High neutrophil-to-lymphocyte ratio and platelet-to-lymphocyte ratio predict poor survival in rectal cancer patients receiving neoadjuvant concurrent chemoradiotherapy.

This study explored the prognostic value of neutrophil-to-lymphocyte ratio (NLR) and platelet-to-lymphocyte ratio (PLR) in rectal cancer patients receiving neoadjuvant concurrent chemoradiotherapy (CCRT).Between January 2006 and December 2016, 184 patients with newly-diagnosed rectal cancer receiving neoadjuvant CCRT were enrolled. Risk of overall survival (OS) and disease-free survival (DFS) were calculated using the Kaplan-Meier method and Cox proportional hazard models. Stratified survival analyses were also performed between post-neoadjuvant pathological (yp) stage.The mean follow-up time was 72.73 ± 36.82 months. High- and low-NLR patients differed significantly in both 5-year DFS (P = .026) and OS (P = .016). High- and low-PLR patients differed significantly in 5-year DFS (P = .011) but not OS (P = .185). Multivariate analyses revealed worse 5-year DFS (adjusted HR [aHR] = 2.8; 95% CI: 1.473-5.41; P = .002) and 5-year OS (aHR = 1.871; 95%CI: 1.029-3.4; P = .04) in the high-NLR group after adjusting for covariates. After adjustments, the high-PLR group had inferior 5-year DFS (aHR = 2.274; 95%CI: 1.473-5.419; P = .038) but not 5-year OS (aHR = 1.156; 95%CI: 0.650-2.056; P = .622). Further stratified analysis indicated that yp stage II and III patients with high NLR had worse 5-year DFS (aHR = 2.334; 95% CI: 1.158-4.725; P = .018) and OS (aHR = 2.226; 95% CI: 1.165-4.251; P = .015). Additionally, yp stage II and III patients with high PLR had inferior 5-year DFS (aHR = 2.012; 95% CI: 1.049-3.861; P = .036).Pre-CCRT NLR and PLR are independent prognostic factors for rectal cancer patients and could be used as a potential biomarker to identify high-risk patients for more intense treatment and care.

Introduction

Colorectal cancer (CRC) is the third most common malignancy and the second leading cause of cancer-related death worldwide.[ Rectal cancer accounts for 30% to 35% of CRCs, and approximately 50% of rectal cancers are diagnosed at the locally advanced stage.[ With advances in the treatment of neoadjuvant concurrent chemoradiotherapy (CCRT),[ followed by total mesorectal excision (TME), outcomes have improved dramatically in recent decades. However, significant differences in survival still exist between post-treatment groups. Useful and practical prognostic biomarkers obtained before treatment are anticipated to forecast the outcome after mainstream treatment. Such biomarkers will help in planning strategies for customized postoperative adjuvant therapy.

Inflammation-driven markers play a crucial role in tumorigenesis and tumor progression.[ More evidences had reported systemic inflammation-based biomarkers could be used to predict tumor behavior.[ Two convenient and economic measures of systemic inflammation, neutrophil-to-lymphocyte ratio (NLR) and platelet-to-lymphocyte ratio (PLR), have been advocated to reflect the interaction between inflammation and host immune status, making them potential prognostic factors for various types of cancers.[ Increased NLR has been advocated to be an independent prognostic factor for poor survival outcomes in pancreatic cancer, CRC, and gastric cancer.[ Elevated PLR has also been reported to be associated with poor prognosis in different cancers.[ Although previous studies used systemic inflammation response such as NLR and PLR to assess the prognostic ability in rectal patients undergoing neoadjuvant CCRT, the results remain inconclusive.[ Thus, this study aimed to investigate the prognostic value of pre-CCRT NLR and PLR in patients with rectal cancer post neoadjuvant CCRT.

Methods and materials

Ethics statement

Written informed consent was not required for this observational study as it was a retrospective medical record review. The study protocol was approved by the Ethics Committee of the Institutional Review Board of Chi Mei Medical Center (IRB: CMFHR10707–012).

Patient demographics and database

The data for this study were collected from the Cancer Registry Dataset of Chi-Mei Medical Center between January 1, 2006 and December 31, 2016. Electronic medical records and the Cancer Registry Dataset were retrospectively reviewed. All patients were regularly monitored after diagnosis until death or last follow up. A total of 184 patients with rectal cancer who underwent neoadjuvant CCRT followed by TME in accordance with National Comprehensive Cancer Network (NCCN) guidelines were identified in this study. Exclusion criteria included the following: previous cancer history, aged <18 years, incomplete patient data or being coded as receiving palliative care. All patients received long course radiation therapy at the dose of 45 to 50 Gray (Gy) in 25 to 28 fractions to the pelvis by NCCN recommendation, and radiation techniques were completed by intensity modulated radiation therapy, arc therapy or tomotherapy. All patients received pre-operation chemotherapy in 6 courses of 5-fluorouracil-based regimens and post-operation chemotherapy in 6 courses of FOLFOX or 5-FU/leucovorin or capcitabine regimens.

Our Cancer Registry Dataset provided the following clinico-pathological characteristics: date of diagnosis, age, gender, clinical stage and pathological stage according to the American Joint Committee on Cancer (AJCC) cancer staging (7th edition), pathologic tumor grade, perineural invasion (PNI), lymphovascular invasion (LVI), circumferential resection margin (CRM), tumor regression grade (TRG), NLR and PLR. TRG was done a standard 5-point scale as initially described by Dworak et al.[ NLR was determined by the neutrophil count divided by the total lymphocyte count. PLR was defined as the absolute platelet count divided by the total lymphocyte count. NLR and PLR values were determined based on complete blood count obtained within 2 weeks before neoadjuvant CCRT. Patients were excluded from the study if they had any acute infection.

Statistical analysis

All statistical operations were performed using SPSS version 22.0 for Windows (IBM Corp., Armonk, NY). All statistical significance levels were set as 2-sided, with P < .05. The primary endpoint was disease free-survival (DFS), defined as the time from the day of disease diagnosis until the day of disease failure. The secondary endpoint was overall survival (OS), defined as the time from the day of disease diagnosis until the date of death. The cut-off value of NLR and PLR were set as the mean value of each. The 5-year DFS and OS were described using the Kaplan-Meier method, and the differences were compared using log-rank statistics. Univariate and multivariate Cox regression models were used to evaluate the effect of NLR and PLR on 5-year DFS and 5-year OS. Stratified survival analyses were also performed between different yp stages.

Results

The demographic, clinical and pathological characteristics of this study are displayed in Table 1. A total of 184 patients were identified, 121 (65.8%) male and 63 (34.2%) female. The mean follow-up time was 72.73 ± 36.82 months. The mean age of all patients at diagnosis was 63.2 ± 11.7 years, with 97 (52.7%) patients ≤65 years and 87 (47.3%) patients >65 years. The mean NLR and PLR values pre-CCRT were 3.5 and 187.9, respectively. In all, 146 (79.3%) patients had pre-CCRT NLR ≤3.5 and 38 (20.7%) patients had pre-CCRT NLR > 3.5. Pre-CCRT PLR was ≤188 in 121 (70.1%) patients and >188 in 25 (29.9%) patients. According to clinical stage, 49 (26.6%) patients were at c stage I, 56 (30.4%) at c stage II and 79 (42.9%) at c stage III. In terms of yp stage, 23 (12.5%) patients were at yp stage 0, 54 (29.3%) at yp stage I, 43 (24.3%) at yp stage II and 64 (34.8%) at yp stage III. Additionally, 23 (12.5%) patients were TRG 4, which was compatible to the proportion of those with complete response (yp stage 0). Moderately differentiated tumor grade (grade 2) was found in 136 (73.9%) patients. In terms of other pathological features, 168 (91.3%) patients had negative PNI, 160 (87%) patients had negative LVI and 153 (95.7%) patients had negative CRM. The relationship between NLR or PLR status and these clinicopathological factors were also analyzed. However, there were no statistically significant difference between NLR or PLR status and clinicopathological factors, such as c stage, yp stage, LN metastasis, Grade, PNI, LVI, CRM, TRG and blood marker CEA (supplementary table 1 and 2).

Table 1

Clinico-pathological characteristics of the patients (n = 184).

Kaplan-Meier survival curves were generated to compare 5-year DFS and 5-year OS. As shown in Figures 1 and 2, the 5-year DFS differed significantly between those with pre-CCRT NLR >3.5 and ≤3.5; and between those with pre-CCRT PLR >188 and ≤188 as well. Log-rank tests showed that groups with pre-CCRT NLR >3.5 and PLR >188 had significantly worse 5-year DFS than those with pre-CCRT NLR ≤3.5 and pre-CCRT PLR ≤188 (P = .011 and P = .011, respectively). Figure 3 shows that the 5-year OS differed significantly between those with pre-CCRT NLR >3.5 and ≤3.5; however, 5-year OS did not differ significantly between those with pre-CCRT PLR >188 and ≤188 (Fig. 4). Log-rank tests showed that those with pre-CCRT NLR >3.5 had significantly worse 5-year OS than those with pre-CCRT NLR ≤3.5 (P = .016).

Figure 1

Kaplan-Meier plot of 5-year disease free survival (DFS) between NLR ≤3.5 and >3.5. NLR = neutrophil to lymphocyte ratio.

Figure 2

Kaplan-Meier plot of 5-year disease free survival (DFS) between PLR ≤188 and >188. PLR = platelet to lymphocyte ratio.

Figure 3

Kaplan-Meier plot of 5-year overall survival (OS) between NLR ≤3.5 and >3.5. NLR = neutrophil to lymphocyte ratio.

Figure 4

Kaplan-Meier plot of 5-year overall survival (OS) between PLR ≤188 and >188. PLR = platelet to lymphocyte ratio.

In univariate analysis, Grade 3 (p = .005, HR = 2.209, 95% CI: 1.182–4.129), PNI (P = .015, HR = 2.587, 95% CI: 1.207–5.543), LVI (P = .005, HR = 2.507, 95% CI: 1.313–4.788), yp stage III (P = .005, HR = 2.083, 95% CI: .802–5.41), TRG 4 (P = .005, HR = 0.276, 95% CI: 0.092–0.825), pre-CCRT NLR >3.5 (P = .029, HR = 1.957, 95% CI: 1.073–3.57) and pre-CCRT PLR >188 (P = .013, HR = 2.019, 95% CI: 1.163–3.503) were significant prognostic factors for poor 5-year DFS. In multivariate analysis, Grade 3 (P = .005, adjusted hazard ratio [aHR] = 2.233, 95% CI: 1.161–4.294), LVI (P = .046, aHR = 2.151, 95% CI: 1.014–4.563), yp stage III (P = .019, aHR = 1.421, 95% CI: 0.524–3.852), TRG 4 (P = .002, aHR = 0.054, 95% CI: 0.005–0.606), pre-CCRT NLR >3.5 (P = .002, aHR = 2.825, 95% CI: 1.473–5.419) and pre-CCRT PLR >188 (P = .038, aHR = 2.274, 95% CI: 1.047–4.937) all remained as independent prognostic factors for 5-year DFS (Table 2).

Table 2

Univariate and multivariate analyses of Pre-CCRT NLR and PLR for 5-yr disease free survival.

Furthermore, univariate analysis showed that PNI (P = .000, HR = 3.378, 95% CI: 1.927–5.921) and pre-CCRT NLR >3.5 (P = .017, HR = 1.748, 95% CI: 1.105–2.765) were significant risk factors for 5-year OS. Multivariate analysis indicated that Grade 3 (P = .04, adjusted hazard ratio [aHR] = 2.353, 95% CI: 0.960–5.768), PNI (P = .001, aHR = 3.416, 95% CI: 1.693–6.895) and pre-CCRT NLR >3.5 (P = .04, aHR = 1.871, 95% CI: 1.029–3.4) remained as independent prognostic factors for 5-year OS (Table 3).

Table 3

Univariate and multivariate analyses of Pre-CCRT NLR and PLR for 5-yr overall survival.

In stratified analysis according to yp stage, only yp stage II-III patients with NLR >3.5 had significantly worse 5-year DFS (P = .018; aHR = 2.334; 95% CI: 1.158–4.725) and 5-year OS (P = .015; aHR = 2.226; 95% CI: 1.165–4.251), after adjustments. In addition, only yp stage II-III patients with PLR >3.5 had worse 5-year DFS (P = .036; aHR = 2.012, 95% CI: 1.049–3.861), after adjustments (Table 4).

Table 4

Stratified analyses of NLR and PLR for 5-yr disease free survival and overall survival by post-neoadjuvant pathological (yp) stage.

Discussion

This observational study evaluated the effects of pre-CCRT NLR and PLR to predict survival in patients with neoadjuvant rectal cancer. Our results found that high NLR patients had significantly inferior 5-year DFS and OS, as compared with low NLR patients, and high PLR patients exhibited significantly poorer 5-year DFS than low PLR patients, even after adjusting for all demographic and clinico-pathologic factors. Stratified analysis showed that patients with elevated NLR with yp stage II-III had significantly poorer 5-year DFS and OS than those with yp stage 0 to I; correspondingly, patients with elevated PLR with yp stage II to III had also had significantly poorer 5-year DFS than those with yp stage 0 to I.

The strengths of this study are as follows. First, patients in our study were treated at a single institution and received neoadjuvant CCRT as routine clinical practice and all regimens were based on NCCN recommendations. Thus, variations in treatment regimen and technique were not likely to have biased results between patients. Furthermore, our database provided crucial information about predisposing factors that could influence survival (ie, age, stage, grade, PNI, LVI, CRM and TRG), thus allowing for an in-depth evaluation of the impact of these factors on outcomes. Third, included were only rectal cancer patients with neoadjuvant CCRT followed by TME. These inclusion criteria facilitated a study under homogeneous patient and treatment conditions. Hence, our conclusion is more specific than that of Ian et al.,[ which included patients diagnosed with all gastrointestinal malignancies under heterogenetic treatment. Finally, the use of stratification analysis made survival prediction more precise than in other studies. Therefore, pre-CCRT NLR and PLR can be recommended as important independent prognostic biomarkers for neoadjuvant rectal cancer.

The association of high NLR and PLR with poor outcomes has been explored before.[ Although the precise mechanisms remain unknown, current consensus is that the hematological markers NLR and PLR may well reflect the host immune activity which is related to cancer progression.[ Generally, neutrophils are considered to not only release cytokines, chemokines and granule proteins in building a favorable microenvironment for tumor progression, but they also act directly in carcinogenesis, assisting in metastasis by suppressing peripheral leukocyte activation.[ In contrast, lymphocytes play a dominant role in restraining tumor growth.[ Therefore, elevated NLR can stimulate a favorable tumor micro-environment which promotes vigorous tumor growth.[ Platelets are thought to take part in tumor progression by regulating neo-angiogenesis and helping circulating tumor cells escape immune surveillance.[ Thus, elevated PLR is assumed to be associated with tumor metastasis.

Numerous published meta-analyses have confirmed that a high NLR and a high PLR are both significant prognostic factors for poor outcomes in locally advanced and metastatic colorectal cancer (CRC).[ Haowen et al[ performed a meta-analysis of 8 studies including data from 1685 patients with CRC and liver metastasis; the results showed that elevated pretreatment NLR was significantly related to poor OS (HR = 2.17, 95% CI: 1.82–2.58) and recurrence-free survival (HR = 1.96, 95% CI: 1.64–2.35). George et al,[ in a pooled analysis of 13 studies with 4,056 CRC patients, showed that a high pretreatment NLR independently predicted poor survival (HR = 2.08,95% CI:1.64–2.64). Guo et al[ conducted pooled analysis of 33 studies containing 15,404 CRC patients; results indicated that elevated PLR was associated with poorer OS (HR = 1.57, 95% CI: 1.41–1.75) and DFS (HR = 1.58, 95% CI: 1.31–1.92). Finally, Gu et al[ performed a meta-analysis of 13 studies with 8601 CRC patients, again finding that increased PLR predicted poor OS (HR = 1.81, 95% CI: 1.42–2.31) and DFS (HR = 1.84, 95% CI: 1.22–2.76). Therefore, high NLR and high PLR have been proven to forecast poor outcomes in locally advanced and metastatic CRC.

Rectal cancer patients with high NLR but not PLR had significant better DFS and OS in those who underwent neoadjuvant CCRT as a previous study performed by Kim et al[. However, Dudani et al[ concluded that neither NLR nor PLR were found to be independently prognostic for DFS and OS in patients with locally advanced rectal cancer undergoing neoadjuvant CCRT. Consistent with other colorectal studies,[ we found that high NLR and PLR were associated with poor DFS. However, NLR was superior to PLR in predicting lower OS in our study. In our stratification analysis, patients with yp stage II to II with high pre-CCRT NLR had significantly poorer DFS and OS; those with yp stage II to II with high pre-CCRT PLR had poorer DFS (HR = 2.012, 95% CI:1.049–3.861, P = .036). However, for yp stage 0 to I, NLR and PLR did not affect outcomes. Unlike the numerous published meta-analyses,[ whose heterogenic populations had wide tumor stages, different treatment methods and included those with colon and rectal cancer, our study population was concentrated on rectal cancer patients treated with neoadjuvant CCRT followed by TME, which makes the results more specific and convincing. In addition, previous studies have advocated discrepancy suggestions for the role of adjuvant chemotherapy for patients with rectal cancer after neoadjuvant CCRT followed by Surgery. Anne et al[ concluded adjuvant chemotherapy did not improve neither OS nor DFS in a meta-analysis. However, Zhifei et al[ addressed a survival benefit in the use of adjuvant chemotherapy in a retrospective study. The challenge of heterogenic survival behind adjuvant chemotherapy after preoperative CCRT for rectal cancer is still existence. Therefore, knowing pretreatment NLR and PLR in rectal cancer patients receiving neoadjuvant CCRT can help physicians refine additional adjuvant therapeutic decisions.

In previous studies, modifying patients’ systemic inflammation response was addressed. Whether altering the pre-CCRT NLR status can influence treatment outcome still need to be unveiled. Some studies have advocated preoperative administration of corticosteroids in patients undergoing operation for cancer can reduce postoperative morbidity.[ Non-steroidal anti-inflammatory drugs (NSAIDS) were also addressed to have effect for against CRC progression.[ However, we still need further clinical trials to discover whether altering the pre-CCRT NLR status via giving corticosteroids or NSAID can improve survival outcome of rectal cancer patients receiving neoadjuvant CCRT.

This study has some limitations. First, it has the potential for selection bias inherent to retrospective analyses. Second, potential confounding factors such as comorbidities and nutritional status could not be acquired for analysis. However, all recruited patients were within performance status 0 to 1, a fact that may eliminate the effect of these unmeasured factors. Third, all patients in this study were enrolled from an Asian population; therefore, our results should be extrapolated carefully to non-Asian populations. Considering these limitations, a well-designed prospective study is still required to verify this result and apply it to clinical practice.

Conclusions

Elevated pre CCRT NLR and PLR are independent prognostic factors for poor outcome in rectal cancer patients. Therefore, pre-CCRT NLR and PLR may be considered as potential biomarkers that need to be further validated by prospective studies, guiding the clinical decision-making process in increasing the intensity of treatment and care for those high-risk patients.

Acknowledgments

The authors express their sincere thanks to the staff of the Cancer Center of the Chi Mei Medical Center for data collection and Gregory Gliemi, Ph.D., for assistance of statistical analyses of the data.

Author contributions

Conceptualization: Te-Min Ke.

Data curation: Te-Min Ke.

Formal analysis: Te-Min Ke.

Investigation: Te-Min Ke.

Methodology: Te-Min Ke.

Project administration: Te-Min Ke.

Resources: Te-Min Ke.

Software: Te-Min Ke, Chun-Che Huang.

Supervision: Te-Min Ke, Li-Ching Lin, Chun-Che Huang, Yu-Wen Chien, Ching-Chieh Yang.

Validation: Te-Min Ke, Chun-Che Huang, Yu-Wen Chien, Wei-Chen Ting, Ching-Chieh Yang.

Visualization: Te-Min Ke, Chun-Che Huang, Yu-Wen Chien, Wei-Chen Ting, Ching-Chieh Yang.

Writing – original draft: Te-Min Ke.

Writing – review and editing: Te-Min Ke, Chun-Che Huang, Ching-Chieh Yang.