The prognostic value of postoperative circulating cell-free DNA in operable hepatocellular carcinoma

Abstract

Background

Tumor biomarkers are eagerly needed in monitoring the recurrence of operable hepatocellular carcinoma (HCC). Circulating cell-free DNA (cfDNA) is a promising noninvasive molecular biomarker for HCC. The current study aimed to evaluate the clinical significance of the postoperative cfDNA in operable HCC.

Methods

This study enrolled 82 HCC patients from January 2018 to June 2019. All patients underwent liver surgery and were pathologically diagnosed with HCC. Postoperative blood samples were collected from each patient. A fluorometric dsDNA assay was used to measure the concentration of cfDNA. We explore the correlation between cfDNA and recurrence. Kaplan–Meier’s curves were used to evaluate the recurrence-free survival (RFS). Univariate and multivariate Cox regression analyses were used for assessing the relative clinical variables in predicting recurrence.

Results

Of the 82 HCC patients, 72 (87%) patients are male and the average age was 52.7 ± 12.8 years. The cfDNA-low and cfDNA-high groups had median recurrence times of 19.5 months and 14.0 months, respectively (p = .023). Multivariate analysis showed that postoperative cfDNA, tumour number and microvascular invasion (p < .050) were independent risk factors for recurrence in operable HCC.

Conclusions

Postoperative cfDNA is still a promising marker to predict prognosis in postoperative HCC patients although prospective and large multicenter clinical study is needed to further validate the relationship between cfDNA and HCC recurrence.

Keywords:

• Hepatocellular carcinoma
• circulating cell-free DNA
• hepatectomy
• recurrence

Introduction

Hepatocellular carcinoma (HCC) is one of the most common malignant tumors worldwide [1]. It accounts for the majority of primary liver cancers [2]. Currently, viral hepatitis and alcohol abuse are the most relevant risk factors for HCC. However, nonalcoholic fatty liver disease (NAFLD) also becomes the non-negligible risk factor for the occurrence of HCC [3,4]. The Barcelona Clinic Liver Cancer (BCLC) algorithm is the most extensively used to classifies HCC patients [5]. Unfortunately, most HCC patients are undetected until a late stage, particularly in China. Previous studies showed that the five-year survival for HCC is less than 20%. HCC is the second most lethal tumor across developing countries [5–7].

There are many approaches to the management of HCC such as surgical therapies, ablation, transarterial chemoembolization (TACE) and systemic therapies [8,9]. Surgical resection, liver transplantation and local ablation are considered as the curative treatment methods for early-stage HCC [10–12]. However, only 30% of HCC patients are able to undergo surgery, as most HCC patients are at a later stage at their first visit to the hospital. Patients with a isolated tumor are the ideal candidates for surgery. For these patients, in 70% of them tumor recurrence occurs at 5 years [10]. Recurrence has become a vital obstacle to prolong the survival time of HCC patients. Moreover, no postoperative adjuvant therapies have been shown to reduce recurrence. Hence, it is urgent and vital to find a biomarker to improve prognostic stratification in HCC patients who have accepted resection.

At present, numerous new diagnostic and prognostic biomarkers like GPC-3 or c-Met have been proposed recently but only a few studies had reported. With the development of next generation sequencing, circulating cell-free DNA (cfDNA) is rapidly maturing to clinical applicability. cfDNA is an extracellular DNA that is commonly thought to be released into the bloodstream by tumor apoptosis and/or dead cells, and is found in many types of cancers such as breast and lung cancer [13–15].

However, almost all studies focused on the diagnostic and predictive function of preoperative cfDNA in HCC, but few studies have evaluated the role of postoperative cfDNA in patients who have accepted hepatectomy [16]. Herein, we perform a retrospective study to measure the postoperative cfDNA and explored its clinical value as a tumor marker in predicting recurrence in HCC patients.

Methods

Patient

From January 2018 to June 2019, 82 HCC patients were enrolled at our hospital. All patients underwent hepatectomy. The principal inclusion criteria were as follows: (1) the definite pathological diagnosis of these patients was HCC; (2) patients underwent radical resection; (3) patients had no extrahepatic metastasis; (4) patients without tumor rupture and bleeding. Patients whose final pathological diagnosis was intrahepatic cholangiocarcinoma or metastatic liver cancer were excluded. Patients were excluded if the data are incomplete. All patients signed informed consent forms. The study was approved by the Ethics Committee of Xiangya Hospital affiliated to Central South University (no. 201703377).

Follow-up and data collection

All patients received regular examinations with analysis of tumour markers, liver function tests and abdominal ultrasound in the outpatient clinic after surgery. Further CT or MRI scans were conducted if necessary. The prognostic endpoint of this study was recurrence-free survival (RFS). Recurrence was defined as intrahepatic recurrence or extrahepatic metastasis based on imaging results, with or without pathological biopsy. Recurrence-free survival was defined as the time interval between resection and the diagnosis of recurrence. Patients with no recurrence were followed up to 30 June 2020.

Moreover, all patients’ preoperative clinical data, including age, gender, preoperative routing blood tests, serum alpha-fetoprotein (AFP), alanine transaminase (ALT), aspartate transaminase (AST) and status of hepatitis B virus (HBV) infection, Child-Turcotte-Pugh (CTP) class, among others, were collected. Besides, postoperative clinicopathological data such as tumour diameter, tumour number, Edmondson grade, macrovascular invasion (VI), microvascular invasion (MVI) and tumor stage were also collected.

Sample processing and DNA extraction

At least 10 mL peripheral blood was collected from each patient using EDTA Vacutainer tubes (BD Diagnostics, Franklin Lakes, NJ) after surgery. There are stabilizing agents to prevent contamination of cfDNA with genomic DNA from white blood cells in the tubes. Then, plasma was purified, and DNA was extracted. Additional detailed information about the experimental procedure was described in a previous article [17].

Statistical analysis

Continuous variables are expressed as the mean (±standard deviation (SD)) and were compared using the Mann–Whitney U test. Categorical variables were compared using the χ² or Fisher's exact test. Univariate and multivariate logistic regression was used to assess the relative clinical variables in predicting recurrence. The results for the predicted rates of recurrence based on contamination of cfDNA were used to plot receiver operator characteristic (ROC) curves, and the area under the curve (AUC) was calculated. The cutoff point corresponding to the maximum of the Youden index was taken as the best critical point for clinical diagnosis. RFS was estimated by the Kaplan–Meier method, and the results were compared with those of the log-rank test. All the tests were bilateral with the significance level of α  =  0.05. Statistical analyses were performed using Prism software (GraphPad Prism Software, La Jolla, CA) and SPSS 21.0 (SPSS Company, Chicago, IL).

Results

Clinicopathological characteristics of patients

A total of 82 HCC patients who underwent hepatectomy were included in this study. The clinicopathological characteristics of the patients are listed in Table 1. The median age of enrolled HCC patients was 52.7 ± 12.8 years. The majority of patients (87.8%) were male. Most of the patients (96.3%) presented the HBV infection and more than three quarters of them (81.7%) had cirrhosis. The mean plasma concentration of AFP was 405.1 ± 459.3 ng/mL. The mean concentration of cfDNA was 5.1 ± 3.7 ng/μL. The mean tumor diameter was 5.7 ± 3.8 cm. Only 16 patients had multiple malignant foci and 19 patients had satellite nodules. Moreover, according to BCLC stage criteria, 59 patients were categorized to stage A while 10 patients to stage B and 13 patients to stage C in this study (Table 1).

Table 1. The clinicopathological characteristics in HCC patients (n = 82).

Variables	Mean ± SD or N (n = 82)
Age (years)	50.6 ± 13.1
Postoperative cfDNA (ng/μL)	5.1 ± 3.7
PLT (×10^9/L)	163.1 ± 63.1
PT (s)	13.9 ± 1.1
AFP (ng/mL)	405.1 ± 459.3
Albumin (g/L)	41.6 ± 4.4
Total bilirubin (μmol/L)	12.9 ± 5.1
ALT (U/L)	46.2 ± 40.3
AST (U/L)	50.3 ± 40.5
Tumor size (cm)	5.7 ± 3.8
Gender
Male	72
Female	10
HBsAg
Positive	79
Negative	3
Edmondson grade
I/II	68
III/IV	14
Tumor number
Single	66
Multiple	16
Satellite nodules
Absence	63
Presence	19
Cirrhosis
Absence	15
Presence	67
Vascular invasion
Absence	69
Presence	13
CTP class
A	80
B	2
BCLC stage
0/A	59
B	10
C	13
TNM stage
I/II	62
III	20
MVI
Absence	40
Presence	42

HCC: hepatocellular carcinoma; PT: prothrombin time; PLT: platelet; ALT: alanine transaminase; AST: glutamic oxaloacetic transaminase; HBV: hepatitis B virus; AFP: alpha-fetoprotein; BCLC: Barcelona Clinic Liver Cancer score; TNM stage: tumor, lymph node and metastasis stage; CTP: Child-Turcotte-Pugh; MVI: microvascular invasion; VI: vascular invasion.

Analyses of risk factors for HCC recurrence

In this section, we enrolled clinicopathologic factors including: age, AFP and other clinicopathologic variables to perform univariate and multivariate analyses to identify independent risk factors of HCC recurrence. Results showed that cfDNA, AFP, tumor diameter, TNM stage, BCLC stage, satellite nodules, tumor number, VI, MVI and cirrhosis were significant in univariate analysis (p < .05). Then, we take these factors into multivariate analysis. It showed that cfDNA (HR, 1.287; 95% CI, 1.035–1.601; p = .023), tumor number (HR, 0.037; 95% CI, 0.004–0.355; p = .004) and MVI (HR, 0.127; 95% CI, 0.054–1.231; p = .005) were the independent factors for recurrence of HCC patients after surgery (Table 2).

Table 2. Univariate and multivariate analyses of independent risk factors of the recurrence in HCC patients (n = 82).

	Univariate analysis		Multivariate analysis
Variable	HR (95% CI)	p Value	HR (95% CI)	p Value
Age (years)	0.979 (0.949–1.011)	.195
PT (s)	1.049 (0.701–1.569)	.818
PLT	1.006 (0.999–1.014)	.091
Postoperative cfDNA (>2.95 vs. ≤2.95 ng/μL)	1.204 (1.044–1.388)	.011	1.287 (1.035–1.601)	.023
Albumin (g/L)	1.011 (0.912–1.121)	.831
ALT (U/L)	0.999 (0.988–1.009)	.787
AST (U/L)	1.000 (0.989–1.011)	.953
AFP (ng/mL)	1.001 (1.000–1.002)	.019	1.001 (1.000–1.003)	.146
Tumour diameter	1.166 (1.032–1.317)	.014	0.771 (0.568–1.047)	.096
HBsAg (negative vs. positive)	0.342 (0.027–3.523)	.342
TNM stage (I/II vs. III)	0.054 (0.014–0.219)	.001	0.321 (0.028–3.699)	.324
BCLC stage (0/A vs. B/C)	0.059 (0.017–0.206)	.001	0.272 (0.031–2.438)	.245
Satellite nodules (absent vs. present)	0.271 (0.093–0.793)	.017	0.704 (0.287–5.031)	.557
Tumour number (single vs. multiple)	0.212 (0.056–0.688)	.011	0.037 (0.004–0.355)	.004
Edmondson grade (I–II vs. III–IV)	0.409 (0.127–1.317)	.134
Vascular invasion (absent vs. present)	0.034 (0.004–0.279)	.002	0.359 (0.011–12.571)	.572
MVI (absent vs. present)	0.144 (0.052–0.401)	.001	0.127 (0.054–1.231)	.005
Cirrhosis (absent vs. present)	4.091 (1.246–13.462)	.021	4.393 (0.517–37.361)	.175

PLT: platelet; ALT: alanine transaminase; AST: glutamic oxaloacetic transaminase; AFP: alpha-fetoprotein; BCLC stage: Barcelona Clinic Liver Cancer stage; CTP: Child-Turcotte-Pugh; cfDNA: circulating cell-free DNA.

cfDNA associated with HCC patient prognosis

In the follow-up period, a total of 32 patients occurred tumor recurrence. Twenty-seven patients occurred intrahepatic recurrence while only five patients occurred extrahepatic metastasis. The median recurrence time of these patients was 12.0 month. In order to investigate the correlation between the concentration of postoperative cfDNA and recurrence, we established the cutoff values of concentration of postoperative cfDNA in HCC patients as 2.95 ng/μL by ROC curve analysis. The ROC area was 0.684, and the 95% CI ranged from 0.566 to 0.802. The sensitivity was 0.875 and specificity was 0.446 (Figure 1(A)).

Figure 1. ROC curve of the cfDNA concentration in HCC patients. (A) The area under the ROC curve was 0.684, and the 95% CI ranged from 0.566 to 0.802. The Youden index is 0.315. RFS in different groups according to the Kaplan–Meier method. (B) Patients in the cfDNA-high group and the cfDNA-low group showed different RFS times. The censored cases was 50 patients.

Then, all patients were divided into two groups according to cutoff value of the concentration of postoperative cfDNA: cfDNA-low group for 54 patients (cfDNA ≤2.95 ng/μL) and cfDNA-high group for 28 patients (cfDNA >2.95 ng/μL). Subsequently, we use the Kaplan–Meier method to analyze RFS in HCC patients. The high-cfDNA group had a significantly poorer RFS compared with the low-cfDNA group (the median recurrence time: 14.0 months versus 19.5 months, p = .0223, Figure 1(B)) based on Kaplan–Meier’s survival analysis.

Moreover, we analyse the relationship between postoperative cfDNA and clinicopathological characteristics. It showed that tumour stage, VI and tumor diameter have correlation with concentration of postoperative cfDNA. The total plasma cfDNA concentration of BCLC stage 0/A and B/C were 4.51 ± 3.13 ng/μL and 6.36 ± 4.64 ng/μL, respectively (p = .017, Figure 1(A)). The total plasma cfDNA concentration of HCC patients with or without vascular invasion was 4.64 ± 3.15 ng/μL and 7.09 ± 5.45 ng/μL, respectively (p = .001, Figure 2(B)). In addition, we found a positive linear correlation between tumor diameter and postoperative cfDNA concentration. The equation of tumor diameter and postoperative cfDNA concentration was Y  =  0.2294×X + 3.713 (p = .028, R  =  0.268, Figure 2(C)).

Figure 2. The relationship between postoperative cfDNA concentration and pathological characteristics in HCC patients. (A) The relationship between BCLC stage and postoperative cfDNA concentration; postoperative cfDNA concentration in the patients with BCLC stages 0/A and B/C was 4.51 ± 3.13 ng/μL and 6.36 ± 4.64 ng/μL, respectively (p=.017). (B) The relationship between VI and postoperative cfDNA concentration; cfDNA concentrations in the absence and presence of VI were 4.64 ± 3.15 ng/μL and 7.09 ± 5.45 ng/μL, respectively (p=.001). (C) Linear correlation between tumor diameter and postoperative cfDNA concentration (the equation was Y = 0.2294×X + 3.713, p=.028, R = 0.28). The 95% CI of slope ranged from 0.024 to 0.434.

Discussion

The use of liquid biopsy and cfDNA profiling has been explored in multiple solid tumors. Herein, we perform a study about the role of postoperative cfDNA in HCC patients who underwent hepatectomy. Results showed that the postoperative cfDNA could independently predict the recurrence. Besides, the concentration of postoperative cfDNA was valuable in estimating prognosis of HCC.

There are a large number of studies on preoperative plasma cfDNA; however, few studies have explored the role of postoperative cfDNA in predicting the prognosis of HCC [18]. cfDNA was originally discovered in 1948. Circulating cfDNA is a short fragment of double-stranded DNA via cell necrosis, apoptosis and the active release of DNA [19]. It can be detected in both healthy individuals and patients with tumors. It has been demonstrated that the total quantity of cfDNA in the tumor patients is higher than the amount in healthy persons [20]. Many studies have demonstrated that minimal residual disease (MRD) plays a vital role in tumour recurrence and metastasis. MRD is considered an important factor in generating and promoting postoperative recurrence [13,21]. Actually, some studies show that cfDNA will drop back to normal when the tumor is removed. Postoperative cfDNA was mainly produced by residual tumor cells. When the concentration of postoperative cfDNA is still high after surgery, it indicated that there is MRD that cannot be detected by imaging examination preoperatively.

The levels of postoperative circulating cfDNA are related to the tumor number and tumor size in patients [22]. It also can reflect in the current study that tumour stage, VI and tumor diameter have relationship with concentration of postoperative cfDNA. At present, AFP is still the routine plasma biomarker for screening and diagnosing HCC recurrence [23,24]. However, the AFP level is elevated in many other diseases, which suggests that it is not specific to HCC. What is more, AFP has limited diagnostic value in HCC because its sensitivity is only 50% [17]. It cannot satisfy the needs of diagnosis in clinical practice. But some studies collectively showed the cfDNA has the better specificity of 70–91% and sensitivity of 50–75% as a diagnostic marker of HCC [25]. Therefore, it is superior to AFP, which is a current diagnostic tool and is often used to confirm the recurrence of HCC.

When it comes to cfDNA, ctDNA has to be mentioned. ctDNA molecules are usually shorter than nonmutant cfDNA molecules [26]. ctDNA only accounts for a small part of the total cfDNA and it is usually diluted by large amounts of noncancerous DNA [27]. With the development of detection technology, especially the technological advances in next-generation sequencing have been used to detect and analyse ctDNA [28]. In principle, ctDNA contains the same genetic defects as the tumour cells from which it originated [29]. However, there is no method to specifically separate ctDNA from other circulating DNA, and only tumor-specific mutations detected on circulating cfDNA indicate the presence of ctDNA [30]. The time to obtain the sequence information for ctDNA is too long to meet the need of rapid diagnosis [31]. In addition, it is costly and provides too much unnecessary and detailed information while the postoperative cfDNA is cost-effective [32]. Postoperative cfDNA is an excellent option for monitoring the recurrence of HCC patients, especially when traditional methods of monitoring cannot provide evidence for diagnosis [33]. This approach can help clinicians identify high-risk groups of HCC and provide effective personalized interventions in a timely manner. Hence, for HCC patients who are with high postoperative cfDNA, we can eliminate cfDNA in the peripheral blood via chemotherapy and molecular therapies to effectively block the formation of recurrence after surgery.

As described above, postoperative cfDNA could predict the prognosis of HCC patients in this study. But larger and independent studies as well as a standard test method, including adequate sample and normalization techniques, are needed. Besides, we did not analyze cfDNA at multiple timepoints. Thus, we could not estimate changes in the concentration of cfDNA over time. Hence, we cannot evaluate time-dependent consecutive analysis of postoperative cfDNA. In some patients at high risk for recurrence, profiling the changes in cfDNA is important for guiding appropriate targeted therapy. Last but not least, there is no in vivo experiments in this study. In the future, we would perform larger and multicenter studies and in vivo experiments. The relationship between changes in cfDNA and postoperative treatment would be assessed in the following research.

Conclusions

In summary, this study demonstrates that postoperative cfDNA is a potential indicator that can be used to predict prognosis in postoperative HCC patients.

Ethical statement

The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The design of this study was approved by the Xiangya Hospital Central South University ethics committee. All clinical data involved in this study were obtained with patient consent.

Author contributions

LeDu Zhou designed the work and wrote the manuscript. Guo Long and Tongdi Fang collected the HCC patients clinical data and completed follow up of the HCC patients after surgery. Wenxin Su and Xingyu Mi analyzed the data. All the authors approved the final version of the manuscript.

Abbreviations
cfDNA	=	circulating cell-free DNA
ctDNA	=	circulating tumor DNA
HCC	=	hepatocellular carcinoma
PT	=	prothrombin time
PLT	=	platelet
ALT	=	alanine transaminase
AST	=	glutamic oxaloacetic transaminase
MVI	=	microvascular invasion
AFP	=	alpha-fetoprotein
BCLC	=	Barcelona Clinic Liver Cancer score
CTP	=	Child-Turcotte-Pugh

Disclosure statement

The author reports no conflicts of interest in this work.

Data availability statement

Research data are not shared.

Additional information

Funding

This study was supported by grants from the National Nature Science Foundation of China (No. 81771932).