Introduction

Carcinoma cervix is the most prevalent disease among member countries of the developing world. The worldwide incidence of carcinoma cervix as per GLOBOCON 2020 data was 604,100 in 2020 and deaths due to cervical cancer was 341,831 in 2020 alone.1 This calls for a need for better prognostication of the patients with cervical cancer. The complete blood count (CBC) is a cost-effective method to study the intensity of the inflammatory process in a known case of malignancy. Hanahan and Weinberg described the hallmarks of cancer as sustaining of proliferative signaling, evasion of growth suppressors, resistance against cell death, enabling of replicative immortality through means of angiogenesis, and activation of invasion and metastasis.2 Inflammation provides for all these through multiple pathways.

Inflammation mediates the neoplastic process in both its early as well as late stages of development. In the early stages of the neoplastic process, inflammatory cells act as tumor promoters. In the late stages of neoplastic process, tumor cells vitiate immune processes to favor tumor development. Distant metastasis is made possible by coating of tumor cells by certain inflammatory molecules like selectins and chemokines, hence enabling receptor recognition which helps in spreading cells via lymphatics and capillaries and evading host defense mechanisms.3 Neutrophils, eosinophils, and monocytes are recruited as macrophages in the inflammatory process and lymphocytes along with various cytokines and chemokines participate in this inflammation-mediated tumor proliferation activity.34 A high neutrophil count is therefore a surrogate marker for tumors with greater proliferative capacity and provides insight into cancer with a higher chance of posing risk of recurrence and metastasis.5

The neutrophil to lymphocyte ratio (NLR) and platelet to lymphocyte ratio (PLR) have been linked to poor prognosis in several solid malignancies like head and neck cancer, esophageal cancer, colon cancer, and gynecological malignancies.67891011

Materials and Methods

This retrospective study included 208 patients treated with definitive chemoradiotherapy at the Institute of Post Graduate Medical Education and Research (IPGME&R), Kolkata, West Bengal, India, from June 2017 to December 2019. Staging was done with the 2018 International Federation of Gynecology Obstetrics (FIGO) staging system.12 The diagnosis was confirmed by histopathological examination of tissue biopsy from cervical growth. The staging workup included clinical examination, review of the histopathological report, contrast-enhanced computed tomography scan (CT scan) of abdomen and pelvis, contrast-enhanced magnetic resonance imaging pelvis, cystoscopy, rectosigmoidoscopy, CT scan thorax, or chest X-ray. The routine blood examinations included CBC, liver function test, and renal function test, done in all patients.

The inclusion criteria were as follows: (1) age more than 18 years, (2) without any history of prior chemotherapy or radiotherapy, and (3) histologically confirmed case of squamous cell carcinoma cervix. Patients were excluded from the study if having cervical carcinoma other than squamous cell carcinoma, distant metastases at presentation, with any chronic diseases, and with any hematological disorder or autoimmune disease or on corticosteroid therapy. Clinical data of all patients were recorded along with CBC test reports, at the time of diagnosis. NLR and PLR were calculated as the ratio of absolute neutrophil or platelet counts, respectively, to the absolute lymphocyte count. All patients received external beam radiotherapy (EBRT) with telecobalt unit (Bhabhatron II, Panacea Medical Technologies, Bengaluru, Karnataka, India). ONCENTRA (Nucletron, Elekta AB, Stockholm, Sweden) EBRT treatment planning system was used for three-dimensional conformal radiotherapy planning. All patients received 50 Gray (Gy) in 25 fractions, 2 Gy per fraction, 5 fractions a week concurrently using weekly injection cisplatin 40 mg/m². After completion of EBRT all patients received intracavitary radiotherapy using the MicroSelectron Brachytherapy Afterloading (Elekta AB) platform equipped with Ir¹⁹² high dose rate brachytherapy. A gap of 5 to 7 days was allowed after the completion of EBRT. The dose prescription of brachytherapy was 7 Gy each fraction once weekly for total of 3 to 4 fractions depending upon the stage of the disease and clinical findings at the end of EBRT. Cervical cancers with stage I the EQD2 (equivalent dose in 2 Gy per fraction) doses of 80 Gy was considered and EQD2 dose of > 80 Gy was considered for stage IIA or more. Follow-up data including clinical findings and imaging every three monthly after treatment completion were also recorded. Local recurrence of the disease was confirmed by biopsy and defined as locoregional recurrence (LRR). Time interval from initiation of treatment to progression of the cervical cancer is used to calculate the progression free survival (PFS).

Results

Optimal Cutoff for NLR, PLR, and ROC Curve

The optimal cutoff points of NLR and PLR were analyzed for LRR of the patients using ROC curve. The results showed that 2.45 was the optimal cutoff value for NLR (AUC = 83.9%; 95% confidence interval [CI] 0.782–0.896; p < 0.001), at which there was significant difference in LRR. The sensitivity and specificity were 82.6 and 77.7%, respectively, at NLR cutoff of 2.45. Patients were divided between high and low NLR risk group on the basis of NLR cutoff value, patients with NLR ≥ 2.45 were included in the high-risk and NLR < 2.45 were included in the low-risk group. The frequency of distribution of NLR risk group, that is, low-risk NLR and high-risk NLR groups were 57.7 and 42.3%, respectively. The high-risk NLR group was significantly associated with LRR (p-value < 0.001). Similarly, the optimal predictive cutoff value of PLR was 140.6 (AUC = 89.2%; 95% CI 0.847–0.936; p < 0.001), at which there was significant difference in LRR. The sensitivity and specificity were 85.5 and 80.6%, respectively, at PLR cutoff of 140.6. Patients were divided between high and low PLR risk group on the basis of PLR cutoff value, patients with PLR ≥ 140.6 were included in the high-risk and PLR < 140.6 were included in the low-risk group (Fig. 1). The distribution of low-risk and high-risk PLR groups were 58.65 and 41.35%, respectively. The high-risk PLR group was significantly associated with LRR (p-value < 0.001) (Table 4).

Close

Fig. 1: Redline—The receiver operator characteristics curve of the cutoff point for NLR; area under ROC curve = 83.9%; sensitivity = 82.6%; specificity = 77.7%. Blue line—The ROC curve of the cutoff point for PLR; area under ROC curve = 89.2%; sensitivity = 85.5%; specificity = 80.6%. NLR, neutrophil to lymphocyte ratio; PLR, platelet to lymphocyte ratio; ROC curve, receiver operating characteristics curve.

Export to PPT

Table 4

Association of NLR, PLR, and locoregional recurrence

		Locoregional recurrence				p-Value
Variable	Risk group	Yes		No
		Count	N %	Count	N %
NLR group	High risk	57	27.4	31	14.9	< 0.001
	Low risk	12	5.8	108	51.9
PLR group	High risk	59	28.4	27	13	< 0.001
	Low risk	10	4.8	112	53.8

Abbreviations: NLR, neutrophil to lymphocyte ratio; PLR, platelet to lymphocyte ratio.

Survival Analysis

Note that 33.2% patients had progression of the disease or LRR at the time of analysis. The median PFS was not reached. The mean PFS was 40.133 months (95% CI 38.290–41.977). Three-year PFS was 100, 71, and 33% in stage I, stage II, and stage III, respectively (Fig. 2).

Close

Fig. 2: Kaplan–Meier survival curve, estimated PFS with statistically significant difference in PFS according to stage of cervical cancer with log-rank p < 0.001. PFS, progression-free survival.

Export to PPT

Discussion

Inflammation plays major decisive role at tumor development including initiation, promotion, malignant transformation, invasion, and metastasis. It also acts as immune surveillance and response to therapy. Elevated inflammatory mediators are related to poor prognosis in patients with cancer.3 Several studies demonstrated raised NLR and PLR as a well-known poor prognostic factor of cervical cancer.13141516171819 Tavares-Murta BM et al were one of the first groups to study hematological markers and their prognostic value in cervical cancer and peculiarly included cases of cervical intraepithelial neoplasia (CIN) as well as all stages of cervical cancer. Their conclusions were rather paving a way for extensive research in this direction. They found out how neutrophilia, leukocytosis, and lymphopenia were more frequently seen in advanced stage cervical cancer and that neutrophilia was more common in early-stage cervical cancer cases when compared with cases of CIN. Also, an NLR value of more than 5 was more common in late-stage cervical cancer.13

Median age of the study population was 50 years. This correlated well with other studies.1417 Stage IIB was the most prevalent stage consisting of 32.2% of the study population followed by stage IIIB (28.8%). A similar finding was seen in a study by Haraga et al where stage IIB (42.1%) was followed by stage IIIB (27.3%). In the same study they found NLR ≥ 2.78 (AUC = 0.635, sensitivity: 63.9%, specificity: 49.2%) and PLR ≥ 172.50 (AUC = 0.597, sensitivity: 58.3%, specificity: 55.9%) as the most accurate cutoff values for predicting PFS.20 This NLR value differ little from our study but a striking similarity in mean value of NLR was found with the study by Prabawa et al where NLR was 2.68 in early-stage group scenarios.21 In some studies PLR correlated with survival outcomes1922 while some found NLR to have correlation with survival outcomes,152023 whereas some found both NLR and PLR to correlate with poorer survival outcomes.162425

In this study, ROC curve analysis of local recurrence yielded the NLR cutoff value as 2.45 (AUC = 0.839) with a sensitivity of 82.6% and specificity of 77.7% and the PLR cutoff value 140.6 (AUC = 0.892) with a sensitivity of 85.5% and specificity of 80.6%. This matched with another study, the sensitivity value for NLR was 82% and specificity 71% with a cutoff of 3.38 using ROC analyses and 72 and 70% with cutoff of 172.05 for PLR, respectively, for predicting stage with NLR and PLR.21

On univariate regression analysis, addiction, FIGO stage, tumor grade, NLR risk group, and PLR risk group were significantly associated with LRR. NLR and PLR were the statistically significantly associated prognostic factor for recurrence on multivariate analysis. NLR also was found to have the most significant hazard ratio with disease recurrence in a study by Lee et al where for every increase of 0.1 in NLR, the risk of progression and death increased by 13 and 19%, respectively.14 Santos Thuler et al performed a study with 1,266 patients in 2021, to find predictors of worse overall survival (OS) in the form of NLR (> 2.57), PLR (≥ 146.70), derived NLR (≥ 1.778), and PLR + NLR in combination but only in locally advanced disease.25 Another study tried to correlate these hematological parameters with stage of cervical cancer and found that patients with higher NLR and PLR have a tendency for higher stage disease and additionally, on ROC curve analyses, only NLR and PLR were feasible to be used as predictive models.21 In 2016, Zheng et al found PLR and serum albumin as prognostic factors in early-stage cervical cancer patients; the cutoff value for PLR was 128.3, high PLR risk groups demonstrated shorter OS and disease-free survival with significant p-values.22

Zhu et al tried to predict parametrial involvement with help of NLR and PLR. Furthermore, they discovered PLR as an independent prognostic factor for PFS and OS on univariate analyses while no such relation was found between NLR and survival outcomes.19 A PLR of more than 322 predicted poorer prognosis for recurrence of cervical cancer in patients who underwent chemoradiotherapy,26 and Mizunuma et al found low NLR to predict complete response to treatment and NLR was the only prognostic factor for PFS and OS.23

NLR and PLR also related to prognostication of patients with early-stage disease. A Chinese study analyzing 460 cervical cancer patients undergoing radical surgery found that NLR was significantly associated with depth of stromal infiltration, both NLR and PLR were significantly associated with lymph node metastasis. A study done by Chen et al found that only PLR was significantly associated with lymph node metastases on early-stage patients having definite surgery. He also showed significant association of both NLR and PLR with reduced recurrence-free survival and OS.16

Conclusion

Inflammation has been established as a mediator of cellular proliferation in neoplastic tissue. High NLR and PLR are markers of inflammation and correlate with disease recurrence in carcinoma cervix as evident in this study. Hence, NLR and PLR can be strongly used as biomarker for prognostication and outcome of patients with carcinoma cervix. However, a well-designed larger study with longer follow-up is warranted.