Prognostic value of pretreatment platelet count, fibrinogen and d-dimer levels in osteosarcoma patients: A meta-analysis

Abstract

Background:

Previous studies explored the prognostic value of pretreatment platelet count, fibrinogen, and d-dimer level in patients with several types of cancer, however, a comprehensive conclusion has not been reached in osteosarcoma patients.

Methods:

PubMed, Web of Science, Embase, and CNKI databases were systematically searched for eligible studies up to May 09, 2023, and pooled hazard ratios (HRs) with corresponding 95% confidence intervals (CIs) were calculated to assess the prognostic impact of these indicators in osteosarcoma patients.

Results:

Twelve studies from China consisting of 1682 patients were finally included. Our findings revealed that an elevated level of pretreatment platelet or d-dimer was associated with a worse outcome of overall survival (platelet: HR = 1.63, 95% CI: 1.18–2.26, P = .003; d-dimer: HR = 2.29, 95% CI: 1.58–3.31, P < .001).

Conclusion:

Based on current evidence, pretreatment platelet count and d-dimer level could be good prognostic biomarkers for Chinese osteosarcoma patients. However, future validation is also needed.

1. Introduction

Current cancer statistics demonstrate that osteosarcoma accounts for approximately 30% of primary bone malignancies with a low annual incidence of 2 to 3 per million, nonetheless, osteosarcoma should not be ignored due to its highly aggressive, metastatic, and lethal characteristics, which resulted in the poor prognosis of many patients.^[1] In recent years, novel therapeutic strategies including adjuvant chemotherapy, preoperative neoadjuvant chemotherapy, and other targeted therapies are burgeoning rapidly,^[2] and their combination with surgical resection has made great advances and benefited patients tremendously with an increase of 5-year survival rate from 10% to 20% to 60% to 70%.^[3,4] Notwithstanding, the prognosis of osteosarcoma is still poor because of its metastatic features, and a majority of patients may die of tumor metastasis, especially metastasis to lung, even without evidence of metastasis at the initial diagnosis.^[1,5] Therefore, the need for some effective and predictive biomarkers for osteosarcoma patients is more and more emphasized, which may help clinicians identify the proper treatment, forecast the prognosis, and prolong the survival of patients.

Previous studies revealed that venous thromboembolism (VTE) is strongly associated with human cancer and orthopedic operations,^[6,7] and the combination of these two factors may substantially increase the risk of VTE.^[8] Moreover, VTE could lead to a poorer prognosis of tumor patients,^[6,9] while some serum indices associated with clotting and thrombosis, such as platelet-to-lymphocyte ratio (PLR) and d-dimer showed good prognostic value in cancer patients.^[10,11] However, the prognostic impact of VTE-related indices including platelet, fibrinogen, and d-dimer, has not been validated in osteosarcoma patients. Therefore, we conducted a meta-analysis to investigate the association between pretreatment platelet count, fibrinogen, and d-dimer level and survival in osteosarcoma patients, with the hope of reaching a conclusion and helping clinical practice.

2. Methods

2.1. Literature search

The study was conducted according to the Preferred Reporting Items for Systematic Review and Meta-analyses.^[12] PubMed, Web of Science, Embase, and CNKI databases were systematically retrieved for eligible publications from database inception to May 9, 2023. The following terms were used in the search process: “osteosarcoma,” “osteogenic sarcoma,” “platelet,” “fibrinogen,” “d-dimer,” “prognosis,” “prognostic,” and ‘survival’. The search strategy in PubMed was as follows: (osteosarcoma OR osteogenic sarcoma) AND (platelet OR fibrinogen OR d-dimer) AND (prognosis OR prognostic OR survival). Specific literature search process in various databases was presented in File S1, Supplemental Digital Content, http://links.lww.com/MD/M773 and 146, 289, 176, and 100 records were identified from the PubMed, EMBASE, Web of Science, and CNKI databases, respectively. Reference lists of relevant studies were also manually reviewed to include extra studies that might be missed during the electronic search.

2.2. Inclusion criteria

Studies were considered eligible if they met the following criteria: Patients were diagnosed with primary osteosarcoma through pathological evidence; The association of pretreatment platelet count, fibrinogen, or d-dimer level was overall survival (OS), disease-free survival (DFS), or cancer-specific survival (CSS) was investigated; Data of hazard ratios (HRs) with corresponding 95% confidence intervals (CIs) for above outcomes (OS, DFS, or CSS) were available or could be calculated by Kaplan–Meier survival curves; and Studies were published in English or Chinese.

2.3. Exclusion criteria

Relatively, studies were excluded if they met the following criteria: Case reports, reviews, meeting abstracts, comments, or editorials; Insufficient data; Studies with overlapping or even duplicated data; and Studies were published in other language.

2.4. Literature selection and data extraction

The steps of literature selection and data extraction were performed by 2 independent reviewers (L.P. and J.L.) with any disagreement resolved through team consensus. Besides, the following data were collected from included studies: the name of the first author, year of publication, country, indicators with thresholds, tumor stage, treatment, and endpoints with HRs and 95% CIs.

2.5. Quality assessment

Due to the nature of included studies, the Newcastle-Ottawa quality assessment scale (NOS) was applied to evaluate the quality of eligible studies in our meta-analysis, and studies were considered high-quality if they had a NOS score of 6 points or higher.^[13]

2.6. Statistical analyses

HRs with corresponding 95% CIs were merged to investigate the association between pretreatment platelet count, fibrinogen, or d-dimer level, and survival outcomes of osteosarcoma patients. I² statistics and Cochrane Q tests were used to evaluate the heterogeneity among studies, wherein I² > 50% or P value < .1 was regarded as significant heterogeneity, and the random effects model was preferred in this case. Otherwise, we chose the fixed effect model.^[14] Besides, if the data of HRs and corresponding 95% CIs were not provided directly, we applied the method by Tierney to calculate the HRs and 95% CIs from survival curves.^[15] The Begg’s funnel plot and Egger’s test were applied to assess publication bias. All statistical analyses above were conducted by STATA 15.0 software.

3. Results

3.1. Literature selection and baseline characteristics

According to our electronic search in multiple databases, a total of 711 records were retrieved, while 583 independent publications were screened after removing duplicates. Subsequently, only 37 studies were further assessed by full-text reviewing. Finally, 12 studies were included in this meta-analysis,^[16–27] and the flowchart of literature selection was presented in Figure 1.

Figure 1.

The PRISMA flow diagram of this study.

As Table 1 displayed, all 12 studies consisting of 1682 patients were conducted in China, with a sample size ranging from 53 to 454. Seven studies reported the tumor stage in the Enneking stage whereas three reported in the TNM stage. The treatment was dominated by surgery or chemotherapy, while OS was the main endpoint for most studies. The studies by Ren and Hu did not provide HRs and 95% CIs directly^[19,20] thus we estimated them using survival curves. All included studies had a NOS score of 6 or higher, suggesting a low risk of bias.

Table 1.

Basic characteristics of included studies.

Author	Year	Country	Sample size	Indicators and thresholds	Stage	Treatment	Endpoints	Source of HR	NOS
Yang[16]	2009	China	53	Platelet count: 300*109L−1	Enneking stage IIB	Surgery	OS	R	6
Wang[17]	2015	China	454	Platelet count: 300*109L−1	Enneking stage IIA-III	Surgery, chemotherapy	OS	R	6
Huang[18]	2016	China	32	D-dimer: 1.16 mg/mL	Enneking stage III	Chemotherapy	OS	R	7
Ren[19]	2016	China	125	Fibrinogen: 3 g/L	Enneking stage IIA-III	Chemotherapy	OS	E	8
Hu[20]	2017	China	106	Fibrinogen: 3 g/L	NR	Surgery, chemotherapy	OS	E	7
Li[21]	2017	China	233	Platelet count: 300*109L−1	Enneking stage I-III	Surgery, chemotherapy	DFS	R	7
Gou[22]	2019	China	188	Platelet count: 233*109L−1	TNM stage I-III	Surgery, chemotherapy	DFS	R	7
Huang[23]	2021	China	103	D-dimer: 3.91 mg/mL	Enneking stage IIB	Surgery, chemotherapy	OS	R	8
Yang[24]	2021	China	70	Platelet count: 266*109L−1	Enneking stage II-III	Mixed	OS	R	6
Ouyang[25]	2022	China	86	Platelet count: NR	TNM I-IV	Surgery	CSS	R	6
Pu[26]	2022	China	145	Fibrinogen: 3.34 g/L; D-dimer: 0.46 µg/mL	TNM I-IV	Surgery	OS, DFS	R	7
Tian[27]	2022	China	87	Fibrinogen: 3.18 g/L; D-dimer: 0.56 µg/mL	NR	Surgery, chemotherapy	OS	R	7

CSS = cancer-specific survival, DFS = disease-free survival, E = estimated, HR = hazard ratio, NOS = Newcastle-Ottawa Scale, NR = not reported, OS = overall survival, R = reported.

3.2. Association between pretreatment platelet count and survival of osteosarcoma patients

Three studies including 577 patients focused on the relation of pretreatment platelet count to OS, and the pooled result showed that an elevated count of platelets was associated with a poorer outcome of OS (HR = 1.63, 95% CI: 1.18–2.26, P = .003, I² = 0.0%, P = .477) (Fig. 2; Table 2). Besides, two cohorts with 421 patients explored the prognostic impact of pretreatment platelet count on DFS, and the association was not significant but with a strong trend (HR = 1.42, 95% CI: 0.98–2.06, P = .065, I² = 6.0%, P = .302) (Table 2).

Figure 2.

Association between pretreatment platelet count and overall survival of Chinese osteosarcoma patients.

Table 2.

Results of meta-analysis.

	No. of studies	Hazard ratio	95% confidence interval	P value	I  2	P value
Platelet count
Overall survival	3	1.63	1.18–2.26	.003	0.0%	.477
Disease-free survival	2	1.42	0.98–2.06	.065	6.0%	.302
Cancer-specific survival	1	1	1–1.01	.015	–	–
Fibrinogen level
Overall survival	4	1.25	0.84–1.87	.277	30.3%	.230
Disease-free survival	1	1.15	0.67–1.97	.624	–	–
D-dimer level
Overall survival	4	2.29	1.58–3.31	<.001	0.0	.895
Disease-free survival	1	1.85	1.02–3.34	.042	–	–

3.3. Association between pretreatment fibrinogen and survival of osteosarcoma patients

463 patients from 4 studies contributed to the analysis, however, no significant association was observed between pretreatment fibrinogen and OS in patients with osteosarcoma (HR = 1.25, 95% CI: 0.84–1.87, P = .277, I² = 30.3%, P = .230) (Fig. 3; Table 2).

Figure 3.

Association between pretreatment fibrinogen level and overall survival of Chinese osteosarcoma patients.

3.4. Association between pretreatment d-dimer level and survival of osteosarcoma patients

Four studies involving 367 patients were merged to assess the prognostic value of pretreatment d-dimer level in osteosarcoma patients, and the result demonstrated that a higher level of pretreatment d-dimer was associated with a worse OS (HR = 2.29, 95% CI: 1.58–3.31, P < .001, I² = 0.0%, P = .895) (Fig. 4; Table 2).

Figure 4.

Association between pretreatment d-dimer level and overall survival of Chinese osteosarcoma patients.

3.5. Publication bias

We conducted publication bias detection for the association between platelet count and OS (Fig. 5, P = .793) and the association between d-dimer level and OS (Fig. 6, P = .168), which indicated non-significant publication bias.

Figure 5.

Begg’s funnel plot for the association between pretreatment platelet count and overall survival of Chinese osteosarcoma patients.

Figure 6.

Begg’s funnel plot for the association between pretreatment d-dimer level and overall survival of Chinese osteosarcoma patients.

4. Discussion

It is known that VTE constitutes one of the leading causes of cancer deaths, and the tumor itself along with the treatment could also increase the risk of VTE.^[6] Therefore, VTE should not be underestimated in tumor treatment. Previous studies indicated that some hematological indices related to thrombosis, such as PLR and d-dimer, could be prognostic factors in cancer patients, with advantages of being easy to access and noninvasive.^[10,11] Therefore, we attempted to validate the prognostic value of pretreatment platelet count, fibrinogen, and d-dimer level in osteosarcoma patients. To the best of our knowledge, this is the first study that comprehensively investigated the clinical utility of VTE-related indices in osteosarcoma patients. Our findings demonstrated that elevated levels of pretreatment platelet count and d-dimer could be independent prognostic factors for OS in Chinese patients with osteosarcoma, while no significant association was observed between fibrinogen and osteosarcoma prognosis. Therefore, pretreatment platelet count and d-dimer level may serve as promising noninvasive biomarkers for predicting the prognosis of osteosarcoma patients, which would assist the treatment selection in clinical practice.

Two previous studies also focused on the relevant index of PLR, however, a limited association was observed between PLR and osteosarcoma prognosis, suggesting a lack of utility for clinical application.^[28,29] However, in our study, we constructed a stricter search strategy, updated more novel studies, and demonstrated the prognostic significance of pretreatment platelet count and d-dimer level, which lent more comprehensive support to our findings.

Interestingly, previous studies indicated that some platelets termed tumor-educated platelets (TEPs) play a critical role in the tumor microenvironment through either direct interaction with tumor cells or indirect effect mediated by platelet-derived microvesicles (PMVs), which contribute to the regulation of inflammation, immune response, and tumor angiogenesis.^[30–32] Moreover, with the help of fibrinogen, these platelets could protect circulating tumor cells from the killing of immune cells, especially NK cells, leading to the immune escape, and tumor metastasis.^[33,34] These studies provided the potential mechanisms by which platelets could promising candidates for cancer biomarkers or intervention targets. Besides, in clinical application, platelets and d-dimer levels are independent risk factors for cancer-associated thrombosis,^[35,36] thus monitoring these indices could not only decrease the risk of VTE and other lethal complications but also improve osteosarcoma patients’ survival outcomes, while a combination of them may be more effective, which needs to be further validated.

There are several limitations in our study, too. First, the included studies for each association were few, and it was difficult to conduct further analyses, which might bring bias to our results. However, the included studies were all of high quality, and the heterogeneity among them was low, suggesting the reliability of our findings. Second, the clinical stage and treatment varied, and we could not perform subgroup analyses due to a lack of information. Third, all studies were conducted in China, and the application of these findings to other ethnicities should be more careful and rigorous.

5. Conclusions

Taken together, pretreatment platelet count and d-dimer level could be good prognostic biomarkers for Chinese osteosarcoma patients, with advantages of noninvasiveness and cheapness. Nevertheless, the application of these indicators should be more prudent, and future prospective cohorts with big sample sizes are needed to validate the prognostic utility of these indicators or their combination in osteosarcoma patients.

Author contributions

Conceptualization: Xianfeng Li, Honghong Ren.

Data curation: Honghong Ren, Lipeng Peng.

Formal analysis: Lipeng Peng, Jie Li.

Investigation: Jie Li.

Methodology: Lipeng Peng.

Software: Jie Li.

Supervision: Xianfeng Li.

Writing – original draft: Xianfeng Li.

Writing – review & editing: Xianfeng Li, Honghong Ren.