Introduction

Breast cancer is the most commonly diagnosed malignancy among women worldwide and represents a growing global public health concern, especially in low- and middle-income countries. In 2020, it accounted for approximately 2.3 million new cases and 685,000 deaths, with projections indicating a continued increase in incidence by 2040 (World Health Organization, 2023)1. This global increase reflects demographic shifts, changes in reproductive behavior, and heightened exposure to environmental and lifestyle-related risk factors.

In Pakistan, breast cancer is the most prevalent cancer among women, accounting for 34.6% of all female malignancies and the highest incidence rate in Asia2. Projections estimate a 60.7% increase in the number of breast cancer cases in Karachi by 2025 compared with 2015, with an alarming 130.6% increase anticipated among women aged 30–34 years3,4. This trend is driven by a complex interplay of demographic expansion, genetic predisposition, and socioeconomic barriers to early detection and treatment.

In the Pakistani population, significant risk factors include advanced age, early onset of menarche, nulliparity, obesity, physical inactivity, and low socioeconomic status5,6. Genetic predispositions, such as BRCA mutations and high rates of consanguinity, further contribute to the disease burden, yet a significant portion of familial risk remains unexplained7. Socioeconomic and cultural determinants, particularly stigma, limited awareness, and inadequate access to diagnostic services, significantly contribute to late-stage presentation and elevated mortality rates6.

Mammographic density (MD), defined as the proportion of fibroglandular to fatty tissue in the breast, is both a major risk factor and barrier to detection. A high MD is associated with a four- to sixfold increase in breast cancer risk8,9 and reduces mammographic sensitivity from 85% in fatty breasts to less than 60% in extremely dense breasts10,11,12,13. This “masking effect” leads to delayed diagnosis14,15,16. This reduced sensitivity complicates screening accuracy and treatment planning.

The prevalence of high MD is notably elevated among Pakistani women. One study reported that 56.9% of women had BI-RADS category 3 or 4 density17, whereas another reported a 62.4% prevalence among women under 45 years of age18. Among patients with breast cancer, 76.7% were classified as BI-RADS category B or C19. Risk factors for high MD include younger age, lower BMI, hormonal exposure, and a history of benign breast disease20,21,22. Biologically, a high MD reflects an abundance of stromal and epithelial tissue, which can promote tumorigenesis through mechanical stiffness, collagen deposition, and a proinflammatory microenvironment23,24. Inflammatory mediators such as MCP-1, TNF-α, and TGF-β are elevated in dense breast tissue, potentially contributing to local immune activation and DNA damage25. Genetic factors, such as polymorphisms in estrogen metabolism and inflammatory pathways, are also implicated in regulating MD26.

Environmental and trace elements have been investigated in relation to breast cancer. Se, an essential micronutrient, has antioxidant, anti-inflammatory, and antiproliferative effects and may modulate oxidative stress and estrogen signaling27,28,29,30, although these findings are context dependent. However, few studies have examined Se and Hg together in relation to breast cancer stage or mammographic characteristics. In contrast, Hg is a known toxicant capable of generating reactive oxygen species (ROS), inducing DNA damage, and disrupting endocrine signaling. Hg exposure has been associated with increased cancer risk, although data specific to breast cancer are still emerging31,32. A significant biological interaction occurs between Se and Hg, wherein Se can form biologically inert complexes with Hg, potentially neutralizing its toxicity33,34,35,36. This antagonism highlights the need to study these elements in tandem, especially in populations with elevated environmental exposure. Nevertheless, evidence from South Asia, particularly Pakistan, remains limited, and most available studies have focused solely on serum concentrations without assessing mammographic density or hematologic parameters. Se exerts its biological effects primarily through selenoproteins that maintain cellular redox homeostasis. In particular, glutathione peroxidase-4 (GPX4) detoxifies membrane lipid hydroperoxides and thereby restrains ferroptosis, a cell death program increasingly implicated in breast cancer biology37. Similarly, thioredoxin reductases (TrxR1/TrxR2) regenerate the thioredoxin system to support nucleotide synthesis and antioxidant defense38. In contrast, Hg binds with high affinity to selenol and thiol groups and directly inhibits these enzymes, most notably TrxR—thereby weakening peroxide detoxification and amplifying oxidative stress39. Furthermore, systemic selenium delivery depends on selenoprotein P (SELENOP) and its receptors LRP8/ApoER2 and megalin/LRP2; disruption of this axis diminishes selenoprotein activity40. Finally, because Hg can sequester Se in inert Hg–Se complexes, the Se: Hg balance is biologically relevant; however, a simple “>1:1 molar ratio” rule is an oversimplification and does not guarantee protection in humans41. Se and Hg interact through complex biochemical mechanisms that influence oxidative balance and detoxification. Mercury binds to selenol and thiol groups, inhibiting antioxidant enzymes such as thioredoxin reductase (TrxR) and glutathione peroxidase (GPx), whereas Se mitigates toxicity by forming inert Hg–Se complexes that promote excretion (Fig. 1).

Fig. 1
figure 1

Selenium–mercury antagonism relevant to this study (author-created). First, Hg can bind cysteine/selenocysteine sites in TrxR and GPx, reducing enzyme activity and antioxidant capacity. Second, Se can form inert Hg–Se complexes, supporting sequestration and excretion. Third, when the Se concentration is low, Hg tends to accumulate, and oxidative stress increases. Together, these interactions explain why the Se: Hg balance is biologically linked to our endpoints (stage, mammographic density, and hematology). Abbreviations: GPx = glutathione peroxidase; TrxR = thioredoxin reductase; Se = selenium; Hg = mercury; Sec = selenocysteine; GSH = reduced glutathione. Adopted from65.

Full size image

Guided by this framework, we aimed to address these gaps. We measured Se and Hg in blood and scalp hair from women with breast cancer and age-matched healthy controls in Sindh, Pakistan, and evaluated associations with (i) clinical stage (T-I to T-IV), (ii) mammographic density (BI-RADS categories), and (iii) hematologic indices (hemoglobin, red blood cell [RBC] count, hematocrit, MCV, MCH, MCHC, ferritin, and TIBC). We hypothesized that lower Se and higher Hg would be associated with a more advanced stage, greater mammographic density, and less favorable hematologic profiles, which is consistent with impaired selenoenzyme activity and heightened oxidative stress.

Materials and procedures

Study design and ethics

This case‒control study was conducted between 2021 and 2023 to evaluate the associations of Se and Hg concentrations in biological samples with breast cancer progression and mammographic density. A total of 285 histologically confirmed female breast cancer patients and 215 healthy age-matched women were recruited from hospitals and screening centers across Sindh Province (Fig. 2). Ethical approval for this study was obtained from the National Center of Excellence in Analytical Chemistry (NCEAC), University of Sindh. Written informed consent was obtained from all participants prior to inclusion. Certified hospital oncologists and radiologists performed the clinical evaluations, diagnoses, and imaging assessments, whereas the analytical procedures were conducted at the NCEAC laboratory.

Fig. 2
figure 2

Distribution of study participants and baseline characteristics. (a) Sample distribution showing the proportions of controls (n = 215) and patients with breast cancer (n = 285). (b) Tumor stage distribution among breast cancer patients (T-I to T-IV). (c) Comparison of the mean age (in years) between controls and breast cancer patients. (d) Comparison of the mean body mass index (BMI, kg/m²) between controls and breast cancer patients. The data are presented as the means ± SDs. p values were obtained via Welch’s t test.

Full size image

Participants and sample size justification

An a priori power analysis (G*Power 3.1) assumed two-group mean differences (independent samples, two-tailed), α = 0.05, power (1 − β) = 0.80, and effect sizes in the medium–large range (Cohen’s d ≈ 0.5–0.8) reported for Se/Hg contrasts in oncology cohorts. Under these assumptions, ≥ 200 participants per group were needed. The final analytic sample (patients = 285; controls = 215) exceeded this threshold, providing ≥ 80% power to detect stagewise differences in elemental and hematologic indices.

Inclusion and exclusion criteria

Women aged 35–79 years with a histologically confirmed diagnosis of breast cancer and no mineral supplementation within the preceding 3.5 months were eligible for the patient group. Healthy women of similar age and regional background with no history of cancer or chronic illness were selected as controls. The exclusion criteria included prior chemotherapy, radiotherapy, or hormonal therapy; coexisting chronic disease (e.g., renal, hepatic, or autoimmune); prior malignancy; or incomplete clinical or biospecimen data.

Sample collection and Preparation

Scalp hair and venous blood samples were collected from participants following a detailed protocol from previous studies29,42,43,44,45,46,47,48,49,50,51. The samples were stored at -20 °C to prevent metal contamination. For biochemical analyses, 4–5 mL of venous blood was used following standard procedures52. Hair samples from the occipital region were washed with nonionic detergent and deionized water and oven dried. Blood was drawn via acid-washed polypropylene tubes to minimize contamination. The apparatus used was based on methodologies from previous studies42,44,51. A Milestone microwave oven (model Rotar MPR-300/12S) and Hitachi model 180–50 were used for the measurements, and the parameters are listed in Supplementary Table S1. Sample preparation was performed in a Pel microwave oven (model PMO23), and the solutions were stored in acid-washed PTFE containers.

Trace element analysis

Selenium and mercury levels were measured via atomic absorption spectrometry (AAS). Selenium was determined via graphite furnace AAS (PerkinElmer AAnalyst 700), whereas mercury was quantified via cold vapor AAS via a Hydride Generation System HGA-900. Calibration was performed with certified multielement standard solutions. All reagents were of analytical grade and procured from Merck and Sigma‒Aldrich.

Clinical evaluation and staging

Cancer staging was performed according to the American Joint Committee on Cancer (AJCC) TNM classification, which is based on tumor size, lymph node involvement, and metastasis. Patients were grouped into four clinical stages: T-I, T-II, T-III, and T-IV. Imaging techniques such as mammography, CT scans, ultrasound, and positron emission tomography (PET) were used for diagnostic confirmation. Mammographic density was classified via the Breast Imaging Reporting and Data System (BI-RADS), with dense breast tissue considered a risk factor for delayed tumor detection. Hospital oncologists and radiologists completed diagnostic procedures and imaging interpretations; the research team analyzed the anonymized clinical reports and elemental data.

Hematological and biochemical analysis

Hematological parameters, including hemoglobin (Hb), red blood cell count (RBC), hematocrit (HCT), mean corpuscular volume (MCV), mean corpuscular hemoglobin (MCH), mean corpuscular hemoglobin concentration (MCHC), and white blood cell (WBC) count, were measured via an automated hematology analyzer. Iron-related markers, including serum ferritin and total iron-binding capacity (TIBC), were assessed via standard immunoassay and colorimetric methods.

Statistical analysis

Group-level summary statistics (means, SDs, n) were calculated, and each cancer stage (T-I–T-IV) was compared with that of controls via Welch’s t tests. We reported the mean difference (Δ), t statistics, degrees of freedom (df), two-sided p value, 95% confidence interval (CI) for Δ, and effect sizes (Cohen’s d and Hedges’ g). The false discovery rate within analytic families (elements and hematology) was controlled via the Benjamini–Hochberg procedure. Group mean ± SD plots and stage-coded scatterplots were generated. All analyses were conducted via SPSS v25 and Python (scipy/statsmodels), with α = 0.05.

Results

Patient demographics and clinical characteristics

We analyzed 285 female breast cancer patients and 215 age-matched healthy controls (Fig. 2). The mean age of the patients was 51.7 ± 9.4 years, whereas that of the controls was 50.8 ± 8.9 years (t = 0.81, p = 0.42, 95% CI − 0.9 to 2.7; Hedges g = 0.10). The tumor stages were T-I (n = 25, 8.8%), T-II (n = 38, 13.3%), T-III (n = 105, 36.8%), and T-IV (n = 117, 41.1%) (Fig. 2). More than 60% of early-stage patients (T-I/T-II) had dense breasts (BI-RADS c/d), which is consistent with reduced screening sensitivity and delayed detection (Supplementary Figure S1). The clinical presentation evolves with stage painless lumps in early disease and nipple retraction/skin dimpling/palpable lymphadenopathy in advanced stages (Supplementary Figure S2). The diagnoses are confirmed by FNAC and/or core biopsy with multimodal imaging (ultrasound, mammography, CT; Supplementary Figure S3).

Elemental analysis of selenium and mercury

We compared the Se and Hg concentrations in the scalp hair and blood of healthy controls and breast cancer patients across stages T-I to T-IV (Fig. 3; Table 1). In the controls, the Se concentration was 1.59 ± 0.29 µg/g in the hair and 220.0 ± 8.19 µg/L in the blood, whereas the Hg concentration was 1.19 ± 0.09 µg/g in the hair and 0.97 ± 0.05 µg/L in the blood. In contrast, patients exhibited a stage-related decline in Se and a concomitant increase in Hg. In stage T-I, the Se concentration decreased to 0.45 ± 0.09 µg/g in hair and 125.1 ± 5.35 µg/L in blood, whereas the Hg concentration increased to 2.85 ± 0.21 µg/g in hair and 2.15 ± 0.35 µg/L in blood. By stage T-IV, Se reached 0.25 ± 0.01 µg/g in hair and 51.3 ± 5.75 µg/L in blood, whereas Hg rose to 4.31 ± 0.15 µg/g in hair and 2.94 ± 0.11 µg/L in blood. These stagewise patterns are displayed in Fig. 3 and summarized in Supplementary Table S2.

Fig. 3
figure 3

Se and Hg levels in blood and scalp hair across clinical stages of breast cancer. (A) Blood Se; (B) Blood Hg; (C) Hair Se; (D) Hair Hg. The bars represent the means ± SDs for controls and patients with stage T-I to T-IV breast cancer. The dashed reference lines indicate functional or toxicological thresholds (70 µg/L = GPx3 sufficiency; 100–120 µg/L = SELENOP zone; 2 µg/g = Hg guidance). The data show progressive Se depletion and Hg elevation with advancing stage. The dashed reference lines indicate commonly cited levels: blood Se 70 µg/L (approximately GPx3 sufficiency), 100–120 µg/L (SELENOP saturation), and hair Hg 2 µg/g (population guidance)55,57,66.

Full size image
Table 1 Selenium and mercury (blood & hair) by stage vs. controls.
Full size table

Inferential comparisons (Table 1) confirmed that these differences were statistically significant and large in magnitude. Compared with that in control hair, Se was lower by − 1.34 µg/g (95% CI − 1.399–−1.281; Welch’s t ≈ − 45.2; p = 1.78 × 10⁻⁶⁶; Hedges g = − 5.05), whereas Hg was greater by + 3.12 µg/g (95% CI 3.05–3.19; t ≈ 91.8; p = 2.88 × 10⁻³¹; Hedges g = + 30.1). Comparable stage-versus-control results for T-I, T-II, and T-III patients are provided in Table 1 and Supplementary Table S2, and all comparisons remained significant after false discovery rate adjustment.

From a clinical perspective, blood Se in advanced disease patients (T-III/T-IV) falls within or below the functional sufficiency range of approximately 70–120 µg/L, which is consistent with potential impairment of selenoprotein activity. In parallel, hair Hg exceeded the 2 µg/g guidance level across all cancer stages, indicating elevated exposure or accumulation. Taken together, these findings clearly reveal a shift toward selenium deficiency and mercury overload with advancement. Consistent with this pattern, the scatterplots in Fig. 4 illustrate a pronounced inverse relationship between Se and Hg across matrices, which is most evident in the higher stages.

Fig. 4
figure 4

Stagewise correlation between selenium and mercury levels in breast cancer patients. Scatter plots showing selenium vs. mercury concentrations in (a) blood (µg/L) and (b) scalp hair (µg/g), color-coded by disease stage (Controls, T-I, T-II, T-III, T-IV). A strong inverse pattern is evident, particularly in advanced-stage patients, supporting antagonistic interactions between these trace elements.

Full size image

Hematological and iron-related parameters

Compared with controls, patients with breast cancer presented substantial abnormalities in hematologic indices (Fig. 5; Table 2; Supplementary Table S3). Hemoglobin decreased from 12.7 ± 1.21 g/dL in the controls to 5.62 ± 0.38 g/dL in stages T-IV (Δ = −7.08 g/dL; Welch’s t = − 78.94; df = 281.5; p < 1 × 10⁻¹⁶; Hedges g = − 7.06). Red blood cells declined from 4.02 ± 0.30 × 10¹²/L to 1.35 ± 0.17 × 10¹²/L (Δ = −2.67 × 10¹²/L; t = − 103.49; df = 329.4; p < 1 × 10⁻¹⁶; g = − 10.18). Additional measures decreased progressively with stage: hematocrit from 41.9 ± 1.05% to 17.9 ± 1.26%, mean corpuscular volume from 93.6 ± 1.45 fL to 58.5 ± 1.99 fL, mean corpuscular hemoglobin from 30.7 ± 0.62 pg to 11.9 ± 0.98 pg, and mean corpuscular hemoglobin concentration from 34.8 ± 0.59 g/dL to 16.5 ± 1.51 g/dL (all p < 1 × 10⁻¹⁶, with large effect sizes; Table 2) In contrast, the levels of iron-related markers increased from 53.9 ± 5.15 ng/mL to 165.0 ± 7.41 ng/mL (Δ = +111.1 ng/mL; t = 144.31; df = 178.3; p < 1 × 10⁻¹⁶; g = 18.35), and the total iron-binding capacity increased from 293.0 ± 9.52 µg/100 mL to 685.0 ± 26.9 µg/100 mL (Δ = +392.0 µg/100 mL; t = 152.51; df = 132.0; p < 1 × 10⁻¹⁶; g = 22.10) (Table 2). All stage-versus-control comparisons remained significant after false discovery rate adjustment. From a clinical perspective, these findings indicate that progressively severe, inflammation-associated, iron-restricted anemia worsens with advancing disease stage.

Fig. 5
figure 5

Hematological and iron metabolism parameters in breast cancer patients versus controls. (A) Hemoglobin (Hb), (B) red blood cell count (RBC), (C) serum ferritin, and (D) total iron-binding capacity (TIBC) are shown as the means ± SDs for controls and patients with breast cancer stratified by stage (T-I to T-IV). The normal physiological ranges are shaded in light color bands (Hb: 12–15 g/dL; RBC: 3.8–4.8 × 10¹²/L; ferritin: 30–150 ng/mL; TIBC: 250–450 µg/100 mL). Hematological indices decreased with advancing stage, whereas ferritin and TIBC increased, which is consistent with inflammation-associated iron-restricted erythropoiesis.

Full size image
Table 2 Hematological indices by stage versus controls (mean ± SD).
Full size table

Frequency and extent of axillary lymph node metastases in female patients with breast cancer on the basis of pathological evaluation

Supplementary Table S4 summarizes multimodal findings relevant to axillary lymph node (ALN) involvement, including mammography, core biopsy pathology, and cross-sectional and nuclear imaging. On mammography, the background parenchyma was predominantly fatty/glandular. A suspicious soft-tissue density in the upper-outer quadrant of the right breast, with associated enlarged axillary lymph nodes, was assessed as BI-RADS 4, prompting core needle (tru-cut) biopsy. Core biopsy confirmed invasive (infiltrating) ductal carcinoma, with minimal tubular formation (3/3), marked nuclear pleomorphism (3/3), and numerous mitoses (3/3). Lymphocytic infiltration, fibrosis, and desmoplastic stroma were present, and the grade was assigned according to the Nottingham (Bloom–Richardson) system. Contrast-enhanced CT revealed heterogeneously enhancing right-breast mass with spiculated margins, architectural distortion, skin thickening, and nipple retraction. Right axillary lymphadenopathy was present (largest node ~ 1.4 cm with loss of fatty hilum), which was consistent with metastatic ALN involvement. Abdominal ultrasound revealed a large complex lesion (13.0 × 12.7 cm) in the umbilical/hypogastric region suspicious for a mesenteric mass; this finding was nonbreast and was documented for completeness. Skeletal scintigraphy (Tc-99 m MDP) revealed focal uptake in the skull requiring correlation and background degenerative changes in the dorsolumbar spine and peripheral joints. Collectively, the imaging and pathology findings support locoregional disease with metastatic right axillary lymphadenopathy and highlight the utility of integrated diagnostic assessment (mammography, histopathology, cross-sectional imaging, and nuclear medicine) in staging and treatment planning (Supplementary Table S4). From a clinical perspective, the concordant evidence of invasive ductal carcinoma with pathologic‒radiologic signs of axillary node involvement favors management along a node-positive pathway (systemic therapy consideration, targeted axillary staging, and individualized locoregional control), whereas nonbreast incidental findings (degenerative changes, possible mesenteric mass) should be correlated separately to avoid confounding breast-specific staging.

Discussion

In this case–control study, we observed a clear stage-related imbalance in trace elements among women with breast cancer. Selenium (Se) levels were progressively lower, and mercury (Hg) levels were markedly higher from stage T-I to T-IV compared with healthy controls, in both blood and scalp hair samples. This trend was consistent across all comparisons and remained statistically significant after correction for multiple testing. These results suggest that breast cancer progression is associated with increasing oxidative stress and reduced antioxidant defense, driven by the opposing biological roles of Se and Hg. Selenium is an essential trace element that supports antioxidant defense and redox balance through its role in selenoproteins, including glutathione peroxidases (GPx) and thioredoxin reductases (TrxR)53,54. These enzymes prevent oxidative damage to lipids, proteins, and DNA, which is crucial for maintaining cellular homeostasis and preventing malignant transformation. Mercury, in contrast, is a toxic heavy metal that binds to selenol and thiol groups in proteins, thereby inhibiting GPx and TrxR activities and increasing reactive oxygen species (ROS) production39. The mechanistic relationship between these elements is illustrated in Fig. 1. Mercury not only inhibits Se-dependent enzymes but also forms stable Hg–Se complexes, such as mercury selenide (HgSe), which reduce mercury toxicity and enhance its excretion. When Se is insufficient, this detoxification pathway becomes less effective, allowing Hg to accumulate in tissues. Thus, Se deficiency may intensify mercury-induced oxidative stress, while Hg exposure may further deplete functional Se reserves, creating a vicious cycle that contributes to cancer progression55. Our statistical findings support these mechanisms. Hair Se decreased from 1.59 µg/g in controls to 0.25 µg/g in stage T-IV patients, while hair Hg increased from 1.19 µg/g to 4.31 µg/g. Similarly, blood Se dropped from 220 µg/L to 51.3 µg/L, and blood Hg rose from 0.97 µg/L to 2.94 µg/L. These large and consistent differences (p < 0.0001; Hedges g > 2 for Se, > 10 for Hg) show a robust association between Se–Hg imbalance and disease severity. The inverse Se–Hg trend observed in both matrices supports the hypothesis that the two elements are biochemically linked and jointly influence redox homeostasis55,56,57. Hematologic parameters also changed significantly with cancer stage. Patients with T-III and T-IV disease had lower hemoglobin and red-cell indices and higher ferritin and TIBC levels than controls, consistent with inflammation-associated, iron-restricted erythropoiesis58. This pattern suggests a systemic response to chronic oxidative stress and inflammation, which are known to suppress erythropoietin production and iron availability. Such anemia contributes to fatigue, poor functional status, and reduced quality of life in advanced breast cancer patients59. The biological plausibility of these findings is further supported by previous studies showing that Hg interferes with antioxidant and immune functions, while adequate Se levels help maintain redox balance and immune competence39,54. Population studies have also reported that Se status varies widely across regions depending on soil content, diet, and supplementation policies. For instance, Se deficiency is common in some parts of China and Eastern Europe, while intake is generally adequate in North America60,61. Conversely, mercury exposure is more frequent in coastal and riverine communities that rely heavily on seafood or engage in artisanal gold mining62,63. These global differences make it difficult to establish universal Se/Hg thresholds for cancer prevention or management. Recent clinical reports also note that overt Se deficiency can be uncommon during therapy and that supplementation may risk overdosing without monitoring64. Differences in timing (pretreatment vs. on-treatment), background exposure, and matrices (hair vs. serum/plasma) may explain various observations. In our study, advanced-stage means fell at or below functional sufficiency, whereas controls were high, supporting context-specific monitoring rather than routine supplementation. Clinically, many patients with advanced disease in our study had Se concentrations at or below the lower functional range (~ 70 µg/L) and hair Hg values above the WHO/UNEP guidance limit of 2 µg/g55. Although this study does not propose definitive diagnostic cut-offs, these levels could serve as practical indicators for closer monitoring, dietary counseling, and environmental exposure assessment. Integrating trace element testing with routine hematology and imaging may improve risk stratification, particularly in low-resource settings where late-stage presentation is common.

Future research should include prospective cohorts and interventional studies that measure dietary intake, exposure sources, and functional selenoprotein activity (e.g., GPX4, TrxR)37,40. Mechanistic studies on ferroptosis and oxidative signaling could clarify whether the Se–Hg imbalance contributes directly to cancer progression. Clinical trials assessing controlled Se supplementation and Hg reduction strategies, with biochemical monitoring, may help determine whether modifying these exposures improves outcomes.

In summary, this study provides evidence of a strong stage-related Se–Hg imbalance in women with breast cancer, supported by both statistical and biological findings. Selenium depletion and mercury accumulation appear to worsen with disease progression, possibly through redox disruption and enzyme inhibition. These results highlight the importance of integrating environmental and nutritional assessments into breast cancer prevention and management, particularly in regions with low Se intake or high Hg exposure.

Limitations and implications

This study has several limitations. First, the case–control design cannot establish causation or the timing of exposure relative to disease development. Second, detailed individual data on diet (including fish intake), environmental exposure, and mercury speciation were not available; therefore, residual confounding may remain, and full multivariable adjustment was not possible. Third, blood and hair samples represent different exposure periods and mercury forms, which limits direct comparison between them. Fourth, the hospital-based sample from a single region may limit generalizability to other populations. Fifth, functional biomarkers such as selenoprotein activity (e.g., GPX4, TrxR) and ferroptosis markers were not measured, which restricts mechanistic interpretation. Finally, the clinical thresholds discussed here should be considered indicative reference points rather than diagnostic cut-offs until validated in prospective studies.

Future research should include prospective cohort studies with standardized exposure assessments, including dietary intake, fish consumption, and mercury speciation. Adjusted regression or survival models are needed to identify independent associations, and validation studies (e.g., ROC analysis) should test whether blood Se < 70 µg/L and hair Hg > 2 µg/g can improve risk prediction. Mechanistic investigations measuring selenoprotein activity and ferroptosis pathways, along with carefully designed interventions to optimize Se status and reduce Hg exposure, may help determine whether modifying these factors can improve patient outcomes.

Conclusions

In this case-control study, Se levels decreased while mercury Hg levels increased from stage T-I to T-IV in blood and scalp hair, accompanied by progressive anemia with lower hemoglobin and RBC indices, and higher ferritin/TIBC. The findings suggest lower Se may reduce antioxidant defenses, while higher Hg increases toxic stress, though remaining associative. Blood Se levels within 70–120 µg/L and hair Hg above 2 µg/g may indicate need for follow-up, counseling, and supportive care. Prospective studies should confirm these thresholds and test whether improving Se status and reducing Hg exposure improve outcomes.