Diagnostic value of peripheral blood inflammatory indices for breast cancer grade and immunohistochemical markers: a retrospective observational study.

Introduction
Breast cancer has emerged as a significant public health concern, posing a serious risk to human life and well-being. As the most prevalent cancer worldwide, according to the GLOBOCAN 2022 report, the Lancet’s Breast Cancer Commission predicts that by 2040, the global incidence of new cases of breast cancer will be more than 3 million annually, with the greatest increase in low-income and middle-income countries [1, 2]. There is also a rapid increase in the incidence of breast cancer in Iran. The estimated results show an age-standardized incidence rate of 32.10 (per 100,000 women), a crude incidence rate of 28.99 (per 100,000 women), and annual percentage changes of 7.54% [3].
Prognosis and postoperative recurrence risk in breast cancer is primarily assessed by clinicopathological characteristics of patients such as tumor size, lymph node status, tumor grade, Ki67 index and molecular subtypes [4, 5]. The levels of estrogen receptor (ER), progesterone receptor (PR), and human epidermal growth factor receptor 2 (HER2) as immunohistochemical (IHC) markers are regularly checked in breast cancer patients to help predict their outcomes and decide on treatments [6]. Besides mammography, the cornerstone of breast cancer screening, powerful combinations of magnetic resonance imaging and molecular testing are adding to the sensitivity and specificity of the breast cancer screening. Despite these developments, challenges exist in ensuring the ability to detect small malignancies and widespread accessibility to screening programs, especially in resource-limited settings [7].
Liquid biopsy, based on the analysis of body fluids, has attracted much attention in the search for cancer biomarkers [8]. Increasing evidence suggests that peripheral blood inflammatory markers reflect the level of local immune inflammation in the tumor microenvironment. Also, systemic inflammation and immunity play important roles in tumor initiation, invasion, and metastasis and may influence tumor response to therapy. Although breast cancer typically produces a few neoantigens, it usually demonstrates significant infiltration by lymphocytes, known as tumor-infiltrating lymphocytes, which varies among different IHC subtypes [9]. These inflammatory pathways can also be manifested through changes in blood cells such as platelets, white blood cells, and red blood cells [10].
Correspondingly, peripheral blood inflammatory indices, such as systemic immune-inflammation index (SII), lymphocytes-albumin to neutrophils ratio (LANR), neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), hemoglobin-to-red cell distribution width ratio (HRR), and glucose-to-lymphocyte ratio (GLR), have become popular topics in the diagnosis, treatment, follow-up management, and prognosis prediction of various solid tumors, including gastric cancer [11], colorectal cancer [12], lung cancer [13], as well as breast cancer [14–16]. These indicators can help cancer patients receive personalized treatment because of their accessibility, reproducibility, non-invasiveness, and cost-effectiveness [17].
Recent studies indicate NLR as a promising novel prognostic marker, which can significantly predict the breast cancer grading and some ICH markers [18, 19]. Studies also show conflicting results regarding the association of PLR and SII with poor prognosis and higher risk of advanced grades, stages, clinicopathological features, and treatment response [15, 20]. Also, there have been insufficient studies to determine the relationship between LANR, HRR, and GLR indices with the clinicopathological characteristics of breast cancer.
This study was designed to explore the association of novel peripheral blood inflammatory indices, including SII, LANR, NLR, PLR, HRR, and GLR, with the clinicopathological features of breast cancer. To our knowledge, this is the first study to propose a diagnostic model for grade and IHC markers of breast cancer based on these indices.

Materials and methods
This study was conducted as a retrospective observational study and was approved by the Ethics Committee of Iran University of Medical Sciences (IR.IUMS.REC.1403.588). The study followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for observational studies.

Sample size calculation
The sample size in this study was computed using Power and Sample software version 3.1.2 by William D. Dupont and Walton D. Plummer for two independent groups. The study from Miao-Feng Wang et al. was adopted as it has the most similar method to this study [21]. With 80% study power, a two-sided α of 0.05, and expected proportions (prevalence of lower grade, including well and moderately differentiated) in high and low NLR values (≤ 2.4 vs. >2.4 in the mentioned study) of 0.28 and 0.5, at least 76 patients were required for the investigation. Based on the anticipated amount of missing and excluded data, the number of 224 patients was determined as the selected sample size.

Study participants
All patients admitted to Rasoul-Akram Hospital for elective breast cancer surgery from September 2022 to 2024 were included in the study.
The inclusion criteria were as follows: (1) All primary localized breast malignancies, including code C50 based on the International Classification of Diseases (2) All the chosen patients underwent elective breast cancer resection.
The exclusion criteria were as follows: (1) patients with a history of any malignancy; (2) patients with any systemic inflammatory condition, infections, hematologic and rheumatologic diseases; (3) patients on antiplatelet therapy or steroid therapy; (4) patients who have recently received platelet products or packed red blood cells; (5) combination of other primary tumors; (6) the patients who did undergo radiotherapy and chemotherapy before surgery.

Sampling and data collection procedure
Variables that were recorded from 114 patients with breast cancer included age, tumor grade (well differentiated, moderately differentiated, and poorly differentiated), PR, ER, and HER2 expression selected as clinicopathological characteristics of the patient according to biopsy reports.
In this study, PR and ER scoring system include 0 as 0% nuclear staining, 1 + as < 20% nuclear staining, 2 + as 20% to 75% nuclear staining, and 3 + as > 75% nuclear staining using the scoring method by Adedayo A Onitilo [22].
Based on the 2023 American Society of Clinical Oncology/College of American Pathologists guideline HER2 scoring system is also defined as [23]:
0 = No staining is observed, or incomplete membrane staining that is faint/barely perceptible and in ≤ 10% of tumor cells.
1 + = Negative. Incomplete membrane staining that is faint/barely perceptible and in > 10% of tumor cells.
2 + = Equivocal. A weak to moderate complete membrane staining is observed in > 10% of the tumor cells.
3 + = Positive. A strong complete membrane staining is observed in > 10% of the tumor cells.
PR, ER, and HER2 was considered positive when scored 3 + or 4 + by immunohistochemistry or confirmed amplification by fluorescent in situ hybridization.
Blood samples were collected 1 week before biopsy, and 6 inflammatory indices were calculated based on patients’ blood samples (Table 1). The association between the indices and the tumor grade, PR, ER, and HER2 expression was analyzed. The flow chart of the sampling and data collection procedure is shown in Fig. 1.

Statistical analysis
Continuous variables with non-normal distributions were expressed as median and interquartile range (IQR). Qualitative variables were expressed as numbers and percentages. The Shapiro-Wilk test was used to check the normality of quantitative variables. The Spearman correlation was used to present the relationship between inflammatory indices. The Mann-Whitney U test was used to determine the association of indices with clinicopathological characteristics. Univariate and multivariate ordinal logistic regression models were applied for analyses of grade, ER, PR and HER2 predictors. In model 1, IHC markers and histological grade are predicted exclusively by markers other than age, IHC markers, and histological grade. In model 2, all clinicopathological characteristics, including ICH markers, age, and histological grade, are included in the analysis. Candidate variables with P < 0.05 in model 2 of multivariate ordinal logistic regression analysis were subsequently entered into a receiver operating characteristic (ROC), and the area under the curve, sensitivity, and specificity values of the diagnostic model were calculated. The ROC curve was used to determine the cutoff value of NLR and GLR in differentiating between low grade (well and moderately differentiated) and high grade (poorly differentiated) groups, and NLR and LANR in differentiating between negative (score 0 and 1+) and positive (score 2 + and 3+) expression of HER2. All statistical analyses were performed using SPSS version 27.0. Two-tailed significance values were used, and P < 0.05 and 95% confidence intervals (CIs) were considered statistically significant.

Results

Clinicopathological characteristics of the patients
This study included 114 patients undergoing tumor resection and final histological investigation. Table 2 represents the clinicopathological characteristics of patients with breast cancer. The median age (range) of patients was 50 (44–60).

More than half of the tumors had a moderately differentiated grade (51.8%). Also, 60.5%, 53.5%, 38.6% of tumors have positive expression of ER, PR and HER2, respectively. Based on results, 20.1% of tumors where ER/PR−, Her2−, 13.1% where ER/PR+, Her2+, 14.9 where ER/PR−, Her2+, and 29.8 where ER/PR+, Her2−.

Correlation between peripheral blood inflammation indices
The results of the correlation analysis showed that all peripheral blood inflammatory indices were significantly correlated with each other, except the HRR, which had no significant correlation with any of the indices. Also, all indices that had a significant correlation with LANR were of the negative correlation type. The correlations between each inflammatory index are shown by correlation matrix in Table 3.

Relationship between peripheral blood inflammation indices and clinicopathological characteristics
Table 4 shows the association between inflammatory indices and clinicopathological characteristics of patients. Among the peripheral blood inflammation indices, all had an insignificant relationship with ER, PR and HER2 expression groups. On the other hand, only NLR had a significant relationship with tumor grade (p value = 0.027); thus, an increasing trend in the NLR value is seen with increasing tumor grade from well differentiated to poorly differentiated (Fig. 2). The only other significant relationship is seen between HRR and age (p value = 0.041), with a trend toward decreasing HRR values ​​as the age groups increase.

Diagnostic value of peripheral blood inflammatory indices for grade, ER, PR and HER2 expression of breast cancer
Table 5 presents univariate and multivariate ordinal regression analyses of grade, ER, PR, and HER2 expression, exclusively in relation to systemic inflammatory indices (model 1). Since no validated SII, LANR, HALP, NLR, PLR, HRR, or GLR cut-off values have been reported in the literature, and the AUC obtained from the ROC curve for most indices was undesirably less than 50%, the medians of the peripheral blood inflammation indices were chosen as cut-off values for grouping, as with similar studies [24–26]: SII (< 434.4 vs. ≥434.4), LANR (< 19.6 vs. ≥19.6), NLR (< 2.10 vs. ≥2.10), PLR (< 99.7 vs. ≥99.7), HRR (< 0.91 vs. ≥0.91), and GLR (< 4.53 vs. ≥4.53). Patients were categorized into either low or high groups. The multivariate analysis indicated no significant associations of SII, LANR, NLR, PLR, HRR, or GLR with grade, PR, or HER2 expression. However, low SII (< 434.4) was significantly correlated with reduced ER expression (p = 0.034). A non-significant trend was also observed for GLR, suggesting a possible decrease in ER expression in the low group.

Table 6 presents model 2, which includes univariate and multivariate ordinal regression analyses of predictors for grade, ER, PR, and HER2 expression in relation to systemic inflammatory indices and clinicopathological characteristics. The results of multivariate analysis show that age, SII, PLR, and HRR have not any effect on the chance of high tumor grade, ER, PR and HER2 expression. NLR and GLR indicates a decrease in the chance of detecting high tumor grade in low NLR (< 2.10) and GLR (< 4.53) values. Fig. 3 illustrates the results of multivariate ordinal regression analysis of the predictors of grade. There also significant results to predict ER expression by the grade of tumor, which show an increase in chance of higher ER expression in low grade tumors, including well and moderately differentiated. The analysis also shows an increase in chance of higher ER expression in low HER2 expression tumors. In contrast, score 0 of PR expression has a decrease in chance of higher ER expression. As same, score 0 and 2 + of ER expression shows a decrease in chance of higher ER expression. NLR and LANR indicates an increase in the chance of detecting positive HER2 expression in low NLR (< 2.10) and LANR (< 19.6) values. There is also an increase by the well differentiated tumors and decrease by negative expression of ER (Score = 0) in chance of higher HER2.

Potential of NLR and GLR for tumor grade diagnosis
Considering the significance of NLR and GLR in the regression model for predicting tumor grade, ROC analyzes were performed on NLR and GLR. Fig. 4 shows ROC analyses pertaining to the NLR/GLR-based diagnostic model. NLR’s best cut-off value for tumor grade diagnosis was estimated to be 2.15 (sensitivity 59.3%, specificity 69.0%, AUC = 65.2%). The diagnostic value of GLR for breast cancer grade was insignificant.

Potential of NLR and LANR for HER2 expression diagnosis
Given the significance of NLR and GLR in the regression model for predicting HER2 expression, ROC analysis was performed on NLR and LANR. Fig. 5 shows ROC analyses pertaining to the NLR/GLR-based diagnostic model. The diagnostic value of both indices for HER2 expression was found to be insignificant.

Discussion
The present study sought to determine whether six routinely available inflammatory indices— SII, NLR, PLR, LANR, HRR, and GLR —could serve as preoperative biomarkers for key clinicopathological features of breast cancer. By relying on measurements from standard blood counts and biochemistry, we aimed to identify cost-effective indicators of tumor biology that might help stratify patients before definitive histopathology and guide personalized management. Although these indices have been widely studied as prognostic markers, their diagnostic value for predicting tumor grade, receptor status and stage before surgery has been less clear [17].
Chronic inflammation is a hallmark of cancer, contributing to tumor initiation, progression and metastasis through mechanisms such as cytokine-driven proliferation, angiogenesis and immune evasion [27]. Indices derived from peripheral counts reflect these processes: elevated neutrophils may promote tumor growth by secreting pro-angiogenic factors, while lymphopenia can signal impaired antitumor immunity [28]. Similarly, platelets facilitate metastatic dissemination by shielding circulating tumor cells, and metabolic markers such as glucose levels may fuel malignant cell proliferation [29].
In our analysis, most inflammatory indices showed no significant associations with tumor grade, PR, or HER2 expression after adjustment in model 1. SII demonstrated a notable relationship with ER status. Patients with low SII had a significantly reduced likelihood of ER expression (OR 0.453, P = 0.034), suggesting that systemic immune-inflammatory burden may influence hormone receptor profiles, consistent with similar studies. However, after including potential confounders in model 2, NLR emerged as the most robust correlate of aggressive tumor features. Patients with NLR above 2.15 were significantly more likely to have high-grade histology (P = 0.027) and advanced clinical grade in multivariable regression (OR 1.32, P = 0.041). ROC analysis yielded an AUC of 0.652 (sensitivity = %59.3, specificity = %69.0), indicating moderate discrimination. Although this performance leaves room for improvement, it is consistent with meta-analytic evidence linking elevated NLR (> 3.0) to worse overall survival (HR 2.56, 95% CI, 1.96–3.35) and disease-free survival (HR 1.74, 95% CI, 1.47–2.07) in breast cancer [30]. Singh et al. likewise reported that an NLR cutoff of 3.417 could distinguish early from advanced grade cases [31]. The lower threshold in our study may capture subtler inflammatory shifts associated with high-grade tumors, suggesting utility for preoperative risk stratification. Recently, Faria et al. validated a prognostic model integrating NLR with clinicopathological factors to stratify early breast cancer patients into distinct relapse-risk groups. Their model, achieving an AUC of 0.74 and externally validated across diverse cohorts, reinforces NLR’s role as a cost-effective biomarker for preoperative risk stratification—aligning with our findings on NLR’s utility in tumor grade assessment [32]. Future work should test combined panels—pairing NLR with other indices or tumor markers such as Ki-67 or circulating tumor DNA—to improve sensitivity and specificity.
LANR showed a strong inverse correlation with NLR (R = − 0.960, P < 0.001), but did not correlate with tumor grade. This index has been proposed as a prognostic marker in cervical and colorectal malignancies [33, 34], and a 2024 study by Wang et al. found that preoperative LANR predicted progression-free survival in operable breast cancer with an AUC of 0.748 [35]. The close relationship between LANR and NLR suggests overlapping inflammatory pathways, but the albumin component may also capture subclinical hepatic or metabolic alterations driven by tumor burden. Larger, prospective cohorts are needed to validate LANR cutoffs and determine whether it adds independent predictive value beyond NLR.
We observed statistically higher HRR values in younger patients (< 35 years) compared with those aged 50 and above (P = 0.041), but no association with histologic grade or receptor status. Prior research identified low HRR (≤ 1.133) as correlating with larger tumor size and advanced stage [36], a finding not replicated here. Discrepancies may stem from differences in patient demographics, nutritional status or prevalence of comorbidities affecting red-cell indices. Given the multifactorial determinants of hemoglobin and red-cell distribution width, HRR may prove more useful as part of a composite score rather than in isolation.
PLR correlated significantly with other indices, yet did not independently predict high grade or hormone receptor status in our regression models. This contrasts with a previous meta-analysis of 20 studies showing that elevated PLR was associated with lymph node metastasis, advanced TNM stage and distant spread [37], as well as Wiranata et al.’s findings linking high PLR to larger tumors and non-luminal subtypes [38]. Variability in cutoff determination, patient selection and tumor heterogeneity likely underlie these conflicting results. Standardization of PLR thresholds and subgroup analyses by molecular subtype may clarify its role.
SII, which integrates neutrophil, lymphocyte and platelet counts, has garnered attention as a prognostic marker in triple-negative and other aggressive breast cancers [39–41]. In our preoperative cohort, however, SII showed only modest associations with tumor grade and receptor status. This may reflect our focus on diagnostic rather than long-term survival endpoints. GLR, reflecting the interplay between hyperglycemia and lymphocyte-mediated immunity, also displayed limited standalone predictive power: although GLR below 4.53 was associated with reduced odds of high grade (OR 0.35, 95% CI, 0.14–0.86), ROC analysis was nonsignificant (AUC 0.579, P = 0.214). Prior studies have linked elevated GLR to shorter progression-free and overall survival through mechanisms of glucose-driven proliferation and lymphocyte depletion [42–44], suggesting that GLR may be most informative in a prognostic context rather than for initial grade prediction.
Accurate prediction of ER, PR and HER2 expression is crucial for treatment selection and have significant implications for prognosis [45]. In our analysis, both NLR and LANR were modest predictors of HER2 positivity, while other indices showed inconsistent associations. Some reports have found links between PLR or SII and HER2 status [46], whereas others have observed no significant associations for preoperative inflammatory markers [47, 48]. These discrepancies underscore the complexity of tumor–immune interactions across molecular subtypes and highlight the need for integrated models combining blood-based biomarkers with imaging or tissue-based assays.
Our findings suggest that NLR—and potentially LANR—could be incorporated into preoperative risk algorithms to inform surgical planning, neoadjuvant therapy decisions or more intensive surveillance, especially in resource-limited settings where access to advanced molecular testing is constrained. A composite inflammatory panel might serve as an early warning system to identify patients at higher risk of aggressive disease who may benefit from expedited management. Cost-effectiveness analyses and decision-curve modelling will be important next steps to evaluate the real-world utility of such an approach.
Key limitations include the retrospective, single-center design and relatively small sample size, both of which constrain statistical power and generalizability. Our cross-sectional analysis lacked longitudinal follow-up, precluding assessment of these indices for predicting survival or treatment response. We also did not evaluate other inflammatory markers such as C-reactive protein or interleukin-6, nor did we explore dynamic changes in indices over time. Future prospective, multicenter studies with larger cohorts should aim to (1) validate optimal cutoffs for each index (2), assess their incremental value over established clinicopathological and molecular models (3), examine combinations of indices and (4) investigate the role of serial measurements in monitoring treatment response.
In conclusion, elevated NLR demonstrated the most consistent association with high tumor grade in our breast cancer cohort, supporting its potential role as a cost-effective preoperative biomarker. LANR showed promise as a complementary indicator, whereas PLR, HRR, SII and GLR exhibited more variable performance. Prospective validation and integration of these indices into composite risk models may enhance patient stratification and guide personalized therapeutic strategies in breast cancer.

Conclusion
Based on the findings of this study, the neutrophil-to-lymphocyte ratio (NLR) emerges as the only inflammatory index with meaningful diagnostic relevance in breast cancer. Elevated NLR was linked to higher tumor grade, while both NLR and LANR showed associations with HER2 expression. Although the diagnostic accuracy of these markers was only moderate, their simplicity, affordability, and availability highlight their potential role in preoperative risk stratification. Incorporating NLR may help identify high-risk patients and guide treatment, especially in resource-limited settings. Due to modest accuracy and lack of validated cut-offs, further multicenter studies are needed to confirm its clinical value and standardize its use.