Nipple aspirate fluid as a proximal breast liquid biopsy platform: advancing precision risk management in breast cancer

Abstract

Background

Current breast cancer (BC) risk management relies primarily on imaging and blood‑based liquid biopsies. However, these approaches are constrained by morphological blind spots, limited sensitivity for early lesions, and an inability to directly capture the biology of the ductal microenvironment where most tumors originate. There is a pressing need for novel strategies that provide direct, local insight into breast tissue risk.

Main body

This review repositions nipple aspirate fluid (NAF) as a proximal, breast-specific liquid biopsy platform. Emerging evidence demonstrates that NAF, obtained by nipple aspiration from non-lactating breasts and not limited to cases of spontaneous nipple discharge, originates from a functionally autonomous ductal microenvironment and provides rich multi-omics information, including proteins, metabolites, epigenetic alterations, extracellular vesicles, inflammatory/oxidative mediators, microbiome-associated signals, and other local regulatory readouts, rather than simply representing a filtrate of blood. We summarize the biological foundations of NAF and its multi-omics landscape, highlighting its potential in three key clinical scenarios: (1) molecular risk stratification for pathologic nipple discharge, (2) individualized biological risk profiling in women with high-risk or dense breasts, and (3) proximal monitoring of treatment response and microenvironmental dynamics in patients with diagnosed BC.

Conclusion

Although standardization and prospective validation remain essential challenges, the development of robust collection protocols, integration of multi‑omics data, and execution of well‑designed clinical trials could enable NAF to shift BC management from a paradigm of “early detection” toward one of “precision risk insight and intervention,” ultimately aiming to improve patient outcomes and net benefit.

Graphical abstract

Introduction: toward precision risk management in breast cancer: redefining the role of nipple aspirate fluid

The global burden and strategic evolution of management goals

Breast cancer (BC) is the most frequently diagnosed malignancy among women worldwide, with its incidence continuing to rise and posing a significant public health challenge [1]. According to the latest estimates from the GLOBOCAN database, approximately 2.3 million new BC cases and 670,000 deaths occurred globally among women in 2022 [1–3]. In response to this pressing reality, the World Health Organization Global BC Initiative has set an ambitious target: to reduce the age-standardized BC mortality rate by 2.5% annually by 2040 [3]. However, a substantial gap persists between this goal and the current trajectory. While mortality rates are declining in some very high Human Development Index countries, few are on track to meet this annual reduction target. Projections suggest that if current trends continue, the global number of new BC cases and deaths will surge by approximately 38% and 68%, respectively, by 2050 compared to 2022 levels [3]. This stark outlook underscores an urgent need for a profound transformation in global BC control strategies. The core objective must evolve beyond merely “early detection” to achieving “maximized net benefit” at the population level. This entails reducing mortality while simultaneously minimizing the harms associated with screening and diagnosis, such as overdiagnosis, false-positive results, and unnecessary invasive procedures [4, 5].

Unmet clinical needs in the era of precision risk management: the imperative for local microenvironment insight

The evolving paradigm in BC management—shifting from a primary goal of “early detection” to one of “maximizing net benefit”—is driving a transition from uniform, population-based screening toward individualized, risk-stratified care. This strategic shift, however, critically exposes the inherent limitations of existing core technologies.

Imaging modalities, which serve as the cornerstone for anatomical localization, are fundamentally constrained by their reliance on macroscopic morphology. This leads to a significant “morphological blind spot.” First, diagnostic performance is heavily influenced by breast density. For the substantial proportion of Asian women with dense breasts (prevalence reaching 70–80% in populations such as South Korea and Thailand), the sensitivity of mammography is markedly reduced, creating an intractable “density blind spot” [6, 7]. Second, approximately 30% of BCs are detected as interval cancers despite regular screening. These tumors often exhibit more aggressive biology, underscoring imaging’s insensitivity to underlying malignant potential [8]. Furthermore, current screening models are associated with considerable harms, including false-positive recalls and overdiagnosis. It is estimated that for every BC death prevented by screening, about three women are overdiagnosed and undergo unnecessary treatment [5, 7]. Even advanced techniques like digital breast tomosynthesis and magnetic resonance imaging, while improving detection rates, face practical challenges of limited resources, high cost, increased false-positive rates, or poor accessibility [9–11].

Conversely, blood-based liquid biopsy offers a unique “systemic” perspective by detecting biomarkers such as circulating tumor DNA (ctDNA) and circulating tumor cells. It has proven invaluable for monitoring treatment response and tracking resistance mechanisms in advanced disease [12, 13]. However, in early-stage or localized disease where tumor burden is low, the concentration of ctDNA in the bloodstream is minimal, resulting in limited sensitivity that precludes reliable early detection [14–16]. More fundamentally, blood analysis reflects a systemic, “averaged” signal. Extensive research confirms that BC is not a homogeneous disease but one with a distinct anatomical origin. Most breast lesions, including ductal carcinoma in situ (DCIS) and lobular carcinoma in situ (LCIS), originate in the terminal ductal lobular unit (TDLU) [17]. Moreover, invasive BCs can be classified by their precise anatomical origin into prognostically distinct subtypes: the more favorable acinar adenocarcinoma (originating from milk-producing acini) and the less favorable ductal adenocarcinoma (originating from major lactiferous ducts) [18–20]. Consequently, blood-based biopsies cannot capture the specific molecular events occurring within the ductal-lobular system—the primary anatomical site of tumor initiation and progression.

Thus, a clear and pressing unmet clinical need exists: a tool capable of providing direct biological insight into the local ductal microenvironment through a non-invasive sampling approach. Such a tool is essential to bridge the informational gaps left by imaging and blood-based biopsies. It would deliver crucial complementary data to inform management decisions for high-risk individuals and those presenting with clinical symptoms, enabling truly precise risk assessment.

A paradigm shift: redefining nipple aspirate fluid as a proximal multi-omics biopsy platform

Nipple aspirate fluid (NAF), a biofluid collected directly from the breast ductal system, has been historically confined to cytological assessment. This traditional role is no longer adequate for population screening, as conventional cytology is limited by low sensitivity (~ 64%) and a high rate of insufficient samples [21, 22]. Advances in multi-omics technologies now compel a paradigm shift in the application of NAF.

We argue that NAF must be fundamentally redefined: not as a screening tool, but as a dedicated proximal liquid biopsy platform for multi-omics interrogation. When collected using standardized protocols, NAF offers a direct “biological window” into the ductal microenvironment, providing a rich molecular portrait that includes proteins, epigenetic markers, non-coding RNAs, exosomes, metabolites, and the local microbiome [23–25].

In this review, NAF refers to ductal fluid obtained from non-lactating breasts by nipple aspiration rather than exclusively to spontaneous nipple discharge. It is typically collected by applying gentle negative pressure to the nipple-areolar complex, either manually or with dedicated aspiration devices; in some protocols, warming, massage, or oxytocin administration may be used to improve secretion and collection success. In addition to its role in the evaluation of pathologic nipple discharge (PND), NAF can also be obtained from asymptomatic women, including those with dense breasts or elevated breast cancer risk, where it serves as a proximal biospecimen for molecular profiling and multi-omics analyses.

Review outline

Guided by this reconceptualization, this review is structured as follows. First, we establish the biological rationale for NAF as a valid proximal biospecimen. We then synthesize the current evidence on its multi-omics landscape. The core of our discussion focuses on its potential to refine clinical decision-making in three high-impact scenarios. Finally, we address the critical requirements for standardization and translation, concluding with an analysis of prevailing challenges and future research directions. This reconceptualization is illustrated in Fig. 1.

Fig. 1.

Redefining Nipple Aspirate Fluid: A Paradigm Shift from Screening Tool to Proximal Multi-Omics Biopsy Platform. This figure illustrates the core argument of this review: through standardized collection, NAF can be redefined as a unique "biological window" that directly reflects the local ductal microenvironment, encompassing multi-omics information such as hormones, proteins, metabolites, and microbiota. This paradigm shift provides a novel biological foundation for enabling precision clinical decision-making in three high-value scenarios: the management of PND, individualized risk assessment for individuals with high-risk or dense breasts, and therapeutic monitoring in patients with confirmed BC

The biological basis of naf: a functionally segregated ductal microenvironment biofluid

Establishing NAF as a robust proximal biopsy platform necessitates a foundational understanding of its distinct biological origin. The functional anatomic unit of the breast is the lobular-ductal system, an arborizing network extending from the nipple into the parenchyma, as visualized through ductography and 3D computational reconstruction [26, 27]. This network is lined by two principal epithelial cell types: secretory luminal cells and contractile myoepithelial cells [28, 29]. NAF is primarily generated by the luminal epithelium, comprising both substances synthesized de novo and released directly into the lumen, and components selectively transported from the systemic circulation [30–32].

Contrary to the historical view of NAF as a passive plasma transudate, converging evidence reveals the ductal system as a functionally segregated local microenvironment. Its biological state is not accurately mirrored in systemic blood.

An autonomous hormonal niche: from cumulative exposure and local synthesis to functional relevance

The comprehension of NAF’s hormonal landscape has evolved from phenomenological observation to mechanistic understanding, revealing direct clinical implications. Early work established that NAF steroid hormone concentrations significantly surpass serum levels, identifying NAF as a biomarker of cumulative exposure—a property that fundamentally distinguishes it from the transient snapshot provided by blood. One study showed that following ovarian suppression by oral contraceptives, the decline in hormone levels within NAF lagged several days behind the drop in serum. Moreover, hormone levels in concurrently sampled NAF and serum correlated poorly [31]. These findings position NAF as a "biological reservoir," integrating the hormonal exposure history of breast tissue and potentially better representing the sustained, biologically effective dose at the target organ than a single blood measurement. This cumulative nature is further evidenced spatially, with significant inter-breast differences in NAF hormone concentrations within the same individual [33].

This cumulative exposure is driven by active local biosynthesis. Sauter et al. (2006) demonstrated that the concentration of bioactive estradiol (E2) in NAF strongly correlated with its precursors (e.g., estrone sulfate, androstenedione); a model incorporating these precursors explained up to 83% of the variance in NAF E2 levels [34]. This robustly supports the active local conversion of circulatory precursors by ductal epithelium, defining an efficient, dedicated estrogen-generating system. Critically, this local synthesis was linked to downstream function: NAF E2 levels strongly predicted the expression of estrogen-regulated proteins (cathepsin D and epidermal growth factor(EGF)) [34]. This confirms the biological activity of ductal estrogen, directly connecting local hormone exposure to proliferative signaling pathways.

The autonomy of this system is further affirmed during physiological transitions. During the menstrual cycle, the NAF E2 peak lags approximately one week behind the serum peak, coinciding with precursor depletion in NAF—a dynamic consistent with local synthesis, not passive equilibration [35]. Most clinically, in postmenopausal women, NAF estrogen concentrations do not fall precipitously alongside serum levels [31, 34]. Coupled with a rise in inflammatory mediators like interleukin-6(IL-6) in postmenopausal NAF [31], this sustained local hormonal activity delineates a pro-proliferative, pro-inflammatory ductal microenvironment, providing a crucial local biological context for postmenopausal BC risk.

In summary, the autonomous hormonal niche delineated by NAF provides a compelling theoretical advantage over systemic blood assays for tissue-specific risk assessment.

A distinct molecular network: the local regulatory landscape from a systems biology perspective

This compartmentalized functionality is further substantiated through a systems biology lens. Definitive evidence emerged from a 2014 metabolomics study, which found that the metabolite profiles of NAF and plasma are fundamentally disparate. Even for metabolites common to both biofluids, their concentrations show minimal correlation, underscoring distinct metabolic milieus [36].

Proteomic analyses reinforce this segregation. A 2016 study revealed a highly coordinated network of specific associations within NAF—linking proteins such as the inflammatory marker chitinase-3-like protein 1(YKL-40) and the pro-angiogenic factor basic fibroblast growth factor (bFGF) with hormones like testosterone. Crucially, this intrinsic network vanished entirely when correlating NAF proteins with serum hormone levels [37]. This demonstrates that the breast duct harbors a sequestered endocrine microniche; its local signaling crosstalk is confined and cannot be extrapolated from systemic circulation, necessitating direct NAF analysis. For example, the observed positive correlation between YKL-40 and testosterone alongside a negative correlation with bFGF reveals complex, potentially bidirectional hormonal regulation within the local microenvironment—offering novel mechanistic insights into hormone-mediated oncogenic risk [37].

Supporting the specificity of this local environment, inter-breast variation in NAF hormone and protein concentrations within the same individual far exceeds the temporal variation within a single breast, highlighting both the spatial heterogeneity and relative temporal stability of the ductal milieu [33].

In summary, across metabolomic, proteomic, and hormonal dimensions, NAF consistently reflects a functionally autonomous unit. This autonomy is rooted in ductal physiology: the mammary duct is a semi-closed system. Its orifices are often physiologically occluded, and active reabsorptive mechanisms maintain a dynamic balance between secretion and reabsorption [38]. This unique anatomical and physiological framework is the essential precondition for the distinct molecular microenvironment characterized above. A summary of these biological principles is provided in Fig. 2.

Fig. 2.

The Biological Basis of NAF: A Functionally Segregated Ductal Microenvironment. This figure illustrates the biological principles that establish nipple aspirate fluid (NAF) as a valid proximal biospecimen. It summarizes key evidence demonstrating that the breast ductal system operates as a functionally autonomous local microenvironment, distinct from systemic circulation. Specifically, hormone levels in NAF reflect cumulative tissue exposure and local synthesis, rather than transient blood levels. Furthermore, the metabolite and proteomic profiles of NAF are fundamentally different from those of plasma. Finally, NAF harbors unique local regulatory networks, underscoring its value as a direct window into ductal biology for risk assessment and mechanistic insight

NAF reflects the broader ductal microenvironment

Endocrine autonomy is a defining feature of the mammary ductal system, but the composition of NAF also reflects a broader range of local biological processes. Recent reviews and overview papers suggest that NAF should be interpreted not simply as a secretion enriched in hormones or soluble proteins, but rather as a composite proximal readout of ductal epithelial activity and the wider ductal microenvironment, including stromal remodeling, inflammatory signaling, and host–microbial interactions [39–41]. In this context, oxidative and inflammatory alterations detected in NAF—including increased protein carbonyls, reduced SOD-1, and changes involving CRP, IL-6, and related inflammatory markers in metabolically or hormonally altered states—are relevant not only as correlative findings, but also as indicators of microenvironmental stress that may accompany early epithelial transformation [31, 42–45]. In parallel, breast cancer research more broadly has increasingly recognized extracellular vesicles as active mediators of intercellular communication, immune modulation, and metastatic programming; accordingly, vesicle-associated cargo identified in nipple fluid should be considered part of the functional signaling network of the ductal niche rather than merely passive cellular debris [25, 46]. Likewise, microbiome-oriented studies support the concept that local microbial communities and their metabolic functions, including β-glucuronidase-related pathways, may shape the biochemical landscape of NAF and influence estrogen-related microenvironmental biology [47, 48]. Taken together, these observations reinforce the view that NAF is a biologically integrative fluid that reflects the broader ductal ecosystem and provides a rationale for multi-omics interrogation.

The multi-omics landscape of nipple aspirate fluid: from complementary molecular dimensions to integrated biological insight

Having established the biological basis of NAF as a proximal readout of the ductal microenvironment, its translational relevance is primarily realized through multi-omics analysis. Different analytical modalities interrogate complementary aspects of breast biology, ranging from cellular context and early molecular alterations to functional states and broader microenvironmental conditions. NAF evaluation has therefore progressed from conventional cytomorphology to a multi-layered molecular platform capable of interrogating the ductal niche through a non-invasive sampling approach. Within this framework, each omics layer contributes a distinct but complementary level of information and is supported by corresponding analytical approaches tailored to cellular, molecular, proteomic, metabolic, and microenvironmental profiling. Representative biomarkers across these layers are summarized in Table 1.

Table 1.

Overview of multi-omics biomarkers in nipple aspirate fluid. This table systematically summarizes representative biomarkers across different omics layers detectable in NAF, including descriptions of supporting evidence, detection methods, and primary clinical utility

Omics layer	Representative marker/pathway	Evidence description & mechanism	Detection method	Primary clinical utility	Ref
Cytomorphology	Ductal Epithelial Cells, Atypical/Malignant Cells	Long-term Risk Warning: The presence of any epithelial cells in NAF signifies a significantly increased future risk of BC; the presence of atypical or malignant cells confers an even higher risk, establishing a clear risk gradient	LBC, CSC, optionally combined with fiberoptic ductoscopy [51, 52]	Initial Risk Screening and Stratification: Applicable for risk stratification of PND and baseline risk assessment in high-risk populations	[49, 50]
		Technological Positioning: Limited sensitivity (~ 64%) and a high rate of insufficient samples preclude its use for population screening, but its value as a tool for initial risk screening and stratification is well-established			[21, 22]
Proteomics (incl. Glycosylation Markers)	sTn Antigen	Ultra-high-specificity diagnosis: expression in PND cytology demonstrates near-100% specificity and is associated with aggressive, HER2-positive subtypes	Lectin Assay, Immunodetection	Precision Diagnosis of PND: For definitive diagnosis of malignant discharge and identification of high-risk cases	[77]
	TF Antigen	Precancerous Lesion Identification: Concentration is highest in NAF from DCIS patients, effectively distinguishing pre-malignant from benign lesions	Immunoassay	PND Triage and DCIS Identification: Aids in distinguishing benign from malignant conditions and identifying the extent of precancerous lesions	[78]
	uPA, PAI-1 (often combined with TF antigen)	Near 100% Predictive Accuracy: a triple-marker panel of uPA, PAI-1, and TF antigen achieves near-100% accuracy in predicting cancer/precancerous lesions in women undergoing surgery for PND	ELISA	Pre-operative Decision-making for PND & Therapeutic/Prognostic Monitoring	[68]
		Validated Prognostic Indicator: Prognostic factors validated by Level 1 evidence; elevated expression is associated with poorer outcomes			[67]
	bFGF & PSA Combination	High-Efficiency Discriminatory Model: Elevation in bFGF coupled with a decrease in PSA indicates a high probability of cancer (94.1%); the inverse pattern strongly suggests benign disease (90.5%)	ELISA, Immunoblot	Risk Stratification & Early Warning; Dynamic Monitoring of Therapeutic Response	[70]
		Prognostic Value of PSA: PSA levels show an inverse correlation with tumor stage and burden, providing prognostic information			[85]
	GCDFP-15 & AAG	Inverse Expression Pattern: Decreased GCDFP-15 alongside elevated AAG is particularly valuable for diagnosing DCIS in premenopausal women	2-DE coupled with MS, ELISA	DCIS Diagnosis and Risk Stratification	[64]
	DJ-1 Protein	Independent Diagnostic Marker: Using a cutoff of 3.0 ng/mL, it shows diagnostic sensitivity of 75% and specificity of 85.9%, functioning as an independent predictor	ELISA	Adjunctive Diagnosis for PND and Independent Prediction	[80]
	BF5 (C-terminal fragment of Alpha-1-Antitrypsin)	Local Microenvironment-Specific Marker: Specifically elevated in NAF from some patients with BC, DCIS, or atypical hyperplasia, associated with increased MMP-7 activity in the tumor microenvironment	Mass Spectrometry, Immunodetection	Reflects Local Microenvironment Dysregulation, used for risk insight in high-risk populations	[71]
	HSP90α	Progressive Diagnostic Marker: Expression in NAF shows a stepwise increase from benign lesions → in situ carcinoma → invasive ductal carcinoma; its diagnostic value surpasses that of CEA	ELISA	Diagnostic Stratification and Progression Monitoring	[72]
	Combined Detection of CA15-3, CA125, CEA, TSGF	Biomarker Enrichment Advantage: The positivity rate of combined detection in NAF is significantly higher than any single-marker assay in serum or nipple discharge	Electrochemiluminescence, ELISA	Disease Monitoring: Provides rich biological information for efficacy evaluation and recurrence monitoring	[88]
	Proteomic Profiles (HomEP/HetEP)	Prognostic Typing Tool: Based on 2D-DIGE protein profiles of bilateral NAF, a "Heterogeneous Expression Profile" (HetEP) is significantly associated with increased tumor aggressiveness and worse prognosis	2D-DIGE, LC–MS/MS	Non-invasive Assessment of Tumor Heterogeneity and Prognosis	[90]
Genomic & Epigenetic Markers	Promoter Hypermethylation (e.g., RASSF1A, p16)	Molecular Biopsy: Hypermethylated genes in BC tissue are consistently detected in matched NAF samples	qMSP, ddPCR	Early Molecular Warning and Monitoring in High-Risk Individuals	[23]
		High-Risk Warning: In BRCA1/2 mutation carriers, RASSF1A methylation—particularly—is associated with short-term BC development			[53]
	mtDNA Mutations	Early Clonal Evolution Tracking: Leveraging their high copy number advantage, they offer a unique window for tracing the earliest clonal evolution within seemingly normal mammary epithelium; somatic mutations are detectable in ~ 20% of NAF from benign breast disease	Sequencing	Risk Screening and Early Molecular Warning	[54, 55]
	LOH/MSI, Chromosomal Aneuploidy (FISH)	Dynamic Genetic Monitoring: The frequency of LOH/MSI in NAF increases significantly with the progression of breast lesions from benign to malignant	FISH, PCR	Risk Re-stratification and Dynamic Monitoring	[57]
		Discrimination of Borderline Cases: FISH provides objective molecular genetic evidence for cytologically ambiguous "mild atypia," effectively distinguishing benign diploid cells from those with malignant potential (aneuploidy)			[56]
Transcriptomics & Exosomes	miR-29c-5p	Driver Associated with Breast Density: Upregulated in NAF from women with extremely high mammographic density, potentially driving the high-density phenotype and associated risk by suppressing pathways such as PTEN	RT-qPCR, Small RNA Sequencing	Risk Assessment for Dense Breasts and Mechanistic Exploration	[58]
	miR-145-5p	Diagnostic Marker for Papilloma: Significantly upregulated in the PND of patients with intraductal papilloma, providing a diagnostic accuracy of 92%	RT-qPCR	Etiologic Differentiation of PND (Papilloma)	[59]
	Exosomes & Their Cargo	Intercellular Communication Vehicle: Exosomes in NAF carry proteins, RNA, DNA, etc., reflecting the state of their cell of origin, offering novel opportunities for diagnostic models	Ultracentrifugation, Immunocapture	Therapeutic Monitoring and Microenvironment Research	[25, 46]
Hormonal, Metabolic & Microenvironmental Exposure Markers	Hormone Profiles (Estradiol/E2, Testosterone, etc.)	Autonomous Local Endocrine Microenvironment: Hormone concentrations in NAF reflect cumulative exposure and local synthesis; E2 in NAF does not fall precipitously postmenopause while IL-6 rises, delineating a pro-inflammatory microenvironment	LC–MS/MS, Immunoassay	Quantifying Local Oncogenic Driving Pressure for risk assessment in high-risk populations	[34]
		Secluded Regulatory Network: NAF harbors intrinsic, specific protein-hormone regulatory networks (e.g., positive correlation between YKL-40 and testosterone) absent in blood			[37]
	Oxidative Stress Markers (Protein Carbonyls, SOD-1)	Damage Progression Gradient: Protein carbonyl levels in NAF show a clear gradient: Healthy < Precancerous < Breast Cancer	ELISA, Colorimetric Assay	Assessment of Microenvironmental Oxidative Damage Extent for Risk Stratification	[43]
		Local Specificity: The key antioxidant enzyme SOD-1 shows decreased levels in the NAF of cancer patients, a change not observed in plasma			[42]
	Inflammatory Marker (CRP)	Linking Metabolism to Local Risk: NAF CRP levels show significant positive correlations with obesity indicators (body fat %, BMI, serum triglycerides)	ELISA	Assessment of Local Inflammatory State and Associated Metabolic Risk	[45]
	Environmental Exposure (Aluminum, Ferritin)	Exposure-Oxidation-Inflammation Axis: Both aluminum and ferritin levels are significantly elevated and strongly correlated (R = 0.94) in the NAF of the cancer group, showing strong linear relationships with oxidative stress and pro-inflammatory factors	ICP-MS (Al), ELISA (Ferritin)	Integrated Assessment of Environmental Exposure and Microenvironmental Risk	[74, 75]
	Microbiome	Microbiota-Estrogen Metabolism Axis: Enrichment of bacterial communities possessing β-glucuronidase activity in the NAF of BC patients, proposing a novel carcinogenic mechanism involving the local remodeling of estrogen metabolism	Metagenomics	Exploration of Microenvironmental Ecology and Risk Mechanisms	[47, 48]
	Pharmacodynamic Markers (PGE₂, GDF-15)	Local Pharmacodynamic Platform: PGE₂ concentration in NAF is approximately 81-fold higher than in plasma. Celecoxib significantly reduces PGE₂ in NAF, validating NAF as a feasible platform for local pharmacodynamic assessment	LC–MS/MS, ELISA	Assessment of Targeted Drug Local Distribution and Pharmacodynamics	[91]
		Revealing Novel Mechanisms: Sulindac induces GDF-15 production, revealing a potential chemopreventive mechanism independent of COX-2 inhibition			[93]

2-DE, Two-dimensional gel electrophoresis; 2D-DIGE, Two-dimensional difference gel electrophoresis; AAG, Alpha-1-acid glycoprotein; BC, Breast Cancer; bFGF, Basic fibroblast growth factor; BMI, Body Mass Index; BRCA1, breast cancer susceptibility gene 1; CA125, Cancer Antigen 125; CA15-3, Cancer Antigen 15–3; CEA, Carcinoembryonic antigen; CRP, C-reactive protein; CSC, Conventional smear cytology; DCIS, Ductal carcinoma in situ; ddPCR, Droplet digital PCR; DNA, Deoxyribonucleic acid; E2, Estradiol; ELISA, Enzyme-linked immunosorbent assay; FISH, Fluorescence in situ hybridization; GCDFP-15, Gross cystic disease fluid protein-15; GDF-15, Growth differentiation factor-15; HetEP, Heterogeneous Expression Profile; HomEP, Homogeneous Expression Profile; HSP90α, Heat shock protein 90α; ICP-MS, Inductively coupled plasma mass spectrometry; IL-6, Interleukin-6; LBC, Liquid-based cytology; LC–MS/MS, Liquid chromatography-tandem mass spectrometry; LOH, Loss of heterozygosity; miRNA, MicroRNA; MMP-7, Matrix metalloproteinase-7; MS, mass spectrometry; MSI, Microsatellite instability; mtDNA, Mitochondrial DNA; NAF, Nipple aspirate fluid; PAI-1, Plasminogen activator inhibitor-1; PCR, Polymerase chain reaction; PGE₂, Prostaglandin E₂; PND, Pathologic nipple discharge; PSA, Prostate-specific antigen; PTEN, Phosphatase and Tensin Homolog; qMSP, Quantitative methylation-specific PCR; RASSF1A, Ras association domain family member 1A; RNA, ribonucleic acid; RT-qPCR, Reverse transcription quantitative PCR; sTn, Sialyl-Tn antigen; SOD-1, Superoxide dismutase 1; TF, Thomsen-Friedenreich antigen; TSGF, Tumor-specific growth factor; uPA, Urokinase-type plasminogen activator; YKL-40, Chitinase-3-like protein 1

Cytological analysis: from diagnosis to risk stratification

Cytological analysis provides the most direct cellular information in NAF, but its current role is primarily risk stratification rather than standalone diagnosis. Prospective studies show that the presence of epithelial cells in NAF, even without atypia, is associated with increased future breast cancer risk, whereas atypical or malignant cytology indicates a still higher-risk state [49, 50]. These observations support cytology as an initial means of defining risk context.

Its limitations remain substantial. Low cellularity and frequent inadequate specimens constrain sensitivity, which is approximately 64% for cancer detection despite relatively high specificity in clearly abnormal samples [21, 22]. Liquid-based cytology and direct discharge-based approaches may improve specimen adequacy and interpretation in pathologic nipple discharge [51, 52], but inconsistent representation of underlying ductal pathology remains the fundamental limitation. Cytology is therefore best regarded as a contextual rather than definitive component of NAF-based assessment.

Genomics and epigenomics: capturing early molecular determinants

Genomic and epigenomic assays extend NAF analysis beyond morphology by detecting molecular alterations associated with early carcinogenesis. In practical terms, this layer has been investigated using qMSP, ddPCR, targeted sequencing, PCR-based LOH/MSI assays, and FISH-based aneusomy analysis, thereby providing a broader methodological base than morphology alone. Promoter hypermethylation is among the most reproducibly reported findings. Hypermethylation of genes such as RASSF1A and p16 has been detected in NAF with concordance to matched tumor tissue, supporting the concept of NAF as a molecular biopsy [23]. In BRCA1/2 mutation carriers, methylation signals in ductal fluid have also been associated with short-term breast cancer development, indicating potential utility in high-risk surveillance [53].

Complementary genomic approaches capture additional dimensions of early transformation. Mitochondrial DNA mutations, facilitated by high copy number, permit sensitive detection of early clonal alterations in breast epithelium [54, 55]. Loss of heterozygosity, microsatellite instability, and chromosomal aneusomy detected by PCR- or FISH-based assays likewise increase with lesion severity and may refine classification in cytologically indeterminate cases [56, 57]. Together, these approaches move NAF analysis toward pre-morphologic risk detection. Their translational potential is tempered by heterogeneity in molecular targets, analytical platforms, and study scale, which currently limits cross-study comparability and clinical standardization.

Transcriptomics and exosomal analysis: revealing active regulatory communication

Transcriptomic and exosomal analyses characterize dynamic regulatory and intercellular signaling states within the ductal microenvironment. Current studies have mainly relied on RT-qPCR, small-RNA sequencing, and exploratory vesicle-isolation workflows, reflecting a field that is biologically promising but still technically maturing. Altered miRNA profiles in NAF have been linked to clinically relevant phenotypes, including increased miR-29c-5p expression in women with extremely high mammographic breast density [58] and the diagnostic association of miR-145-5p with intraductal papilloma in pathologic nipple discharge [59]. Beyond individual biomarker studies, recent NAF-focused commentaries and reviews have highlighted that miRNA-based interrogation of nipple fluid may be particularly attractive because it combines direct breast specificity with relative molecular stability in low-volume samples [60, 61]. This provides a conceptual rationale for using NAF RNA signatures not only for lesion discrimination, but also for biological subclassification and longitudinal monitoring in selected clinical contexts.

At the same time, the translational expansion of this field depends heavily on robust vesicle isolation and characterization workflows, because extracellular vesicles in breast cancer are now understood to participate in immune crosstalk, treatment resistance, and metastatic dissemination across the tumor microenvironment [46]. Exosomes identified in nipple discharge contain proteins, RNA, and DNA cargo reflective of their cells of origin, indicating that NAF contains structured intercellular signaling information in addition to soluble molecules and exfoliated cells [62]. Accordingly, exosomal cargo in NAF should be interpreted as a structured signaling layer that may complement soluble proteins and exfoliated cells, rather than as an isolated analyte category. However, reproducibility remains constrained by limited sample volume, variable nucleic acid yield, and the need for more consistent vesicle-enrichment workflows.

Proteomics: reporting the functional state of the microenvironment

Proteomics is among the most extensively developed molecular applications of NAF and provides a direct readout of functional states within the ductal microenvironment. Early studies using SELDI-TOF identified disease-associated protein fingerprint patterns capable of distinguishing breast cancer from benign disease, establishing initial proof of diagnostic potential in the NAF proteome [63]. Subsequent analyses using two-dimensional gel electrophoresis coupled with mass spectrometry identified candidate markers such as reduced GCDFP-15 and increased AAG in cancer-associated NAF, particularly in ductal carcinoma in situ [64]. The subsequent introduction of LC–MS/MS enabled broader proteome characterization, including a reference map of nearly 2,000 proteins and the identification of numerous breast-specific proteins not observed in plasma, thereby reinforcing the tissue specificity of NAF [65]. Paired analyses of affected and unaffected breasts in unilateral cancer further identified protein alterations linked to glycolysis, immune activation, and proliferative signaling [66].

At the biomarker level, evidence spans both clinically advanced and exploratory candidates. The uPA/PAI-1 system, particularly when combined with TF antigen, has shown near-complete accuracy in predicting atypia or malignancy in women undergoing surgery for pathologic nipple discharge [67, 68]. Other repeatedly reported markers include decreased PSA and increased bFGF in cancer-associated NAF [69, 70], whereas BF5 and HSP90α represent additional exploratory candidates linked to local proteolysis and disease progression [71, 72]. The translational strength of proteomics lies less in isolated marker discovery than in its capacity to define coordinated signatures of epithelial transformation and microenvironmental remodeling. Its remaining limitations include variability in sample volume, dynamic protein abundance, and cross-platform standardization.

From a translational standpoint, recent high-throughput proteomic work suggests that clinically feasible NAF profiling may be achievable without relying exclusively on highly complex discovery platforms. For example, streamlined LC–MS-based workflows have been proposed as a practical bridge between exploratory proteomics and larger cohort validation, illustrating that the field is gradually moving from proof-of-concept biomarker discovery toward clinically scalable assay development [73]. This evolution is important because one of the main barriers to NAF translation has been the historical gap between technically sophisticated discovery studies and real-world reproducibility in larger populations.

Metabolomics, exposure-related markers, and microbiome-associated signals: expanding the microenvironmental perspective

Metabolomic and exposure-related analyses extend NAF profiling beyond cellular alterations to the biochemical state of the broader ductal ecosystem. This layer has been explored primarily with 1H-NMR spectroscopy, GC–MS, LC–MS/MS, ELISA-based oxidative/inflammatory assays, and ICP-MS for exposure-related metals, giving it a methodological depth comparable to other omics layers. Comparative analysis of NAF and plasma has shown that NAF possesses a distinct metabolite composition, supporting its role as a localized metabolic readout rather than a diluted systemic surrogate [36]. Metabolomics therefore captures downstream consequences of epithelial, stromal, endocrine, and inflammatory activity within the ductal niche.

This broader microenvironmental perspective is further supported by oxidative, inflammatory, and exposure-related markers. Increased protein carbonyl levels and reduced SOD-1 in cancer-associated NAF indicate progressive oxidative injury within the local breast microenvironment [42, 43]. Elevated CRP correlates with breast cancer risk indicators and metabolic factors including body fat percentage, body mass index(BMI), and serum triglycerides, linking systemic metabolic dysfunction to local inflammatory states [44, 45]. Higher aluminum levels and altered iron-related markers further support an exposure-oxidation-inflammation axis in NAF [74, 75]. Distinct microbial profiles have also been reported, including enrichment of taxa with β-glucuronidase activity in breast cancer-associated samples [47], a finding of mechanistic interest given the potential influence of microbial enzymatic activity on local estrogen metabolism [47]. However, microbiome-related analysis remains exploratory and requires further methodological standardization.

Integrative perspective

The value of the NAF multi-omics platform lies in the complementarity of its analytical dimensions rather than in any single biomarker class. Cytology provides cellular context; genomics and epigenomics detect early molecular alterations; transcriptomics and exosomal analysis capture active regulatory and intercellular signaling states; proteomics defines functional execution programs; and metabolomic, exposure-related, and microbiome-associated analyses characterize the surrounding microenvironmental landscape. Future progress will depend not simply on the accumulation of parallel biomarkers, but on standardized workflows and integrative models capable of linking these layers into a unified representation of ductal biology. Such integration will be essential if NAF is to serve as a robust proximal liquid biopsy platform for precision breast cancer risk assessment, triage, and monitoring.

Translational application of NAF multi-omics in precision breast cancer management

The growing understanding of NAF biology, together with advances in standardized collection, has shifted its clinical positioning from conventional cytologic assessment toward multi-omics interrogation of the breast ductal microenvironment. Within precision breast cancer management, this transition is most relevant in three scenarios: molecular risk stratification in pathologic nipple discharge, individualized biological risk profiling in women with high-risk or dense breasts, and proximal monitoring of treatment response, residual disease, and recurrence-related molecular change in diagnosed patients.

Application 1: precision risk stratification for pathologic nipple discharge

In the management of pathologic nipple discharge, the central clinical challenge is to distinguish patients who require surgery from those with benign disease who may be safely spared intervention. NAF multi-omics analysis addresses this problem by integrating complementary biological layers, from cellular morphology to molecular dysfunction.

Within the precision risk stratification framework for PND, cytological analysis constitutes the foundational layer of the risk alert system. Although traditional cytology has limited sensitivity for diagnosing invasive carcinoma, its value in long-term risk prediction has been robustly validated by large-scale prospective studies. A key finding is that the mere presence of epithelial cells in NAF itself serves as a significant risk indicator. The landmark 25-year follow-up study by Buehring et al. demonstrated that women with epithelial cells in their NAF had an approximately two-fold increased risk of developing BC. This risk exhibited a clear gradient, escalating with the degree of cellular atypia from "no fluid obtained" (lowest risk) to "atypical cells present" (highest risk) [49, 50]. A subsequent systematic review further confirmed that abnormal NAF cytology is an independent risk factor for BC, conferring a relative risk of 2.1—comparable to the predictive power of a positive family history [76]. Technologically, liquid-based cytology (LBC) offers superior background clarity and higher diagnostic accuracy (AUC = 0.8182) compared to conventional smear cytology (CSC), particularly when combined with ductoscopy, significantly enhancing diagnostic performance [51, 52].

Moving beyond cellular morphology, protein and glycobiological biomarkers provide a higher level of diagnostic precision. The evidence in this domain illustrates an evolution from single-marker analysis to combinatorial approaches. Among glycosylation markers, expression of the sialyl-Tn antigen (sTn) in discharge cytology demonstrates near 100% specificity, with a positive result strongly indicating malignancy and associated with aggressive, HER2-positive subtypes [77]. The Thomsen-Friedenreich antigen (TF) concentration is highest in patients with DCIS, effectively distinguishing pre-malignant from benign lesions [78]. Both TF and Tn antigens—aberrantly glycosylated cancer-associated carbohydrate antigens expressed on cell surface glycoproteins and glycolipids—show significant discriminatory power. In NAF samples from BC patients, TF and Tn were detected in 76% (19/25) and 80% (20/25) of cases, respectively. In contrast, detection rates in NAF from benign conditions or healthy controls were 0% (0/25) for TF and 4% (1/25) for Tn (P < 0.001 for both). Notably, at least one of these antigens was detected in 92% of cancer-associated NAF samples, providing a strong rationale for developing non-invasive diagnostics based on glycobiological markers [79].

Among protein biomarkers, a triple-marker panel of uPA, PAI-1, and the TF antigen demonstrates exceptional diagnostic performance, achieving near 100% accuracy in predicting cancer or precancerous lesions in women undergoing surgery, thereby serving as a robust tool for guiding surgical decisions [68]. The combination of bFGF and PSA also holds significant value: an elevation in bFGF coupled with a decrease in PSA indicates a high probability of cancer (94.1%), whereas the inverse pattern strongly suggests benign disease (90.5%) [70]. Furthermore, a characteristic signature of decreased GCDFP-15 alongside elevated AAG is particularly valuable for diagnosing DCIS in premenopausal women [64]. Emerging markers include the DJ-1 protein, whose levels in NAF are significantly elevated in the presence of carcinoma. Using a cutoff of 3.0 ng/mL, it shows a diagnostic sensitivity of 75% and specificity of 85.9%, functioning as an independent predictor [80]. BF5 (the C-terminal fragment of alpha-1-antitrypsin) is specifically elevated in the NAF of patients with BC, DCIS, or atypical hyperplasia but is undetectable in healthy controls. This elevation is associated with increased matrix metalloproteinase-7 (MMP-7) activity, suggesting its potential as a unique marker of local microenvironmental dysregulation [71].

Epigenetic analysis opens a new chapter in molecular-level diagnosis. DNA methylation profiling enables a veritable "molecular biopsy," as hypermethylated genes in BC tissue are consistently detected in matched NAF samples. Quantitative methylation-specific PCR targeting a panel of genes, including CCND2, p16, RAR-β, and RASSF1A, has been validated as a feasible and accurate methodology [81]. Although the "field effect" can sometimes attenuate the difference between the cancerous and healthy sides, the high specificity of this approach offers definitive value for identifying high-risk patients [82]. Emerging transcriptomic research has identified miR-145-5p as significantly upregulated in the PND of patients with intraductal papilloma, providing a diagnostic accuracy of 92% [59].

Taken together, these findings support a layered risk-stratification strategy for pathologic nipple discharge. Cytology provides an initial risk context, protein and glycobiological markers such as sTn and the uPA/PAI-1/TF panel improve discriminatory precision, and epigenetic assays offer molecular confirmation in selected cases. The clinical value of this framework lies in combining the contextual information of cytology with the higher specificity of molecular markers, thereby supporting more selective surgical decision-making.

Application 2: individualized risk profiling in high-risk and dense breast tissue

In women with high-risk or dense breasts, existing risk models and anatomical imaging do not directly capture the biological state of the ductal microenvironment. The contribution of NAF multi-omics lies in converting abstract risk attribution into localized biological profiling of the breast.

Analysis of the local hormonal microenvironment provides the most robust scientific foundation. A series of studies has not only established the breast as a functionally autonomous endocrine organ [31, 34] but, more critically, directly linked this local hormonal hyperexposure to carcinogenic risk. Estrogen levels in NAF have been shown to strongly predict the expression of local estrogen-regulated proteins (e.g., EGF, cathepsin D). The unique hormone-protein regulatory networks identified within NAF—such as the associations between YKL-40, bFGF, and the androgen receptor—are absent in blood, providing direct in situ evidence for estrogen-driven epithelial proliferation and carcinogenesis [37]. Therefore, measuring the hormonal profile in NAF essentially assesses the actual oncogenic pressure exerted on the target organ, offering superior value over blood tests that merely reflect systemic averages. Furthermore, the discovery of specific proteins reveals unique driving mechanisms. For instance, erythropoietin concentration is significantly higher in the NAF of BC patients than in healthy women and is produced by mammary epithelial cells themselves. Its independence from tumor TNM stage suggests it may function as an early driver in tumor initiation rather than a consequence of progression [83]. Research by Zhou et al. found that BF5 (the C-terminal fragment of alpha-1-antitrypsin) is specifically elevated in the NAF of patients with BC, DCIS, or atypical ductal hyperplasia. Its production is linked to increased MMP-7 activity in the tumor microenvironment, revealing a potential connection between local proteolytic activity and immune regulation [71].

At the protein and metabolic levels, multiple biomarkers have been validated for quantifying abnormal biological states in ductal epithelium, enabling risk stratification. Among these, PSA, a protein highly expressed in normal breast tissue, shows decreased levels in NAF as an early sign of lost epithelial differentiation. This reduction can occur at the precancerous stage and progresses with disease severity, making PSA a predictive indicator of histological malignancy preceding morphological change [69, 84, 85]. Conversely, elevated levels of bFGF reflect sustained pro-angiogenic and proliferative signaling, directly indicating a microenvironment supportive of tumor growth [70]. On a broader scale, proteomic profiling of NAF has revealed a coherent signature in cancerous breasts: significant upregulation of glycolysis-related proteins (Warburg effect) concurrent with markers of immune and stromal activation, vividly delineating the metabolic reprogramming and host response characteristics of the tumor-promoting microenvironment [66].

Oxidative stress and inflammatory markers collectively depict a procarcinogenic microenvironmental "soil." Regarding oxidative stress, biomarkers in NAF outline a clear gradient of damage from health to malignancy. Studies show that levels of protein carbonyls—irreversible markers of protein damage by reactive oxygen species—increase progressively in NAF: lowest in healthy women (median 0.37 nM/mg), significantly elevated in those with precancerous lesions (median 1.09 nM/mg), and peaking in BC patients (median 2.07 nM/mg), with average levels in cancer patients approximately 5.6 times higher than in healthy controls. This supports the hypothesis that local oxidative damage intensifies synchronously with disease progression [43]. Concurrently, the key antioxidant enzyme cytosolic superoxide dismutase 1 (SOD-1) shows an inverse trajectory in the NAF of cancer patients, highlighting NAF’s unique value in reflecting tissue-specific alterations [42]. In terms of inflammation, C-reactive protein (CRP) levels in NAF serve as an effective local indicator. Research has found that in healthy women, NAF CRP levels show a significant positive correlation with BC risk as predicted by the Gail model (p = 0.04) [44]. More importantly, Diana Lithgow et al. discovered that NAF CRP levels are significantly associated with systemic metabolic abnormalities: body fat percentage, BMI, and serum triglycerides all show significant positive correlations with NAF CRP (r = 0.29–0.34, p < 0.05), suggesting that components of metabolic syndrome, like obesity, may elevate BC risk by promoting local breast inflammation [45]. Collectively, these markers suggest that elevated risk in some individuals is associated with a measurable local state of proliferative signaling, inflammation, oxidative damage, and metabolic dysregulation, rather than with statistical risk attribution alone.

Environmental exposure and microbiome research add novel perspectives to risk assessment. Studies on transferrin and ferritin (FTN) suggest that excess free iron may contribute to BC development by inducing local oxidative stress and DNA damage [86]. Detection of aluminum revealed that its average level in the NAF of the cancer group (268.4 μg/L) was significantly higher than in the non-cancer group (131.3 μg/L) and strongly correlated with FTN levels, uncovering a link between environmental exposure, iron homeostasis disruption, and BC risk [74]. Microbiome analysis shows an enrichment of bacterial communities possessing β-glucuronidase activity in the NAF of BC patients, proposing a novel carcinogenic mechanism involving the local remodeling of estrogen metabolism [47].

Genetic and epigenetic analyses capture the earliest molecular abnormalities. Research indicates that the frequency of loss of heterozygosity (LOH) and microsatellite instability (MSI) in NAF increases significantly with the progression of breast lesions from benign to malignant, enabling non-invasive, dynamic genetic monitoring of the carcinogenic process in high-risk individuals [57]. For cytologically ambiguous cases of "mild atypia," fluorescence in situ hybridization (FISH) provides objective molecular genetic evidence, effectively distinguishing benign diploid cells from those with malignant potential (aneuploidy), thereby allowing precise risk re-stratification of borderline cases [56]. In BRCA1/2 mutation carriers, elevated methylation levels of the p16 and RASSF1A genes—particularly RASSF1A methylation—are associated with short-term BC development, offering very early molecular warning for this ultra-high-risk population [53]. At the transcriptomic level, the upregulation of miR-29c-5p in dense breasts may actively drive the high-density phenotype and associated risk by suppressing pathways such as phosphatase and tensin homolog(PTEN) [58].

Overall, these data support a biologically informed framework for risk assessment in high-risk or dense breasts. Within such a framework, hormonal and molecular drivers quantify baseline oncogenic pressure, protein and metabolic markers reflect epithelial and microenvironmental states, oxidative and inflammatory indicators capture tissue damage, and genetic, epigenetic, or microbial signals may reveal early biological deviation. The translational implication is a shift from population-level risk categorization toward more individualized biological stratification.

Application 3: monitoring therapeutic response and prognosis in diagnosed patients

In patients with established breast cancer, particularly during neoadjuvant or adjuvant treatment, accurate assessment of response and residual disease remains challenging. Imaging reflects anatomical change with temporal lag, whereas blood-based liquid biopsy may lack sensitivity in localized disease with low tumor burden. As a proximal biofluid derived from the ductal system, NAF may provide a complementary means of monitoring treatment-associated molecular change within the local microenvironment. At present, however, evidence for this application remains preliminary: large-scale prospective trials proving clinical value for routine response assessment or recurrence detection are not yet available, and current studies support only an exploratory sampling framework rather than a standardized clinical timeline. In published work, NAF has generally been assessed at baseline before treatment, during therapy for pharmacodynamic or biomarker changes, and after treatment or surgery for exploratory evaluation of residual disease or recurrence-related signals.

Dynamic monitoring of protein biomarkers offers critical metrics for efficacy evaluation. Research indicates that PSA levels in NAF show a significant inverse correlation with tumor burden, decreasing systematically with advancing disease stage, lymph node involvement, and distant metastasis. In patients undergoing neoadjuvant therapy, changes in NAF PSA levels may reflect therapeutic response earlier than radiographic imaging [85]. Similarly, uPA and its inhibitor PAI-1—prognostic markers validated by Level 1 evidence—carry prognostic significance when measured in NAF. Studies show that elevated uPA and PAI-1 levels in NAF are associated with poorer outcomes, and monitoring their dynamics during neoadjuvant treatment may help identify high-risk patients and could potentially inform therapeutic adjustment in investigational settings [67, 87]. In explorations of diagnostic and prognostic markers, the combined detection of cancer antigen 15–3(CA15-3), cancer antigen 125(CA125), carcinoembryonic antigen(CEA), and tumor-specific growth factor(TSGF) in NAF demonstrates a significantly higher positivity rate than any single-marker assay in serum or nipple discharge, highlighting NAF’s advantage in enriching tumor-associated biomarkers. While multivariate analysis showed no significant impact of this panel on overall survival, its high sensitivity provides valuable biological information for disease monitoring [88]. Furthermore, levels of tumor-specific antibodies against Mucin 1 in NAF correlate with aggressive tumor features; for instance, immunoglobulin G levels associate with triple-negative BC, and immunoglobulin A levels correlate with lymphovascular invasion and HER2-positive tumors, offering supplementary insights into tumor biology [89].

Proteomic profiling of bilateral NAF reveals prognostically significant heterogeneity. Comparative analysis using two-dimensional difference gel electrophoresis (2D-DIGE) identifies two distinct protein expression patterns: a Homogeneous Expression Profile (HomEP) and a Heterogeneous Expression Profile (HetEP). The HetEP pattern is significantly associated with increased tumor aggressiveness and worse prognosis, establishing NAF proteomics as a novel, sampling-based tool for assessing intratumoral heterogeneity and predicting clinical outcomes [90].

Pharmacodynamic monitoring further illustrates the distinct translational niche of NAF. Research confirms that agents such as the non-steroidal anti-inflammatory drug celecoxib can be delivered to the breast ducts, where they significantly reduce prostaglandin E₂ (PGE₂) levels in the NAF of high-risk postmenopausal women and newly diagnosed patients—an effect more pronounced in NAF than in plasma [91, 92]. Similarly, sulindac and its metabolites distribute into the ductal space and dose-dependently induce growth differentiation factor‑15 (GDF‑15), revealing a potential chemopreventive mechanism independent of COX‑2 inhibition [93]. These findings support the potential of NAF as a pharmacodynamic platform and suggest possible utility for identifying patients most likely to benefit from targeted interventions.

Epigenetic markers provide a sensitive means of monitoring treatment response. The persistence of DNA hypermethylation (e.g., of p16 or RASSF1A) in NAF may indicate minimal residual disease, whereas its clearance after therapy could reflect a favorable molecular response [23, 81].

Taken together, these findings support a provisional framework for NAF-based monitoring in diagnosed patients. Baseline molecular profiling may help define prognostic context, serial protein-based or proteomic assessment may capture early treatment-associated change, pharmacodynamic analyses may document local drug activity, and epigenetic markers may contribute to residual disease surveillance. Although clinical utility remains to be validated prospectively, NAF-based monitoring should currently be regarded as investigational rather than as an established tool for routine assessment of treatment response or recurrence surveillance.

In summary, NAF multi-omics analysis has the potential to address several unmet clinical needs, including more selective triage of pathologic nipple discharge, individualized biological risk assessment in high-risk populations, and proximal monitoring of treatment-related change. Realization of this potential, however, depends on standardized and reproducible methods for sample collection, processing, and analysis. The following section therefore examines the determinants of NAF procurement and recent advances in pre-analytical standardization and quality control. The three translational applications are visualized in Fig. 3. To further contextualize the role of NAF analysis within the existing diagnostic arsenal, a comparative overview of primary modalities for BC detection and risk assessment is presented in Table 2.

Fig. 3.

Translational applications of NAF multi-omics in precision BC care. This figure summarizes the clinical value of NAF multi-omics in three settings: stratifying risk in PND, assessing biological risk in high-risk or dense breasts, and monitoring treatment response in patients with diagnosed BC

Table 2.

Comparison of primary modalities for bc detection and risk assessment. This table compares four major technological modalities—imaging, histopathology (the gold standard), blood-based liquid biopsy, and NAF liquid biopsy—across key dimensions, including their core impact on clinical decision-making, limitations, advantages, sensitivity/specificity, and fundamental value proposition

Detection modality	Core value proposition	Sensitivity & specificity vs. gold standard	Advantages	Limitations	Core impact on clinical decision-making	Ref
Imaging	Anatomical "Detector"	Mammography (DM): • Sensitivity: ~ 77% (significantly reduced in dense breasts) • Specificity: ~ 94% • Interval Cancer Rate: ~ 30% • Overdiagnosis: Accounts for 11%-19% of screen-detected cancers	• Non-invasive • Highly accessible • Provides clear macroscopic localization • MRI offers extremely high sensitivity for dense breasts/high-risk individuals	• Density Blind Spot: Sensitivity greatly impaired by breast density • High risk of interval cancers • High false-positive rate, significant overdiagnosis • Limited resources (MRI), high cost	Abnormal result → Recall and recommendation for biopsy. Serves as the foundation for screening and localization but is insensitive to early biological risk	[5, 8]
		Ultrasound (US): • Can detect an additional 0.7–9.4 cancers per 1000 screenings, but with a significant increase in false-positive rate				[120]
		MRI: • Sensitivity: ~ 100% • Specificity: ~ 87–88% • False-Positive Recall Rate: Addition of MRI increases false-positive recall rate by nearly 25%				[9, 121]
Pathology (Gold Standard)	Diagnostic "Gold Standard" & "Final Arbiter"	Sensitivity & Specificity: Theoretically 100%	• Ultimate basis for diagnosis • Extremely high specificity	• Invasive • Subject to sampling error • Not suitable for repeated dynamic monitoring • Results lag behind molecular changes	Provides definitive diagnosis, guiding treatment plans. Not applicable for risk warning, dynamic monitoring, or detection of early molecular events	——
Blood-Based Liquid Biopsy	Systemic Tumor Burden "Monitor"	Early/Localized BC: • Low sensitivity (due to very low ctDNA abundance)	• Completely non-invasive • Reflects systemic tumor burden and clonal evolution • Valuable for monitoring advanced disease efficacy, prognosis stratification, and guiding immunotherapy [13]	• Insufficient sensitivity for early/localized disease • Cannot reflect breast-specific local microenvironmental events • Standardization pending	A key tool for monitoring efficacy and recurrence in advanced disease. Has limited utility in early diagnosis and local risk assessment, cannot replace local information	——
		Advanced/Metastatic BC: • Sensitivity/Specificity significantly improved, suitable for efficacy monitoring and resistance tracking				[12, 15]
NAF Liquid Biopsy	Breast Ductal Local Microenvironment "Insight Lens"	Traditional Cytology (negated for screening): • Sensitivity: ~ 64% • Specificity: ~ 97% • Inadequate Sample Rate: Up to 38.97%	• Target-Organ Specificity: Directly measures local breast driver signals (e.g., autonomous hormonal synthesis and functional networks [34, 37]; unique metabolome [36]) • Minimally invasive and generally well tolerated[113]) • Amenable to repeated dynamic monitoring • Multi-Omics Integration Potential: Overcomes limitations of single dimensions, providing a panoramic biological profile	• Sampling success rate influenced by genetic, hormonal, and age-related factors [94, 95, 99, 101] • High Pre-analytical QC Requirements: Sample appearance significantly impacts results [117] • Most multi-omics biomarkers are in the research phase, requiring validation of clinical utility in prospective trials	Optimizes decision-making in three high-value scenarios: 1. PND Risk Stratification: Integrates molecular markers to help distinguish benign from malignant, supporting more selective surgical decision-making 2. Insight for High-Risk/Dense Breasts: Provides individualized biological risk profiles (hormones, oxidative stress, inflammation, etc.), translating abstract risk probabilities into concrete biological metrics to guide precision prevention 3. Therapeutic Monitoring for Diagnosed Patients: Provides a proximal, dynamic window into molecular response, potentially assessing treatment efficacy earlier than imaging	[21, 22]
		Multi-Omics Models (High-Potential Direction): • uPA/PAI-1/TF combined model: Near 100% accuracy in predicting cancer/precancerous lesions in women requiring surgery • sTn antigen: Near 100% specificity when positive on cytology, strongly indicative of malignancy • PSA decrease + bFGF increase: Combined prediction of cancer probability 94.1%				[68, 70, 77]

BC, Breast Cancer; bFGF, Basic fibroblast growth factor; ctDNA, Circulating tumor DNA; DM, Digital mammography; MRI, Magnetic resonance imaging; NAF, Nipple aspirate fluid; PAI-1, Plasminogen activator inhibitor-1; PND, Pathologic nipple discharge; PSA, Prostate-specific antigen; QC, Quality control; sTn, Sialyl-Tn antigen; TF, Thomsen-Friedenreich antigen; uPA, Urokinase-type plasminogen activator; US, Ultrasound

Performance values are approximate and context-dependent, derived from representative studies rather than direct head-to-head comparisons across modalities

The key to clinical translation: standardized collection and quality control of nipple aspirate fluid

As established, the clinical utility of NAF is predicated on its accurate reflection of the local ductal microenvironment. Consequently, the success of NAF collection—governed by a complex interplay of intrinsic and extrinsic factors—is paramount. A thorough understanding of these determinants is essential for identifying suitable candidates, optimizing procurement protocols, and accurately interpreting unsuccessful (“dry tap”) attempts.

In practice, NAF collection refers to aspiration-based procurement of ductal fluid from non-lactating breasts and is conceptually distinct from the passive observation of spontaneous nipple discharge, although both materials may be analyzed in relevant clinical contexts. For patient counseling, the procedure can generally be described as a gentle suction-based sampling method applied to the nipple-areolar complex rather than as a tissue biopsy. It is typically performed in the outpatient setting and is generally well tolerated, although sample yield varies among individuals.

Intrinsic biological factors establish the baseline secretory capacity. Genetic predisposition is a key determinant. Early studies observed higher NAF yield in White compared to Asian women, a disparity linked to the wet-type earwax phenotype regulated by the ABCC11 gene [94–96]. This correlation is biologically significant given the shared developmental origin of mammary and apocrine sweat glands. Notably, among U.S.-born Asian women, the wet earwax phenotype is associated with atypical ductal epithelial proliferation, suggesting a confluence of genetic and environmental influences on breast physiology and disease susceptibility [95]. Reproductive history and endocrine status further modulate secretory activity. Younger age, earlier menarche, and a history of parity and lactation are generally associated with higher NAF yield [97, 98]. Within the endocrine axis, serum prolactin (PRL) is a primary physiological regulator. Crucially, BC patients exhibit a heightened responsiveness, yielding aspiratable NAF at lower serum PRL levels than healthy women, indicating amplified local PRL sensitivity within the cancerous microenvironment [99].

This intrinsic secretory potential is modulated by exogenous factors, particularly diet and pharmacologic agents. Regarding dietary influences, high-fat and high-lactose intake are robust predictors of successful NAF collection, an association most pronounced among Black women [100, 101]. Conversely, diets rich in fruits, vegetables, and antioxidants such as carotenoids correlate with a lower probability of obtaining epithelial cells in NAF [102]. The impact of soy isoflavones presents a complex picture: while short-term, high-dose intake may stimulate breast secretion and induce epithelial cell hyperplasia, most studies conducted at intake levels representative of traditional Asian diets have observed no significant effect on NAF yield or local estrogenicity [103–105]. A dedicated study found that high soy intake did not increase the number of hyperplastic epithelial cells in NAF; rather, it showed a trend toward reducing cytologic atypia (24% vs. 14%, P = 0.06), suggesting a potential mild ameliorative effect on breast cell status at habitual dietary consumption levels [106].

Pharmacologic interventions can profoundly alter ductal physiology. For example, oral contraceptives suppress ovarian function, thereby reducing both serum and NAF sex hormone levels [31]. In high-risk women undergoing aggressive risk-reduction strategies—such as bilateral oophorectomy or selective estrogen receptor modulator (SERM) therapy—the mammary ducts often become hypoactive and “dry” due to profound estrogen deprivation. This may limit the applicability of NAF-based cytological monitoring in this specific population [107].

The clinical interpretation of a “dry tap” requires nuanced assessment. It should not be simplistically equated with low risk. Evidence shows that random periareolar fine-needle aspiration (RPFNA) in women with unsuccessful NAF attempts can still detect atypical hyperplasia in a subset of individuals [108]. Thus, a dry tap more likely reflects physical ductal obstruction or intrinsically low basal secretory function rather than the definitive absence of epithelial biological risk.

Transitioning NAF from a research tool to a reliable clinical platform hinges on establishing a standardized, reproducible, and well-tolerated collection protocol. Early methods, reliant on modified breast pumps or manual aspiration, were limited by operator dependence [109]. A key technical advancement was the introduction of oxytocin nasal spray. By stimulating contraction of ductal myoepithelial cells, oxytocin effectively promotes fluid release and improves collection success [110, 111]. Further standardization has been achieved with automated devices (e.g., the HALO system), which use programmed pressure cycles to reduce, although not entirely eliminate, operator dependency and to improve between-center consistency. Their feasibility, safety, and high patient acceptance in large-scale screening studies of asymptomatic women have established a foundation for population-based research [112]. Nevertheless, full standardization has not yet been achieved across studies, as collection methods, use of oxytocin, adequacy criteria, and downstream analytical workflows still vary between centers.

Although NAF sampling is conceptually minimally invasive, its clinical translation depends on standardization at multiple levels, including aspiration technique, oxytocin use, repeatability, specimen adequacy criteria, and pre-analytical handling. Recent reviews have explicitly identified protocol heterogeneity as one of the main reasons why comparisons across NAF studies remain difficult [39–41, 61]. Importantly, patient-centered data indicate that nipple fluid aspiration is generally well tolerated and is experienced as less discomforting than mammography or breast MRI by most participants, supporting its feasibility as an outpatient sampling procedure [113]. In addition, prospective multicenter data demonstrate that repeated nipple fluid aspiration is feasible and safe, which is particularly relevant if NAF is to be positioned as a longitudinal monitoring platform rather than a one-time diagnostic adjunct [114]. Nevertheless, operator technique, variable yield, and low-volume samples still introduce center-to-center variability, underscoring the need for harmonized protocols before broad clinical implementation.

Following collection, rigorous pre-analytical quality control is critical for ensuring the accuracy and reproducibility of downstream molecular analyses. Sample appearance—specifically color and turbidity—serves as a vital, low-cost quality indicator. Darker NAF, particularly bloody or brown samples, is consistently associated with higher concentrations of bioactive molecules and independently correlates with BC presence and progression risk [62, 115, 116]. Furthermore, miRNA expression profiles differ substantially between clear and turbid samples, underscoring the necessity of documenting visual characteristics in standardized protocols [117].

Addressing the challenge of NAF instability at ambient temperature, the Guthrie card (dried blood spot) technique has been successfully adapted for NAF collection and storage. This approach facilitates room-temperature preservation and convenient transport, providing a crucial technical foundation for large-scale, multi-center studies and potential future self-collection models [118]. Notably, the menstrual cycle exerts minimal overall influence on the NAF proteomic profile, greatly simplifying clinical sampling by obviating the need for phase-specific collection in premenopausal women [119].

In summary, NAF procurement is a physiological process regulated by a complex interplay of genetic, endocrine, environmental, and pharmacological factors. Modern collection and quality control systems are now adequately equipped to manage this complexity. The implementation of this standardized pre-analytical framework is essential, as it ensures that NAF can serve as a stable and reliable multi-omics liquid biopsy platform, forming the robust foundation necessary for its application in precision risk stratification and disease monitoring. Ongoing efforts to translate NAF research into clinical practice are exemplified by selected clinical trials listed in Table 3.

Table 3.

Selected clinical trials utilizing NAF

Title of the clinical study	Registration number	Characteristics of participants	Aims related to nipple aspirate fluid samples	Time Period	Country
Breast Imaging Screening Studies in Women at High Genetic Risk of Breast Cancer: Annual Follow-up Study	NCT00006421	Population: Women 25–56 years old, including known BRCA1/2 deleterious mutation carriers or their first-/second-degree relatives Key Criteria: At least one non-irradiated breast, no active infection/inflammation, and able to provide informed consent	1. To use NAF and/or Breast Duct Lavage (BDL) to obtain epithelial cell samples 2. To correlate cytologic findings with imaging findings 3. To gather data on disease progression in BRCA1/2 carriers via NAF/BDL analysis and develop biomarkers	2000–2025	USA
Morphometry of Breast Cells in Ductal Lavage	NCT00259597	Population: Women aged ≥ 18 years with a previous biopsy-confirmed diagnosis of proliferative breast disease Key Criteria: At least 1 month post-biopsy, at least 3 months post any therapy like tamoxifen, and in good general health. Excludes pregnant/lactating women and those with specific chronic diseases	1. To quantify biomarkers related to oxidative stress and dietary intake in NAF obtained prior to ductal lavage 2. To characterize the nuclear morphology of breast epithelial cells in lavage fluid using quantitative image analysis 3. To explore the association between oxidative stress levels in NAF and epithelial cell nuclear morphology	2005–2012	USA
Phase IB Sulindac Study for Women at High Risk for Breast Cancer	NCT00245024	Population: Women at high risk for breast cancer (defined by Gail model, history of LCIS/DCIS, strong family history, or prior personal history), able to produce ≥ 5 µL of NAF Key Criteria: Negative mammogram within the past 10 months	1. To determine the partitioning of sulindac and its metabolites in breast tissue by measuring their levels in NAF after 6 weeks of therapy 2. To determine the effects of the drug on prostaglandin (PGE₂), NAG-1, and C-reactive protein (CRP) levels in NAF	2005–2013	USA
Study of Manual Nipple Fluid Aspiration in Post-Menopausal Women	NCT00291083	Population: Women at high risk for breast cancer, older than 18 years Key Criteria: Excludes those who are pregnant, within 12 months of breastfeeding, or on anticoagulants	1. To examine NAF for secretions that may serve as high-risk indicators for breast cancer 2. To compare hormone levels (e.g., estrogen, prolactin) in NAF with those in serum	2006–2012	USA
Phase IIB Trial of G-2535 (Unconjugated Isoflavones-100) in Women at High Risk for Breast Cancer	NCT00290758	Population: Women at high risk for breast cancer (assessed by Gail/Claus models, or with a history of breast cancer, atypical hyperplasia, LCIS, or BRCA mutation) Key Criteria: No soy intolerance, negative mammogram within 6 months, and any prior cancer therapy completed at least 1 year ago	1. To evaluate the effect of genistein on the breast local microenvironment by analyzing NAF, including measuring changes in concentrations of EGF, estradiol, Cathepsin D, and ps2 protein	2006–2017	USA
Effects of Soy on Estrogens in Breast Fluid and Urine	NCT00513916	Population: Healthy, premenopausal women with regular menstrual cycles Key Criteria: No cancer diagnosis, and not taking birth control pills or other hormones	1. To examine the effects of two daily servings of soy on estrogen levels in NAF and serum 2. To investigate cytologic patterns of epithelial cells from NAF in relation to soy intake 3. To compare estrogen levels in NAF and serum measured during the same luteal phase	2007–2013	USA
Pre-surgical Phase IIb Trial of Transdermal 4-Hydroxytamoxifen vs. Oral Tamoxifen in Women With Ductal Carcinoma In Situ of the Breast	NCT00952731	Population: Women newly diagnosed with hormone receptor-positive DCIS, age ≥ 18 Key Criteria: Normal organ function, not pregnant, and willing to avoid sun exposure to breasts	1. To collect NAF samples on the day of surgery to compare concentrations of tamoxifen and its active metabolites within the breast ducts between treatment groups (topical gel vs. oral), and to explore their relationship with estrogen response markers	2009–2015	USA
Human Breast Tissue Bioavailability of Topically Applied Limonene	NCT01459172	Population: Females 18–65 years of age, with both breasts intact and normal organ function, able to produce NAF Key Criteria: Willing to avoid citrus products. Excludes those with cancer history, skin disorders, or relevant allergies	1. To evaluate the breast tissue bioavailability of limonene by measuring its levels in NAF after 4 weeks of intervention with a limonene-containing massage oil	2011–2016	USA
A Single-Arm, Multi-Laboratory Clinical Trial of the ForeCYTE Breast Aspirator for Sample Collection, Processing, and Cytological Testing of Nipple Aspirate Fluid	NCT02218385	Population: Healthy women aged 20–75 years from whom bilateral NAF samples can be obtained Key Criteria: Non-pregnant, non-lactating, no open breast wounds, and in good general health	1. To evaluate the performance (collection, fixation, transportation) of the ForeCYTE device across multiple independent laboratories 2. To verify that the specimen acceptability rate at each laboratory is at least 90%	2014–2016	USA
Identifying the microRNA Fingerprint in NAF, Serum, and Tissue in Patients With Ductal Carcinoma In Situ (DCIS) or Invasive Breast Cancer	NCT02127073	Population: Patients with a clinical diagnosis of DCIS or invasive breast cancer who are candidates for breast-conserving surgery or mastectomy Key Criteria: Excludes pregnant women, those with prior breast cancer diagnosis, or a history of adverse reaction to oxytocin	1. To determine the feasibility of microRNA profiling using tissue, serum, and NAF (with intranasal oxytocin to enhance yield) 2. To evaluate the effectiveness of intranasal oxytocin in increasing NAF yield in breast cancer patients	2014–2023	USA
The Breast Duct: Pilot Study of Genomic Sequencing of Exfoliated Ductal Cells Obtained Through Endoscopy and Lavage	NCT02121873	Population: Females with at least one intact nipple, > 18 years old. If over 50, a normal mammogram within 12 months prior to enrollment is required Key Criteria: Excludes those pregnant, lactating, or receiving chemotherapy within the past 12 months, and those with surgery near the nipple	1. To utilize nipple aspiration, ductal lavage, and ductoscopy to collect exfoliated ductal epithelial cells 2. To perform whole-genome and transcriptome sequencing on these cells to identify early somatic mutations and investigate clonality across ducts	2014–2023	USA
Phase II Study of Metformin for Reduction of Obesity-Associated Breast Cancer Risk	NCT02028221	Population: Overweight or obese (BMI ≥ 25) premenopausal women aged 21–54 with at least one criterion for metabolic syndrome Key Criteria: Non-diabetic, regular menses for the past 6 months, negative mammogram within 12 months (if ≥ 50 years old)	1. To assess the effect of metformin on the metabolic state of the breast local microenvironment by analyzing changes in the metabolomic profile of NAF samples	2014–2023	USA
Effects of Omega-3 Fatty Acids on Risk Factors for Breast Cancer in Pre-menopausal Women	NCT02816125	Population: Healthy, menstruating, pre-menopausal women Key Criteria: Non-pregnant, non-lactating, non-smoking, and not on birth control pills	1. To analyze the fatty acid composition in NAF cells to examine if omega-3 supplementation replaces omega-6 fatty acids 2. To measure changes in estrogen levels in NAF following omega-3 supplementation and a low-fat diet	2016	Canada
Phi29 Motor Nanopore for Single Molecule Sensing: Breast Nipple Aspirate Fluid	NCT03715959	Groups: • Breast Cancer Group: Patients currently diagnosed with breast cancer • Healthy Control Group: No history of breast cancer and normal imaging findings General Criteria: All participants are non-lactating women ≥ 40 years old	1. To collect NAF samples from non-lactating women ≥ 40 years old 2. To analyze NAF biomarkers using a novel protein nanopore-based platform and evaluate its diagnostic efficacy	2018–2025	USA
A Study on Combined Low-Pass Whole-Genome and Methylome Testing of Bloody Nipple Discharge Specimens for Benign-Malignant Differentiation	NCT07250204	Population: Adults ≥ 18 years with spontaneous, unilateral, single-duct bloody or serosanguinous nipple discharge already scheduled for diagnostic surgery Key Criteria: Excludes physiologic discharge, breast infection, prior breast cancer treatment, and pregnant or lactating patients	1. To develop and validate a novel laboratory test analyzing NAF DNA (copy number variation, fragmentomics, genome-wide methylation) to distinguish benign from malignant lesions 2. To estimate the potential clinical impact of the model in avoiding unnecessary surgeries	2025–2026	China

BDL, Breast Duct Lavage; BMI, Body Mass Index; BRCA1/2, breast cancer susceptibility genes 1 and 2; CRP, C-reactive protein; DCIS, Ductal carcinoma in situ; EGF, Epidermal growth factor; LCIS, Lobular carcinoma in situ; NAF, Nipple aspirate fluid; NAG-1, NSAID-activated gene 1; PGE₂, Prostaglandin E₂

Challenges and future directions: from translational hurdles to breakthroughs in biological understanding

Despite its promising potential, the clinical translation of NAF multi-omics analysis faces significant challenges. The foremost issue is an insufficient evidence base. Although biomarkers like DJ-1 and BF5 have demonstrated diagnostic promise in individual studies, their adoption into practice requires standardized validation and performance confirmation in prospective cohorts. A more profound challenge is the field’s critical lack of high-grade evidence for clinical utility. There is currently no definitive proof that clinical decisions guided by NAF molecular profiling—such as deferring surgery or adjusting surveillance strategies—can improve long-term patient outcomes (e.g., quality of life, survival) without increasing diagnostic risk and while demonstrating cost-effectiveness. Generating this evidence is the definitive requirement for NAF’s transition from research to clinical practice. This is particularly true for treatment-response monitoring and recurrence-related applications, where large-scale clinical validation remains absent.

Secondly, enduring technical limitations persist. While standardized collection and post-processing protocols have advanced considerably, they have not yet achieved the uniformity and broad applicability of blood-based assays. Challenges include operator dependency, interpretation of ‘dry tap’ results, and the multi-omics analysis of minute, heterogeneous samples. More fundamentally, biological constraints also limit the standalone diagnostic power of NAF. Because many samples represent pooled secretions from ducts with uncertain or mixed origins, sensitivity may be reduced for focal, non-secreting, or spatially restricted lesions. In addition, unlike imaging or biopsy-based approaches, NAF cannot localize a lesion to a specific duct or quadrant, limiting its direct use for targeted diagnosis or surgical planning [62]; at present, it should therefore be regarded as complementary to, rather than a replacement for, tissue pathology. A further translational challenge is that concordance between NAF-derived molecular signals and paired tissue pathology remains context-dependent rather than absolute. Recent overview papers stress that NAF should not be judged by a simplistic one-to-one expectation of biopsy concordance, because the fluid may represent pooled biological output from one or more ducts and may capture field effects or microenvironmental alterations that are spatially broader than a focal tissue sample [39–41, 61]. This also means that some degree of discordance with biopsy is biologically plausible, particularly for focal lesions, mixed-duct sampling, or molecular changes that precede overt histopathologic transformation. Therefore, the central future task is not merely to maximize analytical sensitivity, but to define in which clinical scenarios NAF-derived information is complementary to tissue pathology and imaging, and when it can improve risk triage or monitoring without introducing diagnostic ambiguity.

Looking ahead, progress hinges on a paradigm shift from being "technology-driven" to becoming fundamentally "clinically question-driven." Technologically, next-generation analytical methods must focus on resolving the spatial and cellular origins of NAF components. Techniques such as single-exosome analysis or single-cell sequencing will be critical to deconvolute the current population-averaged signal and extract disease-specific information. In parallel, leveraging artificial intelligence (AI) to construct dynamic systems biology models is essential. This goes beyond simple data aggregation; it aims to build a “digital twin” of the ductal microenvironment. By synthesizing NAF multi-omics data with radiological, histopathological, and longitudinal clinical data, such a model could simulate disease trajectories and enable mechanism-based risk prediction.

However, the most pivotal future direction is the initiation of rigorously designed, NAF-stratified clinical decision trials. Such studies must directly validate whether, compared to standard management pathways, the use of NAF molecular profiling to triage PND patients (e.g., avoiding diagnostic surgery for low-risk individuals) or to implement risk-adapted surveillance for high-risk cohorts offers equivalent safety and superior cost-effectiveness. The definitive clinical value of NAF will only be anchored upon the success of such trials.

Ultimately, the highest aspiration for the NAF platform lies in its potential to catalyze a novel interventional paradigm. As NAF analysis advances towards precisely delineating local driver pathways, BC prevention and management could evolve from systemic, whole-body drug administration to precision modulation of the local microenvironment. This could include the development of targeted intraductal therapies or localized immunomodulation strategies. Thereby, NAF would be transformed from a passive diagnostic tool into an active interventional guidance system, providing a scientific foundation for achieving the ultimate goal of BC prevention.

Conclusions: toward a new era of precision medicine mediated by the ductal microenvironment

This review extends beyond presenting the prospective applications of a novel liquid biopsy; it advocates a fundamental reorientation of the paradigm for precision risk management in BC. Confronting the limitations of current diagnostic technologies in providing early insight and tissue specificity, we have systematically established the rationale for redefining NAF as a proximal liquid biopsy platform of the breast ductal system. This reconceptualization is anchored in the recognition of the mammary duct as a functionally segregated local endocrine and immune microenvironment. NAF serves as its direct molecular efflux and functional readout, providing, for the first time, a non-invasive sampling approach to interrogate the biological dialogue within the TDLU—from the cumulative effects of local hormone synthesis to the dynamic crosstalk within epithelial, immune, and microbial networks.

Building upon this premise, we have delineated an integrative framework for applying NAF multi‑omics data across pivotal clinical scenarios. In managing PND, it transforms decision‑making from a rule of “morphologic abnormality mandates surgery” to a principle of “molecular‑function‑based risk stratification.” For women with high‑risk or dense breasts, it advances risk assessment from abstract population‑based probabilities to concrete biological phenotyping of the individual’s ductal niche. For diagnosed patients, it opens a proximal therapeutic monitoring window, capturing pharmacodynamic and microenvironmental responses beyond anatomical imaging.

Yet, the ultimate potential of the NAF platform may reside in reshaping our understanding of BC biology itself. By systematically decoding the mechanisms governing homeostasis and dysregulation within this local microenvironment, we may address a pivotal question: why does carcinogenesis initiate in specific ducts despite shared systemic risk profiles? This knowledge could steer BC control from a reactive strategy centered on “early detection” toward a proactive era focused on “microenvironmental precision prevention.”

Therefore, overcoming the challenges of standardization and clinical validation is not merely about implementing a new test. It is fundamentally about accessing a deeper order of breast tissue biology, establishing the essential framework for achieving genuine precision in BC prevention and therapy.

Abbreviations

2-DE

Two-dimensional gel electrophoresis

2D-DIGE

Two-dimensional difference gel electrophoresis

AAG

Alpha-1-acid glycoprotein

AI

Artificial Intelligence

BC

Breast Cancer

bFGF

Basic fibroblast growth factor

BMI

Body Mass Index

CA125

Cancer Antigen 125

CA15-3

Cancer Antigen 15–3

CEA

Carcinoembryonic antigen

CRP

C-reactive protein

CSC

Conventional smear cytology

ctDNA

Circulating tumor DNA

DCIS

Ductal carcinoma in situ

ddPCR

Droplet digital PCR

DM

Digital mammography

DNA

Deoxyribonucleic acid

E2

Estradiol

EGF

Epidermal growth factor

ELISA

Enzyme-linked immunosorbent assay

FISH

Fluorescence in situ hybridization

FTN

Ferritin

GCDFP-15

Gross cystic disease fluid protein-15

GDF-15

Growth differentiation factor-15

HetEP

Heterogeneous Expression Profile

HSP90α

Heat shock protein 90α

ICP-MS

Inductively coupled plasma mass spectrometry

IL-6

Interleukin-6

LBC

Liquid-based cytology

LCIS

Lobular carcinoma in situ

LC–MS/MS

Liquid chromatography-tandem mass spectrometry

LOH

Loss of heterozygosity

miRNA

MicroRNA

MMP-7

Matrix metalloproteinase-7

MRI

Magnetic resonance imaging

MSI

Microsatellite instability

mtDNA

Mitochondrial DNA

NAF

Nipple aspirate fluid

PAI-1

Plasminogen activator inhibitor-1

PCR

Polymerase chain reaction

PGE₂

Prostaglandin E₂

PND

Pathologic nipple discharge

PRL

Prolactin

PSA

Prostate-specific antigen

PTEN

Phosphatase and Tensin Homolog

QC

Quality control

qMSP

Quantitative methylation-specific PCR

RASSF1A

Ras association domain family member 1A

RPFNA

Random periareolar fine-needle aspiration

RT-qPCR

Reverse transcription quantitative PCR

SELDI-TOF

Surface-enhanced laser desorption/ionization time-of-flight

SERM

Selective estrogen receptor modulator

sTn

Sialyl-Tn antigen

SOD-1

Superoxide dismutase 1

TDLU

Terminal ductal lobular unit

TF

Thomsen-Friedenreich antigen

TSGF

Tumor-specific growth factor

uPA

Urokinase-type plasminogen activator

US

Ultrasound

YKL-40

Chitinase-3-like protein 1

Author contributions

SG wrote the original draft and participated in review & editing. JL contributed to writing review & editing, supervised the study, acquired funding, and conducted formal analysis. HZ participated in writing review & editing. XW wrote the original draft, performed visualization and investigation, conducted formal analysis, and contributed to conceptualization. All authors read and approved the final manuscript.

Funding

This research was funded by the Scientific Research Project of Hubei Provincial Health Commission (WJ2025M107), the Natural Science Foundation of Hubei Province, China (Grant No.2023AFB493), the National Key Clinical Specialty Construction Project Fund of the Breast Center, Hubei Cancer Hospital (Grant No. HBCHBCC‑A02), and Scientific Research Projects of Hubei Cancer Hospital (2024HBCHYN11) awarded to J. Liu; the Hubei Province Medical Youth Top Talent Project (EWT [2023] 65, Oncology), the Talent Project of Hubei Cancer Hospital (2025HBCHHHRC005), Scientific Research Project of Hubei Cancer Hospital (2025HBCHYN24), and the Wu Jieping Fund Project (320.6750.2024‑21‑5 and 320.6750.2025‑21‑13) awarded to H. Zheng; and the Hubei Chen Xiaoping Science and Technology Development Foundation Youth Science Special Fund (CXPJJH123001‑2320), the Subject of Biomedical Research Center of Hubei Cancer Hospital (2022SWZX09), the Talent Project of Hubei Cancer Hospital (2025HBCHLHRC002), and the 2024 Chutian Talents Program Special Fund (CTYC002) awarded to X. Wu.

Data availability

Data sharing does not apply to this article as no new datasets were generated or analyzed during the current study. All discussed information is based on previously published literature.

Declarations

Ethics approval and consent to participate

Not applicable. This is a review article and does not report on any new studies involving human participants, animal subjects, or patient data.

Consent for publication

Not applicable. This review does not contain any person’s data in any form.

Competing interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.