Prognostic value of mean platelet volume/platelet count ratio in patients with resectable esophageal squamous cell carcinoma: a retrospective study

Abstract

Background

Mean platelet volume (MPV) to platelet count (PC) ratio (MPV/PC) is a useful indicator in several cancers. However, the role for MPV/PC ratio in esophageal squamous cell carcinoma (ESCC) is still controversial.

Methods

A retrospective study was conducted including 277 resectable ESCC patients. The optimal cut-off values were calculated by the X-tile program. The receiver operator characteristic (ROC) curves were also created to show the candidate cut-off points. The comparisons between the X-tile plot and ROC curve were performed. The Kaplan-Meier method was utilized to analyze the cancer-specific survival (CSS). Prognostic factors for CSS were calculated with Cox regression univariate and multivariate analyses.

Results

According to the X-tile program, the cut-off values for MPV, PC and MPV/PC ratio were 8.5 (fl), 200 (giga/l) and 0.04, respectively. However, the cut-off values for MPV, PC and MPV/PC ratio by the ROC curves were 8.25 (fl), 243.5 (giga/l) and 0.0410, respectively. The cut-off values were similar between the X-tile and ROC curve. A low MPV/PC ratio level (≤0.04) was associated with poor CSS (22.4% vs. 43.1%, P < 0.001). In multivariate analyses, we found that MPV/PC ratio was an independent predictor for CSS (P < 0.001). When we set the cut-off point using ROC curve, the MPV/PC ratio was still an independent predictor for CSS (P < 0.001).

Conclusion

The MPV/PC ratio is a useful predictive indicator in patients with ESCC.

Introduction

Esophageal cancer (EC) is the 8th most common cancer worldwide and the 6th most common cause of death from cancer (Ferlay et al., 2010). The incidences vary widely in different countries and regions. To date, approximately 53.8% and 51.9% of all ECs occurred and died in China (Siegel, Miller & Jemal, 2015; Ferlay et al., 2010). Esophageal adenocarcinoma is the most common malignancy in the West. In China, however, esophageal squamous cell carcinoma (ESCC) is the predominant subtype (Napier, Scheerer & Misra, 2014). Radical esophagectomy remains the most effective therapy for patients with EC. However, the prognosis for EC remains poor (Bedenne et al., 2007; Domper Arnal, Ferrández Arenas & Lanas Arbeloa, 2015). Therefore, it is very important to find more and more useful and effective prognostic indicators for patients with EC.

Over the past few decades, a number of prognostic factors for EC have been identified, including tumor length, vessel invasion, lymph node status (N stage), depth of invasion (T stage), TNM stage and other serum biomarkers, such as squamus cell carcinoma antigen (SCCA) and carcinoembryonic antigen (CEA) (Peyre et al., 2008; Wijnhoven et al., 2007; Feng, Huang & Zhao, 2013). Inflammation plays an important role in cancer progression and prognosis (Balkwill & Mantovani, 2001; Mantovani et al., 2008). C-reactive protein (CRP), as a most sensitive inflammatory biomarker, has been confirmed in a series of cancers to predict the prognosis, including patients with EC (Shimada et al., 2003; Nozoe, Saeki & Sugimachi, 2001; Platt et al., 2012). In addition, there are other parameters like neutrophil and lymphocyte that are easy-to-measure inflammatory markers (Dutta et al., 2011).

Mean platelet volume (MPV) is recognized as a hallmark for platelet count (PC) activation (Kamath, Blann & Lip, 2001). Several studies showed that MPV and PC are associated with mortality in cardiovascular disease, such as ischemic cardiovascular disease and acute myocardial infarction (Guenancia et al., 2017; Azab et al., 2011). Moreover, recent studies have shown that the ratio for MPV to PC (MPV/PC) is associated with prognosis in some malignancies, such as hepatocellular carcinoma and lung cancer (Cho et al., 2013; Inagaki et al., 2014; Omar et al., 2018). However, the role for MPV/PC ratio in ESCC is still controversial. Furthermore, controversy exists concerning the optimal cut-off points for MPV/PC to predict the prognosis of ESCC. Therefore, the purpose of our study here was to explore the prognostic role of MPV/PC ratio in patients with ESCC.

Patients and Methods

From January 2007 to December 2010 at the Department of Thoracic Surgery, Zhejiang Cancer Hospital, a retrospective study was conducted including 277 resectable ESCC patients. The exclusion criteria were as follows: (1) patients who received preoperative treatment, such as chemotherapy and/or radiotherapy; (2) patients who had any form of acute or chronic inflammatory diseases or infections; (3) patients who had systemic diseases, and (4) those diagnosed with distant metastases. Written informed consent for the collection of specimen and other medical information were obtained from all patients before surgery. The current study was approved by the Ethics Committees of Zhejiang Cancer Hospital (IRB Approval No. IRB-2018-130).

The main clinical characteristics, such as age, gender, tumor location (upper, middle and lower), tumor length, vessel invasion, differentiation (well, moderate and poor) and tumor stage (T stage, N stage and TNM stage), were retrospectively reviewed and collected. The tumor length was defined as the long diameter for pathological specimens. Blood samples were obtained within one week prior to surgery to measure the neutrophil (Neu), MPV, PC, CRP and CEA levels. MPV/PC was defined as MPV to PC ratio. Neu/PC was defined as Neu to PC ratio. The levels of Neu, MPV and PC were measured by automated blood cell counter (Sysmex XE-2100; Sysmex, Kobe, Japan). Serum levels of CRP were determined by latex-enhanced homogeneous immunoassay (Hitachi 917; Skill, Munich, Germany). Serum levels of CEA were measured using enzyme immunoassay kits (Abbott, Chicago, IL, USA). The AJCC/UICC TNM staging system (the 7th edition) was utilized to classify the stage for this study (Rice et al., 2010).

All the above patients were followed-up postoperatively (regularly evaluated every 3–6 months). The assessment included physical examination, blood tumor markers and computed tomography scan. In this study, we conducted a cancer-specific survival (CSS) to analyze the prognosis of patients with ESCC. The mean follow-up for patients was 45 months.

Statistical analyses

In the current study, the optimal cut-off values for Neu, MPV, PC, MPV/PC ratio, and Neu/PC ratio were calculated by the X-tile program (Camp, Dolled-Filhart & Rimm, 2004). The receiver operator characteristic (ROC) curves were also created to show the candidate cut-off points. The comparisons between the X-tile plot and ROC curve were performed. The areas under the curve (AUC) for Neu, MPV, PC, MPV/PC and Neu/PC were calculated and compared by the ROC curve. The chi-squared tests were used to compare the MPV/PC ratio, MPV and PC. The CSS curves were generated by the Kaplan–Meier method. Univariate analyses were performed with log-rank test. Multivariate analyses with cox proportional hazards regression model were utilized to analyze prognostic factors for CSS. SPSS 20.0 (SPSS Inc., Chicago, IL, USA) was utilized to perform the statistical analyses. R 3.2.3 software (Institute for Statistics and Mathematics, Vienna, Austria) was also utilized to conduct the nomogram model by Harrell’s concordance index (c-index) (Iasonos et al., 2008).

Results

There were 37 (13.4%) women and 240 (86.6%) men in all 277 patients with the mean age of 59.2 ± 7.8 years (36-80 years). In the current study, the mean values for Neu, MPV, PC, MPV/PC and Neu/PC were 4.2 ± 1.5 (giga/l) (range 1.5–9.5 giga/l), 9.3 ± 1.3 (fl) (range 6.7–12.9 fl), 232 ± 72 (giga/l) (range 60-473 giga/l), 0.04 ± 0.02 (range 0.02–0.14), and 0.020 ± 0.010 (range 0.0053–0.0667), respectively. The histograms of the preoperative MPV/PC ratio, MPV and PC are shown in Fig. 1.

Figure 1. The histograms of the MPV (A), PC (B) and MPV/PC ratio (C).

According to the X-tile program, the optimum cut-off points for MPV, PC, MPV/PC, Neu and Neu/PC ratio were 8.5 (fl), 200 (giga/l), 0.04, 4.2 (giga/l) and 0.02, respectively (Fig. 2). According to the optimum cut-off points of the above values, patients then were divided into 2 groups (MPV ≤8.5 fl and >8.5 fl; PC ≤200 giga/l and >200 giga/l; MPV/PC ratio ≤0.04 and >0.04). Clinicopathologic characters for the above values (MPV/PC ratio, MPV and PC) were shown in Table 1. The levels of MPV/PC ratio were significantly correlated with the CRP levels (P = 0.029).

Figure 2. X-tile analyses.

X-tile plots of the training sets are shown in A, D, G, J, M, with plots of matched validation sets shown in the smaller inset. The optimal cut-off points highlighted by the black circle in A, D, G, J and M are shown on the histograms of the entire cohort (B, E, H, K, N) and Kaplan-Meier plots (C, F, I, L, O). According to the X-tile program, the optimum cut-off points for MPV (A–C), PC (D–F), MPV/PC (G–I), Neu (J–L) and Neu/PC ratio (M–O) were 8.5 (fl), 200 (giga/l), 0.04, 4.2 (giga/l) and 0.02, respectively.

Table 1. Comparison of baseline clinical characteristics in ESCC.

	Total	MPV (fl)		P value	PC (giga/l)		P value	MPV/PC		P value
		≤8.5	>8.5		≤200	>200		≤0.04	>0.04
Age (years)				0.704			0.221			0.488
≤60	158	51	107		55	103		81	77
>60	119	41	78		50	69		66	53
Gender				0.521			0.271			0.629
Female	37	14	23		11	26		21	16
Male	240	78	162		94	146		126	114
Tumor length (cm)				0.246			0.020			0.087
≤3.0	78	30	48		38	40		35	43
>3.0	199	62	137		67	132		112	87
CRP (mg/l)				0.031			0.152			0.029
≤10.0	200	74	126		81	119		98	102
>10.0	77	18	59		24	53		49	28
Tumor location				0.242			0.096			0.057
Upper	16	7	9		10	6		4	12
Middle	127	36	91		44	83		72	55
Lower	134	49	85		51	83		71	63
Vessel invasion				0.744			0.097			0.111
Negative	232	78	154		83	149		128	104
Positive	45	14	31		22	23		19	26
Differentiation				0.927			0.826			0.454
Well	43	15	28		16	27		25	18
Moderate	179	58	121		70	109		90	89
Poor	55	19	36		19	36		32	23
T stage				0.106			0.313			0.425
T1	50	22	28		18	32		28	22
T2	49	20	29		24	25		21	28
T3	154	44	110		56	98		86	68
T4	24	6	18		7	17		12	12
N stage				0.054			0.720			0.899
N0	150	60	90		61	89		78	72
N1	74	21	53		27	47		39	35
N2	32	7	25		10	22		19	13
N3	21	4	17		7	14		11	10
TNM stage				0.003			0.357			0.546
I	69	31	38		31	38		33	36
II	92	35	57		34	58		52	40
III	116	26	90		40	76		62	54
CEA (ng/ml)				0.818			0.566			0.954
≤5.0	239	80	159		89	50		127	112
>5.0	38	12	26		16	22		20	18
Neu (giga/l)				0.249			0.681			0.186
≤4.2	146	53	93		57	89		72	74
>4.2	131	39	92		48	83		75	56
Neu/PC				0.090			<0.001			<0.001
≤0.02	170	50	120		35	135		114	56
>0.02	107	42	65		70	37		33	74

Notes.

ESCC

esophageal squamous cell carcinoma

CRP

c-reactive protein

MPV

mean platelet volume

PC

platelet count

TNM

tumor node metastasis

CEA

carcinoembryonic antigen

Neu

neutrophil

Kaplan–Meier analyses showed that a low MPV/PC ratio level (≤0.04) was associated with poor CSS (P < 0.001). The 5-year CSS was 43.1% in patients with MPV/PC ratio >0.04, and 22.4% in patients with MPV/PC ratio ≤0.04 (Fig. 3A). There were also significantly different for MPV (42.4% vs. 27.0%, P = 0.010) and PC (41.0% vs. 26.7%, P = 0.009) (Figs. 3B–3C). In multivariate analyses, we found that MPV/PC ratio was an independent predictor for CSS (P < 0.001) (Table 2). In addition, TNM stage (P < 0.001), CEA (P = 0.019), Neu (P = 0.007) and CRP (P < 0.001) were other significant prognostic variables by multivariate analyses (Table 2).

Figure 3. Kaplan–Meier CSS curves.

Patients with MPV/PC ratio >0.04 had a significantly better 5-year CSS than patients with MPV/PC ratio ≤0.04 (43.1% vs. 22.4%, P < 0.001; (A). The 5-year CSS were also significantly different for MPV (42.4% vs. 27.0%, P = 0.010; (B) and PC (41.0% vs. 26.7%, P = 0.009; (C). When we set the cut-off points using ROC curve, the MPV/PC ratio (42.7% vs. 23.5%, P < 0.001; (D), MPV (51.7% vs. 26.7%, P = 0.001; (E), and PC (41.8% vs. 19.3%, P < 0.001; (F) were also associated with CSS.

Table 2. Univariate and multivariate analyses for cancer-specific survival.

	CSS (%)	P value	Univariate analysis	P value	Multivariate analysis	P value
			HR (95% CI)		HR (95% CI)
Age (years)		0.412		0.417	–	–
≤60	33.5		1.000
>60	30.3		1.127 (0.845–1.502)
Gender		0.114		0.120	–	–
Female	45.9		1.000
Male	30.0		1.445 (0.909–2.298)
Tumor length		0.003		0.004	–	–
≤3 cm	42.3		1.000
> 3 cm	28.1		1.642 (1.173–2.297)
Tumor location		0.336		0.342	–	–
Upper/Middle	35.7		1.000
Lower	28.4		1.149 (0.863–1.530)
Vessel invasion		0.003		0.003	–	–
Negative	35.3		1.000
Positive	15.6		1.710 (1.197–2.444)
Differentiation		0.054		0.058	–	–
Well/Moderate	33.8		1.000
Poor	25.5		1.398 (0.989–1.978)
T stage		<0.001		<0.001	–	–
T1-2	45.5		1.000
T3-4	24.7		1.898 (1.382–2.606)
N stage		<0.001		<0.001	–	–
N0	49.3		1.000
N1-3	11.8		2.852 (2.120–3.836)
TNM stage		<0.001		<0.001		<0.001
I	58.0		1.000		1.000
II	38.9		1.966 (1.259–3.067)	0.003	1.825 (1.164–2.861)	0.009
III	13.8		3.799 (2.490–5.736)	<0.001	3.624 (2.362–5.560)	<0.001
Adjuvant therapy		0.085		0.090	–	–
No	35.6		1.000
Yes	24.4		1.297 (0.960–1.753)
CRP (mg/l)		<0.001		<0.001		<0.001
≤ 10.0	39.5		1.000		1.000
> 10.0	13.0		2.066 (1.526–2.798)		1.994 (1.461–2.722)
MPV (fl)		0.019		0.021	–	–
≤ 8.5	41.3		1.000
> 8.5	27.6		1.451 (1.057–1.992)
PC (giga/l)		0.009		0.011	-	-
≤ 200	41.0		1.000
> 200	26.7		1.488 (1.097–2.019)
MPV/PC		<0.001		<0.001		<0.001
> 0.04	43.1		1.000		1.000
≤ 0.04	22.4		1.861 (1.386–2.498)		1.823 (1.347–2.469)
CEA (ng/ml)		0.027		0.031		0.019
≤ 5.0	33.5		1.000		1.000
> 5.0	23.7		1.549 (1.042–2.302)		1.613 (1.082–2.407)
Neu (giga/l)		<0.001		<0.001		0.007
≤ 4.2	43.8		1.000		1.000
> 4.2	19.1		1.945 (1.455–2.600)		1.512 (1.120–2.040)
Neu/PC		0.223		0.229	–	–
≤ 0.02	35.3		1.000
> 0.02	27.1		1.195 (0.894–1.597)

Notes.

ESCC

esophageal squamous cell carcinoma

CRP

c-reactive protein

MPV

mean platelet volume

PC

platelet count

TNM

tumor node metastasis

CEA

carcinoembryonic antigen

Neu

neutrophil

CI

confidence interval

HR

hazard ratio

We also created ROC curves to show the candidate cut-off points. The cut-off values for Neu, MPV, PC, MPV/PC, and Neu/PC ratio by the ROC curves were 4.25 (giga/l), 8.25(fl), 243.5 (giga/l), 0.0410, and 0.0213, respectively (Fig. 4). The candidate cut-off points and the area under ROC curve (AUC) are shown in Table 3. When we set the cut-off points using ROC curve, the MPV/PC ratio (42.7% vs. 23.5%, P < 0.001), MPV (51.7% vs. 26.7%, P = 0.001), and PC (41.8% vs. 19.3%, P < 0.001) were also associated with CSS (D-F) (Figs. 3D–3F). In multivariate analyses, MPV/PC ratio was still an independent predictor for CSS (P < 0.001) (Table 4).

Figure 4. ROC curve analysis.

The cut-off values for Neu, MPV, PC, MPV/PC, and Neu/PC ratio by the ROC curves were 4.25 (giga/l), 8.25(fl), 243.5 (giga/l), 0.0410, and 0.0213, respectively.

Table 3. Comparison of AUC areas for the prognostic factors in ESCC.

	Cut-off	Sensibility	Specificity	AUC	95% CI	P-value
MPV/PC	0.0410	62.9	61.7	0.608	0.548–0.666	Reference
MPV	8.25	84.6	34.8	0.609	0.549–0.667	0.9834
PC	243.5	50.0	74.2	0.648	0.588–0.704	0.0181
Neu	4.25	53.7	76.4	0.689	0.630–0.743	0.1123
Neu/PC	0.0213	38.3	74.2	0.543	0.482–0.603	0.3269

Notes.

ESCC

esophageal squamous cell carcinoma

AUC

area under the curve

MPV

mean platelet volume

PC

platelet count

Neu

neutrophil

Table 4. Multivariate analyses in ESCC with the cut-off values by ROC curve.

	HR (95% CI)	P-value
CRP (mg/l) (>10.0 vs. ≤10.0)	2.060 (1.511–2.807)	<0.001
TNM stage
II vs. I	1.816 (1.160–2.844)	0.009
III vs. I	3.529 (2.298–5.417)	<0.001
MPV/PC ( ≤0.0410 vs. >0.0410)	1.728 (1.275–2.342)	<0.001
CEA (ng/ml) ( >5.0 vs. ≤5.0)	1.636 (1.097–2.438)	0.016
Neu (giga/l) ( >4.25 vs. ≤4.25)	1.553 (1.150–2.096)	0.004

Notes.

ESCC

esophageal squamous cell carcinoma

CSS

cancer-specific survival

CRP

c-reactive protein

MPV

mean platelet volume

PC

platelet count

TNM

tumor node metastasis

CEA

carcinoembryonic antigen

Neu

neutrophil

CI

confidence interval

HR

hazard ratio

Moreover, we wanted to predict the survival risk (CSS) for patients with ESCC, a nomogram model was conducted including age, gender, TNM, CEA, Neu, MPV/PC ratio and CRP for CSS (Fig. 5). From this model, the probability of survival for ESCC patients could be predicted (c-index = 0.72).

Figure 5. Nomogram model for prediction.

The Harrell’s c-index for CSS prediction was 0.72. A nomogram predicts survival prediction based on MPV/PC and other prognostic factors in patients with ESCC. The nomogram is used by totalling the points identified at the top of the scale for each independent factor. This total point score is then identified on the total points scale to determine the probability of risk prediction (A) and survival prediction (B).

Discussion

Our study demonstrated some important findings: (1) MPV/PC ratio was a strong predictor of CSS; (2) MPV/PC ratio, but not MPV or PC, was a useful predictive indicator. This study used X-tile program and ROC curves as candidate cut-off points. The comparisons between the X-tile plot and ROC curve were performed. The cut-off values were similar between the X-tile and ROC curve. Moreover, our study is also the first attempt to predict the survival risk by a nomogram model based on MPV/PC ratio.

Platelet activation has been demonstrated as a common phenomenon in some cardiovascular diseases (Guenancia et al., 2017; Azab et al., 2011). To assess the platelet activation status, MPV and PC are two main aspects. Moreover, studies have shown that MPV/PC ratio is associated with prognosis in some malignancies, such as hepatocellular carcinoma and lung cancer (Cho et al., 2013; Inagaki et al., 2014; Omar et al., 2018). Cho et al. (2013) have shown that the ratio of MPV/PC levels in hepatocellular carcinoma were higher than the control group. Inagaki et al. (2014) have revealed that MPV/PC ratio was significantly different on survival in lung cancer. However, Omar et al. (2018) showed that increased MPV and increased PC were significant higher than the control group. In their study, however, MPV/PC was not an independent predictor in lung cancer.

MPV is an indicator of platelet activation. Shen et al. (2018) demonstrated that reduced MPV is associated with worse survival outcome in EC. The role for MPV/PC ratio in ESCC patients has not yet been well evaluated. A study reported by Sun et al. (2018) showed that the levels of MPV/PC ratio in ESCC were significantly lower than the healthy group, and which were significantly correlated with the tumor length. In our study, however, the MPV/PC ratio was not significantly correlated with the tumor length (P = 0.087). In addition, they revealed that decreased MPV and MPV/PC ratio were significantly associated with locally advanced ESCC. In our study, MPV was not a significant prognostic factor by multivariate analyses. Recently, Zhang et al. (2016) initial conducted a COP-MPV (combination of MPV and PC) model to predict the prognosis in ESCC. They revealed that COP-MPV was a useful independent predictor, but not for MPV or PC. As everyone knows, MPV and PC may be influenced by a variety of other non-cancer related conditions, the potential basis could be decreased by the MPV to PC ratio (MPV/PC). Therefore, the role of the MPV/PC ratio would be more reliable than the effect of either MPV or PC. In the current study, a low MPV/PC ratio level (≤ 0.04) was associated with poor CSS (P < 0.001) and was confirmed by multivariate analyses (P < 0.001).

In previous studies, controversy exists about the optimum cut-off point for MPV/PC ratio to predict prognosis. Cho et al. (2013) demonstrated that 0.0491 might be the optimum cut-off point for MPV/PC ratio in hepatocellular carcinoma according to the ROC curve. Inagaki et al. (2014) and Omar et al. (2018) also conducted the ROC curve analyses to calculate the optimum cut-off point for MPV/PC in lung cancer. They concluded that the optimum cut-off points for MPV/PC ratio were 0.40873 and 0.47424, respectively. Recently, Camp, Dolled-Filhart & Rimm (2004) initial conducted a program to explore the optimum cut-off point (X-tile plot). In our study, according to their method, 0.04 was the optimum cut-off point for MPV/PC ratio. We also created ROC curves to show the candidate cut-off points. When we set the cut-off point using ROC curve, the MPV/PC ratio was also associated with CSS. In multivariate analyses, MPV/PC ratio was still an independent predictor for CSS.

The mechanism between MPV/PC ratio and cancer remains unknown. Inflammation and cancer are closely related (Balkwill & Mantovani, 2001; Mantovani et al., 2008). As is well known, platelets can release a variety of cytokines, such as platelet-derived growth factor (PDGF) and vascular endothelial growth factor (VEGF), which have an important role in regulating angiogenesis (Blair & Flaumenhaft, 2009; Borsig, 2008; Dineen et al., 2009). The inflammation will be inevitably caused by chemotherapy and/or radiation. Therefore, we analyze the role of MPV/PC ratio in ESCC patients without neoadjuvant chemotherapy and/or radiation.

Limitations should be acknowledged in this study. The major limitations of this study are small samples and its retrospective character. Moreover, patients who received preoperative chemotherapy and/or radiotherapy were excluded, which might have influenced the result in the current study. On the one hand, neoadjuvant treatment will have a side effect on MPV and PC. On the other hand, neoadjuvant treatment can improve cancer survival for locally advanced EC, but not for early stage EC (Rawat et al., 2013; Mariette et al., 2014). In addition, we did not set up a validation group to verify the conclusion. Thus, the results of our study are expected more large-sample trials to confirm in future.

Conclusion

In summary, we found that the ratio of MPV/PC is a potential prognostic biomarkers in patients with ESCC.

Funding Statement

This work was supported by the Medical Health Science and Technology Project of Zhejiang Provincial Health Commission (No. 2018KY290, 2019RC129). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.