Contrast-enhanced spectral mammography: what is the 'added value' in a symptomatic setting? Initial findings from a UK centre

Contrast-enhanced spectral mammography (CESM) is a new technology. Dual energy acquisitions during one exposure yield two sets of images: a low energy (LE) set, equivalent to standard full field digital mammography (FFDM); and a recombined set displaying contrast uptake. In our symptomatic breast service, specific patients, including those with a P4/5 clinical abnormality are offered CESM instead of FFDM. Despite encouraging data from Europe and the USA, there are, until now, no UK data to support its use in this setting.

Introduction: Wire localization techniques for impalpable breast tumours require wire placement ideally on the day of surgery. Tumour localization using iodine-125 seeds allows tumour localization to occur prior to surgery, improving both work flow dynamics and the patient experience. Newcastle Hospitals Trust is the first centre in the UK to adopt this technique. Here we present our initial experience of the first 100 patients to undergo wire-free surgery. Methods: From September 2014, data were prospectively collected on all patients undergoing iodine seed tumour localization. Seeds were placed under ultrasound guidance into tumours identifiable on ultrasound between 7 and 14 days preoperatively. Seeds were removed with the tumour after intraoperative localization using a gamma probe. Results: Our first 100 patients are included in this initial analysis. The majority of patients had a wide local excision, with 10 undergoing therapeutic mastectomy. Thirteen patients returned to theatre for positive margins or completion mastectomy, depending on the final pathology. No seeds were lost during use. One patient had a second tumour identified at the time of seed placement which required wire localization. No radiological complications occurred. Introduction of iodine seeds improved radiological workflow, with creation of a planned outpatient 'seed list', remote from the day of surgery and radiological high demand times. Conclusion: Iodine seed tumour localization in the UK is achievable, patient friendly and has great benefits for radiologists in terms of department workflow. Noticeably, patients (and surgeons) appear much more relaxed since the introduction of this technique and initial patient satisfaction surveys have been positive.

O5
Is surgical diagnostic excision always necessary for solid lesions with atypia? Nisha Sharma * , Rebecca Millican-Slater, Eldo Verghese Leeds Teaching Hospital NHS Trust, Leeds, UK Breast Cancer Research 2015, 17(Suppl 1):O5 Introduction: As part of diagnostic work for radiological abnormalities seen in the breast, there has been an increase in use of vacuum-assisted biopsies for diagnosis. This allows more tissue to be sampled and therefore leads to a greater degree of diagnostic accuracy. In addition to diagnosis, in some centres the same procedure has also been used for removal of the entire lesion-vacuum-assisted excision (VAE). This is sometimes offered in place of a diagnostic surgical excision in cases of B3 lesions. We wanted to examine whether VAE can be a safe alternative for B3 lesion that show atypia. Methods: We identified all patients, at Leeds Teaching Hospital NHS Trust, who had undergone a surgical diagnostic excision following a core biopsy which had revealed the following lesions: fibroadenoma, papilloma or radial scar with atypia (FEA, AIDP or ISLN) during the period between 2009 and 2013. We reviewed the slides of the core biopsy and the subsequent excision biopsy to confirm the histological diagnosis. Results: Twenty-nine cases in total satisfied our inclusion criteria. There were nine cases of fibroadenomas with ISLN and/or AIDP. None of the cases showed upgrading of the atypia. There were eight cases of radial scar that had either ISLN, LCIS, epithelial atypia or AIDP, of which two showed DCIS in the surgical excision. There were 12 cases of papilloma with either ISLN or AIDP; of these, five had DCIS on surgical excision. Conclusion: VAE is safe for fibroadenomas with atypia and radial scars with atypia provided the periphery can be adequately sampled, to help diagnose DCIS. Papilloma with atypia requires surgical excision due to complex histological architecture.

O6
Informed choice and consent among women attending for breast screening in the UK: data from a qualitative study Patsy Whelehan 1,2* , Andrew Evans 1 , Gozde Ozakinci 2 1 University of Dundee, Dundee, UK; 2 University of St Andrews, St Andrews, UK Breast Cancer Research 2015, 17(Suppl 1):O6 Introduction: While the concept of overdiagnosis can be difficult to understand, it has been shown that women wish to be informed about it. The latest breast screening information leaflet offers considerable detail about potential benefits and harms of screening, including overdiagnosis. However, it is unknown how much use women attending for screening make of the leaflet. We report qualitative findings on informed choice and consent within the UK breast screening programme. Methods: Participants were clients and mammographers from breast screening units in Scotland and London. Semi-structured, in-depth, individual interviews were conducted and thematic analysis performed. Results: Twenty-two clients were interviewed, aged 50−72: seven firstattenders and 15 subsequent, from a range of deprivation categories. Eighteen mammographer-participants included assistant, registered, and advanced practitioners, with a wide range of ages and lengths of experience. Most clients understood that screening aims to detect breast cancer early to improve the chances of survival. Several were aware of the possibility of false positive results and the risk of mammography inducing a cancer. Others could not name any risks of screening. Women had mostly either skimmed the information leaflet or not read it at all. Several mammographers recounted experiences where women had appeared to attend under pressure from others and where severe challenges existed in ascertaining consent. Conclusion: These qualitative findings that some women attend for breast screening with little knowledge of the balance of risks and benefits, and in some cases may encounter coercion, require further investigation. New methods of communication may be indicated.
Introduction: We investigated our sensitivity for axillary node staging, in patients presenting with symptomatic breast cancer from January to December 2012. Methods: Of 430 patients identified, 288 had first-line surgical treatment, 63 had neoadjuvant therapy first. Seventy-nine women were unfit for surgery, had less aggressive evaluation of the axilla and were excluded from sensitivity calculations. US axilla ± FNA were performed at presentation. Nodal disease prevalence, sensitivity for diagnosis and the NPV of our tests were calculated. In the neoadjuvant cases, pretreatment nodal status was not accurately known. Results: The prevalence of nodal metastases in our surgery first cases was 43% (123/288). Twenty-four per cent of cases were micrometastases (29/123). US sensitivity for macrometastases was 51% (48/94); 41% including micrometastases (50/123 Introduction: Patients with invasive breast cancer undergo axillary ultrasound ± ultrasound fine needle aspiration (US-FNA)/core biopsy for preoperative staging depending on the ultrasound appearance. At our institution, abnormal axillary lymph node assessment includes: a cortical thickness >3 mm, focal or eccentric cortical thickening, nodal shape (spherical) and replaced appearance with loss of echogenic nodal hilum. Our aims were to evaluate the accuracy of preoperative US + US-FNA/core biopsy for detecting axillary metastatic disease. Methods: Excluding those patients who underwent neoadjuvant chemotherapy, we identified 120 patients with invasive breast cancer between January and December 2013, which yielded axillary node metastases on final surgical pathology. We performed a retrospective analysis of the clinical records and used descriptive statistics. Results: Preoperative US correctly identified 60/120 (50%) patients with axillary metastatic disease, 42/60 (70%) had subsequent true positive US biopsies. Of the cases where a biopsy was not performed, 88% (53/60) had one or two positive nodes confirmed after surgery and 12% (7/60) had at least three nodes. Thirty-four of 60 (57%) were from the symptomatic population. Of the total 21 false negative US biopsies from the 18 patients, 81% (17/21) were performed via FNA and 19% (4/21) via core biopsy. Eleven of 18 (61%) were from the symptomatic population. Twenty-nine of 42 (69%) true positive US biopsies were from the symptomatic population. Conclusion: The results highlight the need for a review of our biopsy criteria, which may result in a decrease in our biopsy threshold. An increase in the use of core biopsies may yield greater accuracy in correctly identifying axillary nodal disease.

P3
Evaluation of the use of microbubbles in the ultrasound assessment of the axilla in breast cancer patients Nisha Sharma * , Isobel Haigh, Rebecca Millican-Slater, Benjamin Dessauvagie St James's Hospital, Leeds, UK Breast Cancer Research 2015, 17(Suppl 1):P3 Introduction: Contrast-enhanced ultrasound of the axilla can be used to identify the axillary sentinel lymph node. We introduced this into our practice in 2013. During the study period there was an upgrade of our US equipment. The purpose of our audit was to see the negative predictive value of CEUS biopsy of the SLN. Methods: This was a retrospective audit. In total, 110 patients with invasive breast cancer were identified at the breast MDT. The US core biopsy, surgical sentinel node biopsy and subsequent axillary histology were documented. Results: CEUS was successful in identifying the first draining lymph node in 88.1% (97/111). Eighty-three of 97 cases (86%) had a definitive biopsy (B2−B5) result with 13 being malignant and 69 were benign. Fifteen were non-diagnostic with B1 core biopsy. The prevalence of axillary metastases at surgery was 31% (30/97) (22 macrometastases, six micrometastases and two isolated tumour cells) of which 42% were detected by CEUS, with 100% specificity. Two of the 30 cases were in palpable, non-sentinel nodes. The negative predictive value of CEUS with core biopsy is 80% but 90% if only macrometastases are included. Methods: Prospective audit of data collated at the time of the microbubbles procedure together with multidisciplinary meeting records identified relevant screening and symptomatic patients with primary breast cancer treatment including axillary node surgery between 1 July 2014 and 1 July 2015. Descriptive statistics were performed.
Results: Sixty-four female patients underwent microbubbles injection and axillary node surgery. Overall combined sensitivity and specificity of microbubbles ultrasound/biopsy procedure were 67% (8/12) and 100% (52/52) respectively. Seventy-five per cent of operative sentinel node biopsies (45/60) showed evidence of previous needle biopsy (four axillary clearance specimens excluded). Needle biopsy detection of micrometastatic disease only, shortly after commencing microbubbles use, led to multidisciplinary meeting consideration of size of needle biopsy metastasis and ultrasound appearance of sentinel and surrounding nodes in triage of patients to type of axillary surgery. Results represent the combined learning curve of seven radiologists. The procedure was well tolerated by patients and technically easy to perform. The greatest challenges were optimising ultrasound machines for microbubbles visualisation, and finding time within busy clinics to perform the procedure. Conclusion: In this small patient cohort, introduction of microbubbles has facilitated reliable and effective identification of the sentinel lymph node for assessment of morphology on ultrasound and also biopsy.

P5
Preoperative identification and biopsy of sentinel lymph nodes using contrast- Introduction: Invasive lobular cancer has been associated with an increased risk of multifocal and contralateral disease. The literature suggests an incidence of contralateral disease as high as 15%. Current national (NICE) recommendations are that all patients with lobular carcinoma being considered for breast-conserving surgery have a preoperative breast MRI. The objective was to identify the rate of additional MRI-detected multifocal and contralateral disease in patients with a newly diagnosed lobular cancer to determine whether it is as high as the literature suggests. Based on our findings we hope to further explore whether another imaging alternative should be considered.
Methods: A retrospective search was done on PACS to identify all those patients in Northern Ireland investigated with bilateral breast MRI for a newly diagnosed cancer during a 15-month period. MRI findings were correlated with histopathology records from all regional labs and the data analysed.
Results: A total of 141 patients had an MRI for biopsy-proven lobular carcinoma. Within this regional cohort the incidence of additional contralateral and multifocal disease was 2.13% and 13.4% respectively. Conclusions: The incidence of contralateral lobular disease is 2.13%, within our reasonably large study population, significantly less than the currently cited 15%. Our study does show a significant increase in detection of multifocal disease in the same breast by MRI. Based on our results consideration should be given to exploring the use of tomography or contrast-enhanced mammography prior to MRI to attempt to detect further disease. This could potentially expedite patient care. Introduction: There is a reported increased incidence of contralateral disease at presentation of invasive lobular cancer (ILC). In our unit breast MRI is undertaken to assess the extent of all newly diagnosed ILC. If mastectomy is planned MRI is still carried out to assess the contralateral breast-we set out to evaluate this. Methods: We reviewed 160 reports of consecutive dynamic contrastenhanced breast MRIs of newly diagnosed ILC (January 2010−June 2015). All cases had been double reported according to the BI-RADS lexicon by two trained readers. We looked at the number of cases of BI-RADS MRM scores of 3 or above in the contralateral breast, second-look ultrasound findings, biopsy rate (U/S or MRI guided) and resultant contralateral cancer detection. Results: Of the 160 cases, 23 (14.4%) had an indeterminate or suspicious lesion reported in the contralateral breast. Three of these were contralateral cancers that had already been diagnosed by conventional imaging prior to MRI examination. Seventeen (10.6%) had second-look ultrasound of the contralateral breast: 15 lesions were subsequently biopsied in 11 women. Following negative second-look ultrasounds, two women had MRI-guided biopsy. MRI and subsequent work-up identified three women (1.9%) with previously undiagnosed contralateral malignancies. These were a 5 mm invasive ductal cancer, a 16 mm DCIS and a multicentric ILC.

Conclusion:
The incidence of 'conventional imaging occult' contralateral disease in ILC may be lower than initially reported. The routine use of MRI to assess the contralateral breast is potentially questionable. Introduction: Preoperative assessment of tumour extent is crucial in the management of breast cancer. MRI is currently indicated in cases of invasive lobular carcinoma on histology, a dense breast parenchymal pattern on 2D digital mammography (2DDM) or if there is a discrepancy between the clinical and radiological extent of disease. We compared the imaging characteristics of multifocal breast cancers on MRI, digital breast tomosynthesis (DBT), ultrasound and 2DDM to demonstrate the accuracy of each modality in the assessment of multifocal cancers.

P11
Methods: A retrospective review of 74 cases over a 4-year period was conducted. We included all cases whereby MRI or DBT identified two or more lesions that were considered suspicious or highly suggestive for malignancy. We compared the sign on MRI (including morphology and enhancement characteristics) against the lesion detectability on DBT.
The final histology of these lesions obtained following ultrasoundguided core biopsy, vacuum-assisted MR-guided biopsy or surgical excision was considered.
Results: There were 142 proven malignancies on histology out of the 74 cases, all of which were detected on MRI. The results of the MRI led to a change in surgical management in approximately 50% of cases but overstaged 16% of cases. Conclusion: MRI is more sensitive than the other three imaging modalities combined in accurately identifying multifocal breast cancer; however, DBT is still a useful adjunct in the evaluation of multifocal disease. There was no correlation between the pathological subtype and the non-detectability of multifocal cancer on the combined imaging modalities. Introduction: Breast magnetic resonance imaging (MRI) involves multiple aspects that are unique to a medical environment and may seem frightening and strange to a person from a non-medical background (the tunnel, no credit cards, keys or watches, loud noises, intravenous pump injector). The purpose of an information leaflet is to inform people about what they should expect, and to prepare them for the experience. During public consultation about breast MRI, we discovered that women considered the current information provided by the NHS (from several different hospitals) to be inadequate. They told us that their experience of the process of breast MRI had been more distressing that it would have been had they been better informed. We decided to ask their advice on the design of an information leaflet to see if it could be optimised to better prepare women for the experience. Methods: Public consultation was used to identify aspects of breast MRI that required explanation in an information leaflet and how they would like the information presented. We incorporated their suggestions into our new design and asked for comments at a second public consultation session.

P12
Results: The need for intravenous access, the tunnel, the nature of the loud noises that changed during the scan and knowledge that the radiographers could see and hear them throughout the scan were all emphasised as requiring explanation. The public suggested the use of multimedia including links to videos, sounds and personal accounts of experience.
Conclusion: Our new leaflet has been approved by the public and patients.

P14
Contrast-enhanced spectral mammography: what is the 'added value' in a symptomatic setting? Initial findings from a UK Introduction: Objectives were a comparative study of the radiation dose of two-view digital breast tomosynthesis (DBT) and two-view spot compression views in a symptomatic breast service. Methods: Two hundred patients were included in the study, 100 who had undergone two-view spot compression views and 100 two-view DBT. DBT was carried out using GE Seno Claire with an iso-dose setting and grid system. A retrospective computer-based search of patients in the two categories was undertaken and the accumulative dose for each technique was identified and recorded, as was the thickness of the breast from the original cc mammogram projection. The percentage variance of dose between DBT and spot compression views was calculated according to breast thickness.
Results: The mean accumulative glandular dose for the whole group regardless of breast thickness was 2.84 for DBT compared with 3.50 for spot compression views. In this patient population, the AGD was lower for DBT than for FFDM in 64 % of the patients. When patients were categorized according to breast thickness, the accumulative glandular dose of DBT was on average 22 % less than spot compression mammography with a reduction ranging from 53 to 1 %. There was no evidence in this study that dose reduction with DBT significantly increased with increasing breast thickness. Conclusion: The radiation dose of patients undergoing two-view DBT on average showed a significant reduction compared to two-view spot compression views. The dose reduction may be attributed to the grid and iso-dose technology used in the selected DBT system. Poster presentation: Being able to accurately determine the extent of a possible malignancy on a mammogram is an important task as this can affect the potential follow-up surgical treatment that a woman receives after breast screening. It is known that this can be a difficult task, particularly where the lesion has diffuse abnormalities. A potential computer-aided approach is to employ hierarchical clustering-based segmentation (HCS) and this interactive educational exhibit dynamically demonstrates this technique. HCS is an unsupervised segmentation process that when applied to an image yields a hierarchy of segmentations based on image pixel dissimilarities and so can be used to highlight areas in the mammographic image to aid interpretation. A set of 15 known difficult FFDM mammographic cases were selected from PERFORMS case sets where expert radiologists had previously delineated the extent of various abnormalities. Regions of interest (ROI) around these abnormalities were extracted from the DICOM images and processed using HCS algorithms resulting in a set of paired original mammographic ROI images and related HCS processed ROI images. In the exhibit these paired images are presented and delegates interactively select which of the pair they think best identifies abnormality extent. The original expert delineated abnormality is then provided as feedback. Over the course of the conference, data will be collected on how useful the HCS approach has been found and this information fed back to participants. The learning objectives are to demonstrate the potential of this approach in increasing the perceptual recognition of abnormal appearances. Introduction: Women who inherit a mutated copy of the BRCA-1 or BRCA-2 genes have a higher lifetime risk of developing breast cancer. There have been no large epidemiological studies looking at BRCApositive patients in the UK. Methods: Across the All Wales Genetics Service, individuals with confirmed BRCA mutation, since formal testing began (1995) to 1 January 2015, were included-identified from a prospectively gathered database. Genetics case notes were obtained and retrospective analysis carried out. Results: A total of 419 females with mean age 47 (19−81) were included in the study. Of these, 206 were identified using diagnostic testing with the remaining 213 undergoing predictive testing. Of the predictive group who subsequently had cancer, 18 (78 %) developed breast cancer. Seven (39 %) had wide local excision (WLE), six (33 %) had single mastectomy while the remaining five (28 %) had bilateral mastectomies as their primary operation. Five of the predictive group (22 %) had ovarian cancer. Of these, four (80 %) went on to have prophylactic breast surgery too. Of the 13 patients who underwent WLE or single mastectomy, four (31 %) went on to have completion risk reduction mastectomies (RRM). From the remaining 190 individuals in the predictive group with no cancer diagnosis, 102 (54 %) have had no risk reduction surgery, Breast Cancer Research 2015, Volume 17 Suppl 1 http://www.breast-cancer-research.com/supplements/17/S1 32 (17 %) RRM only, 31 (16 %) BSO only and 25 (13 %) underwent both procedures. Conclusion: There is variation in the surgical management of BRCA positive patients in Wales. This has implications for service allocation and demands for screening for these high-risk patients.

P22
Audit of high prevalent breast screening recall rates: Torbay Hospital Alexander Crowther * , Rebecca Green Torbay DGH, Torquay, UK Breast Cancer Research 2015, 17(Suppl 1):P22 Introduction: The target percentage of women recalled after prevalent round breast screening is <7 % with minimum standards <10 %. Torbay Hospital's prevalent round recall is high at 11.4 %. We plan to assess patterns of recall by category to see if any particular reason for recall could be decreased. Methods: Retrospective audit of 12 months of prevalent round recalls March 2013-February 2014. All age groups were included. Each recall was grouped into one/more of the following categories: calcification, welldefined mass, ill-defined mass, asymmetric density, distortion, clinical, other. We will calculate the proportion of recalls per group that proved to be malignancy and assess to see if any category was a poor predictor of malignancy. All histology proven malignancies from 2012/13 and 2014/15 will also be categorised by group. Results: There were 215 recalls for ages 49-69, 15 proven malignancies. 77% of Ill-defined mass, 22% of distortion and 10% of calcifications recalled proved to be malignant and are the strongest predictors of malignancy. Well-defined mass and asymmetric density had 0% malignancy rates and accounted for 129 (59.4 %) of prevalent recalls. Thirteen clinical recalls (1.4 %) were also 0 % for malignancy but beyond the control of the screening service. Further audit was performed looking at the proven malignancies from 2012/13 and 2014/15, which showed a total of 33 malignancies with 13 calcifications, 17 ill-defined masses, one asymmetry, one distortion and one clinical recall. Conclusion: A high proportion of recalls (60 %) are for well-defined mass and asymmetric density which have poor predictive outcome. These groups are potential areas to decrease recall rates. A total 1.4 % of clinical recalls are beyond the control of the screening service, which would bring prevalent recalls to a compliant level of 10 %.

P23
Outcome of ultrasound of the mammographically normal contralateral breast in patients recalled to the screening assessment clinic Preet Hamilton * , Simon Lowes, Sheetal Sharma, Alicia Wright, Alice Leaver Gateshead Hospitals NHS Trust, Gateshead, UK Breast Cancer Research 2015, 17(Suppl 1):P23 Introduction: Women diagnosed with breast cancer are at increased risk of contralateral breast cancer; some of these cancers will be synchronous and mammographically occult (M1). Ultrasound may detect M1 breast cancers but also benign lesions that necessitate needle testing, conferring additional patient morbidity that could be termed 'over investigation'. Local guidelines for ultrasound of the M1 contralateral breast vary between units. We present a retrospective audit of contralateral M1 breast ultrasound within our screening assessment clinics. Methods: Screening and pathology hospital databases of 2013 and 2014 identified records of 331 women with screen-detected breast cancer. Descriptive statistics were performed. Results: All 331 women underwent ipsilateral breast ultrasound; 288 (87 %) underwent ultrasound of their contralateral mammographically normal (M1) breast. Six contralateral breast lesions were needle sampled: four B2 lesions, two B3 without atypia. No subsequent breast cancer has been detected in any of these patients to date. Conclusion: Two years of routine contralateral ultrasound has yielded no cancers but also very few benign biopsies. Ongoing audit and discussion of risk/benefit to patients is indicated. . For ethnicity, the Caucasian population showed a positive correlation while Asian, a negative correlation. This was more significant in the Pakistani and Bangladeshi groups. Interestingly, when the results were adjusted for deprivation, ethnicity did not show a significant correlation with uptake rates. Conclusions: Our results clearly show that the more deprived an area, the lower the breast screening uptake rate. Moreover, the higher the proportion of Asian in a population, the lower the uptake rates and this is more significant in the Pakistani and Bangladeshi group compared to the Indian and Chinese. Overall the impact is most marked in the prevalent round.

P25
Does tomosynthesis increase confidence in grading the suspicious appearance of a lesion? An audit of cancers diagnosed in the assessment clinic using tomosynthesis: initial experience at Avon Breast Screening Unit Gillian Clark * , Alexandra Valencia Avon Breast Screening Unit, Bristol Breast Care Centre, North Bristol NHS Trust, Bristol, UK Breast Cancer Research 2015, 17(Suppl 1):P25 Introduction: Tomosynthesis is a new technology that is being used increasingly to evaluate the breast for assessment in the UK. It has, however, been approved as a screening tool in the United States, Canada and several European countries. We implemented tomosynthesis in the Assessment Clinic at Avon Breast Screening Unit (ABSU) last year as recommended by the NHSBSP. A retrospective audit of 134 consecutive cancers diagnosed from 9 June 2014 to 31 December 2014 was performed. The aim was to evaluate whether tomosynthesis gives additional information to increase the grading of mammographic features of a lesion seen on initial screening mammography and increase the assessor's confidence. Result: A total of 134 cancers were reviewed. Sixty-six lesions were graded the same on screening mammography and the assessment tomosynthesis. Thirty-six were M5 lesions at screening and assessment. Thirty M3 or M4 lesions remained unchanged. One patient had an M3 lesion that was downgraded. Three patients had incidental cancers found on ultrasound. Sixty-four lesions were upgraded with tomosynthesis. Forty-four of 64 M3 or M4 lesions were upgraded to tomosynthesis 5. Twenty of 64 were upgraded from M3 to tomosynthesis 4. The morphology of the lesions upgraded was spiculated 30/64, 7/64 distortions and 7/64 ill-defined densities. Thirty-one of 44 tomosynthesis 5 lesions measured 10 mm or less.
Breast Cancer Research 2015, Volume 17 Suppl 1 http://www.breast-cancer-research.com/supplements/17/S1 Conclusion: Tomosynthesis is excellent at showing the spiculate nature of lesions, upgrading the appearance of a lesion from M3 and M4 to tomosynthesis 5 which increases the assessor's confidence during the assessment clinic. It is also excellent in helping identify small suspicious lesions of 10 mm or less. However, ultrasound should always be performed in addition to tomosynthesis as lesions may rarely be downgraded.

P26
An audit of marker placement in stereotactic guided biopsy Jane Prady 1,2* , Lucy Hill 1 Introduction: Anecdotal evidence suggests that there is a greater incidence of marker migration using large volume sampling techniques in stereotactic guided breast biopsies. Methods: Prospective study of 130 biopsies with markers done between June and December 2014. Markers more than 10 mm from the target lesion were considered migrated. The aim of the audit was to quantify the number of markers migrating, distance and direction of migration and conditions under which markers migrate. Results: A total of 12.3 % had migrated markers: 10.7 % from use of the Bard Encor system and 1.5 % from use of the Bard Vacora system. The greatest marker migration occurred using a latero-medial approach. The majority of migrated markers were deeper than the target lesion. Marker migration was significantly greater using the Encor system within lucent breast tissue. Firstly, further audit is required incorporating lesion size, routine vacuuming of the cavity before deployment of marker, specific sequencing of marker films, correlation of compressed breast thickness and target depth, clinical impact of marker migration and possible development of expanding marker. Secondly, the breast screening service should provide guidelines regarding distances, thresholds and targets for marker migration. Conclusion: This audit found that marker migration occurred predominantly within lucent breast tissue and using the latero-medial approach when using the Bard Encor system.

P27
Handheld ultrasound-guided 20 mm basket Intact breast lesion excision system biopsy for excision of benign breast lesions Simon Lowes * , Alice Leaver, Alice Townend, Jackie Westgarth, Peter Newton, Dianne Hemming, Alan Redman Gateshead Hospitals NHS Trust, Gateshead, UK Breast Cancer Research 2015, 17(Suppl 1):P27 Introduction: In selected patients, our Unit has recently moved from handheld ultrasound-guided vacuum-assisted core biopsy (VACB) piecemeal acquisition of tissue to the handheld Intact Breast Lesion Excision System (Intact). Intact removes a single piece of tissue, potentially allowing radiologists to excise the entire lesion as well as allowing pathologists to visualise lesion architecture more easily and to calculate margins. We evaluated our early experience of excising benign or likely benign breast lesions using the 20 mm Intact. Methods: Prospective data collection was performed on all patients undergoing handheld ultrasound-guided Intact excision under local anaesthetic in 2014 and 2015, which comprised 19 lesions in 18 female patients aged 29−73. Results: The device was technically straightforward to operate and welltolerated by patients with no significant complications. Handheld needle orientation was difficult within dense glandular tissue (only one acquisition is possible per needle), but improved with increased operator experience. Achieving adequate analgesia required higher quantities of local anaesthetic than for the equivalent VACB. Pathologists found specimens easier to interpret than VACB samples. In all cases adequate excision was completed sonographically at one outpatient appointment, but in six cases a second Intact biopsy and/or a VACB was required to complete that excision, with extra cost implications. In two patients with M3 microcalcification the Intact pathology demonstrated ductal carcinoma in situ, leading to surgical wide local excision. Conclusion: Our early experience shows Intact as a reliable and effective tool for handheld diagnostic and/or therapeutic excision of selected breast lesions.

P28
The 3.5-year to 13.5-year follow up of 137 lesions of uncertain malignant potential (B3 lesions) diagnosed by vacuum-assisted biopsy alone Karen Mullin * , Anne Turnbull, Sharma Puri, Mark Bagnall Royal Derby Hospital, Derby, UK Breast Cancer Research 2015, 17(Suppl 1):P28 Introduction: Vacuum-assisted biopsy (VAB) was introduced in Derby in 2001, as the second procedure after 14g core biopsy. We present 3.5-year to 13.5-year follow up of cases where B3 lesions have been managed with VAB alone. Methods: The NBSS and local BASO databases were searched from January 2002 to December 2011 for all cases with B3 histopathology, a VAB procedure and no surgery. Screening and symptomatic women were included. Results: There were 137 women who met the criteria. The pathologies found are presented in Table 1. The cases where atypia was found were individually discussed at MDT to ensure that the abnormal feature had either been excised or very well sampled. Only one breast cancer has developed at the same site in a woman who had 5 mm calcification excised at VAB. This lobular cancer was identified 4 years later at recall from annual surveillance. Five other cancers have developed in the 137 cases, one contralaterally and four different lesions in different sites in the same breast. Conclusion: This study provides further evidence for the safety of the use of VAB alone in the diagnosis of B3 lesions in the longer term.

P29
Quantitative study: should vacuum-assisted biopsy be the first biopsy approach in breast interventional techniques in stereotactic guided microcalcifications rather than 14 gauge core needle biopsy? Anuma Shrestha * , Louise Wilkinson, Rosalind Given-Wilson, Judi Curtis St George's Healthcare NHS Trust, London, UK Breast Cancer Research 2015, 17(Suppl 1):P29 Introduction: Stereotactic guided 14 gauge core needle biopsy (14GCNB) and vacuum-assisted biopsy (VAB) are the two commonly used biopsy methods for obtaining an accurate diagnosis for microcalcifications. Breast Cancer Research 2015, Volume 17 Suppl 1 http://www.breast-cancer-research.com/supplements/17/S1 Retrospective review of 399 patients who underwent biopsy for breast microcalcification during screening assessment from April 2012 to March 2013 was used to evaluate the performance and cost-effectiveness of both methods. Methods: The repeat biopsy rate, diagnostic accuracy, time taken and cost of both methods was calculated. Microsoft Excel (2010) and SPSS 22 were used for statistical analysis.
Results: The repeat biopsy rate for 14GCNB was 10 % and VAB was 6 %. Specificity, PPV and NPV were all 90 % or higher when compared against post-surgical final diagnosis in both methods. The sensitivity of VAB was 93.75 % vs. 71.88 % for 14GCNB for first biopsy. There was no significant difference in procedure time between two methods (p = 0.291). VAB necessitated almost double the rate of clip deployment compared with 14GCNB. The cost of VAB would be £69,922 greater than 14GCNB if used as the first-line biopsy method in this series.
Conclusion: This study found VAB to have higher sensitivity than 14GCNB. There was also a trend for lower repeat biopsy rate, higher diagnostic accuracy and lower surgical upgrade with VAB. If VAB had been used as the first biopsy method for microcalcifications, the cost would have been significantly higher. 14GCNB is a cost-effective but less sensitive first biopsy method for selected microcalcifications. Introduction: B3 management balances safe treatment of potential malignancy against the morbidity of surgical excision of benign lesions. Vacuum-assisted biopsy (VAB) increases diagnostic accuracy, removing some lesions entirely without surgery. Few follow-up data are available to assess the safety and effectiveness of this approach.
Methods: A total of 215 patients with B3 biopsies without atypia were identified using Labcentre histopathology codes at a single centre. Hospital and NBSS records were analysed to identify patients who were treated with VAB and mammographic surveillance alone and to determine outcome over a follow-up period of 52−149 months (median 85). Local Labcentre and regional Pathlinks histopathology records were independently checked. Mammograms of ipsilateral re-presentations were reviewed by a consultant radiographer and consultant radiologist to determine whether lesions developed at the site of B3 biopsy. Results: Twenty per cent had excision biopsy (42/215) of which <5 % (2/42) contained carcinoma. A total of 144 patients had VAB which identified 30 high-risk cases analysed separately (DCIS, B4 or atypia). In total, 114 B3 lesions without atypia (on either core biopsy or VAB) were followed mammographically after VAB with no surgical intervention. Four patients re-presented to the service with malignancy; 37, 38, 41 and 67 months after VAB. Sixty-one per cent (69/114) of individuals were screened locally 2012−2015.
Conclusion: VAB of B3 biopsies without atypia appears to be safe with no representations in the first 3 years and overall carcinoma and DCIS incidence of 3.5 % over 7 years (4/114). National guidance on B3 lesion management is required. Introduction: B3 management balances safe treatment of potential malignancy against the morbidity of surgical excision of benign lesions. Vacuum-assisted biopsy (VAB) increases diagnostic accuracy, removing some lesions entirely without surgery. Few follow-up data are available to assess the safety and effectiveness of this approach.
Methods: A total of 129 patients with B3 VAB with atypia were identified using Labcentre histopathology codes at a single centre. Hospital and NBSS records were analysed to identify patients treated with VAB and mammographic surveillance alone and to determine outcome over a follow-up period of 52−142 months (median 85).
Nine patients re-presented to the service with invasive carcinoma (six ipsilateral) and two with DCIS (both ipsilateral) between 12 and 80 months. The ipsilateral re-presentation rate was highest for ADH (5/49) and LCIS (2/21). In the absence of ADH or LCIS, the only ipsilateral representation was one low-grade DCIS, 62 months after VAB.
Conclusion: Re-presentation with ipsilateral carcinoma following VAB excision for ADH and LCIS is comparable to surgical excision for ADH and LCIS. National guidance is required.

P32
Use of WHO checklist in interventional breast radiological procedures Introduction: A breast imaging report is a key component of the breast cancer diagnostic process. The report must be clear and concise to avoid ambiguity and confusion. However, substantial variation in the information provided in a breast imaging report is not uncommon to see. We sought to develop a report template containing a summary of all essential information pertinent to the surgeons and the radiologists. Methods: Breast surgeons and radiologists were consulted as to what was required in a report and they stated breast density, correlation with clinical findings, lesion characteristics, R1−R5 category, site and size of lesion, and is clinical area of concern biopsied. A retrospective audit of 30 breast imaging reports of recently diagnosed carcinomas between October 2014 and January 2015 were reviewed to see if these were recorded.
Results: Ten per cent of reports did not mention breast density. The most frequent information provided is lesion size (ultrasound 100 %, mammography 73 %). Correlation with referral was unclear in 10 %, R1 −R5 category not given in 3 %. Site of lesion was not provided in 3 %. Seven per cent of the reports were 3−4 pages long, described as confusing and difficult to read by the two data extractors. Thirty per cent of reports were not separated into mammography/ultrasound/biopsy sections. There were 23 different ways of characterising lesions on mammography and 24 on ultrasound.
Conclusion: The audit highlighted the need for a breast reporting template that met the needs of the clinicians to ensure the relevant facts were included to further improve the patient pathway. Introduction: Incidental findings of breast abnormalities from crosssectional imaging (CT and PET-CT) are a relatively common source of referral for breast assessment at our unit. We sought to describe and quantify our local experience of these referrals and to determine which cross-sectional imaging findings were more predictive of malignancy. Methods: Retrospective review using radiology information system searches for mammography referrals resulting from CT and PET-CT scan findings performed over a 5-year period (July 2010−July 2015) in Oxford University Hospitals NHS Trust. Studies in patients with known active breast malignancy were excluded. Cross-sectional imaging characteristics of the abnormalities were collected including CT enhancement, PET avidity, size and shape. Assessment imaging features, subsequent biopsy and clinical outcomes were recorded. , and more progressed from an initial diagnosis of DCIS, to a final diagnosis of DCIS with an invasive component (12.5 % compared to 2.1 %), p = 0.05. However, differences in the average Van Nuys Prognostic Index score were not statistically significant in Asian (7.13) and Caucasian (7.51) patients, p = 0.236. Interestingly, significantly more Asian women were treated with mastectomy (47.9 %) compared to Caucasian women (22.9 %), p = 0.015. Conclusion: Asian women presented with a larger tumour size, more progressed to a diagnosis of invasive carcinoma, and more had mastectomies compared to Caucasian women. Since fewer Asian women are presenting via the screening programme, education and awareness of breast cancer and screening needs to be increased in Asian women to increase their screening uptake rates.

P38
Comparison of prognostic indices in symptomatic and screen-detected invasive breast cancer in Asian and Caucasian women Andrew Steele 1* , Anil Jain 1,2 , Navneet Randhawa 1 , Philip Foden 1 , Julie and higher proportions with poor prognosis (33.8 %), compared with Caucasian patients (11.9 %) (p <0.001). Multivariable analysis showed invasive grade and tumour size were statistically significant independent discriminators with lymph node status as borderline significant. However, there was no statistically significant difference between the ethnic groups for screen-detected invasive tumours. Conclusion: Prognostic indices in Asian women were worse in symptomatic breast cancer, but similar in screen-detected invasive cancer, compared with age-matched Caucasian women. Greater initiatives need to be implemented to promote breast cancer awareness, education and screening among the Asian ethnic minorities.

P39
Prospective Introduction: Current practice in our unit as agreed with the local Cancer Network Group is for women over 40 years presenting with breast pain and with a normal clinical examination to have a mammogram. NICE recommends no imaging in this group of patients. The aim was to measure workload impact from current practice, and assess diagnostic yield.
Methods: Retrospective audit of imaging and biopsy in female patients over 40 years, presenting with breast pain, and who had normal clinical examination.
Results: A total of 100 patients, aged 40−65, from 30 clinics over 3 months, 2014. Eighty normal mammograms. Seven of these had ultrasound for focal tenderness or probable glandular tissue, all of which were normal. Twenty abnormal mammograms: eight calcifications, six asymmetry, five discrete masses, one implant rupture. Total imaging workload: nine requests for previous imaging from elsewhere, eight further mammographic views, 11 ultrasounds, two stereo core biopsies (benign), one ultrasound-guided FNA followed by core biopsy (malignant). Yield: one cancer (25 mm grade 2 invasive ductal, negative sentinel lymph node). Conclusion: Workload is appreciably impacted by breast pain investigations. The final diagnosis was often delayed because of the wait for pathology results and previous imaging, increasing patient anxiety. The cancer detection rate number is too low for significance, but nevertheless compares favourably to screening. After discussion with clinicians it was decided to keep to our current practice as a means of opportunistic screening, particularly as our unit is in an area of poor screening uptake.