Preoperative color Doppler ultrasonography predicts early recurrence in AFP-positive hepatocellular carcinoma.

The present study aimed to evaluate the hepatic hemodynamics of patients with α-fetoprotein (AFP)-positive hepatocellular carcinoma (HCC) and early recurrence (ER), as determined by color Doppler ultrasound. Screening with color Doppler ultrasound was performed in 150 patients with AFP-positive HCC, yielding 43 cases with postoperative ER. In addition, 35 healthy controls were enrolled in the study for comparison of the blood flow parameters between the healthy control and the HCC groups. Receiver operating characteristic curve analysis was performed to evaluate the predicted effect of Doppler for the ER of HCC. The average mean flow velocity of portal vein in patients with HCC (14.686±5.873 cm/s) was significantly lower in the HCC group compared with the healthy control group (17.631±3.569 cm/s; P=0.005). Additionally, the preoperative portal vein diameter in the HCC group was significantly higher compared with that in the healthy control group (P=0.001). Regarding the effect of surgery, the portal vein diameter was significantly increased in the patients postoperatively compared with preoperatively (P=0.003), while the pulsation and resistance indexes exhibited a decreasing trend in the postoperative group (P=0.001 and P=0.003, respectively). Notably, patients with HCC and ER presented with a higher resistance index (1.163±0.342) compared with the ER-free group (1.023±0.176; P=0.004). Furthermore, a significant difference in hepatic artery pulsation index was also observed between the ER group (0.673±0.075) and the ER-free group (0.624±0.056; P=0.018), indicating that an increased hepatic artery pulsation index may significantly predict the ER of HCC. In addition, areas under the curve of 0.683 and 0.700, respectively, suggested that the hepatic artery resistance and pulsation indexes may be used to diagnose ER in HCC (P=0.009 and P=0.004, respectively). Taken together, Doppler ultrasound provides a reliable and accurate quantification of hepatic hemodynamics for detecting ER in HCC with a good diagnostic accuracy. Copyright: © Chen et al.

Introduction

Hepatocellular carcinoma (HCC) is the sixth most common malignant tumor worldwide and one of the most common causes of cancer-associated mortality worldwide (1,2). Ultrasonography, computed tomography examination and serum α-fetoprotein (AFP) levels are well-characterized approaches for the diagnosis of early HCC (3,4); however, prognosis of patients with HCC remains poor (5). Surgical resection is the preferred treatment for HCC (6). However, postoperative recurrence is the main cause of early mortality in patients with HCC following radical resection (7,8). Poon et al (9) reported that the prognosis of patients with HCC and early recurrence (ER; <1 year after surgical resection) is worse than that of patients with late recurrence (LR; >1 year), indicating that the shorter the interval time prior to recurrence, the worse the prognosis. Portolani et al (10) reported that the disease-free survival rate within 1 year following surgery is only 68.0%. As a result, the prediction of postoperative ER is particularly important, and therefore it was the main aim of the present study.

Ultrasound is considered the first choice for screening HCC due to its noninvasive, convenient and inexpensive imaging characteristics (11). Ultrasound can evaluate the size, number, position and boundary of tumor nodules, as well as monitoring hepatic hemodynamic changes. Cirrhosis is a serious liver disease, and advanced cirrhosis can easily develop into HCC (12). Doppler ultrasound has been used to detect hemodynamic changes that occur during the progression and occurrence of hepatic fibrogenesis associated with cirrhosis (13). Previous research has demonstrated that Doppler ultrasound may be reliably used to assess hemodynamic changes in different stages of liver fibrosis (14). Furthermore, a study conducted by Suk et al (15) demonstrated that changes in hepatic hemodynamics may serve as a predictor for the clinical stages of HCC. Yang et al (16) proposed that the occurrence of HCC is associated with abnormal liver hemodynamics. Combined measurement of AFP levels and Doppler ultrasound has a greater diagnostic accuracy for HCC compared with ultrasonography alone. It has also been recognized that the diagnostic role of AFP together with abdominal ultrasound may effectively predict the risk of HCC development (17). However, to the best of our knowledge, the potential diagnostic value of preoperative Doppler ultrasound for ER of AFP-positive HCC has not been fully investigated thus far.

In the present study, the predicted value of hepatic hemodynamics for the ER of AFP-positive HCC was retrospectively assessed using Doppler ultrasound. Doppler ultrasound parameters were specifically evaluated in AFP-positive HCC and control groups. Furthermore, the present study investigated whether the hepatic artery and portal vein blood flow parameters were associated with postoperative ER of AFP-positive HCC.

Materials and methods

Clinical data

The present study was performed with approval from the Ethics Committee of the Guangxi Medical University Affiliated Tumor Hospital and written informed consent was obtained from all participants. All the 150 patients with HCC received surgical treatment and were diagnosed by pathologists for the first time in the Guangxi Medical University Affiliated Tumor Hospital between September 2015 and December 2016. In total, the 137 AFP-positive patients with HCC (threshold, 400 ng/ml) were followed-up for 1 year following surgical treatment. The patients included 115 males and 22 females with a mean age of 47.82±11.43 years (range, 20–71 years). A total of 13 patients with AFP-positive HCC were lost to follow-up. Subsequently, the patients were divided into different groups according to their ER status, an ER group and an ER free group. The 35 healthy controls were enrolled from the outpatients department of the Guangxi Medical University Affiliated Tumor Hospital between September 2015 and December 2016 and consented to an ultrasound examination of their liver. The healthy controls included 20 males and 15 females, with a mean age of 53.26±8.75 years (range, 35–67 years).

Ultrasound examination

A standard ultrasonographic study was performed utilizing standard equipment (GE Logiq9 color Doppler; GE Healthcare Life Sciences) and a transducer ultrasound probe (2.5 or 3.5 MHz with low acoustic power). The enrolled subjects did not use drugs that affect hemodynamics (including nonselective β-blockers and vasopressin analogues) 5 days prior to the examination, Subjects were fasted for 6–8 h and allowed to rest for 30 min prior to the examination. Two-dimensional ultrasound images were used to evaluate the size, location, boundary, internal echo and blood flow of the liver-occupying lesions. The blood flow parameters of the hepatic artery and portal vein were measured by color Doppler ultrasound. All of the subjects were laid flat and the ultrasonic probe was placed intercostally. The hepatic blood flow of the portal and hepatic arteries was detected between the neck of the gallbladder and the inferior vena cava. The following Doppler ultrasound parameters were measured: i) Portal vein blood flow of the mean flow velocity; ii) portal venous blood flow volume; iii) portal vein diameter; iv) hepatic artery pulsation index (PI); and v) hepatic artery resistance index (RI). To ensure reliability, all parameters were measured ≥3 times simultaneously and the mean value was calculated. When patients with HCC were followed up for 3 months following the operation, results were similar to the pre-operative exam. The tests were performed by two experienced ultrasound physicians, and conflicting opinions were discussed and resolved by a third physician.

Statistical analysis

All statistical analyses were performed using SPSS software (version 22.0; IBM Corp.). The results are presented as the mean ± standard deviation. An independent Student's t-test was used for comparisons between two groups. A paired Student's t-test was utilized to compare hepatic hemodynamics changes in the preoperative and postoperative stages in the same patient. The Chi-squared test was used to analyze the qualitative data including sex, tumor number, tumor thrombi, liver cirrhosis and hepatitis. Receiver operating characteristic (ROC) curves were generated to evaluate the diagnostic accuracy of Doppler ultrasonography for ER in patients with HCC. P<0.05 was considered to indicate a statistically significant difference.

Results

Baseline characteristics

A total of 150 patients (including 128 males and 22 females) with AFP-positive HCC and 35 healthy controls were recruited at the Affiliated Tumor Hospital of Guangxi Medical University (Nanning, China) between September 2015 and December 2016 (Table I). In total, 13 patients were loss to follow-up and 137 patients with AFP-positive HCC (mean age, 48.360±11.482 years) completed the follow-up. Among these patients, 41 cases were diagnosed with postoperative ER (ER group) and 96 cases were free of postoperative ER (ER-free group). The patients were classified into a liver cirrhosis group (n=61) and a non-liver cirrhosis group (n=76), according to their cirrhosis status. In the healthy control group, 20 males and 15 females were included. The characteristics of the patients are summarized in Table II.

Table I.

Baseline pre-operative clinical characteristics of the enrolled patients.

Variables	n
Sex
Male	128
Female	22
Age, years
Mean ± SD	48.36±11.48
Tumor number
Multiple	47
Single	102
Tumor thrombi
Yes	86
No	64
Tumor diameter (cm)
Mean ± SD	7.58±3.92
Liver cirrhosis
Yes	68
No	82
Hepatitis
Yes	143
No	7
Early recurrence
Yes	41
No	96

SD, standard deviation.

Table II.

Clinical characteristics of ER and ER-free patients.

Variables	ER group, n	ER-free group, n	P-value
Sex
Male	37	78	0.216
Female	4	18
Age, years
Mean ± SD	46.34±11.410	48.45±11.439	0.325
Tumor number
Multiple	15	28	0.419
Single	25	68
Tumor thrombi
Yes	28	51	0.131
No	13	45
Tumor diameter (cm)
Mean ± SD	9.25±3.621	6.82±3.628	<0.05[a]
Liver cirrhosis
Yes	22	39	0.190
No	19	57
Hepatitis
Yes	41	90	0.178
No	0	6

P<0.05 was considered to indicate a statistically significant difference. ER, early recurrence; SD, standard deviation.

Change in hepatic hemodynamics between preoperative patients with HCC and healthy controls

The change in hepatic hemodynamics between patients with HCC and the control group was evaluated (Fig. 1). The mean flow velocity (14.686±5.873 cm/sec) in patients with HCC was significantly lower compared with that observed in healthy subjects (17.631±3.569 cm/sec; P=0.005; Fig. 2A). Additionally, the results revealed that patients with HCC presented with significantly higher portal vein diameters (1.085±0.149 cm) compared with the healthy control group (0.991±0.155 cm; P=0.001; Fig. 2B). There was no statistical significance difference in the other parameters, including portal venous flow velocity, portal venous volume, hepatic artery RI and PI, between the preoperative HCC group and the healthy control group (Table III).

Figure 1.

Results from a 50-year-old healthy control. (A) Two-dimensional ultrasound image. (B) Color Doppler blood flow diagram of the portal vein (D, 1.09 cm; V, 14.7 cm/sec; Q, 532.1 ml/min). (C) Color Doppler blood flow diagram of the hepatic artery (white arrowheads; PI, 1.06; RI, 0.63). PS, peak systolic velocity; ED, end-diastolic velocity; VF Diam, portal vein diameter; MD, average flow velocity; VolFlow, portal venous blood flow volume; TAMEAN, time averaged velocity; TAMAX, mean time maximum velocity.

Figure 2.

Comparison of the hepatic hemodynamics between the hepatocellular carcinoma group and the healthy control group. (A) Mean flow velocity. (B) Portal vein diameter.

Table III.

Blood flow parameters of the portal vein and hepatic artery in health controls and patients with HCC.

	Healthy control group			HCC group
Variable	n	Mean	SD	n	Mean	SD	P-value
Hepatic artery
RI	35	0.663	0.048	86	0.654	0.077	0.524
PI	35	1.185	0.217	86	1.1116	0.306	0.442
Portal vein
V, cm/s	35	17.631	3.569	146	14.686	5.873	0.005[a]
Q, ml/min	35	466.349	165.623	150	535.635	229.640	0.094
D, cm	35	0.991	0.155	150	1.085	0.149	0.001[a]

P<0.05 was considered to indicate a statistically significant difference. HCC, hepatocellular carcinoma; SD, standard deviation; RI, resistance index; PI, pulsation index; V, average flow velocity; Q, portal venous blood flow volume; D, portal vein diameter.

Effect of surgery on the hemodynamics of patients with HCC

The hemodynamic changes prior to and following the surgical procedure in patients with HCC were compared using paired Student's t-test. The portal vein diameter significantly increased upon operation (pre-operative, 1.07±0.14 cm vs. post-operative, 1.14±0.19 cm; P=0.003; Fig. 3A). The effect of surgery on the hepatic artery was evaluated. The results revealed that the PI presented the same trend in the post-operative group (pre-operative group, 1.095±36.331 cm; post-operative group, 1.330±0.260 cm; P=0.001; Fig. 3B). Additionally, the RI in the post-operative group (0.690±0.060) was significantly higher compared with that observed in the pre-operative group (0.650±0.080; P=0.003; Table IV; Fig. 3C).

Figure 3.

Changes in the hepatic hemodynamics in patients with hepatocellular carcinoma pre- and post-operatively. (A) Portal vein diameter. (B) Pulsation index. (C) Resistance index.

Table IV.

Changes of the blood flow parameters of the portal vein pre- and post-surgery in patients with hepatocellular carcinoma.

	Pre-operative			Post-operative
Variable	n	Mean	SD	n	Mean	SD	P-value
Portal vein
V, cm/s	107	14.680	6.540	107	13.988	3.748862749	0.287
Q, ml/min	107	526.190	245.810	107	531.600	230.780	0.868
D, cm	107	1.070	0.140	107	1.140	0.190	0.003[a]
Hepatic artery
PI	36	1.100	0.330	36	1.330	0.260	0.001[a]
RI	36	0.650	0.080	36	0.690	0.060	0.003[a]

P<0.05 was considered to indicate a statistically significant difference. SD, standard deviation; RI, resistance index; PI, pulsation index; V, average flow velocity; Q, portal venous blood flow volume; D, portal vein diameter.

Comparison of hemodynamic changes between patients with HCC and postoperative ER and ER-free patients with HCC

The hemodynamic changes between the ER group (Fig. 4) and the ER-free group (Fig. 5) were compared. Notably, the postoperative ER group exhibited a significantly increased hepatic artery PI (0.673±0.075 vs. 0.624±0.056; P=0.018; Fig. 6A; Table V) compared with the ER-free group. Furthermore, a statistically significant difference in RI was observed between the ER and ER-free groups, indicating that the RI in the ER group (1.163±0.342) was significantly lower than that of the ER-free group (1.023±0.176; P=0.004; Fig. 6B, Table V). For the other parameters measured, including average flow velocity, portal venous blood flow volume and portal vein diameter, no statistically significant differences were observed between the ER group and the ER-free group (Table V).

Figure 4.

Results from a 45-year-old patient with hepatocellular carcinoma, positive α-fetoprotein, liver cirrhosis, Child-Pugh A, tumor pathological stage of intermediate differentiation, Barcelona clinic liver cancer stage B and postoperative early recurrence. (A) Conventional ultrasound image revealing a 14 cm-diameter lesion tumor (white arrowhead). (B) Color Doppler blood flow diagram of the portal vein (white arrowheads; D, 1.17 cm; V, 12.0 cm/sec; Q, 516.5 ml/min). (C) Color Doppler blood flow diagram of the hepatic artery (white arrowheads; PI, 1.18; RI, 0.68). PS, peak systolic velocity; ED, end-diastolic velocity; MD, average flow velocity; VolFlow, portal venous blood flow volume; VF Diam, portal vein diameter; PI, pulsation index; RI, resistance index; TAMEAN, time averaged velocity; TAMAX, mean time maximum velocity.

Figure 5.

Results from a 65-year-old patient with hepatocellular carcinoma, positive α-fetoprotein, no liver cirrhosis, Child-Pugh A, tumor pathological stage of intermediate differentiation, tumor emboli, Barcelona clinic liver cancer stage C and no early recurrence following surgery. (A) Conventional ultrasound image revealing no early recurrence following surgery. (B) Color Doppler blood flow diagram of the portal vein (white arrowheads; D, 1.24 cm; V, 19.6 cm/sec; Q, 788.6 ml/min). (C) Color Doppler blood flow diagram of the hepatic artery (white arrowheads; PI, 1.30; RI, 0.71). PS, peak Systolic velocity; ED, end-diastolic velocity; MD, average flow velocity; Vloflow, portal venous blood flow volume; VF Diam, portal vein diameter; PI, pulsation index; RI, resistance index, TAMEAN, time averaged velocity; TAMAX, mean time maximum velocity.

Figure 6.

Differences in hepatic hemodynamics between the ER group and the ER-free group. (A) Pulsation index. (B) Resistance index. ER, early recurrence.

Table V.

Blood flow parameters of the portal vein and hepatic artery in the ER-free group and ER group of patients with hepatocellular carcinoma.

	ER-free group			ER group
Variables	n	Mean	SD	n	Mean	SD	P-value
Hepatic artery
RI	25	1.023	0.176	55	1.163	0.342	0.004[a]
PI	25	0.624	0.056	55	0.673	0.075	0.018[a]
Portal vein
V, cm/s	96	14.725	6.827	37	14.284	3.261	0.687
Q, ml/min	96	520.685	244.144	41	538.344	210.213	0.708
D, cm	96	1.067	0.15	41	1.1	0.14	0.233

P<0.05 was considered to indicate a statistically significant difference. ER, early recurrence; SD, standard deviation; RI, resistance index; PI, pulsation index; V, average flow velocity; Q, portal venous blood flow volume; D, portal vein diameter.

Diagnostic accuracy of Doppler ultrasonography for the diagnosis of ER in HCC

ROC curve analysis was performed to assess the diagnostic accuracy of Doppler ultrasonography for the diagnosis of ER in patients with HCC. The area under the ROC curve (AUC) of preoperative RI value was 0.683 (range, 0.564–0.801; P=0.009). An AUC value for RI of 0.645 corresponded with the highest diagnostic value for ER with a sensitivity of 68% and a specificity of 70.91% (Fig. 7A). As for the preoperative PI, an AUC of 0.700 suggested a high diagnostic accuracy of preoperative PI for the ER of patients with HCC (range, 0.583–0.818; P=0.004). The best cut-off value of pre-operative PI was 1.045, with a sensitivity of 72.00% and a specificity of 67.27% (Table VI; Fig 7B).

Figure 7.

Predicted value of hepatic artery parameters for the early recurrence of hepatocellular carcinoma in patients. (A) Resistance index and (B) pulsation index. AUC, area under the curve.

Table VI.

Receiver operating characteristic curve analysis of PI and RI for the diagnosis of ER in patients with hepatocellular carcinoma.

Variables	Area under the curve (range)	P-value	Cut-off	Sensitivity, %	Specificity, %
RI	0.683 (0.564–0.801)	0.009	<0.6450	68.00	70.91
PI	0.700 (0.583–0.818)	0.004	<1.0450	72.00	67.27

RI, resistance index; PI, pulsation index; ER, early recurrence.

Discussion

Surgical resection is the first choice for the treatment of HCC. However, recurrence subsequent to resection is one of the main factors affecting the prognosis of HCC (18). It has become evident that post-operative recurrence of HCC can be classified into ER and LR (8,9). In particular, ER presents challenges for the treatment of liver cancer. Colecchia et al (19) demonstrated that ER may originate from the liver metastasis of primary tumors, which is mainly associated with biological factors of the tumor, including tumor size, number of lesions and vascular invasion. LR is caused by the residual liver state, and is affected by the history of HCC prior to treatment and the presence of hepatitis and cirrhosis (20). A reliable and effective prediction method to assess the risk of HCC recurrence may shed light on future treatment strategies.

The examination of blood vessels morphology and hemodynamics may directly or indirectly reflect the inherent biological characteristics of tumors (21). As a noninvasive imaging modality, ultrasonography is widely used in routine clinical practice for diagnosis, treatment planning or monitoring of cancer, particularly for detecting changes in hemodynamics of the portal vein and hepatic artery (22–25). Therefore, it was speculated that the hemodynamics of ultrasound detection may be a valuable predictor for the ER of HCC following resection. It has been demonstrated that the combination of ultrasonography and measurement of AFP levels confers the efficient surveillance strategy for early HCC detection (26). Previous studies have demonstrated that the factors affecting the recurrence of HCC include incomplete or absent tumor capsule, preoperative AFP levels >400 ng/l, lesion size and positive pathological margin (25,27–30). Clinically, the size and capsule of HCC tumors may be assessed by various imaging methods prior to surgery, while the AFP levels can be detected through laboratory examination. However, pathological grading of tumors and microvascular invasion cannot be obtained prior to pathological examination. Therefore, the prediction of ER of HCC requires future investigation to improve patient prognosis (3). The present study focused on patients with AFP-positive HCC to investigate the diagnostic value of ultrasonography for the recurrence of HCC (31).

Comparison of the hemodynamics between patients with HCC and healthy controls revealed that the pre-operative portal venous blood flow volume and portal vein diameter in patients with HCC were significantly increased compared with the control group, while the preoperative hepatic artery PI in patients with HCC was significantly decreased compared with the control group. The present study demonstrated that monitoring of the hepatic hemodynamics is important for the early diagnosis of HCC recurrence. To the best of our knowledge, no studies on this topic have been conducted thus far. The results obtained in the present study revealed that pre-operative hepatic artery RI and PI in the ER group were significantly lower compared with in the ER-free group. In addition, the ROC curve revealed that pre-operative PI <1.05 and RI <0.645 may be used as reference values for predicting the occurrence of ER in postoperative HCC. Bonnin et al (32) revealed that the hepatic artery Doppler perfusion index was established earlier during the diagnosis of liver cancer when compared with conventional ultrasonography (37.6±12.7 days), and that increased blood flow in the hepatic artery may be used to monitor the growth of liver tumors in mice, leading to a diagnosis of liver cancer 4–20 weeks earlier.

Blood to the liver is supplied by the hepatic artery and portal vein. Advanced liver cirrhosis is associated with portal hypertension and portal vein widening, with the majority of cases progressing into HCC. The histopathology of HCC revealed the presence of neovasculature and blood pooling in the tumor. Compared with the same diameter hepatic artery, the lower number of elastic fibers in patients with HCC contributes to a lower hepatic artery blood flow resistance and faster flow rate. In conclusion, the pathological changes of HCC are accompanied by changes in the hemodynamics of the hepatic artery and portal vein (33). Therefore, the early dynamic changes of the hepatic artery and portal vein may be used to predict recurrence in HCC. The application of color Doppler ultrasound for hepatic artery and portal venous hemodynamics is particularly useful for the early diagnosis of HCC recurrence.

HCC tumor blood is mainly supplied by the hepatic artery, while the portal vein is involved in the peripheral part of the liver and extends to the center with a small branch (34). When the tumor size is >5 cm, the hepatic artery and portal vein are intertwined to form a vascular lake around the tumor, whereas the tumor blood supply is partially or completely derived from the portal vein when the tumor size is <5 cm (35–37). The portal venous system has multiple collateral circulations and may exhibit a small blood flow in HCC with ER (37,38). Therefore, the hepatic artery affects HCC tumor recurrence more than the portal vein, resulting in no significant changes when the latter is monitored by current ultrasound instrumentation. Hepatic artery PI and RI are the main parameters to reflect the blood flow of the hepatic artery. The hepatic artery hemodynamic changes in liver cancer were previously evaluated by RI and PI in the literature. The PI and RI measurements of the hepatic artery are highly reproducible, easily measured and not affected by the changes in respiration or body position (39). In addition, the ultrasonic detection of PI and RI can be fully quantified. Therefore, the PI and RI of the hepatic artery may be predictors of ER following surgery in HCC.

The present study had several limitations. Firstly, The PI value is affected by heart rate, hemangiectasis ability and the total resistance of peripheral blood vessels (40). Secondly, the limits of the instrument and the operator may lead to discrepancy in the observations. Finally, the retrospective nature and small sample size of the present study may lead to statistical type-II errors. Overall, the sensitivity and specificity of ultrasonic monitoring require improvement. In conclusion, the present study indicated that color Doppler ultrasound detection of hepatic hemodynamics may have significant value for the diagnosis of ER in HCC. It may significantly improve the accuracy of predicting the occurrence of preoperative ER, which may assist clinicians in selecting optimal treatment strategies for patients with HCC.