False-negative rate of ultrasound-guided fine-needle aspiration cytology for identifying axillary lymph node metastasis in breast cancer patients.

Axillary nodal status remains an important determinant of prognosis and of the therapeutic strategy in patients with a newly diagnosed breast cancer. The aim of this study was to assess the false-negative rate of ultrasound (US)-guided fine-needle aspiration cytology (FNAC) in axillary node staging at breast cancer diagnosis. All patients with a newly diagnosed breast cancer who had an indeterminate or suspicious axillary node sampled with an FNAC between 2007 and 2014 were included in the study. FNAC results were compared to the final histopathological results of surgically removed axillary lymph nodes. Patient demographics, tumor, and nodal characteristics were analyzed. Diagnostic accuracy tests were performed using IBM SPSS, version 22. A total of 3515 patients with breast cancer were identified, 675 of whom had ultrasound-guided FNAC of ipsilateral axillary lymph nodes (mean age: 55 years; Range: 26-84). A benign (C2) result was observed in 52% (n = 351) and a malignant (C5) result in 35% (n = 238). C1 was obtained in 11% (n = 76), C3 in 0.6% (n = 4), and C4 in 0.9% (n = 6). Of the 238 patients with a malignant (C5) FNAC, 99.6% had confirmed axillary lymph node metastatic disease on histopathology. Of the 351 patients with benign FNAC (C2), 31% (n = 108) of patients had a positive lymph node on histology. The false-negative rate of preoperative FNAC remains too high (31%) to omit definitive surgical staging of the axilla. The high diagnostic accuracy when a positive FNAC is obtained allows appropriate tailored decisions regarding definitive therapy.

With the Advent of Tomosynthesis in the Workup of Mammographic Abnormality, is Spot Compression Mammography Now Obsolete? An Initial Clinical Experience.

To determine if the routine use of spot compression mammography is now obsolete in the assessment of screen detected masses, asymmetries and architectural distortion since the availability of digital breast tomosynthesis. We introduced breast tomosynthesis in the workup of screen detected abnormalities in our screening center in January 2015. During an initial learning period with tomosynthesis standard spot compression views were also performed. Three consultant breast radiologists retrospectively reviewed all screening mammograms recalled for assessment over the first 6-month period. We assessed retrospectively whether there was any additional diagnostic information obtained from spot compression views not already apparent on tomography. All cases were also reviewed for any additional lesions detected by tomosynthesis, not detected on routine 2-view screening mammography. 548 women screened with standard 2-view digital screening mammography were recalled for assessment in the selected period and a total of 565 lesions were assessed. 341 lesions were assessed by both tomosynthesis and routine spot compression mammography. The spot compression view was considered more helpful than tomosynthesis in only one patient. This was because the breast was inadequately positioned for tomosynthesis and the area in question was not adequately imaged. Apart from this technical error there was no asymmetry, distortion or mass where spot compression provided more diagnostic information than tomosynthesis alone. We detected three additional cancers on tomosynthesis, not detected by routine screening mammography. From our initial experience with tomosynthesis we conclude that spot compression mammography is now obsolete in the assessment of screen detected masses, asymmetries and distortions where tomosynthesis is available.

Examining the impact of COVID-19 disruptions on population-based breast cancer screening in Ireland.

OBJECTIVE: Many population-based breast screening programmes temporarily suspended routine screening following the COVID-19 pandemic onset. This study aimed to describe screening mammography utilisation and the pattern of screen-detected breast cancer diagnoses following COVID-19-related screening disruptions in Ireland. METHODS: Using anonymous aggregate data from women invited for routine screening, three time periods were examined: (1) January-December 2019, (2) January-December 2020, and (3) January-December 2021. Descriptive statistics were conducted and comparisons between groups were performed using chi-square tests. RESULTS: In 2020, screening mammography capacity fell by 67.1% compared to 2019; recovering to 75% of mammograms performed in 2019, during 2021. Compared to 2019, for screen-detected invasive breast cancers, a reduction in Grade 1 (14.2% vs. 17.2%) and Grade 2 tumours (53.4% vs. 58.0%) and an increase in Grade 3 tumours (32.4% vs. 24.8%) was observed in 2020 (p = 0.03); whereas an increase in Grade 2 tumours (63.3% vs. 58.0%) and a reduction in Grade 3 tumours (19.6% vs. 24.8%) was found in 2021 (p = 0.02). No changes in oestrogen receptor-positive or nodal-positive diagnoses were observed; however the proportion of oestrogen/progesterone receptor-positive breast cancers significantly increased in 2020 (76.2%; p < 0.01) and 2021 (78.7%; p < 0.001) compared to 2019 (67.8%). CONCLUSION: These findings demonstrate signs of a grade change for screen-detected invasive breast cancers early in the pandemic, with recovery evident in 2021, and without an increase in nodal positivity. Future studies are needed to determine the COVID-19 impact on long-term breast cancer outcomes including mortality.

High-grade histologic features of DCIS are associated with R5 rather than R3 calcifications in breast screening mammography.

Mammographic calcification is an important radiologic feature of early breast carcinoma whose index of suspicion for malignancy may be reported by a five-tier R-category system. This study aims to describe the histologic diagnoses underlying screen-detected mammographic calcifications using both digital and screen-film mammography, and to correlate these findings with radiologic R-categories. Patients attending the Merrion Breast Screening Unit in Dublin between 2000 and 2011 were identified, who underwent needle-core biopsy for assessment of mammographic calcifications without associated mass or architectural distortion. Radiologic R-category was correlated with biopsy and excision histology reports. A total of 776 cases of calcification were identified, involving 769 individual patients. The radiologic R-categories were as follows: R3 513 (66.1%), R4 192 (24.7%), R5 71 (9.1%). The positive predictive values for malignancy were R3 32.6%, R4 69.8%, R5 95.8%. Several histologic features of DCIS were associated with R5 rather than R3 radiology: high nuclear grade, solid or cribriform architecture, necrosis, periductal inflammation or fibrosis, and associated microinvasive or invasive carcinoma. Mammographic lesions and histologic whole and invasive tumors increased in size from R3 to R5. Radiologic size of calcifications correlated with whole (but not invasive) tumor size, although it tended to underestimate it by several millimeters. Digital-detected calcifications were more likely than screen-film detected to be categorized as R3 and less likely R4 or R5, and there was no significant difference in positive predictive value between the two imaging techniques in any R-category. In conclusion, histologic features of DCIS, in particular those associated with high grade, are associated with R5 radiology. There is no significant difference in positive predictive value for malignancy in any R-category between digital and screen-film mammography.

Interval cancer audit and disclosure in breast screening programmes: An international survey.

OBJECTIVE AND SETTING: Accurate monitoring of interval cancers is important both for quality improvement and education and is a key parameter of breast screening quality assurance. Issues in relation to communication regarding interval cervical cancer in the Irish cervical screening programme were found, prompting interval cancer process review in all cancer screening programmes. An international survey to examine international consensus on interval breast cancer audit processes was conducted to inform Irish processes. METHODS: A survey of 24 international population-based breast screening programmes was done to determine which undertook audit of interval breast cancer; if yes, they were asked (1) how they undertake audit, (2) if they obtain individual consent for audit and inform women of audit results, and (3) if disclosure of audit results occurs. RESULTS: Response was 71% (17/24). Of these, 71% (12/17) have a programmatic audit process to calculate the interval cancer rate (ICR). Of these, ten also carry out radiological reviews, three using a blinded review. Two inform patients that audit is taking place; two provide choice to be in the audit; nine state that routine screening consent covers audit. For two of the five that have an open disclosure policy for medical incidents, this policy applies to screening interval cancers. One other country/region has an open disclosure policy for category 3 interval cancers only. Five have legal protection for interval cancers arising in the screened population. CONCLUSION: While consistency in providing aggregate programmatic audits exists, there is no consistent approach to individual interval cancer reviews or results disclosure.

Screen-detected ductal carcinoma in situ, 2008-2020: An observational study.

OBJECTIVES: The purpose of this study was to evaluate the grade distribution of screen-detected ductal carcinoma in situ (DCIS) diagnosed in Ireland, in the context of the clinical trials currently underway to determine if active surveillance is a feasible management option for low-risk DCIS. SETTING: BreastCheck is the national breast screening programme in Ireland, offering screening to women aged 50 to 69 every two years. METHODS: This study was a secondary analysis of data collected by BreastCheck on all screen-detected DCIS diagnosed in the 12 years of nationwide screening. Incidence and detection rates were calculated. Descriptive analysis of the cases was performed and, for comparative analysis, grade of DCIS was analysed as a binary variable (high vs. low/intermediate) in keeping with the inclusion criteria for active surveillance trials. Analysis was performed in IBM Statistical Package for Social Sciences, version 26. RESULTS: Between 2008 and 2020, 2240 women were diagnosed with DCIS through BreastCheck; 876 (39.1%) were low/intermediate-grade. The overall incidence rate has remained relatively stable during this period. Women with low/intermediate-grade DCIS were younger than women with high-grade DCIS (56 (interquartile range: 56-61) years v 57 (interquartile range: 53-61) years; p < 0.001). They were also more likely to have been diagnosed at an initial screening episode compared with those who had high-grade lesions (42.5% v 29.0%; p < 0.001). CONCLUSION: If current clinical trials recommend active surveillance as a feasible option for DCIS, up to 40% of women with screen-detected DCIS may be eligible. These women are younger and often diagnosed on initial screening episode, so may require longer active follow-up.

Retrospective radiological review and classification of interval breast cancers within population-based breast screening programmes for the purposes of open disclosure: A systematic review.

INTRODUCTION: Interval breast cancers occur following a negative breast screening mammogram and before the next scheduled appointment within screening programmes. Radiological review classifies them as cancers that develop between screens, cancers with no obvious malignant abnormalities on prior screens or cancers not detected at screening. This study aimed to systematically review published literature on the occurrence of open disclosure following interval cancer radiological reviews by breast screening programmes internationally in a retrospective setting and examine methodologies used for radiological reviews for the purposes of disclosure. METHODS: A search for relevant articles published (January 2000 - May 2019) was conducted according to PICO and PRISMA guidelines. The databases Pubmed, Scopus, Google Scholar, Cinahl, Web of Science, Embase, Science Direct and Global Health were searched. Relevant studies were reviewed if they had completed a retrospective review and classification of interval breast cancers. RESULTS: Of 46 relevant articles included, no study was identified that conducted a retrospective review purposely for open disclosure. Retrospective reviews were conducted for audit/quality assurance, and research including for radiologist education and learning. Variation in methodology was found across review type (non-blinded/semi-informed approach), number of reviewers and classification categories. The proportion of false negative cancers classified among the studies ranged from 4 to 40 %. DISCUSSION: Variation among radiological review practices were observed, which likely impacts classification results. To ensure standardised classification of interval breast cancers are employed for the purposes of open disclosure in screening settings, reproducible and consistent methodology is required.

Comparison of digital mammography and screen-film mammography in breast cancer screening: a review in the Irish breast screening program.

OBJECTIVE: Clinical trials to date into the use of full-field digital mammography (FFDM) for breast cancer screening have shown variable results. The aim of this study was to review the use of FFDM in a population-based breast cancer screening program and to compare the results with screen-film mammography. MATERIALS AND METHODS: The study included 188,823 screening examinations of women between 50 and 64 years old; 35,204 (18.6%) mammograms were obtained using FFDM. All films were double read using a 5-point rating scale to indicate the probability of cancer. Patients with positive scores were recalled for further workup. The recall rate, cancer detection rate, and positive predictive value (PPV) of FFDM were compared with screen-film mammography. RESULTS: The cancer detection rate was significantly higher for FFDM than screen-film mammography (6.3 vs 5.2 per 1,000, respectively; p = 0.01). The cancer detection rate for FFDM was higher than screen-film mammography for initial screening and subsequent screening, for invasive cancer and ductal carcinoma in situ, and across all age groups. The cancer detection rate for cancers presenting as microcalcifications was significantly higher for FFDM than for screen-film mammography (1.9 vs 1.3 per 1,000, p = 0.01). The recall rate was significantly higher for FFDM than screen-film mammography (4.0% vs 3.1%, p < 0.001). There was no significant difference in the PPVs of recall to assessment for FFDM and screen-film mammography (15.7% and 16.7%, p = 0.383). CONCLUSION: FFDM resulted in significantly higher cancer detection and recall rates than screen-film mammography in women 50-64 years old. The PPVs of FFDM and screen-film mammography were comparable. The results of this study suggest that FFDM can be safely implemented in breast cancer screening programs.

Breast screening in symptomatic women over 35 years of age: improvements in service efficiency.

INTRODUCTION: The rationalization of cancer services in Ireland saw all women with symptomatic breast problems referred to one of the eight regional cancer centers. A pilot triaging system was introduced in St Vincent's University Hospital to streamline these services. Women over 35 years who do not meet urgent referral criteria are referred for a mammogram prior to a clinic appointment ("image first"). The aim of this study was to retrospectively determine the recall rates, biopsy rates, and rate of breast cancer identification within this cohort of patients. This was compared to a screening population of patients. METHODS: Patients triaged into the "image first" group within a one-year period were identified. Results of the initial mammogram, further imaging and subsequent biopsies were recorded. Data relating to number of recalls, number of patients biopsied and number of cancers identified within the Merrion Unit of the National Breastcheck Screening Program was obtained for comparison. RESULTS: One thousand six hundred eighty-eight referrals were triaged as "image first" over this period. 185 (11%) of patients required a biopsy of an identified lesion. Breast cancer was diagnosed in 65 patients (3.9%). During the same study period, of the 42,099 women who were screened for breast cancer, 496 (1.8%) underwent biopsy and 267 (0.63%) were diagnosed with breast cancer. CONCLUSION: Image first patients, who represent a cohort of "symptomatic" non-urgent women, have a greater rate of breast cancer detection than an asymptomatic screening population. This may have an impact on the appropriate triaging of symptomatic women in a national cancer center.

Evolution of the National Breast Screening Programme in Ireland: Two-year interval analysis (2004-2013) of BreastCheck.

OBJECTIVE: Monitoring breast screening programmes is essential to ensure quality. BreastCheck, the national breast screening programme in the Republic of Ireland, commenced screening in 2000, with full national expansion in 2007, and digital mammography introduced in 2008. We aimed to review the performance of BreastCheck from 1 January 2004 to 31 December 2013. METHODS: Using the customised clinical and administrative database, performance indicator data were collected from BreastCheck and compared with programme and European guideline standards. RESULTS: Over the decade, 972,236 screening examinations were performed. Uptake initially rose following national expansion, but fell in the subsequent years to <70% in 2012-2013. Following the introduction of digital mammography, initial recall rates increased from 5.2% in 2004-2005 to 8.1% in 2012-2013. Subsequent recall rates remained within the target of <3%. On average, invasive cancer detection rates were 6.6/1000 for initial and 4.5/1000 for subsequent women. Small cancer detection rates were for <15 mm 43.4% (initial women) and 51.7% (subsequent) and for ≤10 mm 24.0% (initial) and 29.5% (subsequent). Ductal carcinoma in situ detection as a percentage of all cancers averaged 21.2% for initial and 20.0% for subsequent women. The majority were intermediate or high-grade ductal carcinoma in situ. The positive predictive value was 11.9% for initial and 21.8% for subsequent women. Standardized detection ratios remained above the programme target. CONCLUSION: Revised indicators to reflect the digital mammography era are anticipated in revised European Guidelines on breast cancer screening.

Epithelial displacement during breast needle core biopsy causes diagnostic difficulties in subsequent surgical excision specimens.

BACKGROUND: The use of needle core biopsy (NCB) as part of triple assessment for non-operative evaluation and diagnosis of breast lesions is now routine practice. Trauma to breast tissue during NCB may result in displacement of breast epithelium and may lead to diagnostic difficulty in subsequent excision specimens. METHODS: The cases of seven mammographically detected breast lesions in which epithelial displacement due to NCB was identified and caused problems in confirmation of tumour size, assessment of surgical margins, and interpretation of possible invasive carcinoma and lymphovascular invasion are reported here. CONCLUSION: Previous observations that epithelial displacement is more likely to occur when the interval between NCB and surgical excision is short are supported.

Needle core biopsy in the diagnosis of phyllodes neoplasm.

BACKGROUND: Accurate identification of phyllodes neoplasms without surgical intervention is difficult, reducing the ability to manage "benign" lumps non-operatively and impacting on the open benign biopsy rate. Needle core biopsy is considered to be a highly accurate technique in the diagnosis of breast carcinoma. Its accuracy in the diagnosis of phyllodes neoplasm has not been established. METHODS: A series of 3729 core biopsies performed between January 1999 and July 2005 were examined. All core biopsies followed by surgical excision were identified. Histologic concordance between core biopsy and excision specimen was analyzed. RESULTS: Twenty-three patients had phyllodes neoplasm on excisional biopsy with prior core biopsy findings as follows: phyllodes neoplasm (n=2), "equivocal" for phyllodes neoplasm (n=12), fibroadenoma (n=3), benign (n=6). The false negative rate for phyllodes neoplasm was therefore 39% (n=9/23). Of the total biopsy series, 35 patients had a core biopsy suggesting the possibility of phyllodes neoplasm. Of these, 32% (n=11) were found to be phyllodes neoplasm on excision, 3% (n=1) phyllodes neoplasm with breast carcinoma, 6% (n=2) breast carcinoma, and 3% (n=1) sarcoma. When a preference for phyllodes neoplasm (n=4) was stated on the equivocal core biopsies, excision correlated with the stated preference; this correlation also occurred in 90% (n=9/10) of core biopsies where fibroadenoma was favored. CONCLUSIONS: Needle core biopsy rarely produces a definite preoperative diagnosis of phyllodes neoplasm. A diagnosis of fibroadenoma or equivocal phyllodes neoplasm on core biopsy should not prevent excision if clinical suspicion remains.

Identifying patients at risk of compromised margins following breast conservation for lobular carcinoma.

BACKGROUND: The association of invasive lobular carcinoma with high rates of compromised margins in breast conservation makes choice of operation for these patients difficult. We sought to identify patients at risk of compromised margins following breast conservation surgery. METHODS: We reviewed all patients with invasive lobular and invasive ductal carcinoma over a 5-year period (1999-2004). The imaging, pathology and surgical details of patients with invasive lobular carcinoma undergoing breast conservation were analyzed. RESULTS: A total of 991 patients with invasive ductal carcinoma and 150 patients with invasive lobular carcinoma were identified. Lobular carcinomas had a compromised margin rate of 49% (n = 38/77) in breast conservation compared to 24% (n = 143/588) of ductal carcinomas (P < .0001). Mammographic size (P = .017), pathological size (P = .01), age (P = .03), multifocality (P < .0001), and lymphovascular invasion (P = .015) were significantly associated with compromised margins. CONCLUSION: Invasive lobular carcinoma has a 49% rate of compromised margins following breast conservation. Mammographic size greater than 1.5 cm and young age were preoperative factors predictive of compromised margins.

Magnetic resonance imaging appearances in primary and secondary angiosarcoma of the breast.

Angiosarcomas are malignant tumours of endovascular origin. They are rare tumours accounting for 0.04-1% of all breast malignancies. Two different forms are described: primary, occurring in young women, and secondary angiosarcoma, which occurs in older women with a history of breast-conserving surgery and radiation therapy. Imaging findings on mammography and ultrasound are non-specific, but magnetic resonance imaging with dynamic contrast enhancement is more informative. We present two cases - one of primary and one of secondary angiosarcoma - and review the imaging findings.

Axillary fine needle aspiration cytology for pre-operative staging of patients with screen-detected invasive breast carcinoma.

INTRODUCTION: Fine needle aspiration cytology (FNAC) of radiologically abnormal axillary lymph nodes in patients with breast cancer can identify patients suitable for primary axillary clearance (AC) rather than sentinel node biopsy, enabling surgical axillary staging by a single operation. This study assessed the accuracy of FNAC in predicting positive axillary lymph nodes. METHODS: 161 patients with screen-detected invasive carcinoma and who had pre-operative FNAC of a radiologically abnormal axillary lymph node were identified from two screening units, The axillary FNAC reports were correlated with sentinel node biopsy and AC reports, and sensitivity, specificity, positive (PPV) and negative (NPV) predictive values were calculated. RESULTS: FNAC had a moderate sensitivity (66.3%) and NPV (71.8%), and a high specificity (98.7%) and PPV (98.3%). Most patients (86%) had a single axillary operation. The sensitivity was highest in grade 3 (81.8%) and ductal type (77.8%) tumours. The sensitivity was lower in tumours of special type (34.8%), grade 1 tumours (50%) and those without lymphovascular invasion (LVI) (55.9%). The NPV was highest in pT1 (86.7%) and in grade 1 (84.5%) tumours, and lowest (44%) in tumours with LVI. The PPV was 100% in grade 1 and 3 tumours, stage pT2 and pT3 tumours and those without LVI, and was high (>96%) in all other groups. In lymph-node-positive patients, the mean number of lymph nodes involved was higher in the case of a positive (6.4) than negative FNAC (4.4). CONCLUSIONS: FNAC of ultrasonically abnormal axillary lymph nodes achieved surgical staging by a single operation in most patients with screen-detected invasive breast carcinoma, with moderate sensitivity and high specificity.

Needle core biopsy characteristics identify patients at risk of compromised margins in breast conservation surgery.

Selection of patients for breast-conserving surgery relies on inexact parameters such as the preoperative estimation of lesion size. This study investigates the value of needle core biopsy findings, in particular, the relative quantity of DCIS, in improving patient selection for breast conservation. Patients undergoing breast-conserving surgery for invasive ductal carcinoma from 1999 to 2004 were identified. Only patients who had a preoperative diagnosis of carcinoma (DCIS and invasive) on core biopsy were included. All core biopsies were reviewed by a breast histopathologist to document the quantity and characteristics of the DCIS component. Of a total of 281 patients, 46% (n=129) had invasive disease on core biopsy (group 1) and 54% (n=152) had either invasive disease with an accompanying DCIS component or DCIS only on core biopsy (group 2). The compromised margin rate for group 1 was 23% compared to 39% for group 2 (P=0.004). The rate of compromised margins increased progressively as the core biopsy DCIS component increased until a rate of 75% (n=18/24) was reached in patients with DCIS only on core biopsy. In patients with a DCIS component on core biopsy, the presence of necrosis (P=0.002), solid type architecture (P=0.008), high grade DCIS (P=0.007), calcification (P=0.003), and the relative proportion of DCIS present (P<0.001) were associated with compromised margins on univariate analysis. On multivariate analysis of this subgroup, the proportion of DCIS in this subgroup (P=0.048) was an independent predictor of compromised margins. The presence and relative proportion of DCIS on core biopsy provides important information as to whether patients are at risk of compromised margins. Documentation of these parameters may assist patient selection for breast-conserving surgery or identify patients who may benefit from wider margins at the time of initial operation.

Predictive value of breast lesions of "uncertain malignant potential" and "suspicious for malignancy" determined by needle core biopsy.

BACKGROUND: The optimum management of patients whose needle core biopsy (NCB) results are of "uncertain malignant potential" (B3) or "suspicious for malignancy" (B4) is unclear. This study correlates B3 and B4 NCB findings with excision histology to determine associated rates of malignancy. METHODS: All NCBs categorized as B3 or B4 were identified from a series of 3729 NCBs. Results of biopsies were reported as normal/nondiagnostic (B1), benign (B2), uncertain malignant potential (B3), suspicious but not diagnostic of malignancy (B4), or malignant (B5) according to the B classification system. B3 lesions included atypical intraductal epithelial proliferations (AIEPs), lobular neoplasia, papillary lesions, radial scars, and potential phyllodes tumors. Histological concordance between NCB and excision specimen was analyzed. RESULTS: A total of 211 B3 lesions and 51 B4 lesions were identified during the study period. The open biopsy rate after a B3/B4 finding was 86% (n = 226). The overall rate of malignancy for B3 lesions after excision was 21%. The B3 lesion-specific rates of malignancy were 6% for radial scars, 14% for papillomas, 35% for AIEP, and 44% for lobular neoplasia. Of the patients with a B4 categorization, 90% (44 of 49) were diagnosed with carcinoma after surgery. Those that were "suspicious for ductal carcinoma-in-situ" and "suspicious for invasion" correlated accurately with excision findings in 81% and 89% of patients, respectively. CONCLUSIONS: Management of lesions in the B3 categorization must be tailored to the patient because the specific lesion types are associated with highly variable rates of malignancy. A repeat biopsy or a therapeutic wide local excision should be undertaken in lesions with a B4 NCB categorization because such lesions are associated with a particularly high risk of malignancy at excision.

Factors affecting successful breast conservation for ductal carcinoma in situ.

BACKGROUND: Successful breast-conserving therapy in DCIS is restricted by high rates of residual disease resulting in the need for radiotherapy and/or re-excision. This study identifies patients with DCIS who are most at risk of compromised margins and of residual disease. METHODS: All patients undergoing breast-conserving surgery for DCIS over a 6-year period were included. Method of diagnosis, mammographic size, pathological size, DCIS-margin distance and residual disease on re-excision were analysed. RESULTS: One hundred and thirty-five patients underwent initial breast-conserving surgery for DCIS. The compromised margin rate was 72%, and the rate of residual disease on re-operation was 54%. On univariate analysis, underestimation of pathological size by mammography by >1 cm occurred in 40% of those with compromised margins undergoing a therapeutic operation compared to only 14% of those with clear margins (P = 0.02). However, on multivariate analysis only pathological size (P < 0.0001, OR = 1.0,95% CI 1.037-1.128) and lack of a preoperative diagnosis by core biopsy (P < 0.0001, OR = 5.3,95% CI 1.859-15.08) were predictive of compromised margins. The presence of residual disease on re-excision was associated with increasing pathological size (P < 0.0001, OR = 1.085,95% CI 1.038-1.134) and decreasing DCIS-margin distance (P = 0.03, OR = 6.694,95% CI 1.84-37.855). Twenty-nine percent (n = 13/45) of lesions < or =3 cm compared to 84% (n = 27/32) of lesions >3 cm had residual disease on re-operation (P < 0.0001). Residual disease was present in 62% (n = 34/55), 64% (n = 7/11) and 17% (n = 2/12) of patients with DCIS-margin distances < or =1, 1-2 and 2-5 mm, respectively. CONCLUSION: Considerable underestimation of DCIS extent by mammography occurs in a high proportion of patients with compromised margins in breast conservation. Patients at particularly high risk of residual disease on re-excision are those with lesions >3 cm and those with DCIS-margin distances of < or = 2mm.