Utility of red cell distribution width as a diagnostic and prognostic marker in non-small cell lung cancer

An increasing number of studies have indicated that red blood cell distribution width (RDW) may be a novel biomarker for the diagnosis and prognosis of various malignancies. However, to date, data on the association of RDW with non-small cell lung cancer (NSCLC) are unclear. Our present study aimed to explore the value of RDW in NSCLC patients. A total of 338 NSCLC patients, 109 small cell lung cancer (SCLC) patients, and 302 healthy participants were retrospectively analyzed between January 2016 and December 2018. In the present study, we found that RDW was significantly increased in NSCLC patients. Receiver-operating characteristic (ROC) analysis showed that the area under the ROC curve (AUC) of RDW was 0.753 in discriminating NSCLC patients from healthy participants, the optimal cut-off value of RDW was 12.95, and the specificity and sensitivity were 76.33% and 76.16%, respectively. Further analysis found that RDW can enhance the diagnostic performance of Cyfra21-1 and NSE in discriminating NSCLC patients from healthy participants or SCLC patients. Among NSCLC patients, RDW was significantly correlated with TNM stage, T stage, N stage, M stage, and Cyfra21-1, indicating that RDW may be helpful for predicting the prognosis of NSCLC patients. Our findings suggest that RDW can be used as an auxiliary marker for the diagnosis and prognosis of NSCLC.


Scientific RepoRtS
| (2020) 10:15717 | https://doi.org/10.1038/s41598-020-72585-4 www.nature.com/scientificreports/ However, to date, data about the association of RDW with NSCLC are still scarce. In 2013, Yasuko et al. first found that RDW is significantly associated with clinical cancer stage and higher RDW values suggested that lung cancer patients will have poor clinical outcomes 21 . In 2014, Richard et al. found that RDW plays a significant role in determining potentially curative resection in NSCLC patients 22 . In 2016, Mehmet et al. found that RDW levels are significantly and negatively correlated with survival rate 23 . Recently, two reports suggested that RDW is an independent predictor of worse prognosis in lung cancer 24,25 . These limited reports suggest that RDW may be helpful in predicting prognosis in lung cancer, but there are few studies about the role of RDW in the diagnosis of NSCLC. Additionally, the relationships between histological types of lung cancer and RDW have not been clarified in some studies 21 . Thus, further investigations are needed to analyze the significance of RDW in NSCLC.
In this study, we assessed RDW values to analyze the significance of RDW in diagnosing NSCLC. Moreover, the ability of RDW to indicate prognosis in NSCLC patients was evaluated by analyzing the correlation between clinicopathological features and RDW levels. Our findings suggest that RDW can be used as an auxiliary marker for the diagnosis and prognosis of NSCLC.

Results
RDW levels are increased in nScLc patients. The RDW levels of 338 NSCLC patients, 109 small cell lung cancer (SCLC) patients, and 302 healthy participants were analyzed in this study. The results showed that the RDW levels were significantly higher in the NSCLC group [13.0 (IQR 13.0-14.0)] and the SCLC group [14.0 (IQR 12.85-15.0)] than in the control group )] (P < 0.05). However, no difference was found in terms of RDW between NSCLC and SCLC patients (Fig. 1). Our data demonstrated that the levels of RDW are increased in NSCLC patients.
RDW is a useful blood marker to help diagnose nScLc. ROC curve analysis was used to analyze the power of RDW in the differential diagnosis of NSCLC patients and healthy participants. The results showed that the AUC of RDW was 0.753 ( Fig. 2 and Table 1). When the cut-off value of RDW was 12.95, the sensitivity and specificity were 76.33% and 76.16%, respectively (Table 1).
Serum CEA and Cyfra21-1 are common tumor markers in the clinic, but they cannot be widely used in the diagnosis of NSCLC because they lack specificity 26 . In this study, we found that RDW can enhance the diagnostic sensitivity of CEA and Cyfra21-1 (Table 1). Importantly, in the differential diagnosis of patients with NSCLC and healthy participants, the AUC values of CEA alone and Cyfra21-1 alone were 0.665 and 0.657, respectively. When CEA, Cyfra21-1, and RDW were combined for detection, the AUC values for CEA + RDW and Cyfra21-1 + RDW were 0.816 and 0.808, respectively, which were significantly higher than those for CEA + Cyfra21-1, CEA alone and Cyfra21-1 alone. When combined detection of lung cancer serum markers (CEA and Cyfra21-1) and blood markers (RDW) was applied, the AUC value was further increased, which was 0.834 ( Fig. 2 and Table 1). In brief, these results indicated that RDW can enhance the diagnostic performance of CEA and Cyfra21-1 in the diagnosis of NSCLC.
NSCLC and SCLC are different in terms of their clinical course and response to treatment; thus, it is very important to differentially diagnose NSCLC and SCLC. NSE, a serum tumor marker, is widely used in SCLC patient diagnosis 27 . Next, the results of ROC curve analysis showed that the AUCs of RDW and NSE were 0.581 and 0.798, respectively. When the combined detection of RDW and NSE was applied, the AUC was further increased to 0.824 ( Fig. 3 and Table 2), suggesting that RDW alone did not perform well, but can slightly enhance the diagnostic performance of NSE in the differential diagnosis of NSCLC and SCLC.
Relationships between RDW levels and clinicopathological characteristics in nScLc patients. The patients were categorized according to the RDW cut-off value, which was 12.95%. The cor-  www.nature.com/scientificreports/ relations between RDW levels and clinicopatholotical characteristics are shown in Table 3. RDW levels were associated with TNM stage, T stage, M stage, N stage, and Cyfra21-1. However, no significant difference in age, sex, histological type, neutrophils, C-reactive protein (CRP), procalcitonin (PCT), or albumin (ALB) was shown in NSCLC patients (Table 3).

Discussion
Our findings showed that RDW levels were increased in NSCLC patients and that increased RDW levels can enhance the diagnostic performance of Cyfra21-1 in distinguishing NSCLC patients from healthy participants. Moreover, RDW is closely related to prognostic factors in NSCLC patients, such as tumor stage and tumor markers. Our results showed that RDW may be a useful biomarker in assisting diagnosis and predicting the outcome of NSCLC. For decades, RDW has been routinely used as a useful index for the differential diagnosis of anemia. In recent years, RDW has attracted notable attention 7 . Various studies have found that RDW is a simple, robust, and convenient biomarker in various malignancies 13 . However, reports on the relationship between RDW and NSCLC are still scarce [21][22][23] , and further studies are still needed to investigate the role of RDW in NSCLC.
In the current study, the results showed that the levels of RDW were increased in NSCLC patients, which is consistent with the finding of previous report [21][22][23] . ROC curve analysis showed that RDW was an convenient marker for differentiating NSCLC patients and healthy participants with a cut-off value of 12.95%, sensitivity of 76.33%, and specificity of 76.16%. Cyfra21-1 and NSE, two important serum tumor markers, are often used for the differential diagnosis of lung cancer 26 . However, the low specificity of Cyfra21-1 and NSE limit their clinical application in lung cancer screening. Our results showed that RDW can improve the diagnostic ability of Cyfra21-1 and NSE, suggesting that combining the detection of RDW and Cyfra21-1 or NSE may be a better marker for discriminating NSCLC patients from healthy participants or SCLC patients. To our knowledge, our study is the first to report the clinical utility of the combination of RDW and Cyfra21-1 or NSE for the differential diagnosis of patients with NSCLC.
Next, we investigated whether increased RDW levels might be a potential biomarker for predicting the prognosis of NSCLC patients, and the results showed that reduced RDW was associated with TNM stage, T stage, M stage, N stage, and Cyfra21-1 in NSCLC patients. In contrast, Yasuko et al. reported that RDW is positively correlated with the clinical cancer stage of lung cancer 21 . NSCLC and SCLC have different clinical courses and responses to treatment. However, Yasuko et al. did not analyze the relationships between RDW and NSCLC or SCLC 21 . In addition, the inconsistent findings may be due to the selected populations, small sample sizes, and different RDW cut-off values 21 . Therefore, we need to provide more data to make clarify the associations between RDW and NSCLC. Considering that TNM stage, T stage, M stage, N stage, and Cyfra21-1 are strong prognostic factors for NSCLC, we hypothesize that RDW can be used as a potential prognostic factor in NSCLC patients.
The mechanisms explaining our findings are unclear, but there are some possible explanations. First, some reports have suggested that chronic inflammation plays a significant role in the development and progression of NSCLC, and inflammatory cells release various signaling molecules, which may affect the synthesis or activity of erythropoietin, thus impairing red blood cell maturation and causing immature red blood cells to enter the blood flow 28 . At the same time, inflammation can also lower red blood cell survival, thus resulting in the mixing of red blood cell volumes in peripheral circulation 29 . Previous studies have shown that there is a positive relationship between RDW and inflammatory markers 30 . Second, cancer growth leads to nutritional deficiency, which weakens red blood cell maturation and causes a increased RDW. In this study, we analyzed the relationship between RDW and inflammatory markers (neutrophils, CRP, and PCT) and malnutrition markers (ALB) and found that RDW levels were increased in NSCLC patients with high levels of inflammation markers (neutrophils, CRP, and PCT), but these result were not statistically significant. In addition, RDW levels were not associated with serum ALB levels in NSCLC patients. Therefore, we speculate that RDW is increased and that the relationship between the pathological features and RDW of NSCLC patients may be only partly related to the inflammatory response in NSCLC patients, and the main reasons that may affect RDW levels in NSCLC patients need further study.
Several limitations of the present study remain to be resolved. First, our study was a single-institution, retrospective, relatively small sample size study, which may lead to bias in sample selection and analysis. Second, our study did not assess the association between RDW values and overall survival in NSCLC patients, which limits the generalization of our findings. Therefore, the clinical role of RDW in NSCLC needs to be further verified by other research centers.
Overall, our study indicated that RDW might be used as an auxiliary marker for the diagnosis and prognosis of NSCLC. Table 2. Diagnostic values of RDW alone and NSE alone, and combined detecting in differential diagnosis between NSCLC and SCLC.

Variables
AUC Cut off Sensitivity (%) Specificity (%) 95% confidence interval  www.nature.com/scientificreports/ screening the hospital medical records system. This research was approved by the ethics committee of the Fujian Medical University Union Hospital. Informed consent was obtained from all participants included in the study.
All methods were carried out in accordance with relevant guidelines and regulations.
Laboratory analysis. Peripheral blood samples were collected prior to the start of treatment. Routine blood tests were performed with a Beckman Coulter LH 780 automated hematology analyzer (Beckman Coulter, Brea, CA, USA), the normal range of neutrophils was less than or equal to 7.0× 10 9 /L. The serum carcinoembryonic antigen (CEA), cytokeratin 19 fragments (Cyfra21-1), neuron-specific enolase (NSE), C-reactive protein (CRP), and procalcitonin (PCT) levels were determined by a Cobas 6000 Analyzer (Roche Diagnostics, Basel, Switzerland). According to the manufacturer's instructions, the normal range of CEA is less than 5 ng/mL, the normal range of CYFRA21-1 is less than 3.3 ng/mL, the normal range of NSE is less than 16.3 ng/mL, the normal range of CRP is less than or equal to 8 mg/L, and that for the normal range of PCT is less than or equal to 0.05 ng/mL. Serum albumin (ALB) was detected with a Beckman DXC800 Biochemical analyzer (Beckman Coulter, Brea, CA, USA), and the normal range of the ALB is more than or equal to 35 g/L.
Statistical analysis. Statistical analyses were performed using SPSS software version 21.0 (SPSS Inc., Chicago, IL, USA). RDW values are summarized as the median (interquartile range, IQR), and the Mann-Whitney U-test was used to determine the differences between the two groups. ROC curve analysis was performed to assess the diagnostic value of RDW in NSCLC patients. The optimal cut-off value for RDW was determined by the Youden index (Youden index = specificity + sensitivity − 1). The categorical variables are presented as the numbers of patients and percentages, and the Chi-square test was used to compare the differences in incidence of abnormal RDW data between the different groups. All statistical tests were considered statistically significant at a p-value < 0.05.