Prognostic value of pro-inflammatory markers at the preoperative stage in Algerian women with breast cancer

Breast cancer (BC) is still at the top list among cancers with high incidence and mortality in Algeria^[1]. Indeed, the recent data show that BC takes the first place representing 40% of new cases registered in 2020 among women cancers. Moreover, this neoplasm ranked as the second place in term of mortality in both sexes^[1]. Chronic inflammation, whether occurring before or after cancer, can promote the occurrence, development, and metastasis of cancer^[2]. The systemic inflammation is one of the key hallmarks investigated during the study of the interaction between the tumor cells and the host^[3]. Also, inflammatory proteins, immune cells, and cytokines are all present and readily observable in systemic circulation conducing to dynamic and systemic immune response^[4]. Among them, neutrophils and monocytes (which will differentiate into macrophages) are first immune cells infiltrating the tumor microenvironment site and have a major contribution during the immunoediting process with dual function acting as pro- or anti-tumoral activity in a context-dependent manner^[5]. Thus, their inflammatory mediators like proinflammatory cytokines/chemokines and angiogenic factors in addition to components of the extracellular matrix will thereby influence the systemic inflammatory response in BC patients^[6]. In addition, platelets have been shown to be associated with the tumor growth and angiogenesis through the secretion of plethora of molecules including transforming growth factor β and vascular endothelial growth factor leading to more inflammatory cells migration^[7]. In contrast, it is well reported that the most anti-tumor activity is mediated by cytotoxic T cells, and solid tumors with high tumor infiltrating lymphocytes (TIL) were associated with good prognosis in breast cancer^[8]. This could lead to speculate that the low lymphocyte number could be associated with the poor outcome^[9]. In the same way, a study conducted on breast cancer indicates that patients show different times to disease progression according to their molecular subtype^[10]. For that, biomarkers are required to enhance the prognosis of patients. Thus, various blood markers of inflammation have been evaluated in patients with different malignant tumors. These indicators include neutrophil, monocyte, platelet, and lymphocyte numbers allowing a calculation of several ratios and indexes like neutrophil/lymphocyte ratio (NLR), platelet/lymphocyte ratio (PLR), monocyte/lymphocyte ratio (MLR), systemic immune-inflammation index (SII), and systemic inflammatory response index (SIRI).

Cumulating evidence shows that systemic inflammation markers are associated with an immune response that promotes cancer and are a worse prognostic factor for cancer patients^{[11, 12]}. However, to date, no publications shed the light on their role in Algerian patients with breast cancer. Accordingly, we set two major aims: (i) the selection of the optimal cutoff point of these biomarkers for each endpoint (i.e., RFS, DMFS, OS) and (ii) the evaluation of their relationships with clinical parameters and patient’s survival.

Despite the easy way to obtain them and the affordable cost of these markers, to the best of our knowledge, no study has been performed in Algerian patients, and only one research study reported North African patients with inflammatory breast cancer, which is considered as a rare subtype^[13]. Analyzing systemic inflammatory markers in the blood may be one way to comprehend a role of inflammation in breast cancer risk^[6]. There are more evidence suggesting that the systemic inflammatory response promotes tumor progression by modifying the interactions between neoplastic and non-neoplastic cells. Hematopoiesis is aberrant in individuals with solid tumors, resulting in altered hematopoietic progenitor cell composition and myeloid-biased differentiation^[4]. Thus, the imbalance in neutrophil, monocyte, and platelet counts relative to lymphocytes is observed. Remarkably, most of cutoffs used in publications are issued from Asian, European, or American populations^[14]. Also, several studies showed discrepancy in cutoff selection for inflammatory markers selected at the different endpoints. Moreover, most of studies reported their optimal cutoff based on OS/RFS/DFS instead of DMFS. For that, we aimed to set up our optimal cutoff point selection for these markers at different selected endpoints such as RFS, DMFS, and OS for better relapse-risk stratification. As first observation, these cutoffs were slightly different with respect to those found in literature. Hence, defined cutoff values for these inflammatory markers in our population are crucial for further investigations especially since studies showed disparities in the NLR ratio depending on racial differences^{[15, 16]}. When RFS was selected as endpoint, the NLR cutoff value was 2.11. Various NLR cutoffs were identified in previous reports ranging from 1.34 to 4^{[14, 17]}. Our NLR cutoff is close to those reported by Truffi et al. with 2.21 as NLR optimal cutoff for DFS as endpoint in study including 1806 patients with breast cancer meeting our inclusion and exclusion criteria^[18]. When DMFS was selected as endpoint, the NLR cutoff value was 1.6, which is relatively low compared to that of Orditura’s study with NLR value of 1.97^[19]. Also, a few studies used this endpoint for the estimation of their cutoff values for all five markers.

Respective to the molecular subtype, non-metastatic breast cancer, and DFS/RFS as endpoint, data extracted from meta-analysis showed that the MLR cutoff value ranged from 0.19 to 0.22 in Asian population^[20], whereas our MLR optimal cutoff value was lower defined at 0.16. The decrease observed in this ratio may be attributed to disturbance in monocyte and/or lymphocyte counts where their numbers could be affected by inherited (e.g., age, ethnicity) and environmental factors rather than pathological condition^{[21, 22]}. Thereby, larger studies in the same population are required to minimize these fluctuations. Our data showed that the MLR value of 0.28 was the optimal cutoff value selected for DMFS as endpoint, which is similar to results of Truffi et al. and He J et al. where they reported 0.28 and 0.2632 as the MLR values, respectively^{[18, 23]}.

For SII, a meta-analysis based on seven articles from Asian patients highlighted that SII cutoffs varied from 442 to 624^[24]. In our cohort, the optimal cutoff value of SII was 472 for both RFS and DMFS. In another publication, Li et al determined a best cutoff for SII at 514 and a high value was associated to worse prognosis in breast cancer patients undergoing surgery with 6 years of follow-up^[25].

SIRI is the less investigated index among other markers in breast cancer. Instead, some studies reported the optimal cutoff values of 0.65 for OS^[26] and 0.54 for OS in another study^[27], but no publication has set a SIRI cutoff with RFS and DMFS as endpoint in BC even those including patients receiving NACT^[28]. In our cohort, the cutoff was set at 1.38 for OS which is higher compared to that in literature. Nevertheless, we should be careful in drawing assumptions since the AUC and p value were far from the significance threshold. Overall, in our cohort, for both SII (cutoff value = 472) and SIRI (cutoff value = 0.688), the optimal cutoff value was identical for both RFS and DMFS as endpoints which could be considered as advantageous for further prospective studies compared to other markers.

After that, we sought to determine any correlations between the above-mentioned markers and the clinical parameters. Although many studies associated high values of these indexes with clinical features such as tumor stage, histological and molecular types in breast cancer and other malignancies^{[29, 30]}. In our cohort, we failed to demonstrate any relationships between clinical parameters with above-cited indexes. This result is in line with meta-analysis data showing no significant association between LMR (inverse of MLR) and clinical parameters^[20].

Patient survival can be predicted using systemic inflammatory indexes, which may be connected to a variety of underlying mechanisms. Tumor-induced inflammation and alterations in immune function play a significant role in the link between NLR/MLR and the prognosis of the tumor. NLR is the most inflammatory parameter investigated in multiple diseases including breast cancer. High level of this ratio was associated in most studies to poor prognosis^[14]. However, in our study, we did not observe a statistically significant difference between group with high NLR versus low NLR when assessing patients’ survivals. The prognostic value of NLR could be inconsistent among breast cancer subtypes. Indeed, some studies linked poor prognosis of high NLR to ER+/PR+ subtypes, whereas others linked it to ER‒/HER2‒ or TNBC subtype^{[14, 17, 31,32,33]}. Unfortunately, we did not run this analysis in our cohort due to the small number of patients in each subgroup. In the same way and in a study conducted on non-small-cell lung cancer (NSCLC), it was reported that high NLR is related to cytokines/chemokines production increase such as TNF-α, IL-6, and IL-8. These cytokines were linked to poor prognosis and lower response to immune checkpoint inhibitors^[34]. Considering these data, further investigation combining NLR to cytokines/chemokines profile would strength our understanding of breast cancer pathophysiology.

Even though PLR was reported to play a pivotal role as poor prognosis in different malignancies, it did not have a prognostic role in our cohort. In fact, its prognostic impact had been a subject to controversies in breast cancer. In one hand, many studies reported that PLR was implicated in prognosis, while high values were associated to worst one^{[18, 35, 36]}. In the other hand, PLR was reported to be not associated to poor prognosis in TNBC^[37].

Interestingly, when combining neutrophils with platelets (SII), the prognostic effect was more prominent. Moreover, there was a positive correlation between neutrophils and platelets with p value at the border of significance (P= 0.05; data not shown). Many studies revealed a crosstalk between platelets and neutrophils associated to cancer promotion in part via the induction of NETosis^[38]. Also, platelet–neutrophil network participated in metastatic niche formation through interaction with tumor cells via CXCL-5/CXCL-8/CXCR-2 axis^{[39, 40]}. Together, these complex interactions highlight a potential role of tumor-induced NETosis as a poor prognosis in cancer patients^[41]. Likewise, SII was significantly associated to poor prognosis regardless of the molecular subtypes^[25]. Moreover, a meta-analysis confirmed that SII was a predictive marker of relapses and patient’s survival in breast cancer^[24]. Accordingly, these findings allow us to suggest the great potential of this index as a prognostic parameter instead of calculating NLR and PLR separately.

Importantly, the association of MLR with high degree of distant metastasis in our cohort evokes its contribution to this process. Likely, Gerratana et al. found that high MLR was associated with number of sites at the onset of metastatic disease^[42]. In the same line, Liu et al. reported a significant inverse correlation between LMR in NSCLC patients and brain metastasis especially linked to female gender^[43]. Thus, studying the link between breast cancer and monocytes’ interactions with female hormones could constitute a promising perspective. Moreover, CCL-2 a potent monocyte chemoattractant is involved in various malignancies and is associated to nodal and distant metastasis^[44,45,46]. Thus, mechanisms behind CCL-2-induced metastasis are multiple, in part, via the induction of epithelial–mesenchymal transition, recruitment and/or polarization of immunosuppressive cells like neutrophils, regulatory T cells, and type 2 macrophages^[47]. In a previous study, Wangchinda et al. evoked that different breast cancer subtypes were associated with different times to disease progression with liver and brain as common sites of metastasis associated with early relapse among patients^[10]. Taken together, further studies are crucial to dig the association of MLR values with different molecular subtypes and organs linked to distant metastasis.

SIRI was calculated based on neutrophil, monocyte, and lymphocyte counts. In this study, only SIRI had a significant association with patients’ survivals for three endpoints (RFS, DMFS, and OS). At the cellular and molecular levels, many studies reported interactions between monocytes/macrophages and neutrophils through the release of cytokines mainly IL-17 by γδT cells which in turn regulates G-CSF production enhancing neutrophil recruitment. Thus, IL-17 production is under IL-1β expressed by tumor-associated macrophages (TAM)^[48]. In addition, neutrophils that are releasing IL-8 and TNF-α lead to recruitment and activation of macrophages^[49]. Remarkably, a positive correlation was found between monocytes and neutrophils in glioblastoma rising the importance of these innate cells in maintaining immunosuppressive microenvironment^[50] reflecting the poorer prognosis at the clinical level.

Given this context, we suggest that SII and SIRI depict the equilibrium between the host’s immunological status and the rate of tumor progression, and it may be able to forecast a patient’s prognosis for BC. However, larger investigations are required to validate these suggestions. Additionally, more studies combining these indexes (SII and SIRI) will be the suitable perspectives for better patient’s stratification for treatment options. Also, it appears important to assess these biomarkers with the cellular infiltration into the tumor microenvironment with the aim to see any correlation between the local immune response and the systemic inflammation especially in HER2+-enriched and triple-negative subtypes^{[37, 51]}. Indeed, in a recent paper, Onagi et al. demonstrated that TIL and more strongly regulatory T cells correlated positively with both NLR and PLR in TNBC accompanied with high PDL-1 expression^[37]. This could suggest the existence of local immunosuppressive environment in this subtype.

In our study, the survival curves were more relevant for these markers based on their respective cutoffs at RFS and DMFS but not OS endpoint. This could be explained by the fact that breast cancer management has improved these recent years allowing to more delays regarding relapse and survival among patients which suggest to extend the time of follow-up at least to 20 years.

However, the present study presented several limitations due to the small number of the cohort which limited our analysis with conclusive remarks regarding the association of these inflammatory markers with BC subtypes especially inflammatory breast cancer (IBC) and treatment strategies. Also, multivariate analysis represents another limitation to the study as we could not demonstrate the independent prognostic value of these markers regardless of other co-variables such as molecular subtypes. For example, Wangchinda et al. demonstrated that molecular subtypes were behind early or late recurrence appearance (beyond 5 years of follow-up); with ER+/PR+ and HER2-patients associated to later recurrence than other subtypes^[10]. Moreover, in a meta-analysis, the authors showed that the poor prognosis of NLR was strongly associated with ER-/HER2-subtype^[32].

In summary, our study identified the optimal cutoff point for each inflammatory marker based on three endpoints (RFS, DMFS, and OS) and highlighted the important role of MLR, SII, and SIRI as prognostic tools in monitoring breast cancer patients in our population.