10.1245/ASO.2005.04.037
Annals of Surgical Oncology 12:689-696 (2005)
© 2005 Society of Surgical Oncology
Ductal Lavage Findings in Women With Mammographic Microcalcifications Undergoing Biopsy
Seema A. Khan, MD1,
Judy A. Wolfman, MD2,
Lewis Segal, MD2,
Stephanie Benjamin, MD2,
Ritu Nayar, MD3,
Elizabeth L. Wiley, MD3,
Michele Bryk, PA-C1 and
Monica Morrow, MD1
1 Lynn Sage Comprehensive Breast Center, Feinberg School of Medicine of Northwestern University, 675 North St. Clair Street, Galter 13-174, Chicago, Illinois 60614
2 Department of Radiology, Feinberg School of Medicine of Northwestern University, Chicago, Illinois 60614
3 Department of Pathology, Feinberg School of Medicine of Northwestern University, Chicago, Illinois 60614
Correspondence: Address correspondence and reprint requests to: Seema A. Khan, MD; E-mail: skhan{at}nmh.org
 |
ABSTRACT
|
|---|
Background: We designed a prospective study to assess the likelihood that early lesions presenting as mammographic calcifications could be accessed for cytological diagnosis by ductal lavage (DL).
Methods: Consenting women with calcifications (Breast Imaging Reporting and Data System 4 or 5) underwent DL of fluid-yielding ducts (FYDs) before stereotactic core or excisional biopsy. The DL catheter was used to inject .2 to 1 mL of Isovue 300 into the duct to determine whether the FYD corresponded to the duct containing calcifications (designated overlap). Additional FYDs were injected, if possible, until overlap was identified. DL cytology was compared with histology.
Results: Twenty women were enrolled (mean age, 54.2 years); the mean size of the calcification-bearing area was 190 mm2. The histological findings were as follows: 1 invasive cancer, 9 ductal carcinomas-in-situ (DCIS), 5 atypical hyperplasias, and 5 usual hyperplasias or fibrocystic changes. Four women had no FYD. In 15 women who underwent DL and ductography, overlap of dye and calcifications was seen in 4 (27%): 1 fibrocystic change, 1 hyperplasia, 1 atypical hyperplasia (cytological diagnosis mildly atypical), and 1 DCIS (cytological diagnosis benign). Of the remaining 8 DCIS lesions, 4 had no nipple aspiration fluid, 1 showed extravasation, and 3 were lavaged but the duct did not overlap.
Conclusions: These results are consistent with data from women undergoing mastectomy for larger invasive cancer and DCIS and show that cancer-containing ducts do not yield nipple fluid in most cases.
Key Words: Breast cancer • Detection • Ductal lavage • Microcalcifications • Cytology
|
INTRODUCTION
|
|---|
Despite advances in the early detection of breast cancer, a subset of breast tumors remain troublingly resistant to detection with mammographic screening, and even among regularly screened women, approximately 65% are diagnosed with invasive cancer and one third have axillary node metastases. This is particularly true for young, high-risk women with radiologically dense breasts.1,2 Ductal lavage (DL) is a method of obtaining breast epithelial cells from asymptomatic women by saline lavage of duct orifices, which yield fluid after breast massage and suction/aspiration at the nipple. The results of the initial multicenter study3 on DL were reported in 2001 and demonstrated that this is a feasible, well-tolerated, minimally invasive method of obtaining cytologically assessable samples from some women. Among the 500 high-risk women who entered the study, 299 (59.8%) yielded lavage samples sufficient for cytological diagnosis, 11 underwent surgical resection on the basis of atypical lavage cytology, and 4 received a diagnosis of ductal carcinoma-in-situ (DCIS). This diagnostic yield of almost 1% in the absence of palpable or mammographic findings led to significant enthusiasm about the potential of DL as an early detection tool in high-risk women. However, three studies in women undergoing DL before surgical excision of known lesions suggest that in the presence of a known cancer, DL has low sensitivity for the detection of malignancy.4,5 The failure of DL occurs at two levels: (1) diseased ducts do not produce fluid and are therefore not lavaged, and (2) lavaged ducts produce benign or mildly atypical cytology when in fact in situ invasive cancer is present. Although the low sensitivity seen thus far could be improved by the inclusion of mildly atypical cytology as a positive test, this would introduce the problem of low specificity without increased accuracy.5
These studies of presurgical lavage have been performed in women with already diagnosed and often palpable invasive cancer. However, it is likely that if DL is to detect occult neoplastic lesions, it will be most successful in the setting of early disease, such as DCIS. To test this hypothesis, we performed a pilot study of prebiopsy DL in women with suspicious, mammographically detected microcalcifications to relate duct-specific cytological findings to biopsy histology.
|
METHODS
|
|---|
Women with mammographically detected calcifications (Breast Imaging Reporting and Data System 4 or higher) were offered participation on an institutional review board–approved protocol. Consenting subjects underwent DL before the biopsy. Exclusion criteria included pregnancy or lactation within the past 12 months, prior radiation to the breast, chemotherapy in the past 6 months, a history of periareolar surgery, any use of tamoxifen, or active infection in the breast to be studied. Consenting women were instructed to drink fluids the night before the procedure and to scrub the nipple with a washcloth on the morning of the biopsy.
Ductal Lavage
Subjects were given a warm beanbag to hold over the study breast upon arrival at the Breast Center, where all DL procedures were performed in an office setting adjacent to the radiology suite. The breast was massaged, the nipple was dekeratinized with an abrasive scrub, and aspiration of nipple fluid and cannulation of fluid-yielding ducts (FYDs) was performed with the Cytyc aspirator and catheter (Cytyc Corporation, Boxborough, MA), followed by lavage of the cannulated duct with lactated Ringer’s solution. The catheter was then taped in place in a sterile fashion, the patient was moved to the radiology suite, and a ductogram was performed through the same catheter. If there was more than one FYD during the initial half of the study, we started with the duct least likely to overlap the area of calcifications. For example, if the calcifications were in the upper outer quadrant and FYDs were located in both the upper outer and the lower inner quadrants, we first lavaged the lower inner quadrant (less likely) duct and then lavaged the upper outer quadrant (more likely) duct, left the catheter in place in the upper outer quadrant duct, and completed the ductogram. If the ductogram did not overlap the calcifications on both views, then the lower inner quadrant duct was cannulated, and the process was repeated. The protocol was modified approximately halfway through the study, when we encountered a patient with three FYDs. By the time we were ready to cannulate the third (and most likely to overlap) duct, there was some swelling of the nipple, and cannulation of the third duct could not be accomplished. We therefore changed the protocol so that the most likely duct was cannulated first, the catheter was left in place for the ductogram, and the other FYDs (if present) were cannulated only if the first did not overlap the calcifications. The lavage effluent was placed in Cytolyte and processed by using the ThinPrep 2000 processor (both from Cytyc), processed for cytological examination, and reviewed by the cytopathologist (R.N.).
Radiological Evaluation
Imaging was performed by full-field digital mammography (Senographe 2000D; GE Medical Systems, Milwaukee, WI). Orthogonal mammographic views were obtained immediately before the first contrast injection for documentation of the location of the calcifications before duct opacification. An initial injection of .1 to .2 mL of Isovue 300 (Bristol-Myers Squibb, New York, NY) was performed through the indwelling catheter. If no extravasation was identified or if incomplete opacification of the ductal system was seen in the area of the calcifications, then additional contrast was administered. Digital image postprocessing and electronic magnification at the acquisition station allowed for immediate imaging and evaluation. Additional detailed image analysis was performed at the review workstation.
Stereotactic-guided needle core biopsy was performed by using a prone table stereotactic unit (LoRad, Danbury, CT). Multiple core samples were obtained with an 11-gauge vacuum-assisted biopsy probe (Mammotome; Biopsys/Ethicon-Endo Surgery, Cincinnati, OH). Specimen radiographs confirmed adequate sampling of the microcalcifications. All women with a core biopsy diagnosis of DCIS, invasive cancer, or atypical ductal hyperplasia underwent subsequent preoperative needle localization for surgical excision. In two cases, the microcalcifications were not amenable to stereotactic-guided core biopsy, and preoperative needle localization for excisional biopsy was performed after ductography as described previously.
Cytological Evaluation
The criteria used to classify DL specimens were similar to those established at the National Cancer Institute consensus conference6 in 1997 for interpretation of breast fine-needle aspiration samples and used in the multicenter DL study.3 The following diagnostic categories were used:
- Inadequate cellular material for diagnosis (ICMD) included specimens that were acellular and those that contained fewer than 10 epithelial cells, with or without histiocytes (scant, <100 epithelial cells; moderate, 100–10,000 epithelial cells; abundant, >10,000 epithelial cells). Although the presence of histiocytes was used as a surrogate marker of successful DL, histiocytes were not included in the cellularity/adequacy assessment.
- Benign samples showed no evidence of architectural or nuclear abnormalities.
- Mild atypia showed minimal nuclear enlargement and anisonucleosis, with or without architectural abnormalities.
- Marked atypia was defined as samples with moderate anisonucleosis along with one of more of the features seen in malignancy: nuclear membrane irregularity, prominence of nucleoli, abnormal chromatin distribution, and disorderly architectural arrangement (more than two cell layers, papillary groups, and nuclear overlap). In markedly atypical samples, some, but not all, of the features of malignancy were seen, or there were few abnormal cells; therefore, an unequivocal diagnosis of malignancy was not possible.
- A designation of malignant required clear-cut cytomorphological features of malignancy, with or without additional background features (necrosis and calcifications). The presence of histiocytes and background features, including protein, was noted in a semiquantitative fashion (on a scale from 1+ to 3+). Figure 1
shows examples of dye-containing histological samples and matching cytology.
View larger version (139K):
[in this window]
[in a new window]
|
FIG. 1. Benign cytological sample with microcalcification (arrow) (A) and the corresponding histological sample (B) in a woman with DCIS and overlapping ductogram. DCIS, ductal carcinoma-in-situ.
|
|
Histological Evaluation
All specimens were evaluated by the Northwestern Memorial Hospital Surgical Pathology service according to standard protocols for stereotactic core needle biopsy and excision specimens. Core needle biopsy specimens were received and examined the same day. The tissue was inspected for gross calcifications, necrosis, and other lesions. The cores were counted, and, depending on volume, the tissue was placed in one to eight tissue cassettes for processing overnight. Four hematoxylin and eosin–stained microscopic levels of each tissue specimen were cut. Microscopically, the calcifications were identified along with the associated epithelial disease. For lumpectomy specimens, each specimen was oriented according to the surgeon’s instructions and inked with four separate colors. The specimen was serially sectioned from superior to inferior or from proximal nipple to distal (according to orientation). Sections were inspected for correspondence with the specimen radiograph. Both surgical specimens were <5 cm in largest dimension, and the entire specimen was submitted. The tissue sections were processed overnight, and one microscopic section was cut from each tissue block. Microscopically, the calcifications were identified along with associated epithelial proliferative disease.
|
RESULTS
|
|---|
From August 2001 to May 2003, 20 women entered the study. Their mean age was 54.2 years (range, 44–74 years). All had screen-detected microcalcifications suspicious enough to require biopsy (Breast Imaging Reporting and Data System 4 or worse). Stereotactic core biopsy was recommended for 18, whereas wire localization and surgical excision was recommended for 2 (the location of the lesion was not considered amenable to core biopsy because of the proximity to the nipple).
Results of DL
Four women (20%) did not yield any nipple fluid from the study breast despite warming, massage, dekeratinization of the nipple, and aspiration. If three cycles of massage and aspiration were unsuccessful, the attempt was terminated. Among the remaining 16 women, there were 27 FYDs (1.7 per breast), of which 20 were lavaged (1.3 per breast). The cytological sample contained ICMD in six ducts (30%) among four women. The findings were benign in 10 ducts (50%) and mildly atypical in 4 (20%). Thus, 14 (70%) of 20 samples were cellular. Cytological and histological findings in all 20 women are presented in Table 1. Ductography was performed in 15 women who underwent DL. One woman, who had high duct sphincter tone, showed evidence of extravasation during lavage of the duct that matched the location of the calcifications; she did not proceed to ductography. An overview of the results in terms of study procedures and histological diagnoses is shown in Figure 2.
View larger version (62K):
[in this window]
[in a new window]
|
FIG. 2. Lavage, ductography, and pathologic outcomes in all patients entered. NAF, nipple aspiration fluid; DCIS, ductal carcinoma-in-situ; DL, ductal lavage; ADH, atypical ductal hyperplasia.
|
|
Pathologic Findings in Women Without Nipple Fluid
Of the four women who did not produce nipple aspiration fluid (NAF) from the study breast, all demonstrated DCIS on core biopsy. The DCIS lesions in these four women had a mean size of 11 mm; three were grade 2, and one was grade 3.
Pathologic Findings in Women Who Completed DL
In four women, the lavage cytological diagnosis was ICMD; two showed extravasation on the ductogram, and this suggested duct perforation during the DL procedure. One of these had nonproliferative changes on biopsy, and the second had atypical ductal hyperplasia. In two women with ICMD samples, the ductogram showed no overlap with the calcifications; on biopsy, one of them had DCIS with a small focus of invasion, and the other had atypical ductal hyperplasia.
Among 12 cases with assessable cytology, 8 were classified as benign and 4 as mildly atypical. The ductography findings from these women are shown in relation to cytology in Table 2. Of the four women whose ductogram showed overlap with a lavaged duct, one had DCIS (cytological diagnosis benign), one had atypical hyperplasia (cytological diagnosis mild atypia), and two had nonproliferative changes (cytological diagnosis benign in one and mildly atypical in the other). The cytological and histological findings from the woman with DCIS and an overlapping ductogram are shown in Figure 3. Her ductography images and those of a woman whose lavaged duct did not overlap with the calcifications are shown in Figure 1
.
View larger version (61K):
[in this window]
[in a new window]
|
FIG. 3. Ductogram showing overlap with calcifications: craniocaudal (A) and mediolateral (B) views. Arrows point to calcifications.
|
|
Analysis Including All FYDs
There were four women in whom the ductogram showed overlap with mammographic calcifications. In all four, the quadrant location of the overlapping ducts matched the location of the mammographic lesion (i.e., if the calcifications were in the upper outer breast, then the overlapping duct was also in the upper outer quadrant of the nipple). We therefore examined the locations of the FYDs that were not lavaged in relation to the location of the calcifications on mammography to see whether these were potentially matching ducts if lavage and ductography had been successful in all FYDs.
In six women, there were FYDs that were not lavaged, or the ductogram showed extravasation so that overlap could not be evaluated. Table 3 shows the distribution of FYDs in relation to pathologic findings and the quadrant location of the mammographic lesion. By this analysis, it is possible that two women with DCIS had fluid yield from their DCIS-containing duct. In one, the first duct lavaged matched the quadrant location of the calcifications and showed dye extravasation on ductography. In the other, two FYDs were close together, and both matched the quadrant of the calcifications; the first did not overlap, and the second could not be cannulated. Thus, if all FYDs are considered, it is possible that two additional women with DCIS might have had overlapping ducts with successful cannulation and ductography.
|
DISCUSSION
|
|---|
The concept of cancer detection by cytological examination of exfoliated ductal cells collected at the nipple by aspiration of nipple fluid has a long history; it was described by Papanicolaou et al.7 in the 1950s and was extended by Sartorius et al.8 in the 1970s. It was given new impetus by the initial multicenter study of DL in high-risk women, in which 500 asymptomatic women underwent DL: 26 had severely atypical lavage cytology, and 11 underwent surgical excision. Of these, four were found to have DCIS.3 The criteria for selection of women for surgical excision have not been described, but on the basis of these results, hopes were raised among high-risk women and their physicians that occult DCIS could be detected with DL. We previously completed a study of DL in women with known cancer undergoing DL before mastectomy, and we found that the sensitivity of DL to detect cancer is poor.5 The negative results of that study may be due to the large invasive cancers (mean size, 3.2 cm) in most subjects, although sizable areas of DCIS were also present in most women. Data from the nipple fluid-aspiration studies of Sartorius et al.9 suggest that when invasive cancers are approximately >1 cm, the ductal tree is likely to be obstructed by the invasive disease, and malignant cytology in NAF is uncommon. Thus it seems that if DL is to detect cancer, it is most likely to do so in women with early, noninvasive lesions. Our present study addresses this hypothesis by including only women with microcalcifications, so malignancy, if present, is likely to consist of DCIS.
Although 50% of the lesions in our study were DCIS, with a mean size of 12 mm, we did not encounter any markedly atypical or malignant cytological samples in this study group. Four mildly atypical samples were recovered on lavage; these came from nonoverlapping ducts in three women and from an overlapping duct in one, in whom the histological diagnosis was atypical ductal hyperplasia. Four DCIS lesions were found in breasts without NAF; in another four women with DCIS (one of whom had a small invasive focus), the lavaged ducts did not overlap the calcifications on subsequent ductography, although in two of these patients, additional (possibly matching) FYDs were present. Thus, of the 10 malignancies, 4 were missed on DL because there were no FYDs; in 3, the FYD was not the cancer-containing duct, as shown by ductography. This left three women with DCIS and FYDs concordant in location to the quadrant of the calcifications, in one of whom the lavaged duct overlapped the calcifications. These results challenge the underlying concept of DL that diseased ducts are more likely to produce NAF than nondiseased ducts. Wrensch et al.10 first suggested that standard-risk women who produced NAF were more likely to develop subsequent breast cancer than women who did not. In a subsequent cohort studied by the same investigators, these findings have not been not reproduced,11 and although the women in the second cohort were at overall higher risk, their NAF yield rate was substantially lower (77.3% in the first cohort and 40.4% in the second cohort).
The findings from this study of early lesions are similar to the findings from our earlier study of women with larger invasive lesions undergoing mastectomy.5 In that study, we found that the diagnostic failures were approximately equally explained by a lack of fluid yield from the cancer-bearing breast and fluid production by ducts that did not connect with the malignancy (as demonstrated by injection of a dye/gelatin mixture into the lavaged duct, followed by sub–gross sectioning of the specimen). In that study, we also saw a significant diagnostic failure rate because of mildly atypical cytology from cancer-containing ducts. Mild atypia was encountered in four cytological samples in the present study of early lesions, but none came from ducts with DCIS.
This was a small pilot study, and the accrual goal was 40 women, of whom we expected 30 to be assessable. In fact, we experienced considerable difficulty with recruitment and found that eligible women were often too distressed about their abnormal mammogram and upcoming biopsy to want to submit to an additional procedure before the biopsy. Complete evaluation of all FYDs was also limited by reluctance on the part of patients and physicians to spend additional time attempting to cannulate second and third FYDs when the first duct did not overlap with calcifications. We attempted to compensate for this by looking at the location of FYDs that were not taken to successful ductography and found that an additional two women might have had overlapping ducts if all FYDs had been successfully evaluated. Thus, we were potentially able to identify 3 of the 10 DCIS lesions by NAF production from the corresponding duct orifice. The other seven DCIS lesions were in breasts with no NAF yield whatever (four breasts) or yield from ducts that clearly were not related to the cancer (three breasts).
We have not been able to relate cytological diagnosis to histology with any degree of confidence because of the small number of overlapping ducts (4 of 15 ductograms showed overlap with the mammographic lesion). In patients in whom a ductogram was not performed or in whom there was no overlap, the cytology/histology correlation is meaningless because the lesion-containing duct was not lavaged. Nevertheless, the cytological findings were benign in the one DCIS lesion with overlap and in another in which extravasation was seen on the ductogram but the location of the duct matched that of the lesion. The use of molecular markers in addition to cytological evaluation may be of value in the future but is likely to improve the diagnostic performance only when the correct duct is being lavaged; our present data suggest that most failures are explained by a lack of fluid yield from, and the resultant failure to lavage, the cancer-bearing duct.
These data, although limited by the small size of the study, are in agreement with our previous findings from a study of premastectomy DL5 and those of others who have attempted to compare presurgical DL cytology with surgical pathology findings.12,13 Taken together, these findings suggest that ductal malignancy does not lead to nipple fluid production in a significant proportion of cases and that nipple fluid aspiration and DL should not be presented to patients as a method of early detection of cancer.
|
ACKNOWLEDGMENTS
|
|---|
Supported by the Bluhm Family Breast Cancer Early Detection and Prevention Program, Specialized Program of Research Excellence grant P50-CA89018, and the Cytyc Corporation.
Received for publication April 29, 2004.
Accepted for publication April 8, 2005.
|
REFERENCES
|
|---|
- Kolb TM, Lichy J, Newhouse JH. Comparison of the performance of screening mammography, physical examination, and breast US and evaluation of factors that influence them: an analysis of 27,825 patient evaluations. Radiology 2002; 225:165–75.[Abstract/Free Full Text]
- Porter PL, El Bastawissi AY, Mandelson MT, et al. Breast tumor characteristics as predictors of mammographic detection: comparison of interval- and screen-detected cancers. J Natl Cancer Inst 1999;91:2020–8.[Abstract/Free Full Text]
- Dooley WC, Ljung B-M, Veronesi U, et al. Ductal lavage for detection of cellular atypia in women at high risk for breast cancer. J Natl Cancer Inst 2001;93:1624–32.[Abstract/Free Full Text]
- Khan SA, Rodriguez N, Baird C, et al. Ductal lavage findings in women with known breast cancer undergoing mastectomy (abstract). Breast Cancer Res Treat 2002;76(Suppl 1):S35.
- Khan SA, Wiley EL, Rodriguez N, et al. Ductal lavage findings in women with known breast cancer undergoing mastectomy. J Natl Cancer Inst 2004;96:1510–7.[Abstract/Free Full Text]
- The uniform approach to breast fine needle aspiration biopsy. A synopsis. Acta Cytol 1996;40:1120–6.[Medline]
- Papanicolaou GN, Holmquist DG, Bader GM, Falk EA. Exfoliative cytology of the human mammary gland and its value in the diagnosis of cancer and other diseases of the breast. Cancer 1958;11:377–409.[CrossRef][Medline]
- Sartorius OW, Smith HS, Morris P, Benedict D, Friesen L. Cytologic evaluation of breast fluid in the detection of breast disease. J Natl Cancer Inst 1977;59:1073–80.[Medline]
- Sartorius OW, Morris PL, Benedict DL, Smith HS. Contrast ductography for recognition and localization of benign and malignant breast lesions: an improved technique. In: Logan WW, ed. Breast Cancer. New York: Wiley, 1977: 281–300.
- Wrensch MR, Petrakis NL, King EB, et al. Breast cancer incidence in women with abnormal cytology in nipple aspirates of breast fluid. Am J Epidemiol 1992;135:130–41.[Abstract/Free Full Text]
- Wrensch MR, Petrakis NL, Miike R, et al. Breast cancer risk in women with abnormal cytology in nipple aspirates of breast fluid. J Natl Cancer Inst 2001;93:1791–8.[Abstract/Free Full Text]
- Brogi E, Robson M, Panageas KS, Casadio C, Ljung BM, Montgomery L. Ductal lavage in patients undergoing mastectomy for mammary carcinoma: a correlative study. Cancer 2003;98:2170–6.[CrossRef][Medline]
- King BL, Tsai SC, Gryga ME, et al. Detection of chromosomal instability in paired breast surgery and ductal lavage specimens by interphase fluorescence in situ hybridization. Clin Cancer Res 2003;9:1509–16.[Abstract/Free Full Text]
TOP
ABSTRACT
INTRODUCTION
