cause. When examining reasons for special cause variation, the data are first checked for accuracy. The case mix is then reviewed, followed by processes and resources, and finally the individuals involved.
In Figure 1, the dots represent the proportion of Grade 2 breast cancers reported by West Midlands pathology laboratories. The reporting rate in the laboratory represented by the larger dot lies outside the common cause variation limits when compared with other laboratories. Work will be undertaken with this laboratory to identify the reasons for the special cause variation evident in these data. Assuming that the protocol utilised during the ALMANAC study was continued, the 25% of cases with positive lymph nodes would require a second operation to clear the axilla. This would represent a 20% increase in the number of cases requiring a second therapeutic operation. In addition, as over 80% of these cases had less than five positive nodes found, a full axillary clearance may be overtreatment. Analysis of the variation of lymph node positivity with size and grade demonstrates that these factors could be used to determine which women with a positive sentinel lymph node require a full axillary clearance and which women could be appropriately managed with a level 1 clearance, thus reducing the possible complications of lymphoedema.

50
The Sloane Project: a UK prospective audit of screendetected non-invasive carcinoma of the breast There was a good correlation with Van Nuys Score, supporting the use of this scoring system in routine practice to predict the potential benefits of referral for RT. In the West Midlands, breast screening services are visited every 3 years as part of a rolling programme. Ad hoc visits can occur during this period to a service or one or more disciplines if there are issues or concerns with a particular aspect of a service. Following a QA Team visit, a breast screening service can expect 3-month and longer term recommendations, and occasionally immediate recommendations, which highlight areas that require action to improve service delivery. In order to allow structured follow-up, recommendations are classified into the following disciplines; administration and clerical, radiography, radiology, medical physics/user QA, pathology, surgery, nursing and management. A detailed process has been put into place that tracks the receipt of recommendation responses from the service, which allows completed recommendations to be analysed by discipline, type and time to completion. This detailed analysis can be used to identify which recommendations are completed effectively and within the set timescale, to identify which types of recommendations take longer to complete than anticipated, and to identify ways in which the arrangement and/or wording of recommendations can be improved in order to ensure that recommendations are achieved.  [1,2]. When carefully applied, these techniques can give definitive answers to problems indicated on the original films [3]. However, when seeking extra information it is critical that the correct techniques are applied. If additional views are taken incorrectly or inaccurately, they can produce misleading results or false-negative findings. Therefore, our aim in this poster is to delineate proper applications and to demonstrate -with the help of case studies -just when and how the taking of additional views can prove to be a valuable tool. Equally, we show that their inappropriate use can disguise real pathologies that are intimated on the original films. The study looks at our experience in the first 29 patients who were referred for marker insertion.

To assess the clinical usefulness of radio-opaque marker insertion into malignant breast tumours in patients undergoing neo-adjuvant chemotherapy prior to consideration of breast-conserving surgery SE Evans, RA Crichton
Method The markers were all placed by a radiologist under ultrasound guidance into the central area of the tumour. The technique is described.

Results
Of the 26 patients who had a successful marker placement, 23 went on to have a good response to chemotherapy. In 15 of these the tumour was no longer clinically palpable. These patients underwent standard wire localisation with the hook of the wire placed into the 'tumour bed' using either stereotactic mammographic guidance focusing on the implanted marker (11 patients) or using ultrasound guidance if the tumour was still visible on ultrasound (four patients). This enabled the surgeon to perform a wide local excision centred on the 'marked' tumour bed even though it was no longer palpable. Conclusion We conclude that the insertion of the radio-opaque markers into breast tumours is a useful aid to guide breast-conserving surgery following neo-adjuvant chemotherapy. These initial results require validation by larger scale studies. Ultrasonography is a common follow-up procedure in breast screening to help elucidate the nature of potential abnormalities. Due to the inherent properties of the imaging modality, these images are of fairly low resolution, as compared with mammography, and therefore pose a different interpretation problem for the radiologist or imaging specialist. We describe the steps involved in the development of a PC-based training scheme for improving the interpretation of ultrasound images. Using the existing current literature, ultrasound image features were identified. The key features and the reported difficulty that individuals had in identifying these were then logged. The construction of an image database was then undertaken with the aim of amassing both static images and video clips of imagery that posed particular interpretative challenges. Various optional interfaces that allow participants to make and record decisions are currently being assessed. Once complete, the system will be fully trialled and then will be made available freely available to UK screening personnel. A change is underway, with conventional film-screen mammography being surpassed by the implementation of digital mammography and the associated soft-copy digital film reporting. However, there is an allied period of changeover with implementation that requires careful consideration to ensure optimal performance and efficiency of work tasks.
To examine the human factors implications, a combination of techniques (expert walkthroughs, verbal protocol analysis, workstation assessment) was applied to examine existing working practices during the implementation of digital mammography film reading. Radiologists and advanced practitioners within a UK NHS Breast Screening Unit participated to enable a thorough understanding to be gained of strategies adopted when using: routine conventional roller viewing of analogue (film) cases with analogue priors; routine soft-copy reporting with full-field digital mammography (FFDM) priors; trial FFDM with analogue priors viewed on a multiviewer; and trial FFDM with digitised analogue cases viewed digitally. A variety of changes in working practices were recognised to have occurred with digital implementation. There was an impact upon performance and efficiency of digital soft-copy reporting when viewing analogue priors. Subsequent recommendations for workstation design, working practices and training were produced to assist in improved implementation of digital processes in mammography.  The sensitivity of axillary palpation alone in detecting metastatic nodes was 50%, the specificity was 93%, the PPV was 87% and the NPV was 68%. For axillary ultrasound alone, the sensitivity was 72%, the specificity was 88%, the PPV was 84% and the NPV was 78%. The combination of clinical and ultrasound examination predicted metastases with sensitivity 79%, specificity 84%, PPV 82% and NPV 82%.
Combining clinical examination and ultrasound missed only 10% of metastatic nodes. In another 8% of cases, ultrasound or axillary palpation was falsely positive and FNA of these nodes could have reduced the risk. We recommend that routine axillary ultrasound with FNA is adopted prior to embarking on sentinel node biopsy.
biopsy rate in the prevalent screening round. Failure to meet this standard was of concern because benign biopsies are associated with high healthcare costs and patient anxiety. Method A retrospective review of the records of patients who had undergone benign biopsy (2001)(2002) was conducted to establish reasons for surgical referral and suggest corrective measures to enable the unit to meet the standard in the future.

Results
The review concluded that that the benign biopsy rate achieved ( Fig. 1) did not reflect the performance of the unit and recommended an action plan to improve the standard. This plan was implemented, and as a result the standard has improved in subsequent years. The aim of the study is assessment of the role of non-invasive imaging methods such as mammography, ultrasound and magnetic resonance imaging in the diagnosis of inflammatory breast diseases. The study is mainly focused on the pitfalls in the diagnosis of inflammatory breast cancer and differentiation against another inflammatory lesions, including breast tissue changes after radiation therapy. Inflammatory breast cancer has a mammographic pattern of inflammatory changes, such as skin thickening and stromal coarsening and/or diffusely increased breast density with or without an associated mass and/or malignant-type microcalcifications. Ultrasound is helpful not only in depiction of masses masked by the edema pattern, but also in the demonstration of skin and pectoral muscle invasion and axillary involvement. Magnetic resonance imaging is used to differentiate residual tumor from post-treatment fibrosis and glandular tissue. Older age is a risk factor for developing breast cancer and delayed presentation; however, routine breast screening on the NHSBSP ends at age 70. The aim of this study was to examine how well-equipped older women are to seek help following symptom discovery after their final routine mammogram. We conducted a questionnaire survey among a national sample of 850 women aged 67-73 years (response rate 84%). Fifty per cent of women underestimated their lifetime risk of developing breast cancer as less than 1 in 100, and 75% believed their age decreased the likelihood of them developing breast cancer or made no difference. Identification of 11 possible breast cancer symptoms varied from 93% of women for a breast lump to 14% for nipple rash. Women with no formal education knew fewer breast cancer symptoms than those with at least O-levels (P < 0.001). Poorer knowledge about breast cancer symptoms was associated with less confidence in being able to detect a breast change (P < 0.001) and with decreased intention to seek help in the event of symptom discovery (P = 0.02). Increasing knowledge of breast cancer and the confidence to detect breast cancer symptoms as women leave the routine protection of the NHSBSP may reduce delays in presentation by older women with breast cancer. The tool presented is easy to use and cheap to make. A selection of different sizes and shapes of mask is used to correspond to the various shapes and sizes of implant. The tool may be customised to a suit a specific implant. The tool is placed directly over the implant to mask the light transmitted through it. Subjectively, this gave a great improvement in the quality of the image of the breast tissue demonstrated.

Imaging methods in differentiation between inflammatory breast cancer and post-treatment changes
A study was undertaken to accurately measure the luminance (cdm 2 ) in all areas of the breast demonstrated on the mammograms. A wide range of readings were achieved throughout the breast and varied considerably as we measured through glandular and fibrous tissue as well as through adipose tissue and skin. We conclude that the relative brightness of the visualised breast tissue increases by a factor of 10,000 with the use of the tool to mask the implant. Mammographers and film readers have found this tool a useful aid to the viewing and accurate reporting of mammograms of the augmented breast.  For the West Midlands region, the 949 cases included in the national audit of screen-detected breast cancers and the 2,504 symptomatic cases included in the BCCOM Project were compared. Non-invasive breast cancers formed 5% of the symptomatic cohort, compared with 23% for screening. Symptomatic cancers were larger (mean size 24.5 mm vs 16.6 mm) and more likely to be node-positive (42.7% vs 27.5%). Forty per cent of the symptomatic breast cancers had mastectomy compared with 27% of the screen-detected cohort, and small (diameter <15 mm) invasive symptomatic cancers were more likely to receive a mastectomy than comparable screening cases (31% vs 16%). For invasive cancers with a known Nottingham Prognostic Index score, 20% of the symptomatic cancers fell into the excellent and good prognostic groups compared with 58% of screen-detected cancers. For women aged under 65, the proportion with screen-detected cancers receiving chemotherapy was lower than those with symptomatic cancer (20% vs 36%). Screening cases were also less likely to receive radiotherapy (48% vs 63%). We also analyse the best mammographic predictors of histopathological outcome. Materials and methods One hundred patients were randomly selected with microcalcification (MC) on their screening mammograms who underwent stereotactic core biopsy at our institution between August 2002 and August 2004. All the mammograms were retrospectively read by five readers independently. Each observer noted the various features and final analysis category for all MCs. Interobserver variabilities were calculated using Cohen's kappa statistics, Kilem Gwet's agreement coefficient 1 and the interclass agreement coefficient. The performance of radiologists and mammographers were determined using a logistic regression model. Overall best predictors of histopathology outcomes were also determined.
Results Interobserver agreement was moderate to good for distribution, moderate for the shape, moderate for final analysis category, poor for morphology, poor for variation in density of MC and poor for category on more MC on magnification. There is a significant difference in determining the benign nature of MC and the overall differentiation of MC between radiologists and mammographers favouring radiologists. There is no significant difference between them in determining malignant MC. The best predictors of histopathology were the morphology (P < 0.0001), distribution (P < 0.0042) and number of MCs (P < 0.013). Conclusion There is moderate interobserver variability in assessment of the final analysis category (benign vs malignant). The radiologists are significantly better than mammographers in determining the benign nature and overall assessment but not significantly better at determining definite malignancy. The morphology, distribution and number of MCs are the best predictors of histopathological outcome.
Radiographers expressed many concerns about mammography screening for these women. Issues raised included the necessity of mammography screening, legal consent, the role of carers and the need for education of BreastScreen staff. Wheelchairs, which prevented the mammography equipment from being positioned appropriately, were identified as a major barrier to producing optimally diagnostic mammography films. Importantly, radiographers indicated that the provision of relevant information in appropriate formats for these women would greatly improve the potential for a successful mammography completion. which is contributing to a higher than expected mortality rate from breast cancer. Barriers to successful mammography screening include the mammography procedure, which has been identified as problematic for these women [1]. A large study funded by the Australian National Breast Cancer Foundation investigated these barriers and determined solutions. One part of this study is reported here.
In situ analysis was used to identify barriers as they occur when women with disability have a mammogram. Under scrutiny was the interaction between the woman, the radiographer and the mammography equipment.

Materials and methods
The notes and imaging of women who were screened and assessed between April 1995 and March 2004 but who subsequently developed breast cancer prior to their next routine screen were retrospectively reviewed. A total of 11,341 women were assessed during this time period. Twenty false-negative assessment cases were identified but three were excluded as the abnormality assessed was located at a different site in the breast (one case) or in the contralateral breast (two cases). The remaining 17 cases were examined and the reasons for failure of the assessment process postulated.
Results Six cases were recalled for assessment of stromal deformity, four cases for calcification and seven cases for asymmetry. Some mammographic signs were misinterpreted, especially cases of distortion where the focal compression view was falsely reassuring. Some cancers may have been detected earlier if a core biopsy had been included in the assessment process. However, some cases were challenging and the diagnosis was not made despite undertaking full triple assessment. Conclusion Overall, the proportion of women undergoing falsenegative assessment in this study was very low. Strategies to further improve the accuracy of screening assessment are suggested. There was a significant difference in the 10-year relative survival rates for attenders (SD, IC, lapsed attenders and nonattenders; 85.2% and 53.9%, respectively; RR = 1.58; P < 0.00001). Women with IC had survival rates above those of nonattenders, highlighting the benefits of screening. Available online http://breast-cancer-research.com/supplements/8/S1 Page S19 of S21 (page number not for citation purposes) In order that women continue to receive an invitation for screening at 3yearly intervals when they move locations around the United Kingdom, screening services initiate 'Failsafe' batches at 3-monthly intervals or less. Currently there is a variety of practice in the specification of these failsafe batches within the NHSBSP, and variations in appointment booking practice associated with these batches. The National Co-ordinating Group for Administration and IT in Breast Screening is reviewing these practices with the aim of producing national guidance on the most effective approach, both in terms of accessibility and acceptability for the women involved and costeffective resource management for the service. This poster outlines the results of the review and documents the key points in best practice. The aim of this audit was to document the clinical features most likely to be associated with cancer and to clarify our guidelines for symptomatic recall. We identified 1,075 women with normal screening mammograms who had breast symptoms or signs recorded at screening and were subsequently recalled for assessment between April 1991 and July 2004. The results showed that cancer was found in 13/1,075 women. Of these, nine women had a lump (invasive cancers), two women had a skin dimple (invasive cancers), one woman had an eczematous nipple change (Paget's disease) and one woman had a bloody nipple discharge (DCIS). Our audit shows that the majority of symptomatic recalls did not have cancer (1,062/1,075), and no cancers were detected in women with thickening, nodularity, lumpiness, breast pain/odd sensations or nonblood-stained nipple discharge. More significant symptoms, which in some cases proved to be indicative of an underlying cancer, were a lump in the breast or axilla, a skin dimple, an eczematous/weeping nipple, a bloody nipple discharge and nipple inversion. Barcode data entry for screening mammography offers the attractive possibility of speeding-up a labour-intensive process. Currently, significant time is spent manipulating paper forms with the additional potential for error that this incurs. However, the barcode data entry systems introduced in some centres allow the film reader to enter only a decision. By contrast, film readers using paper forms frequently annotate the forms to record additional information. Concerned by the possible consequences of this information loss, we conducted an ethnographically based study of film reading with paperbased data recording to investigate use of the forms, and contrasted this with barcode data entry. Results reveal that the paper forms provide an important communication tool, fostering collaboration between readers. Specifically, they are used by readers: to indicate an area of concern, and the nature of that concern, to a second reader or arbitration meeting; to indicate to another reader that they have noted a particular feature but decided that it is not of concern; for future reference as a personal learning tool; and to express uncertainty and seek reassurance from another reader. We conclude that barcode systems should either support richer data entry or continue to be supplemented with paper forms. The age extension of the screening programme necessitated an approximately 35% increase in the screening workload. It was not possible to justify the expense of a second mobile unit as this would only be required 1-2 days per week. Finding additional sites for mobile units in small rural villages has become an increasing problem. For these reasons it was decided to increase our capacity on the existing mobile unit by using a shift system and extending the working day. Initial staff consultation showed a willingness to explore this new way of working. A full screening day now extends from 8 am until 8 pm and utilises four radiographers/assistant practitioner radiographers. As a result of the extended day, Somerset Breast Screening has fully implemented the age extension and is achieving a 36-month screening round without the capital expenditure of a second mobile unit. Surveys show a high level of popularity with our clients, especially for early evening appointments, and high acceptance levels from our radiographers and assistant practitioners.