Sage Journals: Discover world-class research

Abstract

BACKGROUND:

Systemic inflammatory biomarkers reflect level of inflammatory response, which have been suggested as prognostic factors in cancer patients.

OBJECTIVE:

To estimate the prognostic value of inflammatory biomarkers in 149 patients with head and neck soft tissue sarcoma (HNSTS).

METHODS:

Pre-treatment neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR), lymphocyte-to-monocyte ratio (LMR), systemic immune-inflammation index (SII), neutrophil-platelet score (NPS) and Aarhus Composite Biomarker Score (ACBS) were analyzed for association with overall survival (OS) and progression-free survival (PFS). Nomograms were consisted of independent predictors for OS and PFS, and evaluated by calibration curve and concordance index (C-index).

RESULTS:

In multivariate analysis, LMR (HR $=$ 0.42, 95% CI: 0.19–0.94, $P=$ 0.035), ACBS (HR $=$ 2.05, 95% CI: 1.02–4.12, $P=$ 0.045) and AJCC stage were independent prognostic markers of OS. Moreover, high NLR (HR $=$ 1.78, 95% CI: 1.07–2.94, $P=$ 0.024) and advanced AJCC stage were independently related with worse PFS. Calibration curves reflected good discriminative ability of prognosis. The nomograms showed better accuracy of predicting OS (C-index: 0.748 vs. 0.690, $P=$ 0.009) and PFS (C-index: 0.644 vs. 0.612, $P=$ 0.028) than stage.

CONCLUSIONS:

Pre-treatment LMR, ACBS and AJCC stage were effective predictor of OS. The NLR and AJCC stage could independently predict PFS. The nomogram might act as a promising prognostic model for OS and PFS in HNSTS.

Keywords

Soft tissue sarcoma head and neck inflammation nomogram prognosis

1. Introduction

Soft tissue sarcoma (STS) is a rare tumor, accounting for approximately 1% of all cancers, with annual incidence around 5 among 100,000 persons [1]. STS comprises over 50 histological types of heterogeneous malignancies arising from various kind of tissue, including smooth muscle, adipose tissue and so on [2]. Head and neck soft tissue sarcoma (HNSTS) is the rarest site, which takes up about 5–15% of sarcoma and 1% of head and neck tumor [3]. However, HNSTS is rare but aggressive, with poor five-year overall survival rate (31–80%) and high rates of local recurrence (22–53%) and distant metastases (10–40%) [4]. Conventional prognostic factors such as tumor size and histologic grade still have imperfections of inaccuracy and inconvenience in clinical practice [5]. Therefore, it is important to estimate novel prognostic markers that identify high-risk HNSTS patients and optimize therapeutic strategy to improve survival outcomes.

Increasing evidences showed that cancer-related inflammation has an important role on tumor microenvironment, which promotes tumor proliferation, angiogenesis and metastasis via different inflammatory cells and cytokines [6]. Therefore, as blood indicators of inflammatory response, the systemic inflammatory biomarkers such as neutrophil-to-lymphocyte ratio (NLR), platelet-to-lymphocyte ratio (PLR) and lymphocyte to monocyte ratio (LMR) were suggested as effective prognostic factors [7, 8]. In previous studies of STS, pre-treatment NLR, PLR, LMR and Aarhus Composite Biomarker Score (ACBS) were practical in predicting overall survival or disease progression [9, 10, 11, 12]. However, to the best of our knowledge, the prognostic value of these inflammation biomarkers has not been separately analyzed for patients with HNSTS. Moreover, the nomogram incorporating inflammatory biomarkers and clinical characteristics could provide an intuitive graphical representation of statistic model, which was considered as a practical method to predict clinical outcome in esophageal cancer, breast cancer and so on [14, 15]. Also, the application of nomograms based on inflammatory biomarkers for HNSTS has not been reported.

Therefore, the purpose of present study was to evaluate the prognostic value of pre-treatment inflammatory biomarkers including NLR, PLR, LMR, systemic immune-inflammation index (SII), neutrophil-platelet score (NPS) and ACBS, then to establish inflammatory biomarkers-based nomograms to reveal the association between inflammatory biomarkers and survival results in patients with HNSTS.

2. Materials and methods

2.1 Patients selection

We searched the electronic medical records of the *BLIND* Hospital, and reviewed all patients with STS between January 2006 and December 2016. The inclusion criteria: 1) Patients with primary STS of head and neck, who were pathological diagnosed and had surgery. 2) Patients had complete laboratory data of inflammatory biomarkers and medical records before initial treatments (surgical resection, neoadjuvant chemotherapy or radiotherapy). 3) Patients received regular and complete follow-up, and the corresponding clinical outcomes were definite. The exclusion criteria: 1) Patients had primary STS of other site, or 2) Lacked pathological reports, pre-treatment blood tests and follow-up data. 3) Patients with some diseases, such as respiratory infection, atherosclerosis and hematologic disease, or medication history that could affect inflammatory markers level were also excluded. This study was approved by the institution Ethics Commission of *BLIND* Hospital of *BLIND* University. The need for consent to participate was waived by Ethics Commission.

2.2 Data collection

The peripheral blood tests were performed within 1 week before first treatment, which contained platelet counts (P), neutrophil counts (N), lymphocyte counts (L), levels of hemoglobin, monocytes, albumin and C-reactive protein (CRP). SII was calculated by this formula: SII $=$ (P $\times$ N)/L. Neutrophil-platelet score (NPS) was calculated based on counts of platelets and neutrophils. A score of 2 refers to elevated levels of neutrophils ( $>$ 7.5 $\times$ 10 ${}^{9}$ /L) and platelets ( $>$ 400 $\times$ 10 ${}^{9}$ /L), a score of 1 indicates only one abnormal biomarker, and score of 0 means that two biomarkers were both normal. Aarhus Composite Biomarker Score (ACBS) was defined by a combination of hemoglobin, neutrophils, lymphocytes, albumin and CRP levels. Patients with abnormal levels of more than one biomarker have a score of 2, one abnormal biomarker was scored 1, and all biomarkers within the normal range were scored 0. Pathological stage was classified according to eighth edition of AJCC-TNM staging system [15], and sarcomas were graded based on FNCLCC grading system [16].

All patients received routine follow-up every three months during the first three years, every six months in fourth to fifth years, and then once a year until December 2018 or their death [17]. Times for overall survival (OS) were calculated from the period from initial histological diagnosis to latest clinical observation or death. The progression-free survival (PFS) was defined as the date of histological diagnosis until first evidence of local relapse, lymph node or distant metastases, death or the latest follow-up.

2.3 Statistical analysis

According to the results of OS, we performed receiver operating curve (ROC) analysis to obtain area under the curve (AUC) of NLR, PLR, LMR and SII. Then we obtained optimal cut-off values via calculating best Youden index. The differences of clinicopathological characteristics between high and low inflammatory biomarkers were compared by the chi-square test. Crude survival analyses were showed in Kaplan-Meier curves and compared by the log-rank test. These factors also were assessed by univariate Cox proportional hazard model, and significant markers would be included in multivariate Cox analysis using the likelihood ratio test. Association between prognostic factors and survival outcome were calculated as the hazard ratio (HR) and 95% confidence interval (CI). The significant $P$ value was of $<$ 0.05, and above analyses were conducted on SPSS statistics 22.0.

Nomograms were established using “rms” package on R software 3.5.3. The significant markers in multivariate Cox analysis were included in nomogram, then the predicted probabilities of 3-year, 5-year and 8-year OS and PFS were presented. We performed bootstrapping method with 1000 resamples to generate calibration curves that verified the deviation between predicted survival and actual observation, with the Hosmer-Lemeshow test. The discriminative abilities of nomograms and traditional markers were compared using Harrell’s concordance index (C-index) and ROC analysis.

3. Results

3.1 Clinical characteristics of patients

There were 149 Asian patients diagnosed with HNSTS in present study, including 90 males and 59 females. All patients were Chinese, which were mostly Han Chinese ( $n=$ 139) with national minorities of Yi ( $n=$ 5), Tibetan ( $n=$ 3) and Qiang ( $n=$ 2). Flow chart for patients’ selection was provided in Supplementary Fig. 1. The median age at diagnosis was 35 years (range, 3–85 years). The histologic subtypes consisted of rhabdomyosarcoma ( $n=$ 39), fibrosarcoma ( $n=$ 30), non-differentiated sarcoma ( $n=$ 20), synovial sarcoma ( $n=$ 13), angiosarcoma ( $n=$ 13), liposarcoma ( $n=$ 11), leiomyosarcoma ( $n=$ 7), malignant peripheral nerve sheath tumor (MPNST; $n=$ 6), extraskeletal mesenchymal chondrosarcoma ( $n=$ 6) and extraskeletal Ewing sarcoma ( $n=$ 4). The most common sites were nasal cavities and sinuses ( $n=$ 34); superficial areas of neck and parotid ( $n=$ 30); orbit and skull base ( $n=$ 27); superficial face and scalp ( $n=$ 25); oral cavity, larynx and pharynx ( $n=$ 19); deep facial space ( $n=$ 10) and ear ( $n=$ 4). Sixty-eight patients underwent tumor resection with adjuvant radiochemotherapy, 10 and 8 patients had surgery with chemotherapy or radiotherapy, and preoperative chemotherapy was given in 3 patients.

The best cut-off value was 2.2 for NLR (AUC $=$ 0.647, $P=$ 0.008, Youden index $=$ 0.285), 92 for PLR (AUC $=$ 0.618, $P=$ 0.033, Youden index $=$ 0.254), 5.3 for LMR (AUC $=$ 0.612, $P=$ 0.043, Youden index $=$ 0.281) and 440 for SII (AUC $=$ 0.638, $P=$ 0.027, Youden index $=$ 0.265). Differences of clinicopathological characteristics between high and low inflammatory biomarkers were listed in Table 1. There were more patients with early stage (I–II), lower SII and normal NPS in low-level NLR and high-level LMR group. Significant differences were also observed in the selection of chemotherapy (LMR, $P=$ 0.044) and radiotherapy (PLR, $P=$ 0.042; LMR, $P=$ 0.042).

Table 1
Clinicopathological characteristics of patients

Variables	Parameters	NLR		$p$	PLR		$p$	LMR		$p$
		$<$ 2.2 (%)	$\geqslant$ 2.2 (%)		$<$ 92 (%)	$\geqslant$ 92 (%)		$<$ 5.3 (%)	$\geqslant$ 5.3 (%)
Age	$<$ 35 years	48 (56.5)	26 (40.6)		35 (57.4)	39 (44.3)		38 (47.5)	36 (52.2)
	$\geqslant$ 35 years	37 (43.5)	38 (59.4)	0.056	26 (42.6)	49 (55.7)	0.117	42 (52.5)	33 (47.8)	0.569
Gender	Male	51 (60.0)	39 (60.9)		41 (67.2)	49 (55.7)		50 (62.5)	40 (58.0)
	Female	34 (40.0)	25 (39.1)	0.908	20 (32.8)	39 (44.3)	0.157	30 (37.5)	29 (42.0)	0.573
Performance status	0–1	62 (72.9)	39 (60.9)		45 (73.8)	56 (63.6)		50 (62.5)	51 (73.9)
	2	23 (27.1)	25 (39.1)	0.121	16 (26.2)	32 (36.4)	0.193	30 (37.5)	18 (26.1)	0.137
Tumor size	$<$ 5 cm	60 (70.6)	43 (67.2)		47 (77.0)	56 (63.6)		56 (70.0)	47 (68.1)
	$\geqslant$ 5 cm	25 (29.4)	21 (32.8)	0.656	15 (23.0)	32 (36.4)	0.081	24 (30.0)	22 (31.9)	0.804
Tumor depth	Superficial	40 (47.1)	26 (40.6)		31 (50.8)	35 (39.8)		31 (38.8)	35 (50.7)
	Deep	45 (52.9)	38 (59.4)	0.434	30 (49.2)	53 (60.2)	0.182	49 (61.2)	34 (49.3)	0.142
FNCLCC grade	G1	20 (23.5)	11 (17.2)		15 (24.6)	16 (18.2)		12 (15.0)	19 (27.5)
	G2	25 (29.4)	19 (29.7)		18 (29.5)	26 (29.5)		26 (32.5)	18 (26.1)
	G3	40 (47.1)	34 (53.1)	0.613	28 (45.9)	46 (52.3)	0.605	42 (52.5)	32 (46.4)	0.166
AJCC stage	I	20 (23.5)	11 (17.2)		15 (24.6)	16 (18.2)		12 (15.0)	19 (27.5)
	II	41 (48.2)	24 (37.5)		26 (42.6)	39 (44.3)		37 (46.3)	28 (40.6)
	III	18 (21.2)	9 (14.1)		12 (19.7)	15 (17.0)		9 (11.3)	18 (26.1)
	IV	6 (7.1)	20 (31.3)	0.002 ${}^{*}$	8 (13.1)	18 (20.5)	0.576	22 (27.5)	4 (5.8)	0.001 ${}^{*}$
Chemotherapy	Yes	39 (45.9)	39 (60.9)		28 (45.9)	50 (56.8)		48 (60.0)	30 (43.5)
	No	46 (54.1)	25 (39.1)	0.069	33 (54.1)	38 (43.2)	0.190	32 (40.0)	39 (56.5)	0.044 ${}^{*}$
Radiotherapy	Yes	39 (45.9)	37 (57.8)		25 (41.0)	51 (58.0)		47 (58.8)	29 (42.0)
	No	46 (54.1)	27 (42.2)	0.149	36 (59.0)	37 (42.0)	0.042 ${}^{*}$	33 (41.3)	40 (58.0)	0.042 ${}^{*}$
SII	$<$ 440	79 (92.9)	4 (6.3)		52 (85.2)	31 (35.2)		24 (30.0)	59 (85.5)
	$\geqslant$ 440	6 (70.6)	60 (93.8)	$<$ 0.001 ${}^{*}$	9 (14.8)	57 (64.8)	$<$ 0.001 ${}^{*}$	56 (70.0)	10 (14.5)	$<$ 0.001 ${}^{*}$
NPS	0	82 (96.5)	45 (70.3)		57 (93.4)	74 (84.1)		66 (82.5)	65 (94.2)
	$\geqslant$ 1	3 (3.5)	19 (29.7)	$<$ 0.001 ${}^{*}$	4 (6.6)	14 (15.9)	0.085	14 (17.5)	4 (5.8)	0.029 ${}^{*}$
ACBS	0	59 (69.4)	42 (65.6)		38 (62.3)	63 (71.6)		56 (70.0)	45 (65.2)
	$\geqslant$ 1	26 (30.6)	22 (34.4)	0.624	23 (37.7)	25 (28.4)	0.232	24 (30.0)	24 (34.8)	0.533

Abbreviations: NLR neutrophil-lymphocyte ratio, PLR platelet-lymphocyte ratio, LMR lymphocyte-monocyte ratio, SII systemic immune-inflammation index, NPS neutrophil-platelet score, ACBS Aarhus Composite Biomarker Score. ${}^{*}$ Statistically significant $p<$ 0.05.

3.2 Univariate and multivariate survival analysis

The median duration of follow-up was 44 months (range, 5–183 months). Totally 113 patients (75.8%) were alive at the latest follow-up with 3-year OS rate of 80.6%, 5-year OS rate of 73.9% and 8-year OS rate of 64.1%. Sixty-nine patients (46.3%) were observed to have disease progression, which included 42 cases (28.2%) with recurrence in situ, 16 patients (10.7%) with distant metastases (mostly in lung and bone) and 9 cases (6.0%) of cervical lymph node metastases. The probabilities of 3-year, 5-year and 8-year PFS were 65.2%, 53.1% and 45.8%, respectively.

As shown in Table 2, NLR ( $P=$ 0.010, $P=$ 0.004), PLR ( $P=$ 0.026, $P=$ 0.009), LMR ( $P=$ 0.005, $P=$ 0.007) and SII ( $P=$ 0.025, $P=$ 0.043) were all significantly associated with OS and PFS in univariate analyses. The poorer curves of OS and PFS were found in elevated NLR and PLR as well as decreased LMR groups (Fig. 1). ACBS had significant prognostic value of OS ( $P=$ 0.004), except for PFS ( $P=$ 0.226). Patients with score 0 of NPS failed to reflect significantly better OS ( $P=$ 0.370) and PFS ( $P=$ 0.113). The Kaplan-Meier survival curves of SII, ACBS and NPS were listed in Supplementary Fig. 2. The univariate analyses also indicated that performance status ( $P=$ 0.006), tumor depth ( $P=$ 0.044), tumor grade ( $P=$ 0.021, $P=$ 0.011) and stage ( $P=$ 0.028, $P=$ 0.023, $P=$ 0.002) were related with OS. In addition, patients with poor performance status ( $P=$ 0.022), deep tumor ( $P=$ 0.044), chemotherapy ( $P=$ 0.005), medium-high tumor grade ( $P=$ 0.001, $P=$ 0.003) and advanced stage ( $P=$ 0.002, $P=$ 0.043, $P=$ 0.005) had worse results of PFS.

Table 2
Univariate Cox analysis of prognostic factors affecting survival

Variables		No. of patients	No. of events	Overall survival		No. of events	Progression-free survival
				HR (95% CI)	$p$		HR (95% CI)	$p$
Age	$<$ 35 years	74	21	1		34	1
	$\geqslant$ 35 years	75	15	0.54 (0.27–1.05)	0.071	35	0.80 (0.49–1.29)	0.360
Gender	Male	90	22	1		40	1
	Female	59	14	0.90 (0.46–1.76)	0.762	29	1.08 (0.67–1.75)	0.752
Performance status	0–1	101	18	1		42	1
	2	48	18	2.51 (1.30–4.83)	0.006 ${}^{*}$	27	1.79 (1.09–2.94)	0.022 ${}^{*}$
Tumor size	$<$ 5 cm	103	21	1		46	1
	$\geqslant$ 5 cm	46	15	1.89 (0.97–3.68)	0.060	23	1.24 (0.75–2.05)	0.397
Tumor depth	Superficial	66	11	1		25	1
	Deep	83	25	2.07 (1.02–4.22)	0.044 ${}^{*}$	44	1.66 (1.01–2.73)	0.044 ${}^{*}$
FNCLCC grade	G1	31	1	1		8	1
	G2	44	12	11.08 (1.43–85.56)	0.021 ${}^{*}$	24	3.86 (1.68–8.86)	0.001 ${}^{*}$
	G3	74	23	13.31 (1.79–98.77)	0.011 ${}^{*}$	37	3.30 (1.51–7.21)	0.003 ${}^{*}$
AJCC stage	I	31	1	1		8	1
	II	65	16	9.66 (1.28–73.04)	0.028 ${}^{*}$	34	3.36 (1.53–7.39)	0.002 ${}^{*}$
	III	27	7	11.33 (1.39–94.44)	0.023 ${}^{*}$	11	2.64 (1.03–6.75)	0.043 ${}^{*}$
	IV	26	12	25.80 (3.30–201.46)	0.002 ${}^{*}$	16	5.09 (2.11–12.30)	$<$ 0.001 ${}^{*}$
Chemotherapy	Yes	78	23	1		44	1
	No	71	13	0.59 (0.30–1.17)	0.131	25	0.49 (0.30–0.80)	0.005 ${}^{*}$
Radiotherapy	Yes	76	22	1		41	1
	No	73	14	0.66 (0.34–1.30)	0.232	28	0.64 (0.39–1.04)	0.069
NLR	$<$ 2.2	85	13	1		30	1
	$\geqslant$ 2.2	64	23	2.44 (1.23–4.83)	0.010 ${}^{*}$	39	2.02 (1.25–3.25)	0.004 ${}^{*}$
PLR	$<$ 92	61	9	1		21	1
	$\geqslant$ 92	88	27	2.37 (1.11–5.05)	0.026 ${}^{*}$	48	2.01 (1.19–3.39)	0.009 ${}^{*}$
LMR	$<$ 5.3	80	27	1		44	1
	$\geqslant$ 5.3	69	9	0.34 (0.16–0.72)	0.005 ${}^{*}$	25	0.50 (0.31–0.83)	0.007 ${}^{*}$
SII	$<$ 440	83	14	1		32	1
	$\geqslant$ 440	66	22	2.15 (1.10–4.20)	0.025 ${}^{*}$	37	1.63 (1.02–2.62)	0.043 ${}^{*}$
NPS	0	131	30	1		57	1
	$\geqslant$ 1	18	6	1.49 (0.62–3.59)	0.370	12	1.67 (0.89–3.13)	0.113
ACBS	0	101	19	1		46	1
	$\geqslant$ 1	48	17	2.64 (1.35–5.13)	0.004 ${}^{*}$	23	1.37 (0.82–2.28)	0.226

Figure 1.

Kaplan-Meier curves of overall survival stratified by NLR (A), PLR (B) and LMR (C). Kaplan-Meier curves of progression-free survival stratified by NLR (D), PLR (E) and LMR (F).

Table 3

Multivariate Cox analysis of prognostic factors affecting survival

Variables	Parameters	Overall survival		Progression-free survival
		HR (95% CI)	$p$	HR (95% CI)	$p$
Performance status	0–1 vs. 2	–	0.489	–	0.372
Tumor depth	Superficial vs. deep	–	0.939	–	0.918
FNCLCC grade	G1 vs. G2 vs. G3	–	0.451	–	0.486
AJCC stage	I vs. II	7.19 (0.94–54.95)	0.057	3.29 (1.51–7.21)	0.003 ${}^{*}$
	I vs. III	9.00 (1.09–74.39)	0.042 ${}^{*}$	2.56 (1.01–6.51)	0.049 ${}^{*}$
	I vs. IV	13.86 (1.70–113.02)	0.014 ${}^{*}$	3.97 (1.61–9.78)	0.003 ${}^{*}$
Chemotherapy	Yes vs. no			–	0.669
NLR	$<$ 2.2 vs. $\geqslant$ 2.2	–	0.761	1.78 (1.07–2.94)	0.024 ${}^{*}$
PLR	$<$ 92 vs. $\geqslant$ 92	–	0.433	–	0.263
LMR	$<$ 5.3 vs. $\geqslant$ 5.3	0.42 (0.19–0.94)	0.035 ${}^{*}$	–	0.648
SII	$<$ 440 vs. $\geqslant$ 440	–	0.663	–	0.621
ACBS	0 vs. $\geqslant$ 1	2.05 (1.02–4.12)	0.045 ${}^{*}$

Abbreviations: NLR neutrophil-lymphocyte ratio, PLR platelet-lymphocyte ratio, LMR lymphocyte-monocyte ratio, SII systemic immune-inflammation index, ACBS Aarhus Composite Biomarker Score. ${}^{*}$ Statistically significant $p<$ 0.05.

Figure 2.

Nomograms for predicting the probability of 3-year, 5-year and 8-year OS (A) and PFS (B). Calibration curves of nomograms for predicting 3-year and 5-year OS (C) and PFS (D).

In multivariate analysis (Table 3), the lower LMR (HR $=$ 0.42, 95% CI: 0.19–0.94, $P=$ 0.035), ACBS $\geqslant$ 1 (HR $=$ 2.05, 95% CI: 1.02–4.12, $P=$ 0.045) and advanced stage (I vs. III, HR $=$ 9.00, 95% CI: 1.09–74.39, $P=$ 0.042; I vs. IV, HR $=$ 13.86, 95% CI: 1.70–113.02, $P=$ 0.014) were unfavorable prognostic markers for OS. In addition, NLR (HR $=$ 1.78, 95% CI: 1.07–2.94, $P=$ 0.024) and AJCC stage (I vs. II, HR $=$ 3.29, 95% CI: 1.51–7.21, $P=$ 0.003; I vs. III, HR $=$ 2.56, 95% CI: 1.01–6.51, $P=$ 0.049; I vs. IV, HR $=$ 3.97, 95% CI: 1.61–9.78, $P=$ 0.003) were independently associated with PFS.

3.3 Establishment and evaluation of nomogram

As independent predictors, AJCC stage, LMR and ACBS were included in nomogram for OS (Fig. 2A). The AJCC stage and LMR were used to establish nomogram for PFS that predicting probability of survival events by calculating total points (Fig. 2B). The calibration curves for 3-year and 5-year OS and PFS reflected good agreement between prediction and observed survival (Fig. 2C and D). The P value of Hosmer-Lemeshow test was 0.746 for OS and 0.262 for PFS. The C-index of nomogram for OS and PFS was 0.748 (95% CI $=$ 0.675–0.820) and 0.644 (95% CI $=$ 0.581–0.701). Moreover, the C-index of AJCC stage was 0.690 (95% CI $=$ 0.610–0.771) for OS and 0.612 (95% CI $=$ 0.546–0.678) for PFS. By comparison, the nomograms showed improved discriminative accuracy than AJCC stage both in OS ( $P=$ 0.009) and PFS ( $P=$ 0.028). Results of ROC analyses showed in Fig. 3, nomograms had enhanced AUCs for OS and PFS than inflammatory biomarkers and AJCC stage.

Figure 3.

ROC curves comparing the discriminative ability of nomograms, clinicopathological features and inflammatory biomarkers for OS (A) and PFS (B).

4. Discussion

First reported by Rudolf Virchow, the cancer-related inflammation is considered to contribute to tumor development and progression via inducing angiogenesis, cell proliferation and tissue infiltration [18]. Systemic inflammatory biomarkers reflected level of inflammatory and immune response, thus were suggested as effective predictors of survival and cancer progression [19]. In present research, we analyzed the prognostic value of six inflammatory biomarkers for first time in HNSTS. The results showed that pre-treatment LMR and ACBS were independently associated with OS, and NLR was an unfavorable independent predictor of PFS. Nomograms cooperating staging system and inflammatory biomarkers also were convenient and practical in predicting OS and PFS in patients with HNSTS.

The allogeneic phenotype of tumor cell stimulates the body to produce inflammatory cells, which leads to the collapse of tumor tissues and results in broad and lasting non-specific inflammatory response [6]. The inflammatory response may lead to enhancement of leukocytes, neutrophils and platelets as well as reduction in lymphocytes [20]. Increased tumor infiltrating neutrophils (TINs) could recruit tumor cell proliferation and promote angiogenesis by secreting various cytokines such as TNF- $\alpha$ , VEGF and IL-6 [21, 22]. On the contrary, tumor-infiltrating lymphocytes (TILs) have antitumor effects of inhibiting tumor proliferation and metastasis by mediating cytotoxicity and releasing cytokines [23]. TINs may suppress immune response via inhibiting activity of natural killer cells and T lymphocytes [24]. Therefore, NLR from peripheral blood can reflect the balance between promote-tumor inflammatory response and anti-tumor immunity. The balance can be broken by increased neutrophils and/or decreased lymphocytes, which indicated damaged immunologic function and poor prognosis of patients [25]. Previous meta-analysis of 14 cohorts (2820 patients with STS) suggested NLR as adverse prognostic biomarker for OS (HR $=$ 1.59, 95% CI: 1.28–1.97, $P<$ 0.001) and PFS (HR $=$ 1.28, 95% CI: 1.12–1.47, $P<$ 0.001) [26]. Our study focused on HNSTS patients, and the results also supported that NLR was an independent prognostic marker for PFS (HR $=$ 1.78, 95% CI: 1.07–2.94, $P=$ 0.024), but not for OS ( $P=$ 0.761).

Platelets were also reported to involve the inflammation and support cancer progression. For instance, platelets can promote angiogenesis by proangiogenic factors, stimulate tumor cells proliferation via TGF- $\beta$ , inhibit apoptosis and increase DNA repair in cancers [27, 28, 29]. In addition, inflammation can recruit monocytes into tumor tissue from peripheral blood, then monocytes would differentiate into tumor associated macrophages (TAMs) [30]. TAMs have a stimulatory effect on cell infiltration, growth, metastases and neovascularization [31]. Thus, LMR can reflect the formation or existence of TAMs, which was related with prognosis in many tumors [32]. There was published article concluded that PLR was a better prognostic factor than NLR in patients with STS [10, 33]. However, multivariate analyses showed no significant difference of OS and PFS in pre-treatment PLR and LMR groups in other studies [11, 34]. Our research failed to verify the independent prognostic role of PLR, but increased pre-treatment LMR may be favorable predictor of OS (HR $=$ 0.42, 95% CI: 0.19–0.94, $P=$ 0.035) in HNSTS.

Systemic immune-inflammation index (SII) is an integrative indicator based on levels of lymphocytes, neutrophils and platelets. Previous studies have suggested SII as a superior predictor of long-term survival outcome compared to NLR and PLR in colorectal cancer [35], lung cancer [36] and so on. It also reported that SII was significantly correlated with the number of circulating tumor cells, which potential promoted the metastases and progression of gastric cancer [37]. However, for all we know, the SII for STS has not been estimated before, and the prognostic role of SII remained unclear in patients with STS. Although the SII was not an independent predictor for OS and PFS in this study, the significant results in univariate analyses indicated the potential prognostic ability of SII, which required further investigation.

Neutrophil-platelet score (NPS) could reflect the levels of neutrophils and platelets, then was suggested as effective predictor of patient’s survival and disease development in many cancers [38]. Evaluation of NPS has not been reported in STS, HNSTS patients with NPS score $\geqslant$ 1 didn’t have significantly worse OS and PFS in present study. ACBS was a new composite biomarker recommended by previous research, which indicated that ACBS was an independent prognostic factor for disease-specific mortality in bone sarcoma and STS of the trunk and extremities [12, 39]. We also found that high ACBS was independently related with worse OS in HNSTS. However, the prognostic role of these inflammatory markers still remains controversial, thereby more studies are needed for full estimation.

The nomogram is a convenient and practical identification method to predict clinical events and prognosis of patients with visualized possible risk factors. It was worth that we established nomograms combined by traditional staging system and inflammatory biomarkers in HNSTS, which was only reported in osteosarcoma and spinal Ewing’s sarcoma expect for STS [19, 40]. The results of C-index, ROC analysis and calibration curve showed improved accuracy of nomograms than AJCC stage and inflammatory biomarkers for OS and PFS. The nomogram-based model could be a novel potential tool to predict prognosis and therapeutic results of HNSTS patients in clinical work.

The present study had several limitations. Firstly, there were significant correlations between inflammatory biomarkers, tumor stage and treatments, which could be potential con-founders of survival analysis. In addition, due to the rarity of STS, especially of head and neck region, the relatively small sample could lead to selection bias. For instance, the median age at diagnosis was lower than that of previous epidemiologic studies [41, 42]. Also, the inevitable heterogeneity of histological types of STS may affect the reliability and generalizability of results. Furthermore, this study was limited by retrospective design in single center, thus it was better to verify the prognostic role inflammatory biomarkers in multi-center prospective research.

5. Conclusions

For patients with head and neck soft tissue sarcoma, pre-treatment LMR, ACBS and AJCC stage were independent prognostic factors of OS. The elevated NLR and advanced AJCC stage were independently associated with worse PFS. The prognostic nomograms integrating NLR, LMR, ACBS and AJCC staging system might be applied to predict clinical outcome and optimize therapeutic strategy.

Footnotes

Conflict of interest

The authors declare that they have no conflict of interests.

Supplementary Figure

Supplementary Fig. 1.

The flowchart of patient selection.

Supplementary Fig. 2.

Kaplan-Meier curves of overall survival stratified by SII (A), NPS (B) and ACBS (C). Kaplan-Meier curves of progression-free survival stratified by SII (D), NPS (E) and ACBS (F).

References

Stiller

C.A.

Trama

Serraino

Rossi

Navarro

Chirlaque

M.D.

Casali

P.G.

and RARECARE Working Group, Descriptive epidemiology of sarcomas in europe: report from the RARECARE project, European Journal of Cancer 49(3) (2013), 684–695.

Fletcher

C.D.

, The evolving classification of soft tissue tumours – an update based on the new 2013 WHO classification, Histopathology 64(1) (2014), 2–11.

Eeles

R.A.

Fisher

A’Hern

R.P.

Robinson

Rhys-Evans

Henk

J.M.

Archer

and Harmer

C.L.

, Head and neck sarcomas: prognostic factors and implications for treatment, British Journal of Cancer 68(1) (1993), 201–207.

Galy-Bernadoy

and Garrel

, Head and neck soft-tissue sarcoma in adults, European Annals of Otorhinolaryngology, Head and Neck Diseases 133(1) (2016), 37–42.

Yang

Zhang

Wang

and Li

, Clinical implications of six inflammatory biomarkers as prognostic indicators in Ewing sarcoma, Cancer Management and Research 9 (2017), 443–451.

Mantovani

Allavena

Sica

and Balkwill

, Cancer-related inflammation, Nature 454(7203) (2008), 436–444.

Kishi

Kopetz

Chun

Y.S.

Palavecino

Abdalla

E.K.

and Vauthey

J.N.

, Blood neutrophil-to-lymphocyte ratio predicts survival in patients with colorectal liver metastases treated with systemic chemotherapy, Annals of Surgical Oncology 16(3) (2009), 614–622.

Diem

Schmid

Krapf

Flatz

Born

Jochum

Templeton

A.J.

and Früh

, Neutrophil-to-Lymphocyte ratio (NLR) and Platelet-to-Lymphocyte ratio (PLR) as prognostic markers in patients with non-small cell lung cancer (NSCLC) treated with nivolumab, Lung Cancer 111 (2017), 176–181.

Choi

E.S.

Kim

H.S.

and Han

, Elevated preoperative systemic inflammatory markers predict poor outcome in localized soft tissue sarcoma, Annals of Surgical Oncology 21(3) (2014), 778–785.

10.

Que

Qiu

Chen

Xiao

Zhou

and Zhang

, Preoperative platelet-lymphocyte ratio is superior to neutrophil-lymphocyte ratio as a prognostic factor for soft-tissue sarcoma, BMC Cancer 15 (2015), 648.

11.

Mirili

Payda

Guney

I.B.

Ogul

Gokcay

Buyuksimsek

Yetisir

A.E.

Karaalioglu

Tohumcuoglu

and Seydaoglu

, Assessment of potential predictive value of peripheral blood inflammatory indexes in 26 cases with soft tissue sarcoma treated by pazopanib: a retrospective study, Cancer Management and Research 11 (2019), 3445–3453.

12.

Maretty-Kongstad

Aggerholm-Pedersen

Keller

and Safwat

, A validated prognostic biomarker score for adult patients with nonmetastatic soft tissue sarcomas of the trunk and extremities, Translational Oncology 10(6) (2017), 942–948.

13.

Liu

Huang

Yang

and Feng

, A nomogram to predict prognostic values of various inflammatory biomarkers in patients with esophageal squamous cell carcinoma, American Journal of Cancer Research 5(7) (2015), 2180–2189.

14.

Cho

Park

H.S.

S.Y.

Yoo

C.Y.

Jung

J.H.

Suh

Y.J.

and Choi

H.J.

, Prognostic value of systemic inflammatory markers and development of a nomogram in breast cancer, PLoS One 13(7) (2018), e0200936.

15.

Cates

J.M.M.

, AJCC eighth edition for soft tissue sarcoma of the extremities and trunk, Annals of Oncology 29(9) (2018), 2023.

16.

Neuville

Chibon

and Coindre

J.M.

, Grading of soft tissue sarcomas: from histological to molecular assessment, Pathology 46(2) (2014), 113–120.

17.

Casali

P.G.

Abecassis

Aro

H.T.

et al., Soft tissue and visceral sarcomas: ESMO-EURACAN clinical practice guidelines for diagnosis, treatment and follow-up, Annals of Oncology 29(Suppl 4) (2018), iv51–iv67.

18.

Balkwill

and Mantovani

, Inflammation and cancer: back to Virchow? Lancet 357(9255) (2001), 539–545.

19.

Liu

Zhang

and Yao

, Prognostic role of elevated preoperative systemic inflammatory markers in localized soft tissue sarcoma, Cancer Biomarkers 16(3) (2016), 333–342.

20.

Xia

Liu

Shen

Lin

and Mao

, Prognostic performance of pre-treatment NLR and PLR in patients suffering from osteosarcoma, World Journal of Surgical Oncology 14 (2016), 127.

21.

el-Hag

and Clark

R.A.

, Immunosuppression by activated human neutrophils, dependence on the myeloperoxidase system, Journal of Immunology 139(7) (1987), 2406–2413.

22.

Jabłońska

Kiluk

Markiewicz

Piotrowski

Grabowska

and Jabłoński

, TNF-alpha, IL-6 and their soluble receptor serum levels and secretion by neutrophils in cancer patients, Archivum Immunologiae et Therapiae Experimentalis 49(1) (2001), 63–69.

23.

Dunn

G.P.

Old

L.J.

and Schreiber

R.D.

, The immunobiology of cancer immunosurveillance and immunoediting, Immunity 21(2) (2004), 137–148.

24.

Mantovani

, Cancer: inflaming metastasis, Nature 457(7225) (2009), 36–37.

25.

Fogar

Sperti

Basso

Sanzari

M.C.

Greco

Davoli

Navaglia

Zambon

C.F.

Pasquali

Venza

Pedrazzoli

and Plebani

, Decreased total lymphocyte counts in pancreatic cancer: an index of adverse outcome, Pancreas 32(1) (2006), 22–28.

26.

Liu

and Sun

, Prognostic value of pretreatment neutrophil-to-lymphocyte ratio in patients with soft tissue sarcoma: a meta-analysis, Medicine (Baltimore) 97(36) (2018), e12176.

27.

Cho

M.S.

Bottsford-Miller

Vasquez

H.G.

Stone

Zand

Kroll

M.H.

Sood

A.K.

and Afshar-Kharghan

, Platelets increase the proliferation of ovarian cancer cells, Blood 120(24) (2012), 4869–4872.

28.

Radziwon-Balicka

Medina

O’Driscoll

Treumann

Bazou

Inkielewicz-Stepniak

Radomski

Jow

and Radomski

M.W.

, Platelets increase survival of adenocarcinoma cells challenged with anticancer drugs: mechanisms and implications for chemoresistance, British Journal of Pharmacology 167(4) (2012), 787–804.

29.

Battinelli

E.M.

Markens

B.A.

and Italiano Jr

J.E.

, Release of angiogenesis regulatory proteins from platelet alpha granules: modulation of physiologic and pathologic angiogenesis, Blood 118(5) (2011), 1359–1369.

30.

Sica

Allavena

and Mantovani

, Cancer related inflammation: the macrophage connection, Cancer Letters 267(2) (2008), 204–215.

31.

Szebeni

G.J.

Vizler

Kitajka

and Puskas

L.G.

, Inflammation and cancer: extra- and intracellular determinants of tumor-associated macrophages as tumor promoters, Mediators of Inflammation (2017), 9294018.

32.

Kano

Homma

Hatakeyama

Mizumachi

Sakashita

Kakizaki

and Fukuda

, Pretreatment lymphocyte-to-monocyte ratio as an independent prognostic factor for head and neck cancer, Head & Neck 39(2) (2017), 247–253.

33.

Suzuki

Yamazaki

Aibe

Masui

Sasaki

Shimizu

Kimoto

Asai

Wada

Komori

Katoh

and Yamada

, Clinical usefulness of the platelet-to lymphocyte ratio in patients with angiosarcoma of the face and scalp, International Journal of Molecular Sciences 18(11) (2017), 2402.

34.

Kobayashi

Okuma

Oka

Hirai

Ohki

Ikegami

Sawada

Shinoda

Akiyama

Sato

Abe

Kawano

Goto

and Tanaka

, Neutrophil-to-lymphocyte ratio after pazopanib treatment predicts response in patients with advanced soft-tissue sarcoma, International Journal of Clinical Oncology 23(2) (2018), 368–374.

35.

Chen

Zhai

Yuan

Peng

Chen

and Cai

, Systemic immune-inflammation index for predicting prognosis of colorectal cancer, World Journal of Gastroenterology 23(34) (2017), 6261–6272.

36.

Hong

Cui

Wang

Yang

Wang

and Xu

, Systemic immune-inflammation index, based on platelet counts and neutrophil-lymphocyte ratio, is useful for predicting prognosis in small cell lung cancer, Tohoku Journal of Experimental Medicine 236(4) (2015), 297–304.

37.

Zheng

Zou

Yang

Chen

Guo

and Xiong

, Inflammation-based indexes and clinicopathologic features are strong predictive values of preoperative circulating tumor cell detection in gastric cancer patients, Clinical & Translational Oncology 19 (2017), 1125–1132.

38.

Watt

D.G.

Proctor

M.J.

Park

J.H.

Horgan

P.G.

and McMillan

D.C.

, The neutrophil-platelet score (NPS) predicts survival in primary operable colorectal cancer and a variety of common cancers, PLoS One 10(11) (2015), e0142159.

39.

Aggerholm-Pedersen

Maretty-Kongstad

Keller

Baerentzen

and Safwat

, The prognostic value of serum biomarkers in localized bone sarcoma, Translational Oncology 9(4) (2016), 322–328.

40.

Lou

Sun

Liu

Huang

Wan

and Xiao

, Establishment of a nomogram-based model for predicting the prognostic value of inflammatory biomarkers and preoperative d-dimer level in spinal ewing’s sarcoma family tumors: a retrospective study of 83 patients, World Neurosurgery 121 (2019), e104–e112.

41.

Wibmer

Leithner

Zielonke

Sperl

and Windhager

, Increasing incidence rates of soft tissue sarcomas? A population-based epidemiologic study and literature review, Annals of Oncology 21(5) (2010), 1106–1011.

42.

Canter

R.J.

Beal

Borys

Martinez

S.R.

Bold

R.J.

and Robbins

A.S.

, Interaction of histologic subtype and histologic grade in predicting survival for soft-tissue sarcomas, Journal of the American College of Surgeons 210(2) (2010), 191–198, e2.

Nomogram based on pre-treatment inflammatory biomarkers predicting survival in patients with head and neck soft tissue sarcoma

Abstract

BACKGROUND:

OBJECTIVE:

METHODS:

RESULTS:

CONCLUSIONS:

Keywords

1. Introduction

2. Materials and methods

2.1 Patients selection

2.2 Data collection

2.3 Statistical analysis

3. Results

3.1 Clinical characteristics of patients

Table 1 Clinicopathological characteristics of patients

Table 2 Univariate Cox analysis of prognostic factors affecting survival

5. Conclusions

Footnotes

Conflict of interest

Supplementary Figure

References

Table 1
Clinicopathological characteristics of patients

Table 2
Univariate Cox analysis of prognostic factors affecting survival