British Society of Breast Radiology Annual Scientific Meeting 2016

Jennifer Ferguson2, Guy Stevens1, Robert Hills2, Kate Gower Thomas1 
1Breast Test Wales, Cardiff, UK; 2Cardiff University School of Medicine, Cardiff, UK 
 
 
Correspondence: Jennifer Ferguson 
Introduction 
 
To determine the rate and classification of interval breast cancer (IC) in the two years preceeding and following the introduction of digital mammography in 2011 in a national screening programme. 
 
Methods 
 
Two 2-year bands were used to ensure completed data for pre and post digital groups. Retrospective analysis of a prospectively collected database of IC was undertaken, identifying women who had undergone screening in the two years prior to and after the introduction of digital screening mammography (1.1.10 until 31.12.11 and 1.1.12 until 31.12.13). Chi squared analysis was used to compare the rate between the two groups and the rate of false negative (FN) and minimal signs (MS) categories considered together. 
 
Results 
 
Of 87,868 screened before digital between 1.1.10 and 31.12.11, 103 developed IC (0.12%) and of 100,672 screened with digital between 1.1.12 and 31.12.13, 145 developed IC (0.14%) (p = 0.12). Those women with FN and MS interval cancer numbered 17 (16.5%) predigital and 41 (28.3%) with digital; (p = 0.04). The number of true IC was similar in the pre (49.5%) and post (46.9%) groups. 
 
Conclusion 
 
The rate if IC is similar to the NHSBSP standard (1.2 per 1000 women screened). Despite earlier groups indicating that digital technology will lessen the number of IC, especially those ‘missed’ (FN and MS groups) which are often asymmetric densities and poorly defined masses, this has not been found in our study. The reasons for this will be discussed and examples will be shown.


Introduction
The low sensitivity of ultrasonography (US) in diagnosing axillary lymph node metastasis in breast cancer patients has sparked the evaluation of various tools in an attempt to increase pre-operative sensitivity. We compared axillary lymph node metastasis probability scores with post-surgical findings, using the Memorial Sloan Kettering Cancer Centre (MSKCC) nomogram and Evidencio, two freely available online predictor tools.

Methods
We retrospectively evaluated 450 breast cancer patients and analysed data from 194 patients. Ultrasound images were evaluated to assess axillary lymph node status. Patients were divided into groups 0, 1 or ≥2 nodes based on the number of post-operative positive nodes. The difference in mean scores across the 3 nodal groups for both nomograms was analysed using the one-way ANOVA test. The Nottingham Prognostic Index (NPI) was also calculated for each patient. Data was analysed using SPSS ver20 and p<0.05 was considered statistically significant. Results There were significant differences in mean scores across the 3 nodal groups when using MSKCC (p= 0.000), and Evidencio (p= 0.000). A strong positive correlation was found between MSKCC, Evidencio and NPI. (r s = 0.671 (MSKCC vs. Evidencio), r s = 0.721 (Evidencio vs. NPI), r s = 0.656 (MSKCC vs. NPI), p= 0.000 (for all)). Conclusion Both MSKCC and Evidencio nomograms can be used to predict axillary node metastasis and guide patient management. Further evaluation is recommended before omitting Sentinel Lymph Node Biopsy or Axillary Node Dissection in patients with very low scores and prompting a 'second look' US in patients with high scores.

Introduction
Pre-operative staging of the axilla is mandated. In 2011 the ACOSOG Z0011 trial indicated that women with small (T1 -T2) breast cancers and ≤2 nodes positive at SLNB may not require axillary clearance, resulting in a change to surgical practice. Pre-operative ultrasound is still routine despite studies showing between 38% and 46% of symptomatic women with positive pre-operative staging had ≤2 nodes positive at surgery and were thus potentially overtreated or eligible for the POSNOC trial. We measured the impact of current staging in one UK breast screening service. Method Data were extracted from the unit's National Breast Screening computer System between 01/04/2008 and 31/03/2015. Axillary staging was compared with final pathology and treatment. Results 164 of 776(21.1%) invasive cancers were node positive. 90 (11.6%) had an axillary biopsy, 54 were positive for cancer (32.9% of the node positive cases). Of these 22 (40.7%) had neoadjuvant treatment, 32 (59.3%) proceeded directly to axillary clearance (mean node count 14.6). 17 (53%) of those who had axillary clearance had ≤2 positive nodes. This compares with 82% of node positive women with a negative biopsy and 74% of node positive women with a normal ultrasound. Summary This small series suggests significant numbers of women are being denied entry into POSNOC or being potentially overtreated because of routine pre-operative axillary staging. A much larger data set is required to confirm this and predict who would benefit from preoperative axillary staging. A bid to use the whole ABS data set is with CRUK.

Introduction
To determine the rate and classification of interval breast cancer (IC) in the two years preceeding and following the introduction of digital mammography in 2011 in a national screening programme. Methods Two 2-year bands were used to ensure completed data for pre and post digital groups. Retrospective analysis of a prospectively collected database of IC was undertaken, identifying women who had undergone screening in the two years prior to and after the introduction of digital screening mammography (1.1.10 until 31.12.11 and 1.1.12 until

Introduction
In the Republic of Ireland, national quality assurance standards state that patients over the age of 35 presenting with breast symptoms should routinely have two-view mammogramography. Our aim was to determine the breast cancer detection rate in women aged 35-39 with low risk factor profiles and normal clinical examinations. Methods A retrospective analysis of all mammograms performed in patients aged 35-39 at our institution from 2011-2015 was performed. Patients with moderate or high familial risk, previous breast cancer or chest radiation, males, GP and internal hospital referrals, and those with abnormal clinical examinations were excluded. Included women had "normal", "benign" or undocumented examinations. Results of imaging including subsequent ultrasound and histopathological information was obtained from EPR and PACS systems. Results 4,087 patients aged 35-39 had bilateral mammograms from 2011-2015. 2,149 patients were excluded from analysis. Of 1,938 included women, 4 (0.21%) were diagnosed with breast cancer confirmed at histology: 2 cases of invasive ductal carcinoma (8 and 2mm) and 2 of DCIS (4.5mm high-grade-DCIS and 2mm lowgrade-DCIS). Other histological findings included 2 B3, 40 B2 and 2 B1 lesions resulting from mammographic screening. Overall, 114 biopsies were performed; biopsy rate of 6.09%. 69 (60.53%) were undertaken due to mammographic findings. Conclusion 2.1 cases of cancer were detected per 1000 women screened. This figure would be below accepted international thresholds to undertake screening mammography and raises radiation protection issues. Additionally, a large number of benign biopsies were performed with inherent anticipated psychological impact. Further studies could inform national guidance.

Introduction
NHSBSP standards for assessment recall in prevalent screen is a minimum of <10% with a target of <7%. Aims: (i) review variation in recall rate (ii) assess differences in practice and learn from units performing well Methods Prevalent recall rate for the last 3 years in units were collated and a questionnaire titled 'NHSBSP Prevalent Round Survey' designed with units invited to partake on Survey Monkey.

Conclusion
Units with lower prevalent recall rates have a higher proportion of consultant radiologists and higher volume readers. Policy for lesions that do not require recall can reduce unnecessary assessment of benign lesions.

Introduction
Breast density is a strong risk factor for breast cancer and has potential use in breast cancer risk prediction, with subjective methods of density assessment providing a strong relationship with the development of breast cancer. This study aims to assess intra-and inter-observer variability in visual density assessment recorded on Visual Analogue Scales (VAS) among 11 trained readers, and hence the reproducibility over time.
Methods 11 readers of varying years of experience estimated the breast density of 120 mammograms on two occasions 3 years apart using VAS. Percent breast density was estimated on VAS score sheets and were scanned using custom software which converted the marks to percentages. Intra-and inter-observer agreement was assessed with intraclass correlation coefficient (ICC) and variation between readers visualised on Bland-Altman plots.

Results
Excellent intra-observer agreement (ICC>0.81) was found in majority of the readers. All but one reader had a mean difference <10 percentage points from the first to second reading. Inter-observer agreement was excellent for consistency (ICC 0.82) and substantial for absolute agreement (ICC 0.69). However, the 95% limits of agreement for pairwise differences were (-6.8 to 15.7) at the narrowest and (0.8 to 62.3) at the widest.

Conclusion
Overall, the readers were consistent in their scores, although some large inter-observer variations were observed. Reader evaluation and targeted training may alleviate this problem. Intra-observer readings are reliable and may be used in monitoring change in breast density over time, for example when assessing the efficacy of chemopreventive therapies.

Introduction
To assess the performance and accuracy of pre-operative Ultrasound evaluation of axillary lymph nodes in patients with breast cancer.

Methods
All patients who underwent wide local excision or mastectomy for invasive primary breast carcinoma in a UK district general hospital between July-Dec 2015 were identified retrospectively and included in the study. Patients who underwent neo-adjuvant chemotherapy were excluded from the study.

Results
Out of 94 patients with invasive breast carcinoma included in the study, 38 patients had abnormal axillary nodes on ultrasound and subsequent ultrasound guided axillary nodal biopsy (UANB), of which 16 had metastatic carcinoma from primary invasive breast carcinoma which was later confirmed on axillary nodal clearance (ANC). The remaining 22 patients underwent sentinel lymph node biopsy (SLNB) and had no malignant cells on histology.
Of the 56 patients who had normal pre-operative axillary ultrasound, 11 patients were found to have axillary nodal metastatic disease and 45 patients had no malignant cells on subsequent SLNB. In total, 27(n=94) patients were found to have axillary nodal metastases in our study. The positive predictive value of preoperative UANB in detecting axillary nodal metastatic disease is 42% (16/38) and the negative predictive value is 80% (45/56). The sensitivity of axillary ultrasound biopsy is 59% (16/27) and specificity is 67% (45/67). There were 22 false positive cases and 11 false negative cases.

Conclusion
The performance and accuracy of pre-operative ultrasound guided axillary nodal biopsy in detecting metastatic axillary nodal disease in our unit remains acceptable within published standards.

PB.2
Preoperative sentinel lymph node identification, biopsy and localization using contrast enhanced ultrasound (CEUS) in patients with breast cancer, a systematic review and meta-analysis

Background
The role of axillary node clearance (ANC) in patients with 1 or 2 positive sentinel nodes is being questioned in the POSNOC trial. Increasing sensitivity of axillary ultrasound has resulted in identifying preoperatively more lymph node positive patients and thereby fewer differences with sentinel node positive patients. The aim of this study was to determine if the number of abnormal nodes seen on preoperative axillary ultrasound correlates with the axillary tumour burden on histopathology after ANC, and whether this information can be used to identify patients with low volume disease who may be offered sentinel node biopsy(SNB) rather than ANC. Methods 66 patients with FNA or core biopsy proven axillary nodal metastasis were included in this prospective study from 4 centres. The number of abnormal nodes at pre-operative ultrasound examination was recorded (score of ≥3). All patients underwent ANC. Results 31 patients had 1 abnormal node and 35 patients ≥2 abnormal nodes on ultrasound scan. The median number of positive nodes found on pathology was 2(range 1-9) in patients with 1 abnormal node and 5(range 1-33) in patients with ≥2 abnormal nodes (p<.0001). 20 of 31 patients(64.5%) with 1 abnormal node on ultrasound had ≤3 positive nodes at ANC. 18 of 31 patients(58.1%) with 1 abnormal node had only 1 or 2 positive nodes at ANC.

Conclusion
Patients with 1 abnormal node seen on ultrasound can be offered SNB rather than ANC as the initial axillary surgery avoiding overtreatment and unnecessary arm morbidity in a significant number of patients.

PB.6
Preoperative axillary staging in breast cancer: a comparison of the sensitivity of fine needle aspiration biopsy and core needle biopsy Ashley Topps 1 , Simon Barr 1 , Panagiotis Pikoulas 1 , Susan Pritchard 1 , Anthony Maxwell 1,2

Introduction
Identifying axillary lymph node metastases preoperatively can inform discussions about neoadjuvant chemotherapy and also allow a patient to proceed to an axillary lymph node dissection (thus avoiding an additional sentinel node biopsy procedure).
A meta-analysis has not shown a statistically significant difference in sensitivity between axillary US-guided fine needle aspiration biopsy (FNA) and core needle biopsy (CNB) 1  In the era of Z0011, current practice of axillary node clearance for preoperative US detected lymph node metastasis will result in overtreatment of the axilla as some of these patients only have 1-2 positive nodes.

Methods
Patients with lymph node positive, newly diagnosed invasive breast cancer, during 2012-2014 at UHCW were selected retrospectively. Abnormal nodes defined by local criteria underwent ultrasound-guided FNA or core biopsy. Patients were grouped into 1-2 or 3 or more nodes (3+) based on their lymph node metastasis burden after ALNC, and tumour characteristics compared. Categorical variables were evaluated with chi-squared and Fisher's exact tests.

Conclusion
Axilla management of patients with pre-operatively diagnosed lymph node metastasis currently leads to overtreatment in 31.3%. There is scope to identify patients with a low nodal burden to select them for axillary staging and possibly avoid ALNC.

PB.9
The use of nomograms to predict additional lymph node metastasis after Sentinel lymph node biopsy: Can they reliably identify those needing no further axillary treatment?

Introduction
Invasive lobular cancer (ILC) attributes to 5-10% of new breast cancer diagnosis. It is often mammographically occult with detection rates of 57-87%. Fibroglandular tissue density is inversely correlated to mammographic detection rates and there is also an increased rate of multifocality in ILC. MRI has higher detection rates of 93% and is recommended in patients with ILC. Theoretically mammographic ILC detection should be better in fatty breasts and therefore we ask whether MRI essential in this subgroup of patients? Aim To ascertain if MRI detects any significant additional lesions in patients with ILC and fatty breasts.

Methods
Retrospective study of breast MRIs conducted within a 4 year period for histological confirmed ILC. Breast densities of 1 were included. All imaging reports recorded.
Results 134 patients identified with ILC, 103 excluded as had breast density 2 or more. 31 identified as having breast density of 1. Primary lesion identified on mammography in 28 patients and on ultrasound in 3 patients. MRI identified additional findings in 9 patients which were occult on mammography. Of these 9 cases, 3 were identified as additional cancers and the remaining 6 were benign. Conclusion MRI detected 3 additional cancers which were otherwise occult on mammogram despite the patient having fatty breasts. These results are congruent with published literature. A study published in 2012 identified 7 additional cancers in 32 cases. We therefore conclude that MRI is necessary in imaging patients with ILC and low density breast tissue on mammography.

PB.11
Is the final post Neoadjuvant chemotherapy MRI scans prior to surgery necessary in patient management in breast cancer? Archita Gulati 1,2 , Furhan Razzaq 2

Introduction
Neoadjuvant chemotherapy is a well-established method of treating large breast carcinomas with the aim of shrinking the tumour to enable breast conservation surgery or to improve outcomes of treatment for patients requiring mastectomy by delivering early systemic therapy to reduce the risk of metastatic disease. In many institutions, MRI scans are performed prior to the commencement of neoadjuvant chemotherapy, at the mid-point of treatment and at the end of chemotherapy prior to surgery. The aim of this audit was to assess the benefit of performing the final MRI scan in the patient's surgical management.

Methods
A list of relevant patients undergoing neoadjuvant chemotherapy was obtained from a database maintained by the breast care nurses. Tumour size on the final MRI scan was compared with tumour size determined histologically from the operative specimen, which was regarded as the gold standard for the purposes of this audit. Concordance was regarded as satisfactory if the tumour size was within 10mm on radiology and histology.

Results
There was 77% concordance between the final MRI scan and the postoperative histology. However the findings at MRI did not alter the surgical management decision and in patients where there was discordance in tumour size between imaging and histology, none required further surgery.

Conclusions
Although felt to be useful by some surgeons and oncologists, MRI scans post neoadjuvant chemotherapy prior to surgery can be safely omitted without adversely affecting the patient's management.

PB.13
Additional significant findings in pre-operative breast MRI in patients with histologically proven breast carcinoma-our experience in one This study aimed to identify pre-NAC MRI and tumour features which could help predict response to NAC. We also investigated the ability of post-NAC MRI to accurately represent the extent of any residual disease which is of importance for surgical planning.

Methods
This retrospective cohort study evaluated forty-six patients with LABC who received NAC between 2008 and 2016 at UHCW. All patients underwent pre and post-NAC DCE-MRI and proceeded to surgical excision. Clinical and imaging data was collected from CRRS and UHCW PACS.

Results
Several tumour features were found to be predictive of pathological response, including: tumour morphology, oedema, shrinkage pattern, HR status, HER2 status and percentage change in MRI long diameter. Circumscribed lesions were more likely respond to NAC than irregular, diffuse and nodular tumours. There was some evidence of correlation between MRI morphological categories and tumour receptor status. The accuracy of post-NAC MRI for predicting residual disease was superior for HR negative tumours.

Conclusions
Pre-treatment MRI features can serve as reliable imaging biomarkers which can be used to predict disease response to NAC. The accuracy of post-NAC MRI varies with tumour biology and imaging features, and this should be considered in decisions about surgical approach.

PB.15
Comparison of digital mammography, ultrasound and MRI in the preoperative size assessment of lobular breast cancer: A district general hospital experience Sachin Kamat 1 , Carla Goncalves 2 , Alan Tan 2 , Asha Eleti 2 , Nithya Vidyaprakash 2

Introduction
Tumour size assessment with imaging has an impact on the diagnosis and treatment planning of breast malignancy. We assessed the correlation between preoperative size of lobular breast cancer on different imaging modalities and postoperative histological size, as the reference standard.

Methods
Retrospective review of lesion size on digital mammography (DM), ultrasound (US) and contrast-enhanced MRI from 73 breast lesions was performed in 63 consecutive women with histological diagnosis of lobular breast cancer. The sizes of different modalities were correlated with histological tumour size using paired T-test analysis and coefficient of determination.

Methods
Three consultant radiologists (C1, C2 and C3 with 7 to 9 years of experience in reading breast MRI) retrospectively and independently reviewed 27 high risk breast MRI studies. All cases were annonymised and review was only performed as per C. Khul's paper, that is MIP images and abbreviated protocol images consisting of pre-contrast, first post-contrast and its subtracted counterpart were only reviewed and reported. Time taken for MIP and abbreviated protocol reading and the sensitivity and specificity for both were calculated for each reader.

Results
Average time to read MIP and abbreviated protocol images were: C1 =

Introduction
There is a lack of collated data on the appearance of cancers in digital mammograms. This information may help in understanding which lesions appearing in mammograms are potentially fatal and which low risk.

Methods
Mammograms have been collected since March 2011 from three breast screening sites alongside the clinical information from National Breast Screening Service (NBSS). Experienced readers have marked a rectangular region of interest (ROI) closely around identified cancers and classified each cancer's conspicuity and appearance. The cancers size was calculated as maximum dimension (height or width) of the ROI. The data were analysed to investigate the appearance of the cancers.

Conclusion
We have characterised the appearance of a large set of cancers.
Once the database is fully linked to data in NBSS, then cancer grade and biology can be included in the analysis.

Results
Eleven relevant studies were identified. These performed the same CESM technology (G.E. SenoBright®) and varied mammography technologies on the same patients with either newly diagnosed breast cancer or suspicious imaging. All studies showed CESM increased sensitivity by 3.1-21.2% and 10 studies showed increased specificity by 5-45.7% (one study found specificity decreased by 4%) compared to mammography. However the differences were not always significant, the data was heterogeneous, and none of the included studies fully demonstrated low risk of bias when quality assessed. Conclusion CESM may improve sensitivity and specificity of breast cancer detection compared to mammography. However further research is required to establish applications and determine which patients would benefit.

Introduction
To analyse the pattern of parenchymal tissue on mammography in women with the densest breasts, variation with age and the effect on recall rates and cancer detection.

Methods
Breast density data (VolparaTM) was obtained in women attending mammographic screening between April 2013 and March 2015. Cases with the densest breasts were selected for visual interpretation of parenchymal pattern. 100 cases were included for non-assessed women age 50, 55, 60, 65 and 69-71. All cases of assessed women with the densest breasts were reviewed. Mammograms were reviewed by 10 film readers. Parenchymal pattern was classified as: smooth; mainly smooth, mixed; mainly nodular or nodular. Average classification was compared by age and assessed v nonassessed. Likelihood of biopsy and cancer diagnosis was analysed by parenchymal pattern.
Results 40760 women were included in the subset, 4331 (10.6%) of these were Volpara4. Proportions in each parenchymal pattern category were similar at all ages and for assessed v non-assessed.
In the assessed group 90 were smooth/mainly smooth; 104 mixed; 106 mainly nodular/nodular. Of women who were subjected to biopsy, 50 were smooth/mainly smooth; 57 mixed; 56 mainly nodular/ nodular. Of women diagnosed with cancer, 7 were smooth/mainly smooth; 10 mixed; 18 mainly nodular/nodular. More cancers were identified in women with nodular breasts.

Conclusion
The ratio of smooth to nodular pattern in women with the densest breasts did not vary with age. The parenchymal pattern of breast tissue did not affect recall rate, but women with nodular breasts were more likely to be diagnosed with cancer.

PB.25
Presentation and follow up of mammographically occult breast cancer: a multicentre audit of 5 years presentation with minimum 5 years follow up Kathryn Taylor  As BMI increased the average FGV decreased with obese women losing on average 2.256 cm 3 . Women who were pre/perimenopausal at first screening had larger drops in FGV than postmenopausal women. Within postmenopausal women 49.4% saw rises in FGV and of those who did the average gain was 8.562 cm 3 . The use of HRT showed an expected positive correlation with FGV.

Conclusion
There was an overall decrease in MD over the 3-year period. However, 48.7% of women demonstrated an increase in MD; in particular women with normal BMI, postmenopausal women and women on HRT. These data show that to fully understand changes in individual MD for both breast cancer risk prediction and risk reducing interventions, confounding factors must be taken into account.

Introduction
Breast screening in the UK is undergoing many workforce changes as more experienced radiologists retire and new readers become engaged. With such major changes it is important to monitor mammographic interpretation skills for national quality assurance purposes. A detailed analysis has been carried out to compare the performance of 'newcomers' to the national performance measures on the PER-FORMS scheme of more experienced readers.

Introduction
There is ongoing concern that silicone implants, particularly those placed in a subglandular position, may reduce the sensitivity of mammography in the detection of malignant lesions. Our aim was to determine whether or not breast implants affected the cancer detection rate in our screening population.

Results
During the period, 310,558 patients attended for screening. Of these, 852 had breast implants. A total of 2,840 malignancies were detected, 6 of which were in patients with implants -1 prevalent and 5 incident. All 6 had subglandular implants. There were 3 diagnoses of DCIS and 3 of invasive carcinoma, with a size range of 10-34mm. None of the patients had nodal disease at the time of diagnosis. The malignancy detection rate in the implant group was 7 per 1000 and 9.2 per 1000 in the non-implant group (P=0.52).

Conclusion
Although there appears to be a trend towards a lower cancer detection rate in women with implants, the results do not reach statistical significance. However our study is limited by small numbers of cancers in women with implants and this also precludes further analysis of data for invasive and small cancer detection rates.

PB.30
Correlation Imaging of breast implants is increasingly performed to investigate suspected leak or rupture. However, there is little opportunity for the radiological findings to be correlated with the pathological findings, providing feedback as to the accuracy of the radiological diagnosis. We reviewed pathological findings in explanted breast implants, in conjunction with the radiological findings, to investigate the accuracy of imaging for breast implant changes. The pathology database was searched for all breast implants removed between July 2000 and January 2015. All patients identified had their imaging history reviewed to identify those with pre-operative imaging. Pathological findings at implant removal and radiological findings on the pre-operative imaging were compared. There were 106 implant related pathology cases identified. From this, 11 patients had explant of breast implants and pre-operative imaging. In 8 cases implants had been placed for cosmetic augmentation, 3 were following implant-based reconstruction. 10 cases had pre operative ultrasound, 1 had MRI and 1 had both. Radiological findings demonstrated intra or extra capsular rupture in 9, silicone granulomas in 1 and an intact implant in 1 case. Imaging findings were consistent with the pathological findings in all cases. Imaging of implants provides an accurate pre-operative assessment of implant pathology. As many more implants are imaged than removed, and many implants removed without any pre operative imaging, it is difficult to get feedback about the accuracy of radiological diagnoses in this setting. Although small, this study shows 100% concordance between radiological and pathological findings following the pre-operative imaging of implant pathology. Following the introduction of this new service we felt it was important to review our practise. This was done in order to understand our recall rates and review the findings at assessment including results of ultrasound and MRI guided biopsies, cancer detection, early recall and interval cancer rates. We include a pictorial review of examples of the recalled lesions.

PB.33
A 15 year review of familial breast cancer screening in Wales -are we offering the right test to the right women? Tom Evans 2 , Guy Stevens 1 , Sian Nisbet 3 , Alex Murray 3 , Kate Gower Thomas 1

Introduction
Since 2001, the Welsh breast cancer screening programme has offered annual mammography to women below age 50, identified as having an increased risk of the disease. Women are referred via the All Wales Genetics Service who analyse client risk and apportion a risk category. Women are screened from either aged 35 or 40 years accordingly. We present results including cancer detection rate, numbers of screening eposodes and try to determine whether this is cost effective.

Method
A prospectively maintained NHSBSP equivalent database for women with a significant family history of breast cancer was analysed retrospectively for uptake of screening and cancer detection, including interval cancer episodes. The costs of this service were estimated. Results 5586 moderate and high risk women have been invited for annual screening. The average prevalent uptake was 82.4% and a total of 22270 mammograms were performed. 118 cancers were detected, 84 were screen detected and 34 were interval cancers. The cancer detection rate in this moderate and high risk population combined was 3.8/1000 screenings, which is significantly lower than 10/1000 seen in population screening (p<0.05). Cancer detection in these moderate and high risk women was only a third of the general population suggesting this screening technique is inappropriate and should be abandoned for reasons of cost, stress and unnecessary radiation exposure. We suggest that other screening methods, with further risk stratification might be more effective. Lobular neoplasia in situ is classified as a B3 lesion of uncertain malignant potential and confers an increased risk of concurrent and future malignancy. In the absence of national guidance there is uncertainty over optimal management and peak malignancy risk may not fall within the 5 years of annual surveillance recommended by some centres. Biopsy reports from 1999-2015 at Southampton General Hospital were screened for code B3 and/or diagnosis of LNIS with collection of a comparison group of patients with B2 diagnoses. Recommended management was noted and screening packets were reviewed to identify subsequent assessments and biopsies. LNIS was diagnosed in 130/12,141 breast biopsies (1.07%) performed between 1999-2015 Concurrent malignancy was present in 44/130 (34.68%) of which 33/44 were infiltrative lobular cancer and a further 7/130 (0.05%) had a B4 or B5a diagnosis. Management of pure B3 lesions was highly varied in those with available follow-up data (58/ 130) but most commonly comprised routine recall following surgical excision biopsy (34.5%). Women with a B3 diagnosis of LNIS were at increased risk of developing invasive malignancy compared to the B2 group (8.62% and 3.13% respectively, p=0.21) and the average time to malignancy was 15 years. The high rate of concurrent malignancy, in particular invasive lobular cancer, may support the theory that LCIS is a precursor lesion. LCIS also confers an increased risk of subsequent invasive neoplasia which in this study occurred on average 15 years from LCIS diagnosis. This timeframe should be considered when recommending management with enhanced mammographic surveillance.

PB.35
The changing risk factor profile of Asian women with breast cancer

Results
There was a decreasing trend of proportion of patients with diabetes mellitus (p<0.001). There were increasing trends for mean age of first full term pregnancy, mean duration of interval between menarche and first full term pregnancy, proportion of women who breastfeed, and BMI. There was a decreasing trend of proportion of multiparous women. However these findings were statistically insignificant.

Conclusions
There are changing trends in certain risk factors for breast cancer which might contribute to the increasing incidence of breast cancer in Asian women. More research has to be done on modifiable breast cancer risk factors and how they can be altered to decrease breast cancer risk. A survey of radiographer film reader's perceptions of workload, performance and job satisfaction in the NHSBSP.

Introduction
Age extension has increased the film reading workload in breast screening. A reported shortage of radiologists plus radiographers double reporting also has the potential to increase reading volumes. There is disagreement on whether performance declines with increasing volumes. There are no recommendations on maximum reporting volumes. This survey aims to identify themes which affect film readers' perceptions of workload, performance and job satisfaction.

Methods
Purposeful sampling was used to select participants. All qualified radiographer film readers were included. Electronic questionnaires were distributed to managers to forward to participants. Thematic analysis was used to analyse results.

Results
The overall response rate was 37%. 77% perceived an increased workload. 60% report a sufficient workforce for reporting. 40% report high volumes. Performance is thought to fluctuate following interruptions, PERFORMS and fatigue. 84% are satisfied in their role.

Conclusion
Radiographers are experiencing an increase in workload but reporting time is frequently interrupted. Audit should assess any effect on performance by high volumes and visual fatigue. More time for CPD is required and increased involvement in research and audit. Job satisfaction is high amongst readers which should aid retention and recruitment. With a potential future shortfall in radiologists, the NHSBSP will be reliant on advanced and consultant practitioners to maintain targets. There is the potential to further develop the role but care should be taken not to compromise performance and job satisfaction with the volume of work.

Introduction
The NHSBSP KPIs require 90% of first offered appointment (FOA) and first actual assessment (AA) to be within 21 days of screening mammography. This leaves little flexibility for women unable to attend FOA. Our unit consistently achieves FOA (monthly rates 93%-100% in last round), but frequently breaches the AA (monthly rates 78%-100%). We investigated why the breaches occur. Patients are advised to report any breast associated symptoms during their screening mammograms in the Warwickshire, Solihull and Coventry Breast Screening Service (based at UHCW NHS trust). Although no specific guidance on whom to recall exists, symptoms such as lump, distortion/change in shape of breast, recent nipple discharge, eczema, or recent inversion, skin tethering and dimpling are routinely recalled for further assessment. In this retrospective study, we investigated the types of symptoms reported and subsequent clinical, imaging and histology findings with specific aim of finding the number of cancers in this cohort of patients. Between 2011 -2015 data was collected from NBBS and hospital RIS, PACS and CRRS on patients reporting symptoms during their routine screening mammogram. This included 709 patients (age range: 47 to77 years, mean age of 56 years). Subsequently, 15 patients were diagnosed with breast cancer, of these 3 patients were excluded as 2 of them were originally referred by the radiographers who noticed the changes at the time of screening and the third patient would have been recalled anyway for abnormal mammography. Among the remaining 12 patients, 6 had cancer on a site or side completely different from the actual location of symptom patient identified. Therefore, 6 out of 709 patients (0.84%) who were recalled were concordant with clinical symptoms. This study shows the positive predictive value (PPV) of patients who reported symptoms turning into a breast cancer is low. Therefore, an alternate pathway of managing these patients should be investigated.

PB.40
The association between regional disposable household income and uptake of breast screening in England between 1999 and Aim To examine possible associations between regional disposable household income and uptake of breast screening and investigate whether any association is related to the type of invitation, categorized as: first invitation; invitation to previous non-attender; invitation to previous attender (last attendance <5 years); invitation to previous attender (last attendance >5 years) and early recall invitation.

Methods
Data on breast screening uptake and Gross Disposable Household Income (GDHI) per head for the English regions for the years 1999-2014 were obtained from the Health and Social Care Information Centre and the Office for National Statistics respectively. Uptake data were adjusted to fit calendar year and GDHI data were adjusted for inflation. The association between these was assessed using multiple linear regression.

Results
There was a significant (p<0.05) positive association between regional GDHI per head and overall breast screening uptake with a coefficient of 0.00066 (95% confidence intervals 0.00027-0.00106). This is equivalent to a 0.66% rise in breast screening uptake for every £1000 increase in regional GDHI per head. No type of invitation alone showed a statistically significant association between GDHI per head and breast screening uptake.

Conclusion
Regional disposable household income has a positive association with uptake of breast screening in the English regions. In this study, there was no significant association between GDHI per head and breast screening uptake for any one type of invitation. Low income populations should be targeted to reduce inequalities and further research should determine which interventions can be cost effectively applied to these populations. The South Asian population has increased across the UK along with increase in breast cancer incidence in them. Yet the uptake rate for breast screening in this group of women remains significantly lower. Barriers to screening, such as language barriers can deter some women from attending.

Materials and Methods
A pilot questionnaire study was carried out to assess the screening experience of 110 women who watched a breast screening video on a tablet device in their choice of language prior to screening. The screening experience was assessed by a five item questionnaire. For advanced understanding of the device and its related issues, supplementary data was collected in the form of a radiographers' feedback questionnaire (n=92).

Results
The majority of the women participating were supportive towards this method of help (85%) and had favourable opinion about the device (90%). In total 93% of the women agreed that the video helped them through the mammography procedure and 79.2% agreed that it made their screening experience better. The radiographers' feedback suggested that time was a sensitive issue in applying this method, with 39% suggesting it increased the clinic times.

Conclusion
The usage of multilingual tablet device was perceived positively by the majority of women attending for breast screening. It was considered helpful in improving the individual's screening experience and possibly will have a positive impact on their compliance in subsequent screening rounds. The tablet device should be considered for national piloting as part of the National Breast Screening Programme. No enhancing abnormality was demonstrated on procedural MRI in six cases, other technical reasons preventing successful biopsy included posterior positioning of the lesion (n=1) and software failure (n=1, successfully repeated).

Conclusions
Our cancer detection rate (32%) is comparable to other UK centres 1 and higher than studies in Europe (21%) and USA (8%) 2,3 , suggesting appropriate case choice. Technical difficulties and complications were part of the learning curve, and practices have evolved to minimize these.

Introduction
Intact is a breast biopsy device that excises a breast abnormality using vacuum and radiofrequency (RF) technology. It is licenced in UK as a mammographic-mounted device, or a hand held device for USS guidance. Research from USA showed complete removal of breast abnormalities occurs frequently during Intact breast biopsy but experience in UK is limited. We present our centre's initial experience using Intact under USS guidance.

Methods
Selection of masses for Intact depended upon the size of mass and the willingness of the patient to undergo the procedure, initially under general anaesthesia (GA) immediately prior to her therapeutic excision, and subsequently under local anaesthetic (LA). For the local anaesthetic cases an additional criteria for U5 masses was that surgical excision under general anaesthetic was contraindicated for the patient due to co-morbidity.

Results
19 selected breast masses in 15 women were biopsied under USS guidance using "Intact".
No complications from the Intact procedures.

Conclusion
Our initial experience suggests it is safe to perform Intact under LA and USS guidance in an outpatient setting and that Intact offers the potential to avoid subsequent excisional surgery for small B3 and carefully selected small B5 masses in the future.

PB.45
Differences in acute and persistent pain following ultrasound and stereotactic guided vacuum-assisted breast biopsy (VABB) -results of a pilot survey Matthew Brown 1,2 , Steve Allen 1 , Liz O'Flynn 1,2 , Nandita deSouza Purpose/background/objectives Vacuum assisted breast biopsy (VABB) is a minimally-invasive modality enabling target lesions identified within breast tissue to be either sampled or removed. A biopsy needle is advanced percutaneously to the target under stereotactic, ultrasound or MRI guidance, whence multiple samples are harvested. This pilot survey explored whether differences in acute and persistent pain intensity occurred between ultrasound and stereotactic-guided VABBs.
Methods A questionnaire-based survey was undertaken; basic demographic and procedural data for patients was recorded at the time of VABB. Participants completed a pain/analgesia diary detailing the intensity of pain experienced and analgesia taken over the 7 day post-VABB period. Participants were contacted at 3 months post-VABB to determine the presence of persistent pain. Results 49 participants were recruited and 38 completed questionnaires were returned (27 US, 11 stereo'). Statistically significant differences were observed in the intensity of pain experienced post-biopsy by patients who underwent US and stereo VABB on day 1; 3.4 (SD 2.8) vs 1.4 (SD 1.6) P=0.04, day 4; 1.3 (SD1.5) vs 0.2 (SD 0.6) P=0.03, day 6; 0.7 (SD 0.9) vs 0 P=0.03 and day 7; 0.7 (SD 1.0) vs 0 P=0.03. No procedural differences existed between the groups. 3 patients (8%) reported persistent pain at the 3-month time point, all had undergone USguided VABB.

Conclusions
Patients undergoing US guided VABB experienced more intense pain in the week following biopsy than those undergoing sterotactic guided VABB and appeared to experience more persistent pain. Further work is required to determine the cause of these findings.

Aim
Largest series comparing the unit current standard-of-care, Pronestereo(PS) Vacuum Breast Biopsy(VAB) with Digital Breast Tomosynthesis (DBT)-guided VAB. Background DBT is an established mammographic technique shown to improve the accuracy of soft tissue lesion characterisation, conspicuity and cancer detection. VAB increases the sensitivity of pre-operative diagnosis, reducing upgrade rates. Method 170 consecutive patients through the NHSBSP and symptomatic breast service undergoing stereo-guided biopsy were recruited into 2 arms: Multicare-Platinum-Prone-Stereotactic-Breast-Biopsy-System Affirm-Breast-Biopsy-Guidance-System with 3D Breast Biopsy using Hologic Selenia Dimensions VAB was performed using the Atec-Sapphire with 9G Eviva needles. Procedure outcomes including patient comfort, time, accuracy, radiation dose and complications were recorded.

Conclusion
We demonstrate DBT-guided biopsy is a good alternative technique to stereo-guided biopsy for soft tissue abnormalities (including those occult on 2D and ultrasound) and microcalcifications. There is an increased risk of vasovagal episodes with upright DBT, which could be avoided with decubitus positioning or Prone-DBT.

Introduction
The management of B3 lesions is becoming increasingly under debate in light of criticism of over-diagnosis within breast screening. New proposed guidelines throughout the UK are suggesting that surgical biopsy for many B3 lesions may no longer be required.
In this audit we review all cases of B3 at initial biopsy over two fiveyear cohorts. 208 14G core biopsies and 256 initial vacuum biopsies were performed. 50% of patients in the first cohort underwent benign surgical biopsy compared to 40.4% in the latter cohort.

Methods
There was a 6% upgrade to invasive malignancy and 18% upgrade to non-invasive malignancy over the 10-year period following surgical biopsy and vacuum excisions. The upgrade rates for each histological category were: Atypical duct hyperplasia 36% Flat epithelial atypia 28% Radial scar/complex sclerosing lesion with atypia 27% and with no atypia 11% Papilloma with atypia 55% and with no atypia 16% Lobular neoplasia in situ 43% Suspected phyllodes tumours 8% Atypical apocrine adenosis 20%.

Conclusions
The results of this audit and upgrade rates are in line with the literature. Upgrade rates remain high even with first line use of vacuum biopsy. Careful consideration is essential prior to changing current practice.

Conclusion
Metastatic disease is common in women presenting with IBC and accurate staging is essential in guiding management and providing prognostic information to patients. 70% of our cohort had an initial staging CT, when all patients should be undergoing this. Of those who received an initial negative staging scan, 32% of these had metastatic disease within 12 months. In view of this, a PET/CT is the recommended staging modality of choice to allow for more accurate staging in this cohort,enabling better management in this select group of patients.

Introduction
Breast cancer mortality is intrinsically linked to stage at diagnosis, so the ability to detect disease in infancy is critical. Sub optimal sensitivity and specificity of current imaging methods such as mammography and ultrasonography result in issues such as false positive and false negative results. Photoacoustic Imaging (PAI) represents a novel approach to breast tissue visualisation. Non-ionizing laser pulses are delivered to tissue and photoacoustic signals generated. Coupled with ultrasound emission, structural imaging and functional analysis of tissue can be determined. This pilot study assesses the potential use of the VisualSonics Vevo LAZR PAI system in the investigation of normal and abnormal breast tissue.

Methods
Ethical approval and informed consent was obtained. Using a 15MHz probe, imaging of healthy and abnormal breast tissue was performed in a tertiary symptomatic breast cancer unit (n=8). Oxygen saturation, haemoglobin concentration and photoacoustic signal was determined for all.

Results
Characterisation of healthy breast tissue, tissue with benign pathology (fibroadenoma) and tissue with previously identified breast cancer was achieved. With locations of interest determined by ultrasound, photoacoustic signals were assessed. Imaging to a depth of 30mm was confirmed. Comparisons in oxygen saturation were made between breast pathology and the controls (disease free contralateral breast).

Conclusion
Current imaging modalities in breast cancer have shortcomings. PAI represents a novel means of both visualisation and assessment of functionality of breast tissue. These preliminary findings offer an insight into the potential of PAI in the role of breast cancer diagnosis. Rates of malignancy varied from 6% in a radial scar with no atypia, to 32% for a papilloma with atypia. Differences in malignancy upgrade rates between atypical and non-atypical lesions were statistically significant (p<0.05). Study heterogeneity could not be explained by differences in core biopsy size or year of publication, however, a significant difference in upgrade rates to malignancy was observed between the US and non US literature. Many studies have assessed the risk of malignancy following diagnosis of B3 lesions, but are often small and lack statistical power. This study is a comprehensive, inclusive assessment of the available literature, on which to base tailored management strategies.

PB.52
The use of supplemental imaging in post-surgical follow-up of women with breast cancer Anthony Maxwell 1,2 , Nisha Sharma 3 , Hilary Dobson 4 , Sarah Vinnicombe 5 , Andrew Evans 5

Introduction
With increasing breast cancer incidence and improved survival, optimised follow-up strategies are required to maximise quality and duration of life. Whilst periodic mammography is adequate for most older women, its sensitivity in younger women and those with dense breasts is poor. Recurrences or new primaries in women whose original cancers were mammographically occult may also be occult. This study investigates the use of supplemental follow-up imaging in these groups. Introduction/Methods Increasing use of body CT has lead to increasing detection of incidental breast lesions, creating two primary concerns within our unit:

Methods
1) that these patients experience longer delays before breast clinic review than standard symptomatic patients. 2) That some cases could avoid clinic altogether following specialist review of the CT/prior imaging.
49 cases over a 20 month period were reviewed to evaluate these concerns.

Results
Time between the triggering CT and clinic attendance varied from 1 to 163 days, inpatient range 1 -19 days, median 9; outpatients range 6 -163 days, median 24. Of the 49 lesions, 36 were focal masses, 5 asymmetries, 5 enlarged nodes, 2 thickened WLE scars and one calcification. 31 biopsies were performed. 21 lesions (68%) were suspected and proven malignant (17 masses, 3 asymmetries, 1 scar recurrence). All lesions with ancillary suspicious findings were malignant. Hounsfield unit measurements were higher for malignant lesions. 5 patients had unchanged screening films or prior images. Breast radiologist review concluded 8 cases could have avoided clinic review as the area of concern was normal or classically benign. Conclusions 1) Outpatients with incidental breast lesions experience greater delays before clinic review than the target 14 days. 2) Specialist current/prior imaging review could reduce unnecessary clinic assessment.
A high percentage (68%) of incidental lesions were malignant. We propose to develop a radiology:breast radiology referral pathway to address both these issues and improve standards of care for these patients at increased risk.

Introduction
Male breast cancer is rare affecting < 1% of the male population. We have observed a marked growth in the number of men referred for imaging and propose that imaging of all men may not be necessary.

Method
Retrospective review of all men referred for imaging between 2010 -2015. Clinical history, imaging and pathology were reviewed.

Results
Total of 452 patients. 3 patients were referred for imaging in 2010 compared with 162 in 2015. In total, 11 patients were diagnosed with breast cancer, 2 with metastases from another primary and 3 with lymphoma. Of these 16 patients, 2 were not given a numerical P value but the clinical history stated suspicious mass. 12 (75%) had a P value of P3-P5. 2 patients were clinically P2 but both had significant history: 1 of blood stained nipple discharge and the other of previous renal cell carcinoma.
The average age of the 16 patients was 75 years, the youngest 53 years.
Conclusion 454 men were referred for imaging over 5 years. Malignancy was detected in 3.5%. Of those patients with malignancy, 88% were clinically suspicious and all were >50 years. We propose an imaging pathway for men based upon clinical findings/age that would be safe and reduce unnecessary imaging and biopsy.

PB.56
Are Asian women more prone to less favourable subtypes of breast cancer? Joseph Wan 1 , Anil Jain 2,3 , Jaanilka Background Longterm follow up mammograms of patients post surgery form a considerable bulk of the reporting in a symptomatic breast unit. Recalls from this group add considerably to patient anxiety, and contribute to workload, taking up valuable appointment time for spot views, ultrasound and biopsy. Aim A decision was made to second read all recalls from the LTFU group; arbitration by a third consultant was used as a deciding vote in case of disagreement. All second reads and arbitrations were recorded in a diary -this has been ongoing for a year. All cases were analysed for presence or absence of malignancy when recalled, retrospectively.

Results
Of 93 LTFU patients recalled from June 2015 to May 2016, 77 patients were recalled after arbitration. 58 of these patients (75.3%) were returned to routine followup following arbitration; 19 patients (24.7%) were recalled following arbitration. 3 malignant cases (16%) were noted in the recall group post arbitration; 16 cases (84%) were benign. The most commonly arbitrated lesion was opacity, in 45%; 37% were recalled for an opacity.

Conclusion
Second reading with arbitration, as in screening, is a valuable tool to reduce number of recalls in the LTFU group, allowing 75% of arbitrated cases to be returned to normal followup. This reduces patient anxiety, and frees up valuable time for assessment of true positive malignant recalls. In the absence of established guidelines, the screening recall rates of 4% for incident round and 7% for prevalent round can be used to compare symptomatic recall rate in an instutution.

PB.61
Impact of Digital Breast Tomosynthesis (DBT) as a standard mammographic investigation compared with Full Field Digital Mammography (FFDM) in a District General Hospital Symptomatic Breast Service Harriet Etheridge 3,2 , Aisling Butler 1,2

Introduction
Digital Breast Tomosynthesis is established as a problem-solving tool within NHS screening. Recent publications have examined the impact of DBT as an assessment tool within the symptomatic setting. To date, literature has not reported on the impact of DBT as a baseline study in this scenario although internationally many centres have adopted this practice. Our centre installed a Hologic Dimensions Tomographic unit in May 2015. Upon installation, DBT replaced Full Field Digital Mammography (FFDM) as the baseline tool for symptomatic and surveillance patients.

Aim
To establish if the number of M3 reports have increased or decreased and to quantify the positive predictive value of M3 reports. Methods DBT mammographic reports over 12months were reviewed and compared with FFDM reports in the previous 12month period, evaluating 'M3' reports and their subsequent up-or downgrading. Cases were evaluated to identify the following -additional mammographic views, ultrasound grading, histopathological classification and final diagnosis.

Results
There were less M3 reports overall. Of the M3 reports, more were true positive than in the previous year. Within this specific cohort we identified a 10.10% increased likelihood of ultrasound upgrade following DBT compared with FFDM, a 25.7% increased likelihood of histological upgrade and a 7.1% increase in ability to predict malignancy using DBT compared to FFDM.

Discussion
We acknowledge the limitations in terms of design and cohort but the results of this small centre study are in line with previous screening studies and substantiate the use of DBT as a baseline tool in the symptomatic setting.

PB.62
Imaging and male breast cancer: Should mammography be employed first-line? How common is breast cancer in men imaged for gynaecomastia? David S J Fenne 1 , Gary Rubin 1,2

Introduction
Traditionally mammography has been used as the first-line imaging test in patients with breast symptoms aged over 40 years of either sex. Some centres advocate 'ultrasound first' in men instead. This study assessed whether mammography correctly identified cancers in a 'mammography first' unit, the imaging modalities used and the relationship of gynaecomastia to imaging and breast cancer.

Methods
We looked at male breast cancers from 2010-2014, in addition to the number of men referred to our breast unit who were imaged with mammography, ultrasound, or both in 2014, noting whether they had gynaecomastia.

Results
From 2010-2014, 19 men had breast cancer, 2 bilaterally. All were detected on imaging, including 2 that were unsuspected clinically. Only 1 occurred in a man with clinically simple gynaecomastia. In 2014, 207 men aged 40+ came to our unit and 143 were imaged. 71 had mammography without ultrasound, 63 had both and 9 just ultrasound. Gynaecomastia was the most common indication and finding, being seen in 95 men, representing 68% of those imaged. Assuming 2014 was representative of 2010-2014, 475 men with gynaecomastia needed imaging to find one unsuspected breast cancer, a lower detection rate than in the NHS breast screening programme.

Conclusions
Neither mammography nor ultrasound missed or misclassified any cancers. This study supports using mammography or ultrasound firstline to image men aged 40+, depending on local resources. Gynaecomastia was the major imaging indicator/finding. There should be stricter adherence to guidelines that limit breast imaging for gynaecomastia, as it rarely masks cancer.

Introduction
The primary imaging workload in the breast unit is within the context of the symptomatic fast track clinic or screening assessment clinic. These usually have defined numbers of patients. In addition there is a significant increasing workload of extra patients who require imaging and invasive procedures via other pathways, which we aimed to quantify.

Methods
The data was identified retrospectively from the Computerised Radiology Information System (CRIS) and recorded on a spreadsheet. Information was collected on the number of extra patients who were imaged and reviewed by the Breast Radiologists in addition to their booked workload during the months of May and June 2016.

Results
Over a two month period there were an additional 60 patients. These patients were referred from a wide variety of departments within the hospital, although the majority were referred through the breast surgeons.
A total of 48 ultrasound examinations and 21 mammograms were performed and reported on these patients. The additional procedures performed were also recorded, comprising 6 core biopsies, 2 fine needle aspirations, 4 seroma aspirations and 3 abscess drainages.

Conclusion
There is a significant "extra" workload performed during the working week by Breast Radiologists which may not be reflected in funding streams or be accounted for in job plans. Extra imaging represents a significant workload; the equivalent of 3-4 symptomatic clinics a month. With escalating trends for imaging and a greater reliance on image-guided seroma and abscess drainage, it is likely that this will increase. Background Breast cancer in much less common in men when compared to women and many cancers occurs in men over 50 years of age. Routine usage of imaging in men presenting via fast track clinic may be not be appropriate and a better stratification of patients for imaging is warranted.

Methods
Retrospective data between 2010 to 2014 from hospital RIS identified a total of 500 male patients of varying ages who were all referred for imaging. Various patient demographic details were recorded and outcomes of triple assessment correlated.

Results
Average age of the patients was 45 years. There was no consistency as to what imaging was used. 130 patients received US only, 12 received mammograms only and 358 received both US and mammograms. 54 US core biopsies and one stereocore biopsy was performed. 488 patients were scored P2 on clinical examination of which none had a cancer and imaging and pathology findings (where necessary) were benign. 8 were scored P3 and there were two cancers in this group. All three P4 cases were benign on imaging and pathology. There was a single P5 case which was malignant on imaging and on pathology. Overall three malignancies were detected which were all in men over 70 years of age.

Conclusion
Routine imaging of men is not warranted. An alternate more cost effective pathway was devised and is being implemented at UHCW.

PB.65
Impact of limiting symptomatic mammograms and use of ultrasound as first line investigation in younger women (35-40years) presenting with benign/indeterminate breast symptoms -Evaluating the local practice Subodh Seth 1 , Archana Seth 1,2

Introduction
Since 2012, in keeping with 'Best Practice Guidelines' Forth Valley breast-team has changed the practice and we don't use mammograms as regular adjunct to assessment in women in age-group 35-40years. Not all centres in Scotland are following this guideline and this results in anxiety over potential missed cancer diagnosis.

Methods
Caldicott approval was obtained. List of breast cancers diagnosed in year 2013 was obtained from the local audit department. Radiology database and clinical portal were searched for relevant breast imaging and clinical information.
Results 3610 women were imaged in year 2013 and 242 cancers were diagnosed. 310 of these women were in the age-group 35-40 years and 11 cancers were diagnosed. Analysing the data for these 11 cases, 9 had abnormal clinical and ultrasound findings. Mammograms were not crucial to reach the diagnosis. 1 case was clinically benign but had suspicious ultrasound findings. Lesion was occult on subsequent mammograms. In the remaining case, only abnormality was incidental lymphnode which was metastatic on biopsy. Malignant calcification was seen on subsequent mammography. This is the only case where potentially we could have missed the diagnosis.

Conclusion
In this study, almost all cancers were picked-up on clinical-examination or ultrasound. There was a single case where without upfront mammography we could potentially have missed the diagnosis of cancer. This was picked-up on diligent scanning of the whole quadrant. In our practice, it is safe to limit the use of mammography in the age-group 35-40years and continue using ultrasound as the first line investigation.
• We accept pre-submission inquiries • Our selector tool helps you to find the most relevant journal Submit your next manuscript to BioMed Central and we will help you at every step: