The Role of Thyroid Disease and Other Biomarkers in Invasive Breast Cancer: A 3-year Retrospective Comparative Study

Worldwide, approximately 2.1 million individuals are affected by breast cancer (BC), making it one of the leading causes of disability and deterioration of health and the fifth leading cause of death among women^1,2. The clinical behavior of BC is highly variable, influenced by a combination of histological characteristics, hormone receptor status, and, increasingly, molecular subtypes. In addition to traditional histological and immunohistochemical classifications, molecular subtypes have been incorporated to enhance prognostic accuracy and guide individualized treatment planning^4,5. According to the molecular classification proposed by Perou et al., BC is categorized into four subtypes: Luminal A, Luminal B, human epidermal growth factor receptor 2 (HER2)-enriched, and triple-negative BC (TNBC)⁵. These subtypes differ in prognosis and response to therapy, with Luminal A associated with the highest overall survival rates and TNBC often demonstrating aggressive clinical behavior and limited treatment options^6,7,8.

In recent years, there has been growing interest in understanding the interaction between endocrine disorders and cancer progression. Thyroid hormones – particularly thyroxine (T4) and triiodothyronine (T3) – are known to influence various cellular processes, including proliferation, differentiation, metabolism, and apoptosis. Dysregulation of thyroid function may alter the hormonal milieu, potentially affecting tumor biology and responsiveness to therapy. For instance, T3 and T4 have been shown to modulate the expression of estrogen receptors (ERs) and progesterone receptors (PRs), activate signaling pathways, such as mitogen-activated protein kinase (MAPK) and phosphoinositide 3-kinase (PI3K), and promote angiogenesis – processes that are often implicated in cancer growth and metastasis⁹.

Menopausal status and hormone therapy are well-established confounding factors in the development of BC. The transition to menopause alters endogenous hormone levels, particularly estrogen and progesterone, which play critical roles in breast tissue biology and carcinogenesis. In addition, hormone therapy, especially combined estrogen–progestin regimens, has been associated with an increased risk of BC in postmenopausal women. Evidence from the Women's Health Initiative demonstrated that estrogen plus progestin therapy significantly raised BC incidence and mortality compared to placebo, highlighting the importance of considering these factors in BC risk assessments and epidemiological studies¹⁰.

Some studies have suggested that patients with hypothyroidism may have a more favorable prognosis in BC, possibly due to a reduced proliferative stimulus from lower circulating thyroid hormone levels. Conversely, hyperthyroidism may contribute to enhanced tumor aggressiveness through increased metabolic activity and hormone-driven proliferation. These findings have sparked interest in evaluating thyroid-related biomarkers – including thyroid-stimulating hormone (TSH), free T4 [fT4], and thyroid antibodies – as potential modulators of tumor behavior or prognostic indicators¹¹.

Despite these hypotheses, the precise relationship between thyroid function and BC outcomes remains unclear. TSH, in particular, is generally considered a marker of thyroid function but not typically associated with cancer aggressiveness. However, its potential role in reflecting systemic endocrine homeostasis and possible cross-talk with other hormonal pathways warrants further investigation.

Given these considerations, the present study aimed to evaluate postoperative TSH levels in patients with invasive BC and assess the potential influence of thyroid disease on tumor aggressiveness, as defined by molecular subtype and other pathological features. By exploring these associations, we hope to contribute to a better understanding of the complex interplay between thyroid function and BC biology.

In this study, we included a total of 97 patients, who were divided into two groups based on the presence or absence of thyroid disease before surgery. Group A comprised patients without any thyroid pathology (n = 49; mean age: 64.26 years), while Group B included all patients with a known diagnosis of thyroid disease before surgery (n = 48; mean age: 64.31 years). The descriptive characteristics of both groups are presented in Table 1. Interestingly, a greater proportion of patients in Group B presented with more favorable molecular subtypes compared to those without thyroid disease.

To assess the distribution of variables, the Kolmogorov–Smirnov (KS) test for normality was applied to each parameter, including age, TSH levels (pre- and postoperative), fT3, fT4, LDH, and calcium levels. In Group A, the normality tests showed that age and postoperative TSH levels followed a normal distribution, as indicated by nonsignificant p-values in both KS (p = 0.200 and p = 0.200) and Shapiro–Wilk (SW) (p = 0.239 and p = 0.819) tests. In contrast, calcium, LDH, and Ki-67 did not follow a normal distribution, as demonstrated by significant p-values in both tests (calcium: KS p = 0.001, SW p < 0.001; LDH: KS p = 0.200 but SW p = 0.004; Ki-67: KS p = 0.004, SW p = 0.002).

In Group B, the normality tests revealed that age, calcium, fT3, and fT4 all followed a normal distribution, as indicated by nonsignificant p-values in both KS (all p = 0.200) and SW (p-values ranging from 0.462 to 0.933) tests. In contrast, LDH, Ki-67, and preoperative TSH levels significantly deviated from normality, with both tests showing p-values <0.01 (LDH: KS p < 0.001, SW p < 0.001; Ki-67: KS p = 0.004, SW p = 0.005; preoperative TSH: KS p < 0.001, SW p < 0.001). Postoperative TSH levels showed mixed results, with KS indicating normality (p = 0.194) but SW suggesting a deviation (p = 0.017). These findings imply that parametric tests can be appropriately used for age, calcium, fT3, and fT4, while nonparametric methods should be considered for LDH, Ki-67, preoperative TSH, and potentially postoperative TSH due to their non-normal distribution. These findings are summarized in Table 2.

The Wilcoxon signed-rank test comparing preoperative and postoperative TSH levels included 18 paired observations. Among these, eight cases showed a decrease in TSH postoperatively (negative ranks) with a mean rank of 8 and a sum of ranks of 64, while 10 cases showed an increase in postoperative TSH (positive ranks) with a mean rank of 10.7 and a sum of ranks of 107. There were no ties. The test statistic Z was 0.936, with a two-tailed p-value of 0.349, indicating no statistically significant difference between the preoperative and postoperative TSH levels in this sample (Table 3).

Furthermore, we conducted a correlation analysis using Spearman's correlation analysis. In Group A, age showed no significant correlation with calcium, LDH, or Ki-67. Calcium had a significant positive correlation with Ki-67 (r = 0.449, p = 0.036), indicating that higher calcium levels are associated with higher Ki-67 proliferation indices. TSH postoperative was strongly correlated with age (perfect negative correlation, r = −1.000), limiting interpretation. LDH and Ki-67 showed no significant correlations with other variables (Table 4a).

In Group B, none of the correlations between age and other variables reached statistical significance. Calcium showed weak, nonsignificant correlations with most variables. TSH postoperative correlated negatively but nonsignificantly with calcium and Ki-67. A notable significant negative correlation was observed between preoperative TSH and fT4 (r = −0.443, p = 0.002), suggesting an inverse relationship between these thyroid-related measures. Ki-67 had weak correlations with other markers, and none of them was significant. Overall, correlations in Group B were generally weak and not statistically significant, except for the relationship between preoperative TSH and fT4 (Table 4b).

Thus, the main meaningful correlations were the positive association between calcium and Ki-67 in Group A and the inverse correlation between preoperative TSH and fT4 in Group B. The rest of the variables mostly showed weak or no significant relationship. Sample size differences and missing data may influence the reliability of some correlations, especially in Group A.

In addition, four further correlation analyses were performed using the JASP statistical program. In Group A, there were clear positive relationships between ER and PR, as well as between these receptors and the molecular subtype, which makes sense because molecular subtypes are based on the hormone receptor status. Ki-67, a marker of cell proliferation, also showed a moderate positive connection with ER and calcium levels, suggesting that higher proliferation might be linked to these factors. However, there was a slight negative trend between ER and age, indicating that ER expression might have decreased as age increases, though this was not statistically strong. In addition, there was a small negative relationship between molecular subtype and HER2, suggesting that certain subtypes may have had lower HER2 expression. Other variables, such as LDH and calcium, did not show strong or consistent correlations. In Group B, positive correlations reveal some key relationships. PR and molecular subtype showed a strong positive correlation, indicating that hormone receptor status was closely linked to subtype classification. Ki-67, a proliferation marker, also positively correlated with molecular subtype, suggesting more aggressive subtypes have higher proliferation rates. In addition, ER was strongly positively correlated with PR and molecular subtype, consistent with their biological connection. Calcium showed a weak positive correlation with PR and Ki-67, but these were not strong. Thyroid hormones (fT4, fT3) and TSH levels (preoperative and postoperative) mostly showed weak or no significant positive correlations with other variables, except for a slight positive association between fT4 and calcium. Overall, this group highlighted strong links between hormone receptors, molecular subtype, and proliferation, while thyroid-related variables appeared less connected. Overall, the key findings highlighted the important links between hormone receptors, molecular subtype, and proliferation in this group (Table 5).

Finally, the Mann–Whitney test results showed no statistically significant differences between the groups for any of the variables tested. Specifically, age (p = 0.862), calcium levels (p = 0.795), postoperative TSH values (p = 0.412), LDH (p = 0.904), and Ki-67 (p = 0.311) – all had p-values well above the common significance threshold of 0.05. This indicates that the distributions of these variables did not differ significantly between the compared groups (Table 6).

In this study, we examined postoperative TSH levels and evaluated the impact of thyroid disease on BC aggressiveness. Our correlation analyses further indicate that preoperative calcium levels are associated with cancer aggressiveness and may serve as potential prognostic indicators.

BC is the most prevalent cancer worldwide and the fifth leading cause of death among women². In 2019, approximately 271,000 new BC cases were estimated in the USA, with around 42,260 expected to result in death. The risk of invasive BC increases with age; for example, the likelihood of diagnosis is one in 25 for women in their 70s compared to one in 567 for those in their 20s¹². According to the American Cancer Society, the lifetime risk of invasive BC diagnosis is approximately 12.8%, with a 2.6% risk of mortality. Incidence rates vary by race and ethnicity, with non-Hispanic White women exhibiting the highest rates and Asian/Pacific Islander women having the lowest¹².

Several factors contribute to the increased incidence of BC, including genetic and non-genetic influences, such as lifestyle, radiation exposure, hormone therapy, and hormone imbalances (e.g., early menopause, miscarriages)¹³. Among these, family history remains the most significant risk factor. Inherited mutations in the BRCA1 and BRCA2 genes account for 3%–10% of all female BC cases and approximately 30% of early-onset cases worldwide^14–15. BRCA1 and BRCA2, located on chromosomes 17 and 13, respectively, act as tumor suppressor genes by repairing DNA damage or inducing apoptosis. Mutations in these genes increase the risk of breast, ovarian, and prostate cancers. Specifically, BRCA1 mutations are implicated in 60%–80% of BC cases, while BRCA2 mutations account for around 35% of BC cases^13,17. BRCA1 mutations are associated with poorer prognosis, characterized by higher mitotic rates, increased lymph node invasion, and lack of expression of ER, PR, and/or HER2 receptors, as well as p53 gene mutations^{18,19,20,21,22}. In addition, carriers of BRCA1 and/or BRCA2 mutations face elevated risks for other cancers, including pancreatic cancer and melanoma^23,24.

BC is classified into three main groups: histological, molecular, and functional. Histologically, the World Health Organization (WHO) categorized BC into several subtypes, including epithelial tumors, invasive carcinoma, rare and salivary gland-type neoplasms, neuroendocrine tumors, epithelial–myoepithelial tumors (such as in situ ductal carcinoma and noninvasive lobular neoplasms), benign epithelial proliferations and precursors (including papillary tumors and adenomas), mesenchymal tumors, tumors of the nipple, malignant lymphoma, fibroepithelial neoplasms, metastatic tumors, and male breast neoplasms^4,5. Among these, breast adenocarcinoma is the most common, accounting for approximately 95% of BC cases²⁵. Invasive ductal carcinoma (IDC), which comprises approximately 55% of all breast carcinomas, originates from the terminal segments of the lobular duct unit^25,26. In contrast, lobular carcinoma arises from the terminal duct lobular unit. Studies, such as that by Toikkanen et al.²⁷, have demonstrated that invasive lobular carcinoma (ILC) generally has a better prognosis than IDC, with 5-year survival rates of 78% for ILC versus 50% for IDC.

Recent advances in BC treatment have integrated molecular parameters to enhance diagnostic and prognostic accuracy^28,29,30. Using microarray technology, Perou et al.⁵ classified BC into four molecular subtypes: Luminal A (ER+ and/or PR+ and HER2-), Luminal B (ER+ and/or PR+, HER2+), HER2+ (ER- and PR-), and basal-like (ER-, PR-, and HER2-).

According to Fallahpour et al.⁷ and Fragomeni et al.⁶, the Luminal A subtype is associated with the highest survival rates, followed by Luminal B, HER2 overexpression, and triple-negative subtypes, which have the lowest survival and highest recurrence rates. In addition, He et al.³¹ reported that molecular subtype significantly influences the benefits of radiation therapy, with patients having the Luminal A subtype gaining the greatest benefit and those with triple-negative and HER2 overexpression subtypes showing less benefit regardless of radiation type.

HER2, encoded by the ERBB2 gene, is a member of the epidermal growth factor receptor family that is often constitutively active due to ligand-independent dimerization. It promotes cell proliferation and inhibits apoptosis, with overexpression found in 15%–30% of BCs. This overexpression correlates with poorer overall and progression-free survival and serves as a prognostic marker^8,32,33. ERs (ERα and ERβ) are intracellular proteins activated by estrogen and regulate gene expression through genomic and non-genomic pathways. ER positivity, observed in approximately 70% of BCs, is generally associated with better prognosis, though it represents a heterogeneous group requiring individualized treatment^{34,35,36,37,38}. Similarly, PR positivity is linked to improved outcomes and progression-free survival, especially when co-expressed with ER^39,40.

Moreover, Ki-67 is a nuclear protein and a biomarker of cell proliferation⁴¹. In BC, Ki-67 is considered a predictive marker^42,43, although its clinical utility remains debated. It is thought to assist in predicting treatment responsiveness and resistance, assessing residual risk during therapy, and serving as an active biomarker for ongoing treatment evaluation⁴⁴.

Non-basal BC has frequently been associated with thyroidopathy, particularly hypothyroidism, which is linked to a more favorable prognosis^45,46. Thyroid hormones typically activate MAPK signaling, leading to phosphorylation of nuclear transactivator proteins and possibly plasma membrane proteins⁴⁷. Moreover, 17β-estradiol promotes phosphorylation of serine-118 on the ERα receptor, a process believed to be MAPK dependent. Thyroid hormones may mimic estrogen's effects in BC, stimulating cell proliferation in both ER-negative and ER-positive tumors⁴⁷. Falstie-Jensen et al.⁴⁸ reported that patients with hypothyroidism and BC had better prognoses. Ortega-Olvera et al.⁴⁹ highlighted the role of body mass index in the relationship between thyroid hormones and BC, although TSH levels were not associated with cancer aggressiveness¹⁰.

Calcium, an essential element in metabolism, has been suggested to inhibit breast tumors in animal studies⁵⁰. In humans, its involvement in vitamin D metabolism and the parathyroid hormone cycle indicates a potential influence on BC^51,52. Almquist et al.⁵³ proposed a link between calcium and increased breast tumor risk, while Sprague et al.⁵⁴ found no such association.

In this study, we observed significant differences in preoperative TSH and fT4 levels between patients with thyroidopathy and those without. Using correlation analyses on the JASP program, we found several trends toward significance among multiple parameters. Notably, there was a positive trend between preoperative TSH levels and PR receptor status, as well as between preoperative LDH levels and PR receptors. A significant positive association was found between Ki-67 and molecular subtypes. Regarding negative correlations, we observed significance between Ki-67 and fT4, and trends toward significance between Ki-67 and LDH levels, age and HER2 receptor status, age and Ki-67, and age and fT3. These results suggest that thyroidopathy may influence cancer aggressiveness; however, the small sample size precludes definitive conclusions.

For patients without thyroidopathy (Group A), positive correlations were found between Ki-67 and calcium levels, suggesting that preoperative calcium could be a potential biomarker for prognosis. Additional correlation analyses performed using the JASP program revealed significant positive associations between Ki-67 and calcium levels and between Ki-67 and ER receptor status. Negative correlations showed trends toward significance between age and molecular subtype and between preoperative LDH and calcium levels. These findings again indicate the potential of calcium and LDH as biomarkers, though conclusions are limited by the sample size.

Study limitations include its retrospective design and single-center setting, which may limit the generalizability of the findings. In addition, Overall Survival and Progression Free Surviavl data were unavailable for analysis. Ki-67 data were only available from late 2018 onward, and some histological reports lacked complete molecular subtype information. The relatively small sample size of 97 patients further restricts the statistical power of the study. Despite these limitations, our findings suggest that thyroidopathy may serve as a favorable prognostic factor in BC and plays a significant role in tumor aggressiveness. In conclusion, elevated calcium levels and Ki-67 expression were associated with higher tumor stages and may act as additional prognostic indicators. Future prospective, multicenter studies with larger cohorts are needed to better define the impact of these factors on OS and PFS and to clarify the roles of calcium, LDH, and thyroid disease in BC aggressiveness.