- Email: [email protected]
Indications for diagnostic open biopsy of mammographic screen-detected lesions preoperatively diagnosed as fibroadenomas by needle biopsy and their outcomes
Clinical Radiology 70 (2015) 507e514
Contents lists available at ScienceDirect
Clinical Radiology journal homepage: www.clinicalradiologyonline.net
Indications for diagnostic open biopsy of mammographic screen-detected lesions preoperatively diagnosed as fibroadenomas by needle biopsy and their outcomes M.A. Sala a, *, R. Dhillon b, D. Brookes a, C. Lagrange a, C. Metcalf c, E. Wylie a a
The University of Western Australia, Stirling Hwy, Crawley, WA 6009, Australia BreastScreen WA, Australia c Department of Anatomical Pathology, Royal Perth Hospital, WA, Australia b
art icl e i nformat ion Article history: Received 21 October 2014 Received in revised form 19 December 2014 Accepted 29 December 2014
AIM: To identify the clinical, radiological, and histopathological factors that resulted in a diagnostic open biopsy of mammographic screen-detected lesions diagnosed preoperatively as fibroadenomas by needle biopsy. MATERIALS AND METHODS: BreastScreen WA data over 10 year period from 1 January 1999 to 31 December 2008 was reviewed. RESULTS: Among the 760,027 women screened in Western Australia between 1999 and 2008, 31 had a fine-needle aspiration (FNA) or a core biopsy (CB) diagnosing a fibroadenoma and subsequently underwent a diagnostic open biopsy (DOB). Three were preoperatively diagnosed as fibroadenoma by initial FNA but subsequent CB showed that these were not fibroadenomas and, therefore, were excluded from the present series. Of the 28 cases, DOB identified 21 fibroadenomas, two cellular fibroadenomas, two benign phyllodes tumours, one malignant phyllodes tumour, one fibroadenoma containing ductal carcinoma in situ (DCIS), and one case of a 40 mm adenosis tumour with a small 5 mm fibroadenoma. The lesions ranged from 5e100 mm in size with an average size of 28 mm. DOB and CB results were concordant in 25 (89%) of the cases. The primary clinical indications for undergoing DOB included indeterminate histopathological findings of cellular fibroadenomas versus phyllodes tumour (n ¼ 10), enlarging size (n ¼ 4), large size (n ¼ 5), fibroadenomas with atypia (n ¼ 1), discordant radiological and pathological findings (n ¼ 3), patient preference (n ¼ 1), association with a second screen-detected lesion requiring excision (n ¼ 2), and an unknown indication (n ¼ 1). CONCLUSION: CB diagnosis of fibroadenomas is a safe diagnosis unless it has atypical clinical, radiological, or pathological features. Ó 2015 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
* Guarantor and correspondent: M.A. Sala, University of Western Australia, 35 Stirling Hwy, Crawley, WA 6009, Australia. Tel.: þ61 403 760 448. E-mail address: [email protected] (M.A. Sala). http://dx.doi.org/10.1016/j.crad.2014.12.021 0009-9260/Ó 2015 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
508
M.A. Sala et al. / Clinical Radiology 70 (2015) 507e514
Introduction Fibroadenomas are benign tumours composed of stromal and epithelial components.1 They are a common incidental finding in a mammographic screening programme. The reported prevalence of fibroadenomas requiring surgery in a screening population is 0.02%, and a prevalence of 2.9% in a predominantly symptomatic patient group.2 As they are benign tumours, fibroadenomas usually do not require excision. The majority of fibroadenomas have typical clinical and radiological features, and the diagnosis can be easily identified on core biopsy. In cases where clinical, imaging, or histological features make the diagnosis non-definitive, diagnostic open biopsy (DOB) is indicated to confirm the diagnosis and exclude the possibility of malignancy. The objective of the present study was to identify the proportion of screen-detected fibroadenomas in the BreastScreen Western Australia (BSWA) Programme undergoing DOB and the features and indications that lead to DOB. The interval and subsequent cancer rates were also investigated. These outcomes were then compared with currently available evidence and guidelines.
Materials and methods Between January 1999 and December 2008 the BSWA database was searched to identify women with screendetected lesions diagnosed as fibroadenomas and who subsequently underwent DOB. BSWA is a population-based free screening mammographic service for women over 40 years. Women who are between 50e69 years are invited every 2 years to attend BSWA as part of the Australian national screening programme BreastScreen Australia. The service comprises both screening and assessment components. Mammography in two standard imaging planes (mediolateral oblique and craniocaudal) was performed using dedicated film screen mammographic equipment. Mammograms were performed at six metropolitan and four rural mobile screening sites. All films were initially read by at least two radiologists in a blinded fashion. For the purpose of the study, screen-detected fibroadenoma was defined as fibroadenomas diagnosed on either fine-needle aspiration (FNA) or core biopsy (CB) of screendetected abnormalities in the BSWA programme. The database revealed that 760,027 women were screened over this 10-year period and 36,129 of these women were recalled to the assessment clinic. Of the women who attended the assessment clinic, there were 1391 cases of fibroadenomas detected on either FNA or CB. Thirty-one of these 1391 cases of screen-detected fibroadenomas underwent DOB. Of the 31 cases, three were preoperatively diagnosed as fibroadenomas by initial FNA but subsequent CB showed that these were not fibroadenomas and, therefore, these cases were excluded from the present series, leaving 28 cases for analysis. A radiologist reviewed the case notes, radiology films, and pathology reports. The age of the women, size of the lesion, imaging characteristics, clinical
features (palpable or impalpable), reasons for undergoing DOB, and final pathology results of the lesion were recorded. Cases of interval cancers and subsequent cancers in the screen-detected fibroadenoma group were also reviewed by linking to the Western Australia Cancer Registry. All women screened at BSWA gave written consent that their de-identified data can be used for research and publication.
Results Twenty-eight (2.01%) of the lesions diagnosed preoperatively as a fibroadenoma on needle biopsy underwent DOB between 1999 and 2008 in the BSWA programme. The age range of the women was 40e72 years with an average age of 51 years. The size range of lesions was 5e100 mm with an average size of 28 mm. Of these cases, 22 underwent CB only, two underwent both CB and FNA, and four underwent FNA only. The main or primary indication that led to DOB of a lesion (summarized in Table 1) was indeterminate pathology (usually cellular fibroadenomas versus phyllodes tumour), large size, enlarging size, discordant imaging, and histopathological findings, DOB with excision of an associated screen-detected lesion, fibroadenomas with cellular atypia, sampling issues, and patient choice. In one instance, the reason(s) for undergoing DOB was not recorded. Some lesions had more than one indication that lead to their excision. In such cases, the indication that was most concerning has been used as the primary indication. The commonest indication for undergoing DOB was where needle biopsy gave a differential diagnosis of cellular fibroadenomas or phyllodes tumour occurring in 10 cases. Fibroadenoma was found at DOB in six cases, cellular fibroadenomas (Fig 1) in two cases, benign phyllodes in one case (Fig 2), and one malignant phyllodes tumour. The malignant phyllodes tumour was detected in a 46-year-old woman. She had a 100 mm palpable lesion, which showed features of a low-grade phyllodes tumour or a cellular fibroadenoma on direct FNA. The DOB identified a lowgrade phyllodes tumour with a focal area of malignant phyllodes tumour (Fig 3). Four cases were primarily excised because of an enlarging lesion on imaging. Three of these cases found at Table 1 A summary of the main indications for diagnostic open biopsy (DOB) of lesions with needle histopathology consistent with fibroadenomas. Indication for DOB
Primary indication
Cellular fibroadenomas versus PT Enlarging size Large size (>30mm) Discordant imaging/path Personal requesta Associated lesion Sampling issues Fibroadenomas with atypia Unknown Total
10 4 5 3 1 2 1 1 1 28
PT, phyllodes tumour. a Exclusive from any other clinical indication.
M.A. Sala et al. / Clinical Radiology 70 (2015) 507e514
509
Figure 1 A 51-year-old woman with a palpable 30 mm mass in the right breast and a 12 mm impalpable mass in the left (aeb). Right and left craniocaudal mammography (a) demonstrating bilateral non-calcified masses (arrows). Ultrasound (b) showed a well-defined oval-shaped hypoechoic mass in the right breast. CB was consistent with fibroadenomas but phyllodes tumour and stromal neoplasm could not be excluded. DOB histopathology identified a cellular fibroadenoma with sclerosis.
DOB were fibroadenomas and one was a fibroadenoma with benign breast change. These lesions ranged in size from 22e40 mm on imaging. Five cases were primarily excised because the lesion was >30 mm. This group included a benign phyllodes tumour measuring 40 mm that had no suspicious findings on needle biopsy and a sclerosed fibroadenoma measuring 35 mm with an adjacent focus of DCIS (Fig 4). The remaining three cases were all found to be fibroadenomas or cellular fibroadenomas. There were three cases with discordant histopathology and imaging. The lesions were identified as fibroadenomas on needle biopsy but were suspicious on imaging. One case was an irregular hypoechoic mass at ultrasound with ultrasound-guided FNA showing fibroadenoma on cytology with a differential diagnosis of hormonal mastopathy. DOB identified a fibroadenoma together with benign hormonal mastopathy. The second case was in a 46-year-old woman with a 40 mm new impalpable area of localized increased stroma which, on core biopsy, showed a fibroadenoma. The DOB found a 40 mm adenosis tumour with a 5 mm fibroadenoma at the edge of the specimen. The third case was in a 47-year-old woman with an ill-defined hypoechoic mass at ultrasound and a CB result of fibroadenoma. The final DOB identified a fibroadenoma. One case was excised because of a sampling issue that involved a 41-year-old woman with a screen-detected 5 mm solitary calcified mass. The patient underwent CB on two separate occasions with results consistent with a fibroadenoma but no retrieval of calcification. DOB showed a fibroadenomas with calcifications. In two cases, DOB of the fibroadenomas was performed as part of surgery to excise radial scars. The diagnosis of fibroadenoma and radial scar was confirmed on DOB. One lesion was excised because CB showed a fibroadenoma with atypia. This occurred in a 58-year-old woman
with a screen-detected 5 mm non-calcified mass in the right breast. An ultrasound-guided CB was consistent with fibroadenoma with lobular epithelial proliferation, cellular atypia, and abundant myxoid stroma. DOB identified a fibroadenoma with sclerosing adenosis and benign breast change. One lesion was removed with no documented indication. A 15 mm lesion was noted on the mammogram of a 43year-old woman who underwent CB on two different occasions. The first was non-diagnostic but suspicious of a fibroadenoma with the final CB confirming fibroadenoma. Subsequent DOB confirmed fibroadenoma. Overall, CB was effective for definitively excluding malignancy. The histopathological findings on DOB are given in Table 2. Of all screen-detected fibroadenomas undergoing DOB, 89% of the needle biopsy results are concordant with final diagnoses. Four women underwent FNA only. The malignant phyllodes tumour was sampled by direct FNA, which gave a differential diagnosis of cellular fibroadenoma or a lowgrade phyllodes tumour. The malignant features were only identified at DOB. The three other cases had FNAs consistent with fibroadenoma. One lesion was excised because the patient requested the excision exclusive of any other indication and subsequent DOB found a fibroadenoma with benign mammary dysplasia. The other two cases were previously discussed. One lesion was a fibroadenoma that was excised with an associated investigation for a radial scar and the other lesion was excised because of discordant ultrasound and FNA findings that DOB later identified as a fibroadenoma with hormonal mastopathy. Three cases were excluded in the final analysis because although FNA was suggestive of a fibroadenoma, a subsequent CB showed unusual or suspicious findings. In a case with screen-detected increased localized stroma and a corresponding irregular hypoechoic mass on ultrasound,
510
M.A. Sala et al. / Clinical Radiology 70 (2015) 507e514
Figure 2 A 63-year-old woman with a new impalpable 11 mm mass in the right breast (aed). (a) Right mediolateral oblique mammography demonstrates a new round non-calcified mass in the upper outer quadrant of the right breast on screening mammogram of 2006 compared to 2005. Ultrasound (b) showed a well-circumscribed hypoechoic mass with internal vascularity. CB was indeterminate with a differential diagnosis of fibroadenomas or phyllodes tumour. The DOB (ced) showed a well-demarcated biphasic tumour on low power (c) and on higher power (d) variable mildly hypercellular stroma with only occasional mitotic figures, focal periductal stromal condensation, and a predominantly intracanalicular epithelial pattern in keeping with a benign phyllodes tumour.
the CB identified intraductal papilloma with squamous metaplasia with subsequent DOB consistent with a radial scar. In a second case, CB of a poorly defined solid mass revealed epithelial hyperplasia and unusual spindle cell stroma that DOB later identified as atypical ductal hyperplasia with unusual myofibroblast differentiation. In a third case, CB of a palpable nodule different to that of screendetected abnormality in the same breast detected atypical ductal hyperplasia with subsequent DOB consistent with a radial scar. Overall, the DOB identified four lesions as high risk including two benign phyllodes tumours, one malignant phyllodes tumour, and one case of DCIS occurring in a fibroadenoma. The negative predictive value for identifying high-risk lesions using CB alone was at 99.8% in the present case series. Preoperatively, all of these lesions had either imaging or histopathological findings that indicated the lesions were suspicious and should undergo excision.
However, the final diagnosis of most of the excised lesions in the present case series was benign.
Interval and subsequent cancers Within the 1391 women with a screen-detected fibroadenoma, interval cancers were defined as cancers diagnosed within 24 months of the previous diagnosis and subsequent cancers, cancers diagnosed beyond this time frame. There were three interval cancers in the screendetected fibroadenoma group over the 10 year period from 1999 to 2008. Two cases of interval cancers occurred in the same breast as the screen-detected fibroadenomas but at different sites to the initial lesion. One case developed multifocal cancer ductal carcinoma and DCIS in the contralateral breast 18 months later when she presented with a palpable mass. None of these cases were related to a missed diagnosis.
M.A. Sala et al. / Clinical Radiology 70 (2015) 507e514
511
Figure 3 A 46-year-old woman with a palpable 100 mm solitary mass in the right breast presented for mammographic screening (aed). Right and left medial lateral oblique mammography (a) demonstrating a large non-calcified mass in the right breast. Direct FNA gave a differential diagnosis of cellular fibroadenomas or benign phyllodes tumour. On DOB, the lesion was predominately composed of low-grade phyllodes elements, including leaf-like structures growing into the cystic spaces (b). In one focus (c), hypercellular stroma extending into adjacent fat (arrow) and necrosis (lower third of photomicrograph) were present. On high power (d), there were large markedly pleomorphic stromal cells. These are all features of a malignant phyllodes tumour.
There was one case of a subsequent cancer. This occurred in a woman who was assessed for screen-detected localized cluster of calcification with an initial stereotactic FNA finding of inadequate epithelium and a subsequent CB finding of benign breast change with some features of fibroadenomas, but no calcification was identified at histopathology. Ten years later she developed cancer. The detailed notes were not available for this case. It is not known if this occurred in the same breast or not. The current sensitivity of the present method at identifying malignant lesions in screen-detected fibroadenomas is at 100%.
Discussion The management of fibroadenomas is typically conservative as they are benign lesions but their diagnosis can
represent a challenge if clinical, radiographic, sonographic, or histopathological features make the diagnosis nondefinitive. To reduce the risk of cancer associated with these non-definitive lesions, surgical removal may be recommended. Some women may also elect to remove these lesions if they represent a source of anxiety or discomfort. It is important to examine the indicators that lead to surgical removal to assess any mortality, morbidity, and costs associated with unnecessary surgery or any risks associated with not performing surgery. Markopoulos et al.2 examined the surgical outcomes of histologically verified fibroadenomas in 255 women presenting with palpable symptomatic lesions, 133 (52%) of the lesions were excised out of patient choice, 23 (9%) for enlarging size, 18 (7%) for large size, and 16 (6%) for suspicious cytology.2 These indications were similar to the present case series but vary notably in frequency as the present case-series was performed in a screened and predominately
512
M.A. Sala et al. / Clinical Radiology 70 (2015) 507e514
Figure 4 A 63-year-old woman presented for mammographic screening with a palpable 35 mm mass in the left breast (aed). Right and left mediolateral oblique mammography (a) demonstrated a well-circumscribed mass with calcification in the left breast. Ultrasound (b) showed a well-circumscribed oval-shaped hypoechoic mass. CB was consistent with fibroadenomas containing benign dystrophic calcification. On DOB (ced), the lesion was a fibroadenoma (black arrow) with sclerotic stroma, solitary dystrophic calcification, and (white arrow) a focus of DCIS (central). High power (d) showed low-grade DCIS with a predominantly solid pattern.
asymptomatic population rather than a symptomatic one. To date, the authors are not aware of any assessment of the indications that lead to excision of needle biopsy diagnosed fibroadenomas in an asymptomatic screened population. Fibroadenomas, similar to phyllodes tumours, are biphasic tumours consisting of hypocellular to cellular Table 2 Diagnostic open biopsy (DOB) findings in lesions consistent with fibroadenomas on needle biopsy. Histopathology on DOB
Cases
Benign fibroadenomas Cellular fibroadenomas Benign phyllodes Malignant phyllodes DCIS within fibroadenomas Adenosis tumour with small fibroadenomas Total
21 2 2 1 1 1 28
DCIS, ductal carcinoma in situ.
stroma and benign epithelium. The two lesions are distinguished by the cellularity of their stroma. However, this presents two problems on CB. One is that phyllodes tumours often have quite variable cellularity throughout the tumour so that a core biopsy may sample only hypocellular stroma.3 The second problem is that cellular fibroadenomas may have a similar degree of cellularity to low-grade phyllodes tumours on CB.3 Although there are no specific features on needle biopsy to distinguish cellular fibroadenomas from phyllodes tumours, fibroadenomas may have either a pericanalicular epithelial pattern (elongated, compressed ducts with cleft-like structures) or an intracanalicular pattern (open elongated ducts and lobules), whereas phyllodes tumours have an intracanalicular pattern with, at most, a focal pericanalicular pattern. Multiple cores should demonstrate the typically variable stromal cellularity of phyllodes tumours.3 Other features, such as periductal condensation of the stroma, adipose tissue
M.A. Sala et al. / Clinical Radiology 70 (2015) 507e514
within stroma, and stromal overgrowth (10 field with no epithelium), may be seen. Alternatively, the uniformity of the stromal cellularity of fibroadenomas without periductal stromal condensation or stromal overgrowth and infrequently adipose tissue may be seen.3 Mitotic activity and stromal atypia are not usually seen, but if present, certainly suggest phyllodes tumors.3 Cores of phyllodes tumours are often fragmented but this can also be seen in fibroadenomas. Another problem with CBs is classifying phyllodes tumours as benign, borderline, or malignant. Stromal cellularity, stromal overgrowth, nuclear pleomorphism, mitotic rate, and infiltrating margins are all used to classify the tumours and these features are best seen in an excision biopsy.3e6 Attempts have been made to diagnose and classify phyllodes tumours using cytology and needle biopsy alone, but have been unsuccessful; although immunohistochemistry may improve the diagnostic accuracy in the future.4e7 Of 1883 radiographically-detected category 3B lesions in the Breast Screen South Australia programme most lesions with indeterminate cytology had a final diagnosis of fibroadenoma.8 Only 0.6% of the lesions were diagnosed as phyllodes tumours and none were malignant. Farshid et al.8 suggested that if a malignant phyllodes tumour does occur in these screening groups other imaging features are likely to exclude it from the 3B category suggesting that radiography with cytology is reasonably specific for fibroadeSeventy-eight percent of indeterminate noma.8 hypercellular lesions in the present study had a final diagnosis of fibroadenoma but this is not considered to be an acceptable level of uncertainty, and therefore, excision is recommended in these cases. Most fibroadenomas are composed of polyclonal epithelial and polyclonal stromal cells, whereas phyllodes tumours have monoclonal stromal cells and, usually, polyclonal epithelial cells. Monoclonal stromal fibroadenomas have been identified on very rare occasions and these have subsequently been shown to develop into phyllodes tumours. Fibroadenomas are reported to have a 0.0125e0.3 incidence of transformation into malignancy.9e11 This is believed to be the result of somatic mutations in monoclonal fibroadenomas, which cause recurrence and transformation into phyllodes tumours.9,10 This occurrence was not document in the present screening population as none of the interval and subsequent lesions in the present study were associated with a pre-existing or removed fibroadenoma. Lesions with diagnostic features of fibroadenoma are unlikely to be malignant or subsequently transform into malignancy as confirmed by both in-vitro and now population-based studies; however, a high index of suspicion is needed to differentiate between hypercellular fibroadenomas and neoplastic phyllodes tumours.4e10 For needle biopsy to be diagnostic, sampling must be representative and this is more difficult to achieve in larger lesions that may contain malignant foci. At mammography, if these lesions are categorized as 3B, Farshid et al.12 noted they are associated with a reduced risk of malignancy as larger malignant lesions are likely to show suspicious radiological features and not be classified into this group.
513
The incidence of malignancy of excised fibroadenomas has previously been reported to be at 1e3%.2,11,13,14 In the present series an incidence of 0.14% was reported. As a lesion’s size increases, the incidence of malignancy in screendetected fibroadenomas increases. The only two cases of malignancy that arose in the present series were in lesions >3 cm with needle biopsy also failing to identify malignancy. This supports the recommendation of Gordon et al.15 that lesions >3 cm should be surgically excised to account for sampling errors that become clinically significant in larger fibroadenoma-like lesions. Fibroadenomas are commonly found to change in size but enlarging lesions tend to reduce the confidence in the needle biopsy diagnosis, and are therefore, more likely to be excised.2,15 Gordon et al.15 reported that 17.5% of 1070 patients with breast lesions given the diagnosis of probable fibroadenoma by FNA showed a notable volume increase. For women over 50, the 95th percentile in volume of growth was reported to be at 22% per month. Of 194 enlarging lesions, two were found to be phyllodes tumours and one was a hamartoma. Gordon et al.15 suggested that growth rates by volume in excess of 15% per month should undergo excision but otherwise long-term surveillance should be safe. Farshid et al.12 noted that enlarging grade 3 lesions did not have a statistically significant increase in relative risk of malignancy. In the present sample, four lesions were indicated primarily because they were enlarging and all were fibroadenomas. Three other lesions in the present series were also identified to be enlarging but were excised because of another concerning feature. Two were fibroadenomas and one of these lesions was a large (40 mm) and enlarging benign phyllodes tumour. In order to identify phyllodes tumours, an upper limit on the size and rate of growth is pertinent as an indication for excision, but this supports the notion that changes in lesion size may not be associated with a significant increase in risk and that other features are likely to suggest malignancy. CB is reliable in the absence of suspicious clinical, cytological, and histological features. Previous studies have shown FNA alone to be insufficient to reliably exclude malignancy2e6,8,15 and FNA in isolation was found to be inadequate for investigating suspicious features in the present study. There were three cases excluded from the present series where initial FNA suggested a diagnosis of a fibroadenoma but a subsequent CB indicated another diagnosis that was more concordant with DOB. In the present series, four cases underwent FNA only. Early on in the present study review period, FNA was considered sufficient to investigate screen-detected lesions in concordance with best practice and unit policy at the time. In the case of the large phyllodes tumour, it was clear that malignant features were difficult to identify on FNA. In the present series, final histopathological results on DOB were concordant with CB in the majority of the cases. Therefore, it is not recommended that FNA is used in assessing suspected fibroadenomas; CB is considered to be more representative and specific in identifying pertinent histopathological findings. The current policy of the unit is to perform CB as a standard practice.
514
M.A. Sala et al. / Clinical Radiology 70 (2015) 507e514
Our investigation of interval cancers and subsequent cancers in the screen-detected fibroadenomas group found that the interval cancer rate was 2.1 per 10,000 women/year over the 10 year study period, which is comparable to the BSWA interval cancer rate of 7.5 per 10,000 women/year between 2006 and 2010.16 None of the lesions in the present series, including the interval and subsequent cancers, were associated with a pre-existing screen-detected fibroadenomas. Although patients with fibroadenomas have been shown to have an increased relative risk of breast carcinoma by 1.3e2.1 for typical fibroadenomas, and 3.1e3.72 for fibroadenomas with complex features,11,17,18 the women in the present study did not show an increased risk of developing malignancy. In conclusion, in the majority of screen-detected fibroadenomas that underwent DOB in the BSWA programme, the CB diagnosis was concordant with DOB histopathology. However, to accurately diagnose and manage patients using only CB, it is important to correlate the clinical, radiological, and pathological findings, which may indicate, repeat CB or DOB. In the programme, most of the cases who underwent DOB had atypical imaging, clinical and/or pathological findings. This practice was safe as demonstrated by a very low interval cancer rate, which was comparable to the background interval cancer rate in the BSWA WA programme. Data published by Farshid et al. supports this practice where the study concluded that surgical biopsy was not indicated in well-sampled grade 3 masses with benign needle biopsy results.8 CB diagnosis of an uncomplicated typical fibroadenoma is an extremely safe diagnosis. DOB should be reserved for atypical fibroadenomas diagnosed on CB and/or lesions with suspicious clinical features.
Acknowledgements The authors thank Janette Tresham for assistance with data and records and Barbara Taylor for assistance with image manipulation and upload.
References 1. Rosen PP. Fibroepithelial neoplasms. In: Rosen PP, editor. Breast pathology. Philadelphia, PA: Lippincott-Raven; 1997. p. 143e75. 2. Markopoulos C, Kouskos E, Mantas D, et al. Fibroadenomas of the breast: is there any association with breast cancer? Eur J Gynaecol Oncol 2004;25:495e7. 3. Dabbs D. Breast pathology. 1st edn. Pittsburgh, PA: Elsevier Saunders; 2012. p. 233e48. 4. Scoyler RA, McKenzie PR, Achmed D, et al. Can phyllodes tumours of the breast be distinguished from fibroadenomas using fine needle aspiration cytology. Pathology 2001;33:437e43. 5. Tan PH, Thike AA, Tan WJ, et al. Predicting clinical behavior of breast phyllodes tumours: a nomogram based on histological criteria and surgical margins. J Clin Pathol 2012;65:69e76. 6. Tan PH, Jayabaskar T, Chuah KH, et al. Phyllodes tumours of the breast d the role of pathologic parameters. Am J Clin Pathol 2005;123:529e40. 7. Maity PP, Chatterjee S, Das RK, et al. Finding an optimum immunohistochemical feature set to distinguish benign phyllodes from fibroadenoma. Micron 2013;48:34e41. 8. Farshid G, Downey P, Gill P. Assessment of 1183 screen-detected, category 3B, circumscribed masses by cytology and core biopsy with longterm follow up data. Br J Cancer 2008;98:1182e90. 9. Noguchi S, Motomura K, Inaji H, et al. Clonal analysis of fibroadenoma and phyllodes tumor of the breast by means of polymerase chain reaction. Cancer Rrs 1993;53:4071e4. 10. Noguchi S, Yokouchi H, Aihara T, et al. Progression of phyllodes tumour demonstrated by clonal analysis. Cancer 1995;76:1779e85. 11. Sklair-Levy M, Sella T, Alweiss, et al. Incidence and management of complex fibroadenomas. AJR Am J Roentgenol 2008;190:214e8. 12. Farshid G, Walker A, Battersby G, et al. Predictors of malignancy in screen-detected breast masses with inderminant/equivocal (grade 3) imaging features. Breast 2011;20:56e61. 13. Kuijper A, Mommers EC, van der Wall E, et al. Histopathology of fibroadenoma of the breast. Am J Clin Pathol 2001;115:736e42. 14. Ozello L, Gump FE. The management of patients with carcinomas in fibroadenomatous tumors of the breast. Surg Gynecol Obstet 1985;160:99e104. 15. Gordon PB. Gagnon fibroadenomas, Lanzkowsky L. Solid breast masses diagnosed as fibroadenoma at fine-needle aspiration biopsy: acceptable rates of growth at long-term follow-up. Radiology 2003;229:233e8. 16. Breast screen WA statistical report 2006e2010. Perth: Breast Screen WA; 2012 Oct.. 17. Dupont WD, Page DL, Parl FF, et al. Long-term risk of breast cancer in women with fibroadenoma. N Eng J Med 1994;331:10e5. 18. El-wakeel H, Umplelby HC. Systematic review of fibroadenoma as a risk factor for breast cancer. Breast 2003;12:302.
Contents lists available at ScienceDirect
Clinical Radiology journal homepage: www.clinicalradiologyonline.net
Indications for diagnostic open biopsy of mammographic screen-detected lesions preoperatively diagnosed as fibroadenomas by needle biopsy and their outcomes M.A. Sala a, *, R. Dhillon b, D. Brookes a, C. Lagrange a, C. Metcalf c, E. Wylie a a
The University of Western Australia, Stirling Hwy, Crawley, WA 6009, Australia BreastScreen WA, Australia c Department of Anatomical Pathology, Royal Perth Hospital, WA, Australia b
art icl e i nformat ion Article history: Received 21 October 2014 Received in revised form 19 December 2014 Accepted 29 December 2014
AIM: To identify the clinical, radiological, and histopathological factors that resulted in a diagnostic open biopsy of mammographic screen-detected lesions diagnosed preoperatively as fibroadenomas by needle biopsy. MATERIALS AND METHODS: BreastScreen WA data over 10 year period from 1 January 1999 to 31 December 2008 was reviewed. RESULTS: Among the 760,027 women screened in Western Australia between 1999 and 2008, 31 had a fine-needle aspiration (FNA) or a core biopsy (CB) diagnosing a fibroadenoma and subsequently underwent a diagnostic open biopsy (DOB). Three were preoperatively diagnosed as fibroadenoma by initial FNA but subsequent CB showed that these were not fibroadenomas and, therefore, were excluded from the present series. Of the 28 cases, DOB identified 21 fibroadenomas, two cellular fibroadenomas, two benign phyllodes tumours, one malignant phyllodes tumour, one fibroadenoma containing ductal carcinoma in situ (DCIS), and one case of a 40 mm adenosis tumour with a small 5 mm fibroadenoma. The lesions ranged from 5e100 mm in size with an average size of 28 mm. DOB and CB results were concordant in 25 (89%) of the cases. The primary clinical indications for undergoing DOB included indeterminate histopathological findings of cellular fibroadenomas versus phyllodes tumour (n ¼ 10), enlarging size (n ¼ 4), large size (n ¼ 5), fibroadenomas with atypia (n ¼ 1), discordant radiological and pathological findings (n ¼ 3), patient preference (n ¼ 1), association with a second screen-detected lesion requiring excision (n ¼ 2), and an unknown indication (n ¼ 1). CONCLUSION: CB diagnosis of fibroadenomas is a safe diagnosis unless it has atypical clinical, radiological, or pathological features. Ó 2015 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
* Guarantor and correspondent: M.A. Sala, University of Western Australia, 35 Stirling Hwy, Crawley, WA 6009, Australia. Tel.: þ61 403 760 448. E-mail address: [email protected] (M.A. Sala). http://dx.doi.org/10.1016/j.crad.2014.12.021 0009-9260/Ó 2015 The Royal College of Radiologists. Published by Elsevier Ltd. All rights reserved.
508
M.A. Sala et al. / Clinical Radiology 70 (2015) 507e514
Introduction Fibroadenomas are benign tumours composed of stromal and epithelial components.1 They are a common incidental finding in a mammographic screening programme. The reported prevalence of fibroadenomas requiring surgery in a screening population is 0.02%, and a prevalence of 2.9% in a predominantly symptomatic patient group.2 As they are benign tumours, fibroadenomas usually do not require excision. The majority of fibroadenomas have typical clinical and radiological features, and the diagnosis can be easily identified on core biopsy. In cases where clinical, imaging, or histological features make the diagnosis non-definitive, diagnostic open biopsy (DOB) is indicated to confirm the diagnosis and exclude the possibility of malignancy. The objective of the present study was to identify the proportion of screen-detected fibroadenomas in the BreastScreen Western Australia (BSWA) Programme undergoing DOB and the features and indications that lead to DOB. The interval and subsequent cancer rates were also investigated. These outcomes were then compared with currently available evidence and guidelines.
Materials and methods Between January 1999 and December 2008 the BSWA database was searched to identify women with screendetected lesions diagnosed as fibroadenomas and who subsequently underwent DOB. BSWA is a population-based free screening mammographic service for women over 40 years. Women who are between 50e69 years are invited every 2 years to attend BSWA as part of the Australian national screening programme BreastScreen Australia. The service comprises both screening and assessment components. Mammography in two standard imaging planes (mediolateral oblique and craniocaudal) was performed using dedicated film screen mammographic equipment. Mammograms were performed at six metropolitan and four rural mobile screening sites. All films were initially read by at least two radiologists in a blinded fashion. For the purpose of the study, screen-detected fibroadenoma was defined as fibroadenomas diagnosed on either fine-needle aspiration (FNA) or core biopsy (CB) of screendetected abnormalities in the BSWA programme. The database revealed that 760,027 women were screened over this 10-year period and 36,129 of these women were recalled to the assessment clinic. Of the women who attended the assessment clinic, there were 1391 cases of fibroadenomas detected on either FNA or CB. Thirty-one of these 1391 cases of screen-detected fibroadenomas underwent DOB. Of the 31 cases, three were preoperatively diagnosed as fibroadenomas by initial FNA but subsequent CB showed that these were not fibroadenomas and, therefore, these cases were excluded from the present series, leaving 28 cases for analysis. A radiologist reviewed the case notes, radiology films, and pathology reports. The age of the women, size of the lesion, imaging characteristics, clinical
features (palpable or impalpable), reasons for undergoing DOB, and final pathology results of the lesion were recorded. Cases of interval cancers and subsequent cancers in the screen-detected fibroadenoma group were also reviewed by linking to the Western Australia Cancer Registry. All women screened at BSWA gave written consent that their de-identified data can be used for research and publication.
Results Twenty-eight (2.01%) of the lesions diagnosed preoperatively as a fibroadenoma on needle biopsy underwent DOB between 1999 and 2008 in the BSWA programme. The age range of the women was 40e72 years with an average age of 51 years. The size range of lesions was 5e100 mm with an average size of 28 mm. Of these cases, 22 underwent CB only, two underwent both CB and FNA, and four underwent FNA only. The main or primary indication that led to DOB of a lesion (summarized in Table 1) was indeterminate pathology (usually cellular fibroadenomas versus phyllodes tumour), large size, enlarging size, discordant imaging, and histopathological findings, DOB with excision of an associated screen-detected lesion, fibroadenomas with cellular atypia, sampling issues, and patient choice. In one instance, the reason(s) for undergoing DOB was not recorded. Some lesions had more than one indication that lead to their excision. In such cases, the indication that was most concerning has been used as the primary indication. The commonest indication for undergoing DOB was where needle biopsy gave a differential diagnosis of cellular fibroadenomas or phyllodes tumour occurring in 10 cases. Fibroadenoma was found at DOB in six cases, cellular fibroadenomas (Fig 1) in two cases, benign phyllodes in one case (Fig 2), and one malignant phyllodes tumour. The malignant phyllodes tumour was detected in a 46-year-old woman. She had a 100 mm palpable lesion, which showed features of a low-grade phyllodes tumour or a cellular fibroadenoma on direct FNA. The DOB identified a lowgrade phyllodes tumour with a focal area of malignant phyllodes tumour (Fig 3). Four cases were primarily excised because of an enlarging lesion on imaging. Three of these cases found at Table 1 A summary of the main indications for diagnostic open biopsy (DOB) of lesions with needle histopathology consistent with fibroadenomas. Indication for DOB
Primary indication
Cellular fibroadenomas versus PT Enlarging size Large size (>30mm) Discordant imaging/path Personal requesta Associated lesion Sampling issues Fibroadenomas with atypia Unknown Total
10 4 5 3 1 2 1 1 1 28
PT, phyllodes tumour. a Exclusive from any other clinical indication.
M.A. Sala et al. / Clinical Radiology 70 (2015) 507e514
509
Figure 1 A 51-year-old woman with a palpable 30 mm mass in the right breast and a 12 mm impalpable mass in the left (aeb). Right and left craniocaudal mammography (a) demonstrating bilateral non-calcified masses (arrows). Ultrasound (b) showed a well-defined oval-shaped hypoechoic mass in the right breast. CB was consistent with fibroadenomas but phyllodes tumour and stromal neoplasm could not be excluded. DOB histopathology identified a cellular fibroadenoma with sclerosis.
DOB were fibroadenomas and one was a fibroadenoma with benign breast change. These lesions ranged in size from 22e40 mm on imaging. Five cases were primarily excised because the lesion was >30 mm. This group included a benign phyllodes tumour measuring 40 mm that had no suspicious findings on needle biopsy and a sclerosed fibroadenoma measuring 35 mm with an adjacent focus of DCIS (Fig 4). The remaining three cases were all found to be fibroadenomas or cellular fibroadenomas. There were three cases with discordant histopathology and imaging. The lesions were identified as fibroadenomas on needle biopsy but were suspicious on imaging. One case was an irregular hypoechoic mass at ultrasound with ultrasound-guided FNA showing fibroadenoma on cytology with a differential diagnosis of hormonal mastopathy. DOB identified a fibroadenoma together with benign hormonal mastopathy. The second case was in a 46-year-old woman with a 40 mm new impalpable area of localized increased stroma which, on core biopsy, showed a fibroadenoma. The DOB found a 40 mm adenosis tumour with a 5 mm fibroadenoma at the edge of the specimen. The third case was in a 47-year-old woman with an ill-defined hypoechoic mass at ultrasound and a CB result of fibroadenoma. The final DOB identified a fibroadenoma. One case was excised because of a sampling issue that involved a 41-year-old woman with a screen-detected 5 mm solitary calcified mass. The patient underwent CB on two separate occasions with results consistent with a fibroadenoma but no retrieval of calcification. DOB showed a fibroadenomas with calcifications. In two cases, DOB of the fibroadenomas was performed as part of surgery to excise radial scars. The diagnosis of fibroadenoma and radial scar was confirmed on DOB. One lesion was excised because CB showed a fibroadenoma with atypia. This occurred in a 58-year-old woman
with a screen-detected 5 mm non-calcified mass in the right breast. An ultrasound-guided CB was consistent with fibroadenoma with lobular epithelial proliferation, cellular atypia, and abundant myxoid stroma. DOB identified a fibroadenoma with sclerosing adenosis and benign breast change. One lesion was removed with no documented indication. A 15 mm lesion was noted on the mammogram of a 43year-old woman who underwent CB on two different occasions. The first was non-diagnostic but suspicious of a fibroadenoma with the final CB confirming fibroadenoma. Subsequent DOB confirmed fibroadenoma. Overall, CB was effective for definitively excluding malignancy. The histopathological findings on DOB are given in Table 2. Of all screen-detected fibroadenomas undergoing DOB, 89% of the needle biopsy results are concordant with final diagnoses. Four women underwent FNA only. The malignant phyllodes tumour was sampled by direct FNA, which gave a differential diagnosis of cellular fibroadenoma or a lowgrade phyllodes tumour. The malignant features were only identified at DOB. The three other cases had FNAs consistent with fibroadenoma. One lesion was excised because the patient requested the excision exclusive of any other indication and subsequent DOB found a fibroadenoma with benign mammary dysplasia. The other two cases were previously discussed. One lesion was a fibroadenoma that was excised with an associated investigation for a radial scar and the other lesion was excised because of discordant ultrasound and FNA findings that DOB later identified as a fibroadenoma with hormonal mastopathy. Three cases were excluded in the final analysis because although FNA was suggestive of a fibroadenoma, a subsequent CB showed unusual or suspicious findings. In a case with screen-detected increased localized stroma and a corresponding irregular hypoechoic mass on ultrasound,
510
M.A. Sala et al. / Clinical Radiology 70 (2015) 507e514
Figure 2 A 63-year-old woman with a new impalpable 11 mm mass in the right breast (aed). (a) Right mediolateral oblique mammography demonstrates a new round non-calcified mass in the upper outer quadrant of the right breast on screening mammogram of 2006 compared to 2005. Ultrasound (b) showed a well-circumscribed hypoechoic mass with internal vascularity. CB was indeterminate with a differential diagnosis of fibroadenomas or phyllodes tumour. The DOB (ced) showed a well-demarcated biphasic tumour on low power (c) and on higher power (d) variable mildly hypercellular stroma with only occasional mitotic figures, focal periductal stromal condensation, and a predominantly intracanalicular epithelial pattern in keeping with a benign phyllodes tumour.
the CB identified intraductal papilloma with squamous metaplasia with subsequent DOB consistent with a radial scar. In a second case, CB of a poorly defined solid mass revealed epithelial hyperplasia and unusual spindle cell stroma that DOB later identified as atypical ductal hyperplasia with unusual myofibroblast differentiation. In a third case, CB of a palpable nodule different to that of screendetected abnormality in the same breast detected atypical ductal hyperplasia with subsequent DOB consistent with a radial scar. Overall, the DOB identified four lesions as high risk including two benign phyllodes tumours, one malignant phyllodes tumour, and one case of DCIS occurring in a fibroadenoma. The negative predictive value for identifying high-risk lesions using CB alone was at 99.8% in the present case series. Preoperatively, all of these lesions had either imaging or histopathological findings that indicated the lesions were suspicious and should undergo excision.
However, the final diagnosis of most of the excised lesions in the present case series was benign.
Interval and subsequent cancers Within the 1391 women with a screen-detected fibroadenoma, interval cancers were defined as cancers diagnosed within 24 months of the previous diagnosis and subsequent cancers, cancers diagnosed beyond this time frame. There were three interval cancers in the screendetected fibroadenoma group over the 10 year period from 1999 to 2008. Two cases of interval cancers occurred in the same breast as the screen-detected fibroadenomas but at different sites to the initial lesion. One case developed multifocal cancer ductal carcinoma and DCIS in the contralateral breast 18 months later when she presented with a palpable mass. None of these cases were related to a missed diagnosis.
M.A. Sala et al. / Clinical Radiology 70 (2015) 507e514
511
Figure 3 A 46-year-old woman with a palpable 100 mm solitary mass in the right breast presented for mammographic screening (aed). Right and left medial lateral oblique mammography (a) demonstrating a large non-calcified mass in the right breast. Direct FNA gave a differential diagnosis of cellular fibroadenomas or benign phyllodes tumour. On DOB, the lesion was predominately composed of low-grade phyllodes elements, including leaf-like structures growing into the cystic spaces (b). In one focus (c), hypercellular stroma extending into adjacent fat (arrow) and necrosis (lower third of photomicrograph) were present. On high power (d), there were large markedly pleomorphic stromal cells. These are all features of a malignant phyllodes tumour.
There was one case of a subsequent cancer. This occurred in a woman who was assessed for screen-detected localized cluster of calcification with an initial stereotactic FNA finding of inadequate epithelium and a subsequent CB finding of benign breast change with some features of fibroadenomas, but no calcification was identified at histopathology. Ten years later she developed cancer. The detailed notes were not available for this case. It is not known if this occurred in the same breast or not. The current sensitivity of the present method at identifying malignant lesions in screen-detected fibroadenomas is at 100%.
Discussion The management of fibroadenomas is typically conservative as they are benign lesions but their diagnosis can
represent a challenge if clinical, radiographic, sonographic, or histopathological features make the diagnosis nondefinitive. To reduce the risk of cancer associated with these non-definitive lesions, surgical removal may be recommended. Some women may also elect to remove these lesions if they represent a source of anxiety or discomfort. It is important to examine the indicators that lead to surgical removal to assess any mortality, morbidity, and costs associated with unnecessary surgery or any risks associated with not performing surgery. Markopoulos et al.2 examined the surgical outcomes of histologically verified fibroadenomas in 255 women presenting with palpable symptomatic lesions, 133 (52%) of the lesions were excised out of patient choice, 23 (9%) for enlarging size, 18 (7%) for large size, and 16 (6%) for suspicious cytology.2 These indications were similar to the present case series but vary notably in frequency as the present case-series was performed in a screened and predominately
512
M.A. Sala et al. / Clinical Radiology 70 (2015) 507e514
Figure 4 A 63-year-old woman presented for mammographic screening with a palpable 35 mm mass in the left breast (aed). Right and left mediolateral oblique mammography (a) demonstrated a well-circumscribed mass with calcification in the left breast. Ultrasound (b) showed a well-circumscribed oval-shaped hypoechoic mass. CB was consistent with fibroadenomas containing benign dystrophic calcification. On DOB (ced), the lesion was a fibroadenoma (black arrow) with sclerotic stroma, solitary dystrophic calcification, and (white arrow) a focus of DCIS (central). High power (d) showed low-grade DCIS with a predominantly solid pattern.
asymptomatic population rather than a symptomatic one. To date, the authors are not aware of any assessment of the indications that lead to excision of needle biopsy diagnosed fibroadenomas in an asymptomatic screened population. Fibroadenomas, similar to phyllodes tumours, are biphasic tumours consisting of hypocellular to cellular Table 2 Diagnostic open biopsy (DOB) findings in lesions consistent with fibroadenomas on needle biopsy. Histopathology on DOB
Cases
Benign fibroadenomas Cellular fibroadenomas Benign phyllodes Malignant phyllodes DCIS within fibroadenomas Adenosis tumour with small fibroadenomas Total
21 2 2 1 1 1 28
DCIS, ductal carcinoma in situ.
stroma and benign epithelium. The two lesions are distinguished by the cellularity of their stroma. However, this presents two problems on CB. One is that phyllodes tumours often have quite variable cellularity throughout the tumour so that a core biopsy may sample only hypocellular stroma.3 The second problem is that cellular fibroadenomas may have a similar degree of cellularity to low-grade phyllodes tumours on CB.3 Although there are no specific features on needle biopsy to distinguish cellular fibroadenomas from phyllodes tumours, fibroadenomas may have either a pericanalicular epithelial pattern (elongated, compressed ducts with cleft-like structures) or an intracanalicular pattern (open elongated ducts and lobules), whereas phyllodes tumours have an intracanalicular pattern with, at most, a focal pericanalicular pattern. Multiple cores should demonstrate the typically variable stromal cellularity of phyllodes tumours.3 Other features, such as periductal condensation of the stroma, adipose tissue
M.A. Sala et al. / Clinical Radiology 70 (2015) 507e514
within stroma, and stromal overgrowth (10 field with no epithelium), may be seen. Alternatively, the uniformity of the stromal cellularity of fibroadenomas without periductal stromal condensation or stromal overgrowth and infrequently adipose tissue may be seen.3 Mitotic activity and stromal atypia are not usually seen, but if present, certainly suggest phyllodes tumors.3 Cores of phyllodes tumours are often fragmented but this can also be seen in fibroadenomas. Another problem with CBs is classifying phyllodes tumours as benign, borderline, or malignant. Stromal cellularity, stromal overgrowth, nuclear pleomorphism, mitotic rate, and infiltrating margins are all used to classify the tumours and these features are best seen in an excision biopsy.3e6 Attempts have been made to diagnose and classify phyllodes tumours using cytology and needle biopsy alone, but have been unsuccessful; although immunohistochemistry may improve the diagnostic accuracy in the future.4e7 Of 1883 radiographically-detected category 3B lesions in the Breast Screen South Australia programme most lesions with indeterminate cytology had a final diagnosis of fibroadenoma.8 Only 0.6% of the lesions were diagnosed as phyllodes tumours and none were malignant. Farshid et al.8 suggested that if a malignant phyllodes tumour does occur in these screening groups other imaging features are likely to exclude it from the 3B category suggesting that radiography with cytology is reasonably specific for fibroadeSeventy-eight percent of indeterminate noma.8 hypercellular lesions in the present study had a final diagnosis of fibroadenoma but this is not considered to be an acceptable level of uncertainty, and therefore, excision is recommended in these cases. Most fibroadenomas are composed of polyclonal epithelial and polyclonal stromal cells, whereas phyllodes tumours have monoclonal stromal cells and, usually, polyclonal epithelial cells. Monoclonal stromal fibroadenomas have been identified on very rare occasions and these have subsequently been shown to develop into phyllodes tumours. Fibroadenomas are reported to have a 0.0125e0.3 incidence of transformation into malignancy.9e11 This is believed to be the result of somatic mutations in monoclonal fibroadenomas, which cause recurrence and transformation into phyllodes tumours.9,10 This occurrence was not document in the present screening population as none of the interval and subsequent lesions in the present study were associated with a pre-existing or removed fibroadenoma. Lesions with diagnostic features of fibroadenoma are unlikely to be malignant or subsequently transform into malignancy as confirmed by both in-vitro and now population-based studies; however, a high index of suspicion is needed to differentiate between hypercellular fibroadenomas and neoplastic phyllodes tumours.4e10 For needle biopsy to be diagnostic, sampling must be representative and this is more difficult to achieve in larger lesions that may contain malignant foci. At mammography, if these lesions are categorized as 3B, Farshid et al.12 noted they are associated with a reduced risk of malignancy as larger malignant lesions are likely to show suspicious radiological features and not be classified into this group.
513
The incidence of malignancy of excised fibroadenomas has previously been reported to be at 1e3%.2,11,13,14 In the present series an incidence of 0.14% was reported. As a lesion’s size increases, the incidence of malignancy in screendetected fibroadenomas increases. The only two cases of malignancy that arose in the present series were in lesions >3 cm with needle biopsy also failing to identify malignancy. This supports the recommendation of Gordon et al.15 that lesions >3 cm should be surgically excised to account for sampling errors that become clinically significant in larger fibroadenoma-like lesions. Fibroadenomas are commonly found to change in size but enlarging lesions tend to reduce the confidence in the needle biopsy diagnosis, and are therefore, more likely to be excised.2,15 Gordon et al.15 reported that 17.5% of 1070 patients with breast lesions given the diagnosis of probable fibroadenoma by FNA showed a notable volume increase. For women over 50, the 95th percentile in volume of growth was reported to be at 22% per month. Of 194 enlarging lesions, two were found to be phyllodes tumours and one was a hamartoma. Gordon et al.15 suggested that growth rates by volume in excess of 15% per month should undergo excision but otherwise long-term surveillance should be safe. Farshid et al.12 noted that enlarging grade 3 lesions did not have a statistically significant increase in relative risk of malignancy. In the present sample, four lesions were indicated primarily because they were enlarging and all were fibroadenomas. Three other lesions in the present series were also identified to be enlarging but were excised because of another concerning feature. Two were fibroadenomas and one of these lesions was a large (40 mm) and enlarging benign phyllodes tumour. In order to identify phyllodes tumours, an upper limit on the size and rate of growth is pertinent as an indication for excision, but this supports the notion that changes in lesion size may not be associated with a significant increase in risk and that other features are likely to suggest malignancy. CB is reliable in the absence of suspicious clinical, cytological, and histological features. Previous studies have shown FNA alone to be insufficient to reliably exclude malignancy2e6,8,15 and FNA in isolation was found to be inadequate for investigating suspicious features in the present study. There were three cases excluded from the present series where initial FNA suggested a diagnosis of a fibroadenoma but a subsequent CB indicated another diagnosis that was more concordant with DOB. In the present series, four cases underwent FNA only. Early on in the present study review period, FNA was considered sufficient to investigate screen-detected lesions in concordance with best practice and unit policy at the time. In the case of the large phyllodes tumour, it was clear that malignant features were difficult to identify on FNA. In the present series, final histopathological results on DOB were concordant with CB in the majority of the cases. Therefore, it is not recommended that FNA is used in assessing suspected fibroadenomas; CB is considered to be more representative and specific in identifying pertinent histopathological findings. The current policy of the unit is to perform CB as a standard practice.
514
M.A. Sala et al. / Clinical Radiology 70 (2015) 507e514
Our investigation of interval cancers and subsequent cancers in the screen-detected fibroadenomas group found that the interval cancer rate was 2.1 per 10,000 women/year over the 10 year study period, which is comparable to the BSWA interval cancer rate of 7.5 per 10,000 women/year between 2006 and 2010.16 None of the lesions in the present series, including the interval and subsequent cancers, were associated with a pre-existing screen-detected fibroadenomas. Although patients with fibroadenomas have been shown to have an increased relative risk of breast carcinoma by 1.3e2.1 for typical fibroadenomas, and 3.1e3.72 for fibroadenomas with complex features,11,17,18 the women in the present study did not show an increased risk of developing malignancy. In conclusion, in the majority of screen-detected fibroadenomas that underwent DOB in the BSWA programme, the CB diagnosis was concordant with DOB histopathology. However, to accurately diagnose and manage patients using only CB, it is important to correlate the clinical, radiological, and pathological findings, which may indicate, repeat CB or DOB. In the programme, most of the cases who underwent DOB had atypical imaging, clinical and/or pathological findings. This practice was safe as demonstrated by a very low interval cancer rate, which was comparable to the background interval cancer rate in the BSWA WA programme. Data published by Farshid et al. supports this practice where the study concluded that surgical biopsy was not indicated in well-sampled grade 3 masses with benign needle biopsy results.8 CB diagnosis of an uncomplicated typical fibroadenoma is an extremely safe diagnosis. DOB should be reserved for atypical fibroadenomas diagnosed on CB and/or lesions with suspicious clinical features.
Acknowledgements The authors thank Janette Tresham for assistance with data and records and Barbara Taylor for assistance with image manipulation and upload.
References 1. Rosen PP. Fibroepithelial neoplasms. In: Rosen PP, editor. Breast pathology. Philadelphia, PA: Lippincott-Raven; 1997. p. 143e75. 2. Markopoulos C, Kouskos E, Mantas D, et al. Fibroadenomas of the breast: is there any association with breast cancer? Eur J Gynaecol Oncol 2004;25:495e7. 3. Dabbs D. Breast pathology. 1st edn. Pittsburgh, PA: Elsevier Saunders; 2012. p. 233e48. 4. Scoyler RA, McKenzie PR, Achmed D, et al. Can phyllodes tumours of the breast be distinguished from fibroadenomas using fine needle aspiration cytology. Pathology 2001;33:437e43. 5. Tan PH, Thike AA, Tan WJ, et al. Predicting clinical behavior of breast phyllodes tumours: a nomogram based on histological criteria and surgical margins. J Clin Pathol 2012;65:69e76. 6. Tan PH, Jayabaskar T, Chuah KH, et al. Phyllodes tumours of the breast d the role of pathologic parameters. Am J Clin Pathol 2005;123:529e40. 7. Maity PP, Chatterjee S, Das RK, et al. Finding an optimum immunohistochemical feature set to distinguish benign phyllodes from fibroadenoma. Micron 2013;48:34e41. 8. Farshid G, Downey P, Gill P. Assessment of 1183 screen-detected, category 3B, circumscribed masses by cytology and core biopsy with longterm follow up data. Br J Cancer 2008;98:1182e90. 9. Noguchi S, Motomura K, Inaji H, et al. Clonal analysis of fibroadenoma and phyllodes tumor of the breast by means of polymerase chain reaction. Cancer Rrs 1993;53:4071e4. 10. Noguchi S, Yokouchi H, Aihara T, et al. Progression of phyllodes tumour demonstrated by clonal analysis. Cancer 1995;76:1779e85. 11. Sklair-Levy M, Sella T, Alweiss, et al. Incidence and management of complex fibroadenomas. AJR Am J Roentgenol 2008;190:214e8. 12. Farshid G, Walker A, Battersby G, et al. Predictors of malignancy in screen-detected breast masses with inderminant/equivocal (grade 3) imaging features. Breast 2011;20:56e61. 13. Kuijper A, Mommers EC, van der Wall E, et al. Histopathology of fibroadenoma of the breast. Am J Clin Pathol 2001;115:736e42. 14. Ozello L, Gump FE. The management of patients with carcinomas in fibroadenomatous tumors of the breast. Surg Gynecol Obstet 1985;160:99e104. 15. Gordon PB. Gagnon fibroadenomas, Lanzkowsky L. Solid breast masses diagnosed as fibroadenoma at fine-needle aspiration biopsy: acceptable rates of growth at long-term follow-up. Radiology 2003;229:233e8. 16. Breast screen WA statistical report 2006e2010. Perth: Breast Screen WA; 2012 Oct.. 17. Dupont WD, Page DL, Parl FF, et al. Long-term risk of breast cancer in women with fibroadenoma. N Eng J Med 1994;331:10e5. 18. El-wakeel H, Umplelby HC. Systematic review of fibroadenoma as a risk factor for breast cancer. Breast 2003;12:302.