Decoding the Complexity of Systemic Inflammation Predictors in Locally Advanced Cervical Cancer, with Hemoglobin as the Hidden Key (the ESTHER Study)

Simple Summary We explored whether specific factors, like inflammation indicators in the blood, could help predict treatment outcomes for locally advanced cervical cancer (LACC). LACC is generally treated with a combination of chemotherapy and radiation. We wanted to see if these factors could help physicians personalize treatments for better results. Our study involved looking at various aspects, including inflammation indices in the blood and various clinical treatment details, in LACC patients. While some factors, such as age and hemoglobin levels, seemed to predict outcomes, there was no clear connection between inflammation indicators in the blood and results. These findings challenge previous ideas and highlight the importance of considering multiple factors to predict the prognoses of LACC patients. Abstract Locally advanced cervical cancer (LACC) is treated with concurrent chemoradiation (CRT). Predictive models could improve the outcome through treatment personalization. Several factors influence prognosis in LACC, but the role of systemic inflammation indices (IIs) is unclear. This study aims to assess the correlation between IIs and prognosis in a large patient cohort considering several clinical data. We retrospectively analyzed pretreatment IIs (NLR, PLR, MLR, SII, LLR, COP-NLR, APRI, ALRI, SIRI, and ANRI) in 173 LACC patients. Patient, tumor, and treatment characteristics were also considered. Univariate and multivariate Cox’s regressions were conducted to assess associations between IIs and clinical factors with local control (LC), distant metastasis-free survival (DMFS), disease-free survival (DFS), and overall survival (OS). Univariate analysis showed significant correlations between age, HB levels, tumor stage, FIGO stage, and CRT dose with survival outcomes. Specific pretreatment IIs (NLR, PLR, APRI, ANRI, and COP-NLR) demonstrated associations only with LC. The multivariate analysis confirmed Hb levels, CRT dose, and age as significant predictors of OS, while no II was correlated with any clinical outcome. The study findings contradict some prior research on IIs in LACC, emphasizing the need for comprehensive assessments of potential confounding variables.


Introduction
Cervical cancer stands as a prevalent global malignancy [1].Locally advanced cervical cancer (LACC) patients are commonly treated with concurrent chemoradiation (CRT), based on the simultaneous administration of chemotherapy (CHT) and radiotherapy (RT), and represent the standard therapeutic approach.Though CRT proves effective in achieving substantial local tumor control [2], a notable subset of patients, around one third, experience treatment failure [3,4].Over the past years, interest in oncology has gradually grown in the development of predictive models of outcome.In fact, predictive models can offer the potential for clinicians to foresee clinical outcomes following specific treatments, thereby enabling personalized medical interventions based on individual stages, recurrence risks, and demographic attributes.
Several predictive factors have been scrutinized and incorporated into these models, specifically within the context of LACC.Elements such as tumor size, histological type, lymph node metastases, and Federation of Gynecology and Obstetrics (FIGO) stage have been identified as significant prognosticators linked to overall survival (OS) [5,6].Moreover, anemia has long been acknowledged as an unfavorable prognostic determinant among LACC patients [7][8][9][10].
However, many of these studies have primarily focused on a single index [11,13,16,18,20,26,27] or a limited array of indices [12,14,15,19,23,25], often without a thorough assessment of potential confounding variables [13,16,17,22,23,25,26].Therefore, the principal objective of this study is to comprehensively analyze the predictive capabilities of a spectrum of systemic IIs within an extensive cohort of LACC patients.This analysis will incorporate pertinent clinical prognostic factors, encompassing clinical, tumor-related, and treatmentrelated data.The final aim of this study is to evaluate whether the prognosis of LACC patients can be improved by modulating the treatment based on the values of the IIs.

Study Objective and Design
This study aimed to investigate how various systemic IIs are linked to the prognosis of LACC by examining their effects on key outcomes: local control (LC), distant metastasis-free survival (DMFS), DFS, and OS.
More precisely, our main objective in this analysis was to externally validate the predictive significance of the IIs' pretreatment values, as well as the associated thresholds proposed in the existing literature, within the context of LACC [11][12][13][14][15][16][17][18][19][20][21][22][23][24][25][26][27].Additionally, we pursued a secondary goal of conducting an exploratory assessment regarding the predictive influence of IIs' values observed at the end of CRT, considering the relatively limited information available on the impact of post-treatment IIs' values [15].Lastly, our aim was to assess whether the differences between pre-and post-treatment values (Deltaindexes) showed significant correlations with the outcomes under examination.Also, this evaluation was conducted due to the limited data existing in the literature on this aspect [15,23].
We conducted a single-center retrospective analysis of patients treated at our institution between July 2007 and July 2021.These patients were part of an approved observational study (ESTHER study, code CE 973/2020/Oss/AOUBo) overseen by our local Ethical Committee.All patients provided informed consent to participate, and no exclusions were made to maintain the study's real-world applicability.However, patients were excluded from this analysis in case of treatment performed with palliative intent and if essential data from clinical records were unavailable.This includes situations where necessary blood-test results for calculating the IIs were missing.

Staging, Treatment, and Follow-Up
The retrospective classification of LACCs was based on the 2018 FIGO staging system [28].Patients underwent definitive concurrent CRT, which involved a combination of external beam RT (EBRT) targeting the pelvic area (doses of 45-50 Gy, delivered in fractions of 1.8-2 Gy) and intracavitary interventional RT (brachytherapy-BRT), administered as either pulsed or high dose rate.The goal was to achieve a total equivalent dose of 80-90 Gy for the gross tumor volume (GTV).The clinical target volume (CTV) encompassed the GTV, uterus, upper third of the vagina, parametria, and pelvic nodes (internal, external, common iliac, obturator, and presacral nodes), with a 7 mm expansion.Treatment of para-aortic lymph nodes was considered only if there were nodal metastases in that region.The planning target volume was defined as the CTV with an additional 10 mm expansion in all directions.Metastatic or suspicious pelvic nodes received an additional radiation boost in a sequential or simultaneously integrated timing, reaching a total equivalent dose of 55-65 Gy.Patient alignment was monitored daily using electronic portal imaging devices until 2015 and then shifted to onboard cone-beam CT [29].Concurrent CHT involved intravenous Cisplatin (40 mg/m 2 weekly).Patients received follow-up through physical examinations every three months for two years and subsequently every six months for three years.Thoracic-abdominal-pelvic computed tomography scans were performed as needed clinically or every six months during the first two years and annually in the subsequent three years.

Patient, Tumor, and Treatment Information
This analysis encompassed patient-related details such as age and hemoglobin (Hb) level, measured in g/100 mL).Additionally, we considered tumor-related information, including histological type, FIGO stage, clinical tumor stage, clinical nodal stage, and maximum tumor diameter.Moreover, treatment-related data comprised RT technique, EBRT dose and fractionation applied to the pelvic region, BRT boost dose, total tumor dose, and overall treatment time (including both EBRT and BRT, measured in days).
Considering the primary aim of our analysis, which involves externally validating a range of IIs assessed in cases of LACC, along with the varying threshold values outlined in the existing literature, we undertook the process of dichotomizing the index-related data.Specifically, we utilized the published cut-off points, particularly focusing on those associated with significant clinical outcomes.Additionally, we conducted an exploratory analysis on the prognostic significance of the II values assessed at the conclusion of concurrent CRT and of pre-post-treatment variations of the indices (Delta-indices).In this scenario, we dichotomized the parameters using the median value, owing to the relatively limited availability of published data in this context.

Statistical Analysis
Patient and tumor characteristics, as well as treatment data, were depicted using descriptive statistical methods.Categorical data were presented using numbers and percentages, while continuous data were expressed in terms of medians and ranges.LC was computed as the time elapsed from the start of concurrent CRT until the evidence of local-regional recurrence, as identified through imaging studies or clinical observations, or until the last follow-up in patients without pelvic recurrence.DMFS was computed as the time span from the initiation of CRT until the occurrence of distant failure, detected through imaging studies or clinical observations, or until the last follow-up in patients without extrapelvic recurrence.DFS was calculated as the period from CRT initiation until any treatment failure or until the last follow-up in patients without a recurrence of LACC.OS was defined as the interval between CRT initiation and the time of death or the most recent follow-up date.
For each of these four endpoints, survival curves were generated using the Kaplan-Meier method, and a univariate analysis (log-rank) was carried out, encompassing all the specified variables.Furthermore, a multivariate Cox's regression analysis was conducted, involving variables with a p-value of less than 0.1 in the univariate analysis.A significance level of 5% was employed (p-value < 0.05).If the assessment of various cut-off points for a particular II indicated statistical significance for only one specific cut-off, that particular value was exclusively integrated into the multivariate analysis.Moreover, when multiple cut-off values exhibited statistically significant results, only the cut off linked to the lowest p-value was integrated into the multivariate analysis.The analysis was performed using SPSS for Windows (version 20.0; SPSS Inc., Chicago, IL, USA).

Patient Demographics
A total of 173 patients were included in this analytical study.The characteristics of the patients are comprehensively presented in Table 1.The median age at the time of diagnosis stood at 56 years, ranging from 27 to 85 years, while the median duration of follow-up was 36 months, spanning from 3 to 151 months.

Treatment Aspects
All patients underwent concurrent CRT with weekly administration of Cisplatin.Detailed treatment-related characteristics can be found in Table 1.For patients with positive lymph nodes (57 cases), an additional dose was administered either sequentially or simultaneously, resulting in a median total dose of 57.5 Gy (ranging from 52.5 to 61.0 Gy).BRT was administered to all patients, with a median dose of 37 Gy for pulsed-dose-rate BRT (ranging from 23 to 39 Gy) and 28 Gy for high-dose-rate BRT (ranging from 4 to 42 Gy).[30]).

Treatment-Related Parameters
Concerning treatment factors, higher total RT doses were significantly associated with improved OS (p-value = 0.012), while no significant variations were noted based on treatment duration (Table 2).

Tumor-Related Parameters
Our analysis showed significant correlations as follows: patients with lymph node metastases experienced worse DMFS (p-value = 0.045), and individuals with more advanced FIGO stages exhibited worse LC (p-value = 0.005), DMFS (p-value = 0.021), DFS (p-value = 0.003), and OS (p-value = 0.032).However, no significant differences were observed concerning maximum tumor diameter and histological type.

Treatment-Related Parameters
Concerning treatment factors, higher total RT doses were significantly associated with improved OS (p-value = 0.012), while no significant variations were noted based on treatment duration (Table 2).

Post-Treatment Inflammatory Indices
The sole statistically significant correlations were the most favorable LC in patients with NLR < 5.66 (p-value = 0.037) and the most favorable DMFS rate in patients with a systemic inflammatory response index (SIRI) < 3.50 (p-value = 0.018) (Supplementary Table S1).

Delta Indices
The dynamic assessment of the IIs did not exhibit any significant correlation with the considered outcomes (Supplementary Table S2).

Discussion
We conducted a validation study to assess whether IIs and their associated cutoff values could serve as indicators of prognosis in LACC.Our study included an analysis of patient, tumor, and treatment-related clinical parameters.The univariate and multivariate analyses confirmed that established clinical factors, such as age, FIGO stage, Hb levels, and RT dose, influence clinical outcomes [31].The univariate analysis also showed a potential impact, primarily on LC, of IIs, like NLR, PLR, APRI, ANRI, and COP-NLR.The multivariate analysis including only IIs showed the significance of APRI in predicting LC.However, our examination of post-treatment IIs showed a correlation only between SIRI with DMFS and DFS.This discrepancy suggests that the prognostic relevance of different IIs might vary based on their evaluation either before or after treatment, as previously reported [15].
Furthermore, the lack of correlations between Delta-IIs and outcomes suggests the limited utility of the dynamical assessment of IIs over time.Moreover, the primary clinical interest resides in pretreatment IIs values due to their potential to guide personalized treatment adjustments.Conversely, after CRT and BRT, there is no evidence to support additional treatments, particularly CHT, to improve clinical outcomes [32,33].
The results of this analysis diverge from our prior study [30], in which a multivariate analysis indicated a notable correlation between higher SII values and poorer DMFS.In contrast, our present multivariate analysis reveals a significant association between APRI and LC.This discrepancy can likely be attributed to differing analytical approaches in the two studies.
In fact, in our prior study, the primary objective was to identify which IIs warranted further scrutiny.Consequently, we assessed IIs as continuous variables in statistical analyses.Conversely, our current analysis is focused on external validation of IIs and their established cut-off values from the scientific literature.As such, we evaluated IIs as dichotomized data, precisely aligned with the published thresholds.
Moreover, our analysis distinguishes itself by comprehensively addressing a broader spectrum of potential confounding factors compared to prior studies (Table 6).In fact, certain analyses omitted the consideration of important factors, such as the FIGO stage [17,26] or nodal stage [13,16,22,23,25].Additionally, treatment-related variables were frequently omitted in many studies [12,14,15,17,19,20,[22][23][24][25][26].Notably, only our study, and the paper of Koulis et al. [10], integrated Hb levels into the analysis.Interestingly, even in their analysis, no significant correlation was observed between the chosen II (NLR) and the outcome of interest, and only anemia emerged as a factor significantly associated with lower OS rates in the multivariable analysis.
On the basis of these considerations, we hypothesize that, compared to other analyses, our study has a lower risk of confounding bias.In fact, when confounding factors are not adequately controlled for in the study design or analysis, they can distort the observed relationship between the independent and dependent variables.This bias can either exaggerate or mask a real effect.
Our study has limitations that warrant acknowledgment.Certain parameters were not included in our analysis.For instance, squamous cell carcinoma antigen data, with known prognostic significance [34,35], were not incorporated due to their unavailability in most of our patient cohort.Moreover, certain other IIs, such as platelet-to-neutrophil ratio, monocyte-to-neutrophil ratio, platelet-to-white blood cell ratio, platelet-to-monocyte ratio, lymphocyte-to-monocyte ratio, eosinophil-to-lymphocyte ratio, and eosinophil-tomonocyte ratio, were not assessed in our analysis [36].On the other hand, it was our opinion that the most common IIs, and especially those most correlated with prognosis in previous studies, were included in our analysis, and that, therefore, it was preferable not to further burden our evaluation.
Additionally, the retrospective nature of our study and lack of preliminary power analysis introduces potential limitations to the precision of outcome evaluations.However, it is important to underscore the strengths of our study, including the substantial number of cases examined and the comprehensive analysis of a wide array of clinical parameters.Furthermore, our study offers potential utility by validating results already published within the scientific community.Legend: ALRI: aspartate aminotransferase-to-lymphocyte ratio index; ANRI: aspartate transaminase-to-neutrophil ratio index; APRI: aspartate aminotransferase/platelet count ratio index; BLR: basophil/lymphocyte ratio; BMI: body mass index; cN+: clinical positive nodes; COP-NLR: combination of platelet count and neutrophil-to-lymphocyte ratio; CR: complete response; CRT: chemoradiation; DFS: disease-free survival; ELR: eosinophils-lymphocyte ratio; FIGO: International Federation of Gynecology and Obstetrics; Hb: hemoglobin; LLR: leukocyte-to-lymphocyte ratio; MLR: monocyte-to-lymphocyte ratio; N: nodal; NLR: neutrophil-to-lymphocyte ratio; OS: overall survival; PFS: progression-free survival; PLR: platelet-to-lymphocyte ratio; RT: radiotherapy; SII: systemic immune inflammation index; SIRI: systemic inflammatory response index; T: tumor; * COP-NLR scored as follows: 0: NLR < 3 and PLT < 300; 1: NLR > 3 or PLT > 300; 2: NLR > 3 and PLT > 300.#: different cut off from published studies.

Conclusions
In summary, our analysis, along with other studies, presents conflicting outcomes that currently do not support the routine use of IIs as prognostic tools in patients with LACC.To enhance the clarity and reliability of future investigations, a comprehensive inclusion of potential confounding variables is needed.Based on our findings and those of Koulis et al. [10], the consideration of Hb cannot be overlooked, and that the correction of anemia is a key element in LACC patients treated with CRT.
Furthermore, advanced statistical methodologies and collaborative efforts could enhance the accuracy of prognostic models.Constructing large databases through cooperative initiatives can provide a robust foundation for developing predictive models.Notably, comprehensive evaluations encompassing multiple IIs may prove more effective in prognostication than the analysis of individual parameters.As illustrated by the study by Lee et al. [14], the combined assessment of pretreatment NLR and PLR values demonstrated a stronger association with worse OS.

Figure 1 .
Figure 1.Impact of pretreatment hemoglobin level on actuarial local control.Figure 1. Impact of pretreatment hemoglobin level on actuarial local control.

Figure 1 .
Figure 1.Impact of pretreatment hemoglobin level on actuarial local control.Figure 1. Impact of pretreatment hemoglobin level on actuarial local control.

Figure 1 .
Figure 1.Impact of pretreatment hemoglobin level on actuarial local control.

Figure 2 .
Figure 2. Impact of pretreatment hemoglobin level on overall survival.

Figure 2 .
Figure 2. Impact of pretreatment hemoglobin level on overall survival.

Table 2 .
Univariate analysis of patients, tumors, and treatment characteristics; survival outcomes are expressed in percentages.

Table 3 .
Univariate analysis of inflammatory indices (pretreatment values); survival outcomes are expressed in percentages.

Table 4 .
Multivariate Cox's analysis of inflammatory indices (pre-treatment values); survival outcomes are expresses in percentages.Only statistically significant values (and values with trend for significance) are shown.

Table 5 .
Multivariate Cox's analysis of inflammatory indices (pre-treatment values) and clinical parameters.Only statistically significant values (and values with trend for significance) are shown.

Table 6 .
Comparison between the results of previous analyses and those of our series (adapted fromFerioli et al., 2023 [30]).