Platelet count as a predictor of vascular invasion and extrahepatic metastasis in hepatocellular carcinoma: A systematic review and meta-analysis

Abstract

Background

Vascular invasion (VI) indicates highly invasive tumor biological behavior and is a major determining factor of poor survival and high risk of metastasis in hepatocellular carcinoma (HCC). Epidemiological evidence of the association between pretherapeutic platelet count (PLT) and the risk of VI and extrahepatic metastasis in HCC remains controversial.

Methods

A systematic retrieval was executed in databases of PubMed, Embase, and Web of Science until Dec 2022. Effect size and 95% confidence interval (CI) were extracted or estimated to synthetically investigate the effects of pretherapeutic PLT on VI and extrahepatic metastasis. Meta-analyses were performed by using a random or a fixed effects model.

Results

Finally, the current meta-analysis included 15 studies with a total of 12,378 HCC patients. It was shown that, patients with a higher pretherapeutic level of PLT had a significantly increased risk of VI (11 studies,8,759 patients; OR = 1.44, 95%CI: 1.02–2.02) and extrahepatic metastasis (6 studies,8, 951 patients; OR = 2.51, 95% CI: 2.19–2.88) in comparison with patients with a lower PLT. Funnel plots and Begg's tests indicated that there were no significant publication biases.

Conclusion

This meta-analysis shows that pretherapeutic elevated PLT is associated with an increased risk of VI and extrahepatic metastasis in HCC.

1. Introduction

Hepatocellular carcinoma (HCC) is the most common primary liver malignancy and one of the most lethal tumors worldwide [1]. In recent years, the prognosis of HCC has been improved with the rapid advancement of various treatment modalities, such as precision partial hepatectomy, ablation therapy, targeted and immunotherapy [2]. However, the total 5-year survival rate of HCC remains below 20% due to the high incidence of recurrence and metastasis after treatment [3].

Vascular invasion (VI) is a common finding of HCC patients. In the course of VI, malignant cells invade vessels and spread to normal liver tissue and distant organs. The presence of VI has been considered as one of the crucial risk factors for recurrence and metastasis of HCC [[4], [5], [6]]. Portal vein tumor thrombus (PVTT) is a common form of VI, the proportion of which is ranged from 13% to 45% [7]. A large number of HCC patients presents macrovascular invasion (MaVI) at diagnosis or posttherapeutic recurrence [8]. In contrast, microvascular invasion (MVI) can only be determined by microscopic examination of the resected tumor [9]. Failure to identify VI before surgery may result in incomplete surgical resection and high probability of recurrence. Although imageological examination is useful for preoperative assessment of VI, it provides insufficient information [10].The preoperative models are potential effective tools in the prediction of VI and prognosis of HCC, but the calculations are relatively complex and the predictive value is limited [11,12]. Therefore, use of simple indicators is expected to predict VI, recurrence and metastasis.

Previous studies have suggested that platelets play multiple roles other than hemostasis and coagulation. Platelets have been reported to be involved in immunity and inflammation [13,14]. In addition, platelets are indispensable in all steps of tumorigenesis including tumor growth, angiogenesis, and metastasis [15,16]. However, in HCC, the roles of platelets are complicated and conflicted. On the one hand, as a marker of liver fibrosis, thrombocytopenia increases the incidence of HCC [17]. We recently showed that, preoperative thrombocytopenia is associated with poor survival and high risk of recurrence in HCC [18,19]. On the other hand, several antiplatelet drugs significantly reduce the incident and recurrent risk of HCC [20]. Moreover, thrombocytosis is found to be associated with large tumor size, multiple tumor, high level of AFP, poor performance status, high risk of VI, and increased lymph node involvement and extrahepatic metastasis in HCC [21,22]. However, epidemiological evidence of the associations of platelet count (PLT) with VI and extrahepatic metastasis remains controversial [23,24]. To our best knowledge, this is the first meta-analysis to explore the roles of PLT in VI and extrahepatic metastasis in HCC.

2. Methods

2.1. Search strategy and selection criteria

Three databases (PubMed, Embase, and Web of Science) were retrieved to seek potential literatures for inclusion. The search date was up to Dec 2022. The search words were: (platelets OR platelet OR PLT) AND (“hepatocellular carcinoma” OR HCC) AND (“vascular invasion” OR “microvascular invasion” OR MVI OR “macrovascular invasion” OR “portal vein tumor thrombosis” OR PVTT OR metastasis).

Two authors screened the eligible publications based on the following selection criteria. Inclusion criteria: (a) cohort study; (b) patients were pathologically or clinically diagnosed as HCC; (c) estimated the roles of pretherapeutic PLT in VI or extrahepatic metastasis; (d) Odds ratio (OR) value and 95% confidence interval (CI) was provided or could be estimated. Exclusion criteria: (a) only estimated the role of PLT in recurrence, intrahepatic metastasis, mortality, or other outcomes; (b) only estimated the role of posttherapeutic PLT; (c) effect size was not reported and could not be estimated; (d) reviews, cases report, editorials, letters and correspondence, or conference abstract. To avoid the inclusion of duplicate literatures with the same population, we only included the study with the maximum sample size or the latest publication. Disagreement was discussed with another author.

2.2. Data abstraction

The extracted information in the eligible studies included: first author, publication year, country of population, design method, study period, cut-off of PLT, etiology of HCC, HCC cases, therapy method, cases with VI or extrahepatic metastasis, demographic data (sex, mean or median age), OR value (adjusted or unadjusted value) and 95% CI. The scale of Newcastle-Ottawa Scale (NOS) was adopted to estimate the study quality of each included publication. Studies with the scores of four to six were considered as moderate quality and seven to nine were high quality. This meta-analysis was reported according to the guideline of preferred reporting items for systematic reviews and meta-analyses (PRISMA).

2.3. Statistical analysis

The main outcomes were VI and extrahepatic metastasis. The pooled results were assessed by using Stata 12.0 software (StataCorp, Texas, USA). Heterogeneity between studies was considered to be low when I² < 50%, moderate when I² was between 50% and 75%, and high when I² > 75%. Random or fixed effects model was adopted depend mainly upon the magnitude of heterogeneity. Galbraith's plot was used to analyze outliers and Begg's funnel plot was used to evaluate publication bias. P < 0.05 was considered to be statistically significant.

3. Results

3.1. Article characteristics

Fig. 1 depicts the flow chart of the literature search and selection process. 709 publications were initially screened out in the three databases after removing duplicate literatures. 34 articles were assessed for eligibility through full-text search and 15 were further eliminated. A total of 16 articles [8,13,[18], [19], [20], [21], [22], [23], [24], [25], [26], [27], [28], [29], [30]] met the selection criteria, and their basic characteristics were presented in Table 1. Lee et al. analyzed duplicate population in two articles [35,36], and the article with relatively small sample size was excluded [36]. Finally, fifteen studies with a total of 12,378 HCC patients were included in this meta-analysis. Agreement between the two authors on which articles to select was satisfactory (kappa = 0.907).

Fig. 1.

Flow chart of search strategy and study selection.

Table 1.

Baseline characteristics for studies included in meta-analysis.

Author year	Area	Design	Study period	Etiology	Therapy	Cut-off of PLT, (ⅹ109/L)	Tumor size (cm)a	Number, n	VI or EHM, n(%)	Endpoints	OR (95% CI)	NOS
Poon 2001 [25]	China	Pro	1998–1999	NR	Hepatectomy	150	>/≤ 5: 58/42	100	42 (42%)	VI	UOR, 1.517(0.676–3.403)	8
Hagiwara 2006 [26]	Japan	Retro	1999–2003	HCV 251 HBV 44	Mixed	130	≥5/<5:211/101	312	25 (8.0%)	PVTT	UOR, 4.035 (1.682–9.679)	7
Lei, 2016 [23]	China	Pro	2004–2011	HBV	Hepatectomy	100	3.27 ± 1.15	707	211 (29.8%)	MVI	UOR, 0.444(0.303–0.654) AOR, 0.541(0.347–0.840)	9
Okamura 2018 [27]	China	Retro	2008–2016	Viral/non-viral 257/166	Mixed	151	MVI+ 5.1 (1.1–17.5) MVI-3.4 (0.6–18)	425	76 (17.9%)	MVI	UOR, 2.37 (1.39–4.04) AOR, 2.42 (1.33–4.41)	8
Scheiner 2019 [28]	Austria	Retro	1995–2013	Alcohol 293 HCV 200	Mixed	150	5.5 (0.6–25.0),	626	86 (14%)	MVI	UOR, 1.677 (1.164–2.416)	6
Wang 2019 [29]	China	Retro	2013–2016	HBV	TACE	222	>/≤10: 101/216	317	98 (30.9%)	PVTT	UOR, 1.049(0.580–1.900)	7
Xu 2019 [22]	China	Retro	2014–2016	HBV/HCV 438 Others 57	Hepatectomy 479 LT 16	100	MVI+8.7 (4.2) MVI-4.9 (3.3)	495	149 (30.1%)	MVI	UOR, 1.744 (1.093–2.782)	8
Liu 2020 [21]	Taiwan	Retro	2002–2017	HBV 40% HCV 22% Others 38%	Mixed	150	4.3 (2.5–8.8)	3561	855 (24%)	VI	UOR, 2.729(2.315–3.216)	7
Liu 2020 [21]	American	Retro	2008–2017	HCV 21% Alcohol 15% Mixed 46%	Mixed	150	4.1 (2.4–8.7)	1145	332 (29%)	VI	UOR, 1.672(1.289–2.169)	7
Wang 2021 [30]	China	Retro	2016–2019	Hepatitis 317 Others 64	Hepatectomy	125	≥5/<5: 198/183, NR	381	198 (52.0%)	MVI	UOR, 1.151 (0.708–1.871)	7
Fan 2022 [31]	China	Retro	2013–2014	Hepatitis 225	Hepatectomy	220	7.6 ± 4.2	279	106 (38%)	MVI	UOR, 1.101(0.534–2.272)	6
Gu 2022 [32]	China	Retro	2018–2020	HBV 41 Others 400	Hepatectomy	125	5.6 (3.5, 7.5)	411	208 (50.6%)	MVI	UOR, 0.673 (0.452–1.002)	7
Bae 2011 [33]	Korea	Pro	2005–2008	HBV 337 HCV 13 Others 4	Mixed	130	≥/< 5: 93/261	354	76 (21.5%)	EHM	UOR 2.319 (1.454–3.698) AOR, 1.911 (1.187–3.076)	9
Morimoto2014 [24]	Japan	Retro	1997–2012	HCV 586 HBV 125 Co-infected 17	TACE 637 HAIC 97	100	2.4 (0.5–20)	803	71 (8.8%)	EHM	UOR, 2.17 (1.31–3.72)g, AOR 1.73 (0.99–3.14)g	8
Xue 2016 [34]	China	Retro	2007–2011	HBV 140 HCV 9	TACE	171	>10	178	44(24.7%)	EHM	UOR, 2.43(1.210–4.865)g AOR, 2.18(1.074–4.443)f	8
Lee 2019 [35]	Taiwan	Retro	2000–2009	Training HBV 57.8% HCV 44.9% Validation HBV 41.7% HCV 61.6%	Mixed	200	>3 54.7%	Training 1387 Validation897	Training 228(16.4%)	EHM	UOR, 3.13(2.19–4.47) AOR Training, 1.97(1.28–3.04) Validation, 2.56(1.72–3.81)	8
Scheiner 2019 [28]	Austria	Retro	1995–2013	Alcohol 293 HCV 200	Mixed	150	5.5 (0.6–25.0), NR	626	86 (14%)	EHM	UOR, 2.399 (1.510–3.811)	7
Liu 2020 [21]	Taiwan	Retro	2002–2017	HBV 40% HCV 22% Others 38%	Mixed	150	4.3 (2.5–8.8)	3561	392 (11%)	EHM	UOR, 2.907 (2.299–3.677)	7
Liu 2020 [21]	American	Retro	2008–2017	HCV 21% Alcohol 15% Mixed 46%	Mixed	150	4.1 (2.4–8.7)	1145	206 (18%)	EHM	UOR, 2.859(2.100–3.892)	7

Abbreviations: NR not reported, HBV hepatitis B virus, HCV hepatitis C virus, VI vascular invasion, EHM extrahepatic metastasis, TACE transcatheter arterial chemoembolization, HAIC hepatic artery infusion chemotherapy, LT liver transplantation, PLT platelet count, PVTT, portal venous invasion, MVI macrovascular invasion, AOR adjusted odds ratio, UOR unadjusted odds ratio, NOS Newcastle-Ottawa Scale, OR odds ratio.

^aMean or median (range or interquartile range).

Among the fifteen publications, thirteen were retrospective studies and the other two [23,25] were prospective design. The recruited population were from China in ten publications, from Japan in two publications, from Austria in one publication, and from America in one publication. HCC patients received hepatectomy in six studies [[21], [22], [23],[30], [31], [32]], interventional therapies in three studies [24,29,34], and mixed therapies in other studies. The cut-off of PLT ranged from 100ⅹ10⁹/L to 222ⅹ10⁹/L. Eleven studies reported the cases of VI, the incidence rate of which was 27.2% (2386/8759) with the range from 8% to 52%. The incidence rate of metastasis ranged from 8.8% to 24.7%, which were reported in six studies [21,24,28,[33], [34], [35]] and the total proportion was 13.7% (1103/8054). Six publications [23,24,27,[33], [34], [35]] provided adjusted OR values after controlling for confounding variables. NOS scale revealed that the study quality score ranged from 6 to 9 with a median of 8, and 88.2% of the studies were high quality (Table 1).

3.2. Effect of PLT on VI

Eleven studies with a total of 8759 HCC patients reported the effect of PLT on VI. According to the types of VI, MVI were recorded in seven studies [22,23,27,28,[30], [31], [32]], PVTT in two studies [26,29], and any VI in two studies [21,25]. Meta-analysis demonstrated that, there was a significantly elevated risk of VI in patients with a high level of PLT (11 studies, 8759 patients; OR = 1.44, 95%CI: 1.02–2.02) (Fig. 2). However, the heterogeneity was relatively high between the included studies (I² = 87.7%, P < 0.001). Furthermore, Galbraith's plot identified three outliers, which may be the possible sources of heterogeneity (Supplementary Fig. 1). After the removal of the outliers, the heterogeneity between studies was low (Supplementary Fig. 2; I² = 28.4%, P = 0.192).

Fig. 2.

Forest plots of meta-analysis on association between platelet count and vascular invasion in patients with hepatocellular carcinoma.

As high inter-study heterogeneity was displayed, subgrouped and meta-regression analyses were further performed to explore the potential sources of heterogeneity (Supplementary Table 1). Subgrouped analysis showed that, among the eight potential factors, cut-off value of PLT, proportion of patients with VI, type of VI, therapeutic method, and type of survival analysis were found to be statistically significant (P < 0.05). In addition, therapeutic method was shown to be the potential source of heterogeneity following meta-regression. After the introduction of the regression model with the factor of therapeutic method, the estimate of inter-study variance decreased from 0.2910- to 0.1275. Thus, this factor explains 56.2% of the heterogeneity.

3.3. Effect of PLT on extrahepatic metastasis

In meta-analysis of the effect of PLT on extrahepatic metastasis, a total of 8951 cases of HCC patients from six studies [21,24,28,[33], [34], [35]] were included. The pooled analysis showed that, patients with a high pretreatment PLT had a significantly high risk of extrahepatic metastasis (6 studies, 8951 patients; OR = 2.51, 95% CI: 2.19–2.88) (Fig. 3). There was low heterogeneity between these studies (I² = 0%, P = 0.489).

Fig. 3.

Forest plot of meta-analysis on association between platelet count and extrahepatic metastasis in patients with hepatocellular carcinoma.

3.4. Sensitivity analysis and publication bias

To further verify the viability of the above results, we performed sensitivity analysis. In the meta-analyses of VI and extrahepatic metastasis, the usages of random and fixed effects models indicated similar results (Supplementary Table 2). Influence analysis showed that, none of the studies could significantly influence the pooled results (Supplementary Fig. 3).

Publication bias was estimated by using funnel plots, the visual check of which showed basic symmetry (Fig. 4A and B). Begg's tests further verified that there were no significant publication biases (P > 0.05).

Fig. 4.

Begg's funnel plots of studies that assessed the effect of platelet count in vascular invasion (A) and extrahepatic metastasis (B) in patients with hepatocellular carcinoma.

4. Discussion

Despite tremendous advances in the treatment of HCC, tumor recurrence and metastasis remain a huge challenge. VI represents a highly invasive form of HCC and facilitates to posttherapeutic recurrence and metastasis. In this meta-analysis, the proportion of VI was up to 27.2% in HCC patients and high level of PLT was found to be associated with VI in HCC. We also firstly showed that, HCC patients with high level of PLT had a 1.51-fold increased risk for extrahepatic metastasis.

Besides serving as the important component of blood system, platelets play a critical function in tumor biological behavior. Thrombocytopenia is a hallmark of fibrosis and is associated with an increased risk of morbidity and mortality in cirrhotic patients [35]. Thrombocytopenia is also thought to be a risk factor for HCC incidence [28]. Recently, the role of platelets in tumor biology attracted extensive attention [15,[36], [37], [38]]. Several studies showed that thrombocytosis, not thrombocytopenia, is associated with an elevated risk of distant metastasis in various kinds of tumors. Meta-analysis with ten studies containing 2995 cases of endometrial cancer suggested that pretreatment thrombocytosis was significantly associated with high tumor stage, lymph-vascular space invasion, and lymph node metastasis [37]. Thrombocytosis was also a predictor of metastasis in ovarian tumor, urothelial carcinoma, colorectal cancer, and pulmonary adenocarcinoma [[38], [39], [40]]. Consistently, we meta-analysis showed that thrombocytosis contributes to extrahepatic metastasis in HCC.

The exact mechanisms behind the link between high level of PLT and VI as well as extrahepatic metastasis in HCC are not yet fully elucidated. Platelet lysates from HCC patients after RFA promoted the proliferation, migration, invasion and vasculogenic mimicry of HCC cells [41]. Platelets may promote tumor metastasis via platelet-tumor cell interactions. The interactions create favorable conditions for tumor cell cross-endothelial migration [42]. Tumor cells activates platelets, which can enhance the formation of platelet-tumor cell aggregates, facilitating subsequent distant metastasis [43]. Activated platelets release extracellular vesicles and particles containing exosomes, microparticles and some pro-angiogenic contents, such as serotonin [44], PDGF [45], VEGF [46] and GPIbα [47], which promote HCC proliferation, metastasis, and angiogenesis. Via NK cells and T cells, platelets assist tumor cell immune escape, which is the key to achieve tumor cell metastasis [42].

There are several limitations in this meta-analysis. First, all the included publications were designed as retrospective studies, and study qualities were relatively low. Second, the therapy of HCC was various and the cut-off of PLT was inconsistent, which could be confounding factors. Third, subgrouped analysis in term of the subtype of VI was not performed due to the limited included. Lastly, as a lack of consensus for the definition of MVI, there may be interobserver and intraobserver variability in the evaluation of MVI in HCC.

5. Conclusion

Based on these findings, pretherapeutic high level of PLT is associated with an increased risk of VI and extrahepatic metastasis in HCC. As PLT is readily available in all HCC patients, the presence of thrombocytosis can serve as a marker identifying patients who may have poor outcome. Our study also suggests a potential role for antiplatelet therapy, and future research should focus on whether the antiplatelet therapy could contribute to better outcomes.

Funding statement

This study was supported by the Science and Technological Fund of Anhui Province for Outstanding Youth [Grant No. 2008085J37], the Key project of Natural Science Research in Universities of Anhui Province [Grant No. 2023AH053369, 2022AH052337 and KJ2021A1267], the First Affiliated Hospital of Bengbu Medical College Science Fund for Distinguished Young Scholars [Grant No. 2019byyfyjq05], and the 512 Talent Training Program of Bengbu Medical College [Grant No. by51201318]

Data availability statement

Data will be made available on request.

CRediT authorship contribution statement

Qing Pang: Writing – original draft, Investigation, Funding acquisition. Xuankun Gong: Writing – original draft, Methodology, Data curation. Hongtao Pan: Validation, Funding acquisition, Data curation. Yong Wang: Software, Methodology, Investigation. Xiaosi Hu: Validation, Investigation. Huichun Liu: Writing – review & editing, Project administration, Conceptualization. Hao Jin: Writing – review & editing, Project administration, Funding acquisition, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article: