Patterns of reduced nipple aspirate fluid production and ductal lavage cellularity in women at high risk for breast cancer

Nipple aspiration is a noninvasive suction technique used to obtain breast fluid from the duct openings at the nipple. The procedure elicits droplet formation from one or more ducts in 48 to 94% of healthy, nonlactating women, and these fluids often contain exfoliated ductal epithelial cells [1–10]. The nipple aspiration technique has been studied for several decades as a potential screening tool to detect cytologic abnormalities associated with proliferative breast lesions [1, 5, 7, 11–13]. Long-term follow-up of women undergoing nipple aspiration revealed that the risk of developing breast cancer was twofold to fivefold greater for women diagnosed with cytologic atypia relative to those who did not yield any fluid [11–13]. Interestingly, these studies revealed a continuum of relative risk that was lowest for women who did not yield any fluid, higher for women yielding acellular fluid, higher still for fluid-yielders with normal or benign cells, and highest for fluid-yielders diagnosed with atypia. Thus, the mere elicitation of acellular nipple aspirate fluid (NAF) was associated with increased relative risk.

More recently the ductal lavage technique, which employs a microcatheter to cannulate an individual duct opening and flush the associated ductal system with saline, has been used to collect higher numbers of exfoliated cells. In a multicenter trial that we participated in to evaluate this approach, nipple aspiration was performed on all subjects for the purpose of locating fluid-yielding ducts (presumed to contain higher numbers of cells) before the lavage procedure [3]. Fluid-yielding ducts were identified in 84% of the subjects overall, and in 84% of the trial subjects enrolled at our center. However, this rate fell markedly to 36% (P = 0.001) when we performed the procedure on a similar series of 33 high-risk women after the conclusion of the trial. Here we present data suggesting that the reduced NAF yields observed in the second series of subjects may be related to the use of risk reduction therapies including selective estrogen receptor modulators (SERMs; tamoxifen and raloxifene) and/or oophorectomy.

In the multicenter ductal lavage trial [3] only NAF-yielding ducts were lavaged, in accordance with the rationale that productive ducts would be more likely to contain large numbers of cells. In this study the same protocol was followed, but lavage was also performed on nine subjects who did not yield fluid; all nine of these lavage specimens were classified as ICMD. Of the 22 lavage specimens collected in this series, 4 of 22 (18%) were classified as mildy atypical, 5 of 22 (23%) as benign, and 13 of 22 (54%) as ICMD. ICMD was observed in nine of nine non-fluid-yielding ducts versus 3 of 13 (59%) fluid-yielding ducts (P = 0.00046). The proportion of ICMD in the multicenter trial was 84 of 383 (22%) [3].

The overall percentage of fluid-yielding subjects in the present series (36%) was relatively low in comparison with those reported in other studies (48 to 94%) [1–3, 5–10, 14] and at our site during the multicenter trial (84%) [3]. Reduced NAF production has consistently been associated with postmenopausal status and/or increasing age in other studies [1, 3, 10, 15, 16]. In the multicenter trial, fluid-yielding ducts were observed in 90% of the premenopausal women, compared with 78% of the postmenopausal women (P < 0.001) [3]. In the present study, NAF-yielding ducts were identified in only 22% of the postmenopausal women. The marked difference between these similar high-risk postmenopausal groups (P < 0.001) suggests that the overall decline observed in the present series is associated with additional variables. The subject population in the present study differed in having more women with BRCA germline mutations, many of whom had undergone BSO and/or were using SERMs. Seven of 23 postmenopausal women in this series were less than 49 years old and had undergone surgically induced menopause as the result of BSO. When NAF production was analyzed among the 33 subjects in relation to these variables, we found it to be significantly reduced in association with BRCA mutations and risk reduction therapies (Table 3). Additional analyses revealed lower NAF production in association with risk-reducing therapies in BRCA-negative women (P = 0.045) and a similar, but insignificant, trend for BRCA-positive women.

Although NAF and ductal lavage have been evaluated in women at high risk for breast cancer development, there have been only a handful of studies describing the application of intraductal approaches to high-risk women with BRCA germline mutations [17, 18]. The largest study, conducted by Kurian and colleagues [18], included 75 women with high inherited risk, 56 of whom had BRCA1/2 mutations. The overall NAF rate in this study was 48%. Although no direct correlation between fluid production and risk reduction therapies was observed, there were significantly fewer fluid-yielding subjects among women with a previous history of breast or ovarian cancer. The authors suggested therapy-related anti-hormonal mechanisms including chemotherapy-induced amenorrhea as possible explanations for the trend. Mitchell and colleagues, who performed intraductal approaches on a series of 52 women with germline BRCA1/2 mutations, observed NAF production in 60% of all subjects (Mitchell G personal communication). Although no correlation between NAF production and previous cancer was noted in this study, significantly reduced levels of NAF secretion and ductal lavage cellularity were seen in association with natural and surgically induced menopause. Although we cannulated and lavaged only nine dry ducts, all of them were also characterized as acellular. Kurian and colleagues [18] also reported higher rates of ICMD in association with previous breast or ovarian cancer, chemotherapy, and SERM usage.

The risk-reducing benefits of SERMs and oophorectomy for breast cancer patients and women at high risk for breast cancer are well documented. Clinical trials have demonstrated that tamoxifen significantly decreases recurrence rates in breast cancer patients and markedly reduces the incidence of invasive and noninvasive breast cancer development in high-risk women [19–26]. Breast cancer risk reduction has also been observed in postmenopausal women undergoing treatment with raloxifene for osteoporosis [27]. Similarly, ovarian ablation has been shown to reduce recurrence and mortality in breast cancer patients [23, 24, 28, 29], and prophylactic oophorectomy has been shown to significantly reduce breast cancer incidence in high-risk women, including BRCA1 and BRCA2 mutation carriers [30–33]. Our series included eight BRCA1 mutation carriers, and although most BRCA1 tumors are estrogen receptor negative, both oophorectomy and tamoxifen have been shown to protect against the development or recurrence of breast cancer in this group [24, 25, 30, 32, 33].

The biologic regulation and significance of breast duct fluid production in nonlactating women are not currently understood. Our data showing that risk reduction therapies are associated with diminished breast fluid secretion complement previous epidemiologic studies showing that NAF production is linked to increased relative risk for breast cancer development [11–13]. The fact that NAF is reduced in association with therapies that eliminate or block ovarian steroids suggests that the relationship between risk and breast fluid production is related to hormones. Although SERMs modulate the actions of estrogens, we did not analyze the relationship of NAF to SERMs and BSO separately. We therefore cannot speculate on the relative contributions of estrogens and progesterone to the observed effect. However, earlier studies did not find associations between serum estrogen levels and NAF production [9, 10, 34]. One of these studies reported an association between NAF production and breast fluid prolactin levels, but not estrogen levels [34]. It is possible that the relationship between NAF production and ovarian hormones involves indirect systemic and/or paracrine mechanisms.

Because NAF production and cytologic atypia are both associated with increased relative risk for breast cancer development, it is tempting to speculate that NAFs reflect the secretory activity of proliferative ductal epithelial cells. However, several recent studies have generated paradoxical results concerning the relationship between breast fluid, cytologic atypia, and breast cancer risk. These studies have revealed the presence of cytologic atypia in non-fluid-yielding ducts and in NAF-negative women [9, 17, 18, 35]. Although seemingly at odds with the body of data on NAF, the atypical epithelial cells associated with non-fluid-yielding ducts may constitute biologically distinct variants and may hold clues for stratifying the prognostic significance of atypia. Our data suggest that reduced NAF production reflects lower endogenous hormone levels and therefore that atypical cells proliferating in non-fluid yielding ducts may represent hormone-independent variants. Kurian and colleagues suggested the possibility that women with atypia in non-fluid-yielding ducts might be at higher risk for developing estrogen-receptor-negative tumors [18].

An analysis of high-risk women in the present series revealed patterns of reduced NAF production and ductal lavage cellularity in comparison with a previous multicenter trial. The present series included more BRCA-positive women, many of whom had undergone BSO and/or were using SERMs. Our data suggest that endocrine mechanisms associated with these risk-reducing therapies may be related to patterns of diminished breast fluid production. Although NAF production does not seem to be obligatorily linked to cytologic atypia, it may be a general indicator of endogenous reproductive hormone levels that are widely implicated in breast cancer development. More importantly, NAF seems to reflect the impact of the systemic hormonal milieu on the breast tissue itself. Thus, its production during the NAF procedure could potentially provide additional information to a woman's risk profile, and could serve as a valuable endpoint in chemoprevention trials.