Surg Endosc (2011) 25:1712–1722 DOI 10.1007/s00464-010-1465-4

REVIEWS

Mammary in the current management of disease

Sarah S. K. Tang • Dominique J. Twelves • Clare M. Isacke • Gerald P. H. Gui

Received: 4 May 2010 / Accepted: 5 November 2010 / Published online: 18 December 2010 Ó Springer Science+Business Media, LLC 2010

Abstract terms ‘‘ductoscopy’’, ‘‘duct ’’, ‘‘mammary’’, Background The majority of benign and malignant ‘‘breast,’’ and ‘‘intraductal’’ were used. lesions of the breast are thought to arise from the epithe- Results/conclusions Duct endoscopes have become lium of the terminal duct-lobular unit (TDLU). Although smaller in diameter with working channels and improved modern , ultrasound, and MRI have optical definition. Currently, the role of MD is best defined improved diagnosis, a final pathological diagnosis cur- in the management of SND facilitating targeted surgical rently relies on percutaneous methods of sampling breast excision, potentially avoiding unnecessary surgery, and lesions. The advantage of mammary ductoscopy (MD) is limiting the extent of surgical resection for benign disease. that it is possible to gain direct access to the ductal system The role of MD in breast-cancer screening and breast via the nipple. Direct visualization of the duct epithelium conservation surgery has yet to be fully defined. Few allows the operator to precisely locate intraductal lesions, prospective randomized trials exist in the literature, and enabling accurate tissue sampling and providing guidance these would be crucial to validate current opinion, not only to the surgeon during excision. The intraductal approach in the benign setting but also in breast oncologic surgery. may also have a role in screening individuals who are at high risk of breast cancer. Finally, in spontaneous nipple Keywords Breast Á Cancer Á Duct endoscopy Á discharge (SND), as biopsy instruments improve and Mammary ductoscopy Á intraductal therapeutics, such as intraductal excision and laser ablation, become a possibility, normal or benign ductoscopic findings may help minimize surgery in selec- The majority of benign and malignant breast lesions are ted patients. As MD technology is rapidly advancing, a thought to arise from the epithelium of the terminal duct- comprehensive review of current practice will be a valu- lobular unit (TDLU), the biologically active unit of the able guide for clinicians involved in the management of breast. Although modern mammography, ultrasound, and breast disease. magnetic resonance imaging (MRI) have improved diag- Methods This is a review of current ductoscopic practice nosis, a final pathologic diagnosis currently relies on per- based on an exhaustive literature search of Pubmed, Go- cutaneous methods of sampling breast lesions. The ogle Scholar, and conference proceedings. The search advantage of mammary ductoscopy (MD) is that it enables direct access to the ductal system via the nipple. Direct visualization of the duct epithelium allows the operator to locate the intraductal lesion precisely, thereby enabling S. S. K. Tang Á D. J. Twelves Á G. P. H. Gui (&) Academic Breast Unit, Royal Marsden NHS Trust, Fulham accurate tissue sampling. Visualization also aids mapping Road, London SW3 6JJ, UK of the ductal anatomy in relation to the lesion and can e-mail: [email protected] therefore be a guide to the surgeon during excision. The intraductal approach may additionally have a role in C. M. Isacke Breakthrough Breast Cancer Research Centre, Institute of screening individuals at high risk for breast cancer who Cancer Research, London, UK form a unique motivated group of women. Finally, in 123 Surg Endosc (2011) 25:1712–1722 1713 spontaneous nipple discharge (SND), as biopsy instruments and rhomboids) to be passed through the scope for locali- improve and intraductal therapeutics such as intraductal zation and to obtain specimens (microbiopsy forceps and excision and laser ablation become a possibility, normal or cytology brushes). benign ductoscopic findings may help target, locate, and The duct orifices usually are identified by nipple fluid minimize surgery for selected patients. This review inves- expressed through a combination of warming, massage, tigated current ductoscopic practice based on an exhaustive and gentle pressure. The orifice is dilated with a probe or literature search of Pubmed, Google Scholar, and confer- funnel-shaped dilator. A metal introducer, sometimes ence proceedings. The technology and techniques required called a working shaft, is inserted with the aid of a metal to perform ductoscopy and microbiopsy as well as the lumen expander and an in situ plastic obturator. Once in the limitations of MD are examined. The current role of the duct, the obturator and lumen expander are removed, intraductal approach, MD and (DL) in leaving the introducer in place to secure the duct for pas- screening, nipple discharge, and breast cancer is discussed. sage of the endoscope. When associated with a standard surgical procedure, MD often is performed with the patient under general Technology and technique anesthesia. Mammary ductoscopy also can precede stan- dard minor procedures performed with the aid of intrave- Intraluminal microendoscopic technology has improved nous sedation in addition to local anesthesia. In addition, over the past 20 years, and earlier limitations of poor MD can be performed easily as an office or out-patient optical resolution and access restriction of large-caliber diagnostic procedure with application of a local anesthetic scopes have been overcome. The technology of entering (commonly lidocaine or xylocaine) by topical gel to the large orifices such as the respiratory, gastrointestinal, and nipple or by periareolar infiltration or infusion down the urinary tracts has now evolved to entering smaller-diameter cannulated nipple duct. Instillation of a local anesthetic ducts such as the salivary gland and the breast. The into the nipple ducts after cannulation facilitates relaxation development of working channels within microendoscopes of the major duct muscle sphincters [1–3]. makes it possible to biopsy lesions and perform therapeutic An understanding of the breast’s three-dimensional procedures. anatomy is important to prevent loss of orientation and to Earlier techniques of air insufflation have been super- ensure that all possible ducts are explored and once seded by saline infusion for distension of mammary ducts explored, not reexplored. Markers using color dyes, to obtain a superior picture quality. Current scopes can be marking wires, or clips may be helpful for precise identi- flexible or rigid, with diameters ranging from 0.7 to fication of lesions during MD to ensure excision at open 1.2 mm (Fig. 1). Microendoscopes magnify tissues up to surgery and for accurate pathologic assessment [4]. More 60 times normal size to produce high-quality images. The crude ways to localize the cannulated duct can be achieved presence of a working channel allows accessories (hooks using transillumination [5] or by simply inserting a

Fig. 1 The dissembled components of a 0.9-mm LaDuscope (PolyDiagnost GmbH, Pfaffenhofen, Germany) comprising a A fiberoptic scope, B disposable two-port cannula, C cannula sheath, D shifter, and E protective metal sheath for the fiberoptic to be used during sterilization

123 1714 Surg Endosc (2011) 25:1712–1722 lacrimal probe into the recently endoscoped lesion-con- secondary or long-lasting effects of breast duct perforation taining duct. When disposable scopes are used, the sheath have been reported, although this may be a transient cause of the MD may be sutured in place and if not dislodged, of postprocedure discomfort. will enable the pathologist to identify accurately the lesion of interest to the endoscopist [3]. It also is possible to inject methylene blue dye through the endoscope to mark the Intraduct appearances target duct for surgical excision. Potential complications of MD are uncommon and General consensus exists concerning the intraductal include pain, inflammation, and infection. Occasionally, appearances of common lesions, with studies providing MD fails as a result of lumen occlusion from scarring and histologic correlation (Table 1, Fig. 2). Malignant lesions sclerosis. Access to peripheral lesions may be limited by are more likely to display hemorrhagic characteristics than the scope length (6 cm). Perforation of the duct by the benign lesions [9, 10], although papillomas can be friable scope creates a false passage into the breast parenchyma and seen to bleed at endoscopy. and usually can be recognized by transition of the visual Despite ductoscopic characteristics, it is not always image of ducts from a white shiny smooth surface to a gray possible to make a final diagnosis based on visual ragged surface (fibrous parenchyma) or to a yellow cav- appearances alone. A number of studies have evaluated ernous honeycomb (adipose tissue). Acute angulation of endoscopic clinical features compared with histologic ductal branches may predispose to duct perforation. No outcomes that give an indication of diagnostic accuracy

Table 1 Published articles describing typical intraductal appearances correlated with final histologic diagnosis Ductoscopic appearance Histologic diagnosis

Japanese Association of Mammary Polypoid Solitary or multiple Benign papillary lesions Ductoscopy and Makita et al. [6] Superficial Continuous luminal irregularity with no obvious Carcinoma Combinedelevations Carcinoma Okazaki et al. [7] Normal Smooth lustrous white/pale-yellow walls sometimes with visible capillary veins and ring-folds on duct walls Fibrous bridging Inflammatory tissue Solid nodules Benign papillary lesions (red lesions were of high vascularity) Superficial Lustrous and smooth surface with 3 different colors DCIS spreading (yellow, red, or ash-grey) lesions Slightly elevated with rough and smooth areas Shen et al. [8, 9] Circumferential Irregular fungating masses protruding into lumen DCIS obstruction Polypoid Hemorrhagic Benign papillary lesions Smooth-bordered lesions arising from sessile or narrow stalks Matsunaga et al. [10] Hemispheric or Benign papillary lesions papillary Carcinoma lesions Flat protrusions Moncrief et al. [11] Red patches DCIS Fronds 2/4 Cases of hyperplasia 2/4 Cases of nonproliferative pathology Papillomatous 36/49 Cases of papillomas lesions 5/49 cases DCIS 6/49 Cases of hyperplasia 2/49 Cases of nonproliferative pathology DCIS ductal carcinoma in situ

123 Surg Endosc (2011) 25:1712–1722 1715

Fig. 2 Images from mammary ductoscopy showing a A normal duct, B close-up view of branching ducts, C large papilloma, D flat papilloma, E calcified lesion sitting in the duct, F branching ducts with a patch of ductal carcinoma in situ (DCIS), G, DCIS and invasive disease in branching ducts, with normal epithelium in the forefront of the image, and H positive tumor margin, with surgical cavity from recent excision visible in distance

based on visual appearances alone. Moncrief et al. [11] incorrectly described as papillomas, five were ductal car- conducted a study comparing ductoscopy-guided duct cinoma in situ (DCIS), three were atypical ductal hyper- excision with conventional terminal duct excision for 117 plasia (ADH), three were hyperplasias of the typical type, women with SND. In the ductoscopy-guided excision and two had a nonproliferative pathology. Conversely, of group, if ductoscopy identified a lesion, the extent of the the 37 papillomas diagnosed pathologically, 36 (97.3%) disease was marked out on the skin by transillumination at were correctly identified by ductoscopy. The single papil- the most proximal and most distal lesions. The outer can- loma missed was in a duct with a stricture too tight to allow nula of the scope was left in place, and resection of the the scope to pass. One patient who had DCIS confirmed by diseased duct was performed. For the 59 women who final pathology had red patches seen at ductoscopy. underwent ductoscopy-guided excision, 49 lesions were Louie et al. [12] evaluated 188 women who underwent described as papillomas, but only 36 (73%) were confirmed ductoscopy-guided excisional biospy for SND. An intra- to be papillomas at final histology. Of the lesions areolar incision was made, and a standard wedge biopsy

123 1716 Surg Endosc (2011) 25:1712–1722 was performed around the tip of the scope positioned at the [14]. In one study, cytology brushes were able to collect up to most distal intraluminal defect. The final pathology was 33,000 cells per sample from pure duct epithelium, whereas DCIS for 12 patients and invasive breast cancer for 2 16% had inadequate cell yield when ductal lavage alone was patients. Ductoscopy identified intraluminal growths in 6 performed [15]. A cell-rich endoscopy specimen allows (43%) of these 14 patients. Other pathologies visualized better discrimination by the cytopathologist between likely included strictures (2 patients), obstruction (2 patients), and benign atypia (related to papilloma and hyperplasia) and wall irregularities (2 patients). The remaining two patients more sinister atypia related to a potential neoplastic process had unremarkable ductoscopies. (ADH, DCIS, and invasive cancer). The authors also found that for several patients with Most studies show that a high percentage of surgically either normal ductoscopies or visualization of a benign biopsied and histologically sampled lesions are associated pathology, final histology showed that the carcinoma was with known intraductal pathology. These findings are so adjacent to the main duct but not directly continuous with prevalent that the few cases in which an endoscopic it. In one patient, lobular carcinoma in situ was found abnormality was found but no histopathologic lesions within a papilloma. The methods used to mark lesions identified could be explained simply by a submillimeter ductoscopically vary between operators and may be a lesion missed with standard approaches to histology. This potential source of discrepancy between what the endos- has led some authors to conclude that any intraluminal copist has seen and what the pathologist samples for defect seen at ductoscopy can be considered pathologic. histology. One study found that a normal endoscopic ductal lining It may be possible to improve the ability of MD to is rarely associated with abnormal cytology and that an indicate a diagnosis based on visual appearance alone using abnormal endoscopic appearance of the ductal lining has a autofluorescence technology [13]. When tissues are illu- much higher association with abnormal cytology [1]. Using minated with short-wavelength light (380–430 nm), the intraluminal brush cytology, 36% of ducts with intralumi- absorbed energy is emitted as light at a longer wavelength nal defects showed abnormal cytology, whereas only 6% of (475–800 nm) and observed as fluorescent light of a dif- endoscopically normal ducts showed abnormal cytology. ferent color. Tissues illuminated with regular light emit a The findings showed 20% of patients to have severe and small amount of different-colored fluorescent light, which malignant atypia, 24% to have papillary and hyperplastic often is not seen because the overall illuminating white lesions, 31% to have benign cytology, 44% to have benign light is so much brighter. By applying specific filters to the infection-related cytology, and 18% to have acellular or illuminating light, the amount of fluorescent light emitted inadequate samples. can be maximized. Using observation filters, the large Some authors have combined MD with ductal lavage amount of illuminating light can be filtered out, allowing cytology to increase diagnostic sensitivity. Findings have the small amount of colored fluorescent light to be more shown cytology from ductal lavage during MD to be more easily seen. Because connective tissues and surface epi- useful than nipple squeeze or aspirate alone [2]. If the thelia have background autofluorescence, pathologic ductoscope does not reach smaller ducts, lavage fluid could lesions that grow on the surface of an epithelial layer may be obtained from the TDLU within the same duct system. stand out compared with normal tissue when viewed in this A combination of MD and ductal lavage may thus be a manner by having a different light pattern. The end result is means of optimizing direct visualization of the ductal an enhanced image that allows the operator to distinguish system with cellular and biologic assessment. potentially between benign and malignant lesions or to In a large study of 415 women with SND, Shen et al. [8] identify early premalignant lesions such as DCIS. found an overall positive predictive value of 80% with MD alone in detecting DCIS, which improved to 100% when combined with cytologic analysis of ductoscopic lavage Histologic specimens fluid. In a separate study, Shen et al. [9] concluded that because intraductal lesions in DCIS cases were localized Cytology alone cannot fully differentiate between benign more distally (3.3 cm) than their benign papilloma coun- lesions (such as papillomas and unusual type hyperplasias) terparts (2.7 cm), combining MD with lavage cytology and more sinister findings (such as DCIS and atypical ductal provided a useful diagnostic tool for DCIS in women with hyperplasia). A variety of devices have been developed to SND. Breast cancer in the presence of SND is discussed in pass down the working channel to obtain a cytologic or greater detail later. histologic specimen. Fine brushes can be used to scrape The development of intraductal biopsy is essential if intraluminal lesions under direct vision, with the resultant MD is to fulfil its diagnostic potential and enable a histo- exfoliated cells analyzed for cytology. This results in a logic diagnosis to be reached without surgical excision. greater cellular yield than obtained by lavage cytology alone Progress in the development of intraductal biopsy 123 Surg Endosc (2011) 25:1712–1722 1717 techniques has been limited by the small dimension of the direct visualization and biopsy where indicated. This working channel, which could not support the passage of approach carries no additional radiation risk even if con- biopsy instruments. ventional mammography is included and can be performed Recent developments have been described for the suc- with the patient under local anesthesia in an outpatient cessful application of intraductal (IDBB) setting [1]. Its use would be particularly feasible for high- using the novel instrument design [8, 16, 17] of a 0.7-mm risk women, who would form a unique group of highly gradient index endoscope covered by an external metal motivated individuals. Because breast cancer arises in the sheath containing a side-opening aperture near its tip. To terminal duct-lobular epithelium, detection by the intra- perform biopsy, the scope is withdrawn by a few milli- ductal approach is possible several years before the lesion meters, thereby creating a chamber within the sheath. The is clinically palpable or visible mammographically. specimen then is guided through the side aperture into the chamber under direct vision and then removed with gentle Ductal fluid in screening suction. Between 2002 and 2006, Hunerbein et al. [17] per- Ductal fluid comprises varying proportions of cells derived formed 111 ductoscopies. For 89% of nipple discharge from the duct-lobular unit, namely, columnar epithelial patients, IDBB yielded diagnostic material, with histologic cells, macrophages, foam cells, leukocytes, and cellular correlation in 100% of cases. In 41% of breast cancers, MD debris. This fluid can be acquired through nipple aspiration identified intraductal lesions, seen as red patches, micro- (achieved using either a breast suction device or direct calcifications, or ductal obstruction. The authors also found expression by breast massage and areolar squeezing) or by that patients with abnormal ductoscopic findings had a ductal lavage. Ductal lavage can be collected using lavage significantly higher risk of an extensive intraductal com- catheters to cannulate specific ductal systems. Ductoscopy ponent to their breast cancer (71% vs 16%). specimens also may be collected at MD from saline used for insufflation or after the endoscopic procedure. Most studies have concentrated on ductal fluid cytology Screening for screening high-risk ducts, in which nipple aspiration and ductal lavage were found to be feasible. Cytologic Breast cancer remains the most common cancer in the yield can be further improved in combination with duc- United Kingdom despite the fact that it is rare in men. In toscopy, in which sampling under vision can be used to 2005, in the United Kingdom, 45,947 new cases of breast provide samples of relatively pure ductal epithelial content cancer were diagnosed. A woman in the United Kingdom (in excess of 91% purity) [23]. Other studies have found has a lifetime risk of 1 in 9.5 of developing breast cancer that ductal lavage of cancerous and high-risk pro- [18]. duced a low cytology yield and therefore may not be of Mammography currently is the gold standard for breast value for screening [24–26]. cancer screening, but it has a poor positive predictive value With the recognized limitations of ductal fluid cytol- of only 25%. Mammography has reduced sensitivity and ogy, other biomarkers of breast cancer are being investi- specificity for smaller lesions in women younger than gated (Table 2). Much work has been done in profiling 50 years and women with a family history of breast cancer protein differences in lavage fluid from normal and can- (who generally are younger) [19]. cerous breasts, with limited success. More recently, the Regular mammography carries a cumulative risk due to emergence of epigenetics as a major player in cancer cell radiation [20]. Dose and age at exposure are the two most molecular biology has resulted in a search for epigenetic important determinants of this risk, and hence the risk is changes in breast cancer. In particular, hypermethylation theoretically greater for younger women. In addition, those of the promoter regions of tumor suppressor genes can be who have an inherited predisposition to cancer may be detected by methylation-specific polymerase chain reac- more susceptible to environmental carcinogens such as tion (PCR). Because ductal lavage collects breast fluid in radiation [21–23]. direct contact with ductal epithelium, it offers a promising Consequently, a need exists to improve additional medium for biomarker studies compared with other screening methods for younger patients and those at higher physiologic media such as serum or tissue implying risk for breast cancer. Selective use of the intraductal that neoplastic change has already become established. approach as a screening tool for these groups of moderate- Because tumor suppressor gene hypermethylation is likely to high-risk women may be feasible. This approach offers to represent a field change in the ductal epithelial cells of the unique opportunity to sample the breast duct fluid in high-risk individuals [32], evaluation of ductal fluid could direct contact with epithelial cells that potentially undergo overcome limitations of the sensitivity of MD and ductal early malignant change, with the additional benefit of lavage as a screening method. 123 1718 Surg Endosc (2011) 25:1712–1722

Table 2 Summary of proteomic and epigenetic biomarker identification using ductal fluid Biomarker type Benefits Technology Pros/cons of technology Potential markers identified

Proteomics/protein Protein activity determines Protein fractionation by Slow throughput Gel free Human neutrophil expression cellular function. Differences in 2-D PAGE peptides 1–3 [27] protein expression in healthy vs Protein chip arrays High throughput Gel free Alpha 2 HS- cancer cells may be more ICAT High throughput Requires glycoprotein, lipophilin relevant than DNA/RNA B, betaglobulin, Protein identification smaller amount of protein differences, which are not hemopexin, vitamin D- by mass spectrometry Naturally occurring forms of always expressed binding protein and protein peptide and proteins Recent advances in technology precursor [28] fingerprinting e.g. comparatively profiled have allowed high-throughput SELDI-TOF MS LC- analysis of thousands of Not quantitative MS proteins Used following ICAT Quantitative Ductal fluid is protein rich Only proteins containing cysteine are analyzed Tumor-suppressor Tumor-suppressor gene- Methylation-specific Quantitative analysis superior to RASSF-1A, TWIST1, gene-promoter promoter-region PCR e.g. QM-MSP conventional MS-PCR HIN-1 [29] region- hypermethylation may because gene methylation also P16, RAR-b, RASSF-1a hypermethylation represent field change in high- can be present but in smaller [30] using candidate risk individuals; therefore, DL degrees in benign or adjacent predictive primers from any available duct should normal tissue. QM-MSP has provide sufficient screening been shown to double material (Euhus et al. [32]) detection of tumor cells in Subjective breast ductal fluid [31] Choice of candidate primers for hypermethylated genes based on previous experiments and authors’ choice Inefficient method of biomarker discovery DL ductal lavage, PAGE polyacrylamide gel electophoresis, ICAT isotope-coded affinity tagging, SELDI-TOF MS, surface-enhanced laser desorption ionization time of flight mass spectrometry, LC-MS liquid chromatography mass spectrometry, PCR polymerase chain reaction, RASSF-1A RAt sarcoma association domain family 1A tumor suppressor gene, TWIST twist homolog 1 (Drosophila) gene, HIN-1 high in normal 1 gene, QM-MSP quantitative multiplex methylation-specific PCR, RAR-b retinoic acid receptor-beta gene, MS-PCR methylation specific polymerase chain reaction

Limitations of mammary ductoscopy of ducts may not necessarily be associated with the affected cancer segment. Every surgical technique has its limitations, and those of The length and outer diameter of the scope may limit MD should be recognized. Unlike other applications for access to the TDLU, thereby missing more distal lesions. microendoscopy, such as the salivary gland system in Moncrief et al. [11] reported that 37.3% of patients which a sizeable single duct orifice exists, the breast ductal undergoing ductoscopy had lesions more than 5 cm from system opens via multiple tiny orifices at the nipple surface the nipple. Kapenhas-Valdes et al. [3] found that lesions or just below it. The number of ducts and duct orifices were visualized from depths of 3 to 8 cm (average, varies, with one anatomy study demonstrating 29 ducts 4.4 cm). Dooley [5] reported that the majority of lesions arising from 15 orifices [33]. This study also found that were found within 3 cm from the nipple, although some some ducts branch within the nipple itself and may not be were identified as deep as 7.5 cm from nipple. Shen et al. readily accessed by lavage cannulas or the endoscope. [9] found that intraductal lesions in their DCIS cases were Although it would be possible to examine more ducts, located more distally (average, 3.3 cm) than their benign most studies on duct endoscopy report detailed visualiza- papilloma counterparts (average, 2.7 cm). At histologic tion of one or two ducts per breasts, thereby potentially examination, 18% had papillary lesions peripherally that missing a significant number of other ducts. Although this were not seen at ductoscopy. The authors questioned may not be as relevant in pathologic nipple discharge when whether failure to identify these lesions was because the MD is directed at a single discharging duct, most breast distal ducts were not accessed by the ductoscope or cancers are not associated with discharge, and cannulation because the papillomas may have exfoliated off during the

123 Surg Endosc (2011) 25:1712–1722 1719 procedure. In a study of galactograms, Hou et al. [34] 36]. Recognized causes of pathologic nipple discharge found that 70% patients with carcinomas presenting with include duct ectasia, fibrocystic breast changes, intraductal nipple discharge had lesions farther than 2 cm from the papillomas, and carcinoma. The most common cause of nipple. bloody nipple discharge is intraductal papilloma (35–48%). Breast lesions farther than 6 cm from the nipple usually This discharge usually is spontaneous and elicited from a are beyond the reach of duct endoscopes. Conventional single duct. In their study of nipple discharge and ductos- surgery by major duct excision or also is copy, Okazaki et al. [36] found that up to 70% cases of likely to miss these lesions because only the proximal nipple discharge result from papillomas, followed by ade- ducts, usually closer than 3 cm, are removed. nomatous or papillary epithelial proliferations (14%). In a study of patients with spontaneous hemoccult Traditional investigations such as mammography and positive pathologic nipple discharge, 26 of 27 patients had cytology may not accurately diagnose the cause of dis- positive ductoscopic findings, and 70% had multiple charge, but the presence of an associated underlying mass lesions [5]. The authors concluded that these findings increases the risk of cancer from 2.9 to 12.5%. Although no suggest an underestimation of the true causes of bleeding, longer used routinely, galactograms may provide a visual which may arise from deeper lesions (DCIS or ADH) and image of the location of pathology, but this is an indirect would not be identified by excising the proximal 2–3 cm of guide to the surgeon and has not been shown to alter sur- the offending duct. In this study, DCIS was identified in gical management. two patients who also had papillomas more proximally in In contrast, direct visualization and localization of the same ductal system. In 33% of the cases, ADH was pathology using MD guides surgery to the affected duct, found coexisting with papillomas. resulting in a more targeted excision. Furthermore, visu- In a feasibility study combining MD with DL, Danforth alization of ‘‘remaining’’ ducts can rule out coexisting et al. [23] found that they were able to examine ducts to pathology. Unnecessary dissection of otherwise normal 4 cm or more in most subjects, but further advancement ducts is of particular significance in younger women who was limited by narrow lumen or fibrous bridging across the may wish to breastfeed in the future. duct. It is relatively common to have a situation in which A separate study of 54 women with spontaneous per- cytology obtained by SND, nipple aspiration, or massage is sistent unilateral uniductal nipple discharge by Denewer abnormal but all imaging including MRI is normal. In this et al. [2], no palpable lesions and negative mammography situation, MD may be a useful additional tool to guide yielded the following results. Of the 54 women, 13 had further management. Normal MD in this setting would normal ductoscopy, but these included two false-negatives reinforce a policy for observation, whereas an abnormality in which the final histology demonstrated the presence of shown at MD would lead to further action. papillomas missed at ductoscopy. The overall diagnostic Spontaneous nipple discharge is ideally suited to duc- sensitivity was therefore 85%. The authors speculated that toscopic investigation because it enables accurate identifi- the operator’s learning curve and experience may have cation of the ductal orifice containing the causative accounted for the limited sensitivity, and in particular, pathology. Coexisting duct ectasia or dilation may further some papillomas may have been present in small ducts aid scope passage. As a result, the majority of published beyond the fourth branch and therefore beyond the reach of data on ductoscopy has reported investigation of nipple the scope. As longer, thinner, and more flexible scopes are discharge, with most authors describing high cannulation developed, false-negative rates should fall. Interestingly, rates and diagnostic yields [2, 5]. 19.5% had positive MD but no papilloma found at histol- Conventional surgery for pathologic nipple discharge ogy. Denewer et al. [2] speculated that small and fragile has consisted of major duct excision or microdochectomy. papillomas were being extruded or exfoliated during the Major duct excision often is an undirected procedure in procedure. It also is possible that small lesions were missed which a disc or wedge of retroareolar tissue is excised that by the pathologist, and it may well be that the false-positive includes a considerable amount of tissue that might not rate will fall when marking systems become more reliable need resection for a benign process. Microdochectomy for small lesions. usually is lacrimal probe guided, targeting the distal offending duct at the level of the nipple orifice but not offering any guidance for the surgeon to the location of Spontaneous nipple discharge more proximally sited lesions, thus leaving to chance whether the surgical specimen contains the cause of the Spontaneous nipple discharge is the third most common discharge. The greater the amount of tissue resected, the presenting symptom of breast disease (after lumps and higher the risk of surgical complications including altered mastalgia), representing 3–8% of breast symptoms [35, sensation, hemoseromas, poor aesthetic outcome from 123 1720 Surg Endosc (2011) 25:1712–1722 volume deficit, and nipple ischemia. In this era with its 188 patients had a final diagnosis of DCIS (12/14) or possibilities of targeting small lesions in the breast for invasive breast cancer with DCIS (2/14). In 8 of these 14 needle biopsy by stereotaxis, ultrasound, and MRI and cases, duct wall irregularities or intraluminal growths were locating them when necessary for excision, it seems visualized by MD [12]. unthinkable that surgeons would perform undirected sur- A large study of 415 women with SND had 11 cases of gery on presumed lesions in ducts that in the many cases DCIS. All these DCIS cases were detected by either a show normal conventional imaging. Duct endoscopy combination of MD and biopsy or MD and ductal lavage without surgery is unlikely to be adequate for troublesome cytology [8]. Six of these patients had normal physical copious spontaneous nipple discharge because transection examination results and negative mammograms. The of the major ducts disconnects these ducts from the nipple authors concluded that although nipple discharge is an orifices to stop the symptom. Mammary ductoscopy of unusual way for DCIS to present, MD plus ductal lavage otherwise normal ducts or duct ectasia alone provides cytology is a useful technique for diagnosing DCIS before reassurance that occult atypical or malignant changes definitive surgery. within the major ducts will remain undetected, thus enabling major duct transaction with minimal tissue resection. Treatment of pathologic nipple discharge with MD

In the study by Matsunaga et al. [16], 107 of 295 patients Spontaneous nipple discharge and breast cancer with SND were found to have papillomas. The authors performed IDBB for these patients with a resultant diag- Breast carcinoma may occasionally present with blood- nostic sensitivity of 76.2% and a specificity of 100%. Of stained nipple discharge but also is associated with non- these patients, 70 were followed up more than 3 years. bloody discharge. The reported incidence of cancer among During the follow-up period, 36 stopped having nipple women with bloody nipple discharge ranges from 4 to 31% discharge after intraduct biopsy, and a further 13 stopped [37]. Nipple discharge may be associated with carcinoma having discharge after a second intraduct procedure (70%). despite a normal mammogram. A white, yellow, green, or Nipple discharge continued for five patients in whom no brown-black discharge usually points to benign pathology, intraductal masses were found. The authors were therefore whereas clear, serous, serosanguinous, bloody discharge is able to conclude that they had a total therapeutic efficacy of more suspicious of malignant pathology. 77.6% (54/70 patients) for intraductal papillomas. Matsunaga et al. [10] performed MD for 315 patients Interestingly, IDBB was found to be less effective at with SND. There were 47 cases of carcinoma, 38 of which collecting material from carcinomas. In 27 patients, 30 were seen at ductoscopy (81%). Mammary ductoscopy ductal carcinomas were found. Intraductal breast biopsy successfully identified 115 of 119 papilloma cases (96.6%). was performed 36 times among these 27 patients and However, in a separate study by the same authors, IDBB yielded 21 diagnostic specimens (58.3%). Breast cancer was found to be significantly less effective at collecting was diagnosed for 9 of these 21 specimens, suspected tissue from papillomas [16]. carcinoma (atypical papillary neoplasia) in 7 cases, and In the study by Denewer et al. [2], MD detected lesions intraductal papilloma in the remaining 5 cases. in 41 (76%) of 54 patients. For 11 cases in which intra- The authors speculated that carcinomas generally were ductal thickening was found, the final histology showed located more peripherally in the TDLUs and exhibited hyperplasia in five cases, papilloma in two cases, ADH in weaker tissue cohesiveness than papillomas. Ductal carci- two cases, DCIS in one case, and invasive cancer in one nomas also were difficult to access ductoscopically as only case. The prospective review by Kapenhas-Valdes et al. [3] the tip was visible because lesions tend to extend beyond of 93 patients undergoing ductoscopy to evaluate nipple TDLU. The authors suggested that surgical excision or close discharge found that 67 patients had abnormal ductoscopic follow-up assessment should be recommended for patients findings and went on to have ductoscopically guided duct with insufficient diagnostic material or atypical papillary excision of 77 ducts. Visible pathologies included duct lesions diagnosed at intraductal breast biopsy. Among 89 ectasia and hyperplasia in 17 ducts, papilloma in 48 ducts, patients who had negative MD results, 3 had cancer detected ADH in 6 ducts, and cancer in 6 ducts. The cancers pre- during a minimum follow-up period of 3 years. In all three sented with luminal obstruction, streaking, and irregularity patients, nipple discharge ceased after ductoscopy and breast of the ductal walls together with nodularity and debris, cancer was found to be located in an area different from that luminal mass, and blood-tinged tissue. previously identified on . In a retrospective review of ductoscopy-assisted exci- Bender at al. [38] performed 22 endoscopic papillom- sion biopsy for spontaneous nipple discharge, 14 (7%) of ectomies among SND patients found to have solitary 123 Surg Endosc (2011) 25:1712–1722 1721 papillomas at MD and 5 ductoscopy-assisted microdoch- pathologic nipple discharge, and 14 patients were found to ectomies among patients found to have multiple papillo- have invasive or in situ disease at the final histologic mas. The discharges in all but one patient stopped during a analysis [12]. For 6 (43%) of these 14 patients, ductoscopy mean follow-up period of 11.5 months, giving endoscopic showed intraluminal tumor growths. The margin status for papillomectomy a therapeutic efficacy of 95.4% (21/22). 10 of these lesions was assessed, and 7 were found to be positive. Two of the margins came to within 1 mm of the edge, and only one lesion had negative margins. The Breast conservation surgery authors therefore concluded that ductoscopy was not an aid in estimating the required extent of resection in cancer The third area in which MD has been studied is breast cases [12]. conservation surgery for malignant disease, both for inva- sive cancer and for in situ disease. The principles of breast conservation surgery are to achieve a balance between a Conclusion good cosmetic outcome and clear margins (important for reducing potential recurrence). Adjuvant radiotherapy has The role of mammary ductoscopy has evolved with tech- become a routine component of breast-conserving treat- nological advances to overcome limitations imposed by ment, with significant survival benefit where breast cancers older technology. Duct endoscopes have become smaller in are completely excised. diameter, with working channels and improved optical Current strategies to achieve adequate margins include definition. The intraductal approach may have a clinical ultrasound marking, needle localization, use of frozen sec- role to play in screening selected patient groups with tions, and re-excision of margins if necessary. Mammary moderate to high breast cancer risk, as determined from ductoscopy could potentially be a useful tool to help the family history or a breast cancer predisposition gene surgeon define the boundaries of the lesion and may be mutation. The ability to perform intraductal biopsy and the particularly useful in excising DCIS. However, few pub- developments in autofluorescence techniques may offer lished studies have specifically investigate this role for MD. significant improvements in diagnostic capability. In an early study, Dooley et al. [39] examined 55 women The role of MD currently is best defined in the man- with a formal diagnosis of ductal hyperplasia, DCIS, or agement of pathologic nipple discharge as facilitating tar- invasive breast cancer, as shown by screening and percu- geted surgical excision, potentially avoiding unnecessary taneous biopsy histology. A target orifice was identified by surgery, and limiting the extent of surgical resection for the presentation of discharge after massage in the affected benign disease. The benefit of MD as an adjunct to breast quadrant of the breast, and successful cannulation was conservation surgery for cancer, particularly to reduce achieved for all but eight patients. The endoscopically re-excision rates for positive margins, remains to be identified lesions were catalogued and excised using the tip defined. The potential for intraductal therapeutic endolu- of the endoscope to direct the surgeon. A target lesion was minal procedures and delivery of topical chemotherapy is a confirmed for 41 patients (75%), and in 21 cases (38%), particularly exciting future direction in microendoscopic ductoscopy showed unexpected extensive intraluminal dis- technology. Few prospective randomized trials exist in the ease necessitating a more extensive resection than originally literature, and these are crucially needed to validate current planned. Although cancers may obstruct the duct lumen and opinion, not only concerning the benign setting, but also restrict passage of the scope, the authors found that the need concerning breast oncologic surgery. for excision was reduced on the side of the specimen where ducts were visualized. Ductoscopically guided excisions did Disclosures Sarah S. K. Tang, Dominique J. Twelves, Clare M. not result in any positive margins within 5 mm of the nipple- Isacke, and Gerald P. H. Gui have no conflicts of interest or financial ties to disclose. side boundary of the excised tissue. In a subsequent publication, Dooley et al. [40] reported successful MD for 150 of 201 patients (74.6) and found References additional lesions outside the planned excision site in 83 cases (41%). They therefore concluded that MD identified 1. Dooley W, Francescatti D, Clark L, Webber G (2004) Office- more disease than conventional preoperative investigations based breast ductoscopy for diagnosis. Am J Surg 188:415–418 (mammography and ultrasound) and in these cases 2. Denewer A, El-Etribi K, Nada N, El-Metwally M (2008) The role decreased the potential positive margin rate from 23.5 to and limitations of mammary ductoscopy in management of pathologic nipple discharge. Breast J 14:442–449 5%. 3. Kapenhas-Valdes E, Feldman SM, Cohen J-M, Boolbol SK In a more recent retrospective review, 188 patients (2008) Mammary ductoscopy for evaluation of nipple discharge. underwent ductoscopically guided excision biopsy for Ann Surg Oncol 15:2720–2727 123 1722 Surg Endosc (2011) 25:1712–1722

4. Escobar PF, Baynes D, Crowe JP (2004) Ductoscopy-assisted lavage and ductal endoscopy for the evaluation of the high-risk microdochectomy. J Fertil Women Med 49:222–224 breast: a feasibility study. J Surg Oncol 94:555–564 5. Dooley WC (2002) Routine operative breast endoscopy for 24. Wood ME, Stanley MA, Crocker AM, Kingsley FS, Leiman G bloody nipple discharge. Ann Surg Oncol 9:920–923 (2009) Ductal lavage of cancerous and unaffected breasts: pro- 6. Makita M, Akiyama F, Gomi N, Ikenaga M, Yoshimoto M, Kasumi cedure success rate and cancer detection. Acta Cytol 53:410–415 F, Sakamoto G (2002) Endoscopic classification of intraductal 25. Carruthers CD, Chapleskie LA, Flynn MB, Frazier TG (2007) lesions and histological diagnosis. Breast Cancer 9:220–225 The use of ductal lavage as a screening tool in women at high risk 7. Okazaki A, Okazaki M, Asaishi K, Satoh H, Watanabe Y, Mi- for developing breast carcinoma. Am J Surg 194:463–466 kami T, Toda K, Okazaki Y, Nabeta K, Hirata K, Narimatsu E 26. Khan SA, Lankes HA, Patil DB et al (2009) Ductal lavage is an (1991) Fiberoptic ductoscopy of the breast: a new diagnostic inefficient method of biomarker measurement in high-risk procedure for nipple discharge. Jpn J Clin Oncol 21:188–193 women. Cancer Prev Res 2:265–273 8. Shen KW, Wu J, Lu JS, Han QX, Shen ZZ, Nguyen M, Barsky 27. Li J, Zhao J, Xiodong Y, Lange J, Kuere H, Krishnamurthy S, SH, Shao ZM (2001) Fiberoptic ductoscopy for breast cancer Schilling E, Khan SA, Sukuma S, Chan DW (2005) Identification patients with nipple discharge. Surg Endosc 15:1340–1345 of biomarkers for breast cancer in nipple aspiration and ductal 9. Shen KW, Wu K, Lu JS, Han QX, Shen ZZ, Nguyen M, Shao lavage fluid. Clin Cancer Res 11:8312–8320 ZM, Barsky SH (2000) Fiberoptic ductoscopy for patients with 28. Pawlik TM, Hawke DH, Liu Y, Krishnamurthy S, Fritsche H, nipple discharge. Cancer 89:1512–1519 Hung KK, Kuerer HM (2006) Proteomic analysis of nipple 10. Matsunaga T, Ohta D, Misak T, Hosokawa K, Fujii M, Kaise H, aspirate fluid from women with early-stage breast cancer using Kusama M, Koyanagi Y (2001) Mammary ductoscopy for diag- isotope-coded affinity tags and tandem mass spectrometry reveals nosis and treatment of intraductal lesions of the breast. Breast differential expression of vitamin D binding protein. BMC Can- Cancer 8:213–221 cer 6:68 11. Moncrief RM, Nayar R, Diaz L, Staradub V, Morrow M, Khan S 29. Fackler MJ, Rivers A, Teo WW, Mangat A, Taylor E, Zhang Z, (2005) A Comparison of ductoscopy-guided and conventional Goodman S, Argani P, Nayar R, Susnik B, Sukuma S, Khan S surgical excision in women with spontaneous nipple discharge. (2009) Hypermethylated genes as biomarkers of cancer in women Ann Surg 241:575–581 with pathological nipple discharge. Clin Cancer Res 15:3802– 12. Louie LD, Crowe JP, Dawson AE, Lee K, Baynes DL, Dowdy A, 3811 Kim JA (2006) Indentification of breast cancer in patients with 30. Zhu W, Qin W, Hewett JE, Sauter ER (2010) Quantitative pathologic nipple discharge: does ductoscopy predict malig- evaluation of DNA hypermethylation in malignant and benign nancy? Am J Surg 192:530–533 breast tissue and fluids. Int J Cancer 126:474–482 13. Jacobs VR, Paepke S, Schaaf H, Weber B-C, Kiechle-Bahat M 31. Fackler MJ, Malone J, Zhang Z, Schilling E, Garrett-Mayer E, (2007) Autofluorescence ductoscopy: a new imaging technique Swift-Scanlan T, Lange J, Nayar R, Davidson NE, Khan SA, for intraductal breast endoscopy. Clin Breast Cancer 8:619–623 Sukumar S (2006) Quantitative multiplex methylation-specific 14. Khan SA, Baird C, Staradub VL, Morrow M (2002) Ductal PCR analysis doubles detection of tumor cells in ductal fluid. Clin lavage and ductoscopy: the opportunities and the limitations. Clin Cancer Res 12:3306–3310 Breast Cancer 3:185–191 32. Euhus DM, Bu D, Ashfaq R, Xie X-J, Beitch AM, Lewis CM 15. Johnson-Maddux A, Ashfaq R, Cler L, Naftalis E, Leitch Am, (2006) Atypia and DNA methylation in nipple duct lavage in Hoover S, Euhus DM (2005) Reproducibility of cytologic atypia relation to predicted cancer risk. Can Epid Biom Prev 16:1812– in repeat nipple duct lavage. Cancer 103:1129–1136 1821 16. Matsunaga T, Kawakami R, Namba K, Fujii M (2004) Intraductal 33. Rusby JE, Brachtel EF, Michaelson JS, Koerner FC, Smith BL biopsy for diagnosis and treatment of intraductal lesions of the (2007) Breast duct anatomy in the human nipple: three-dimen- breast. Cancer 39:863 sional patterns and clinical applications. Br Cancer Res Treat 17. Hunerbein M, Dubowy A, Raubach M, Gebauer B, Topalidis T, 106:171–179 Schlag P (2007) Gradient index ductoscopy and intraductal 34. Hou MF, Huang TJ, Liu GC (2001) The diagnostic value of biopsy of intraductal breast lesions. Am J Surg 194:511–514 galactography in patients with nipple discharge. Clin Imaging 18. Office for National Statistics (2008) Cancer Statistics registra- 25:75–81 tions: registrations of cancer diagnosed in 2005, England. Series 35. Baitchez G, Gortchev G, Todorova A, Dikov D, Stancheva N, MB1, no. 36, National Statistics, London Daskalova I (2003) Intraductal aspiration cytology and galac- 19. Kerlikowske K, Grady D, Barclay J, Sickles EA, Ernster V tography for nipple discharge. Int Surg 88:83–86 (1996) Effect of age, breast density, and family history on the 36. Okazaki A, Hirata K, Okazaki M, Svane G, Azavedo E (1999) sensitivity of first screening mammography. JAMA 276:33–38 Nipple discharge disorders: current diagnostic management and 20. Lucassen A, Watson E, Eccles D (2001) Evidence-based case the role of fibre-ductoscopy. Eur Radiol 9:583–590 report: advice about mammography for a young woman with a 37. Devitt JE (1985) Management of nipple discharge by clinical family history of breast cancer. BMJ 322:1040–1042 finding. Am J Surg 149:789–792 21. Mettler FA, Upton AC, Kelsey CA, Ashby RN, Rosenberg RD, 38. Bender O, Balci F, Yuney E, Akbulut H (2009) Scarless endo- Linver MN (1996) Benefits versus risks from mammography: a scopic papillectomy of the breast. Onkologie 32:94–98 critical reassessment. Cancer 77:903–909 39. Dooley WC (2000) Endoscopic visualization of breast tumors. 22. Feig SA, Hendrick RE (1997) Radiation risk from screening JAMA 284:1518 mammography of women aged 40–49 years. J Natl Cancer Inst 40. Dooley WC (2003) Routine operative breast endoscopy during Monogr 22:119–124 . Ann Surg Onc 10:38–42 23. Danforth DN, Abati A, Filie A, Prindiville SA, Palmieri D, Simon R, Reid T, Steeg PA (2006) Combined breast ductal

123