Liver cancer is one of the most frequent causes of cancer related death worldwide, with hepatocellular carcinoma (HCC) accounting for about 90% of malignant liver tumors(1–3). Treatment varies in different pathological types of focal liver lesions (FLLs). For malignant lesions, surgical resection, chemotherapy, radiotherapy or targeted therapy are recommended methods(4,5). However, for most benign lesions, conservative treatment is first considered(6). Thus, correct preoperative diagnosis plays an important role in clinical decision making for FLLs. With the advantages of noninvasiveness, real-time nature and no radiation exposure, ultrasound (US) is now used as the first-line imaging modality for detecting and diagnosing hepatic lesions(7). The resistance index (RI) of spectral Doppler imaging (SDI) in ultrasound examination could reflect the resistance of distal vessels, which may be helpful in the differential diagnosis of FLLs(8–11). The purpose of this study was to assess the value of RI in the differential diagnosis of different FLLs.
This retrospective study was approved by the ethical board of our institution. Informed consent was waived.
From January, 2018 to June, 2021, a total of 576 patients met the criteria and were retrospectively included. Inclusion criteria were as follows: (1) Patients with FLLs detectable on B-mode ultrasound (BMUS), (2) arterial blood flow signals detectable within lesion or peritumoral area, (3) obtainable surgery/biopsy and histopathological diagnosis of FLLs. No exclusion criteria were defined.
Ultrasound equipment included Siemens ACUSON Sequoia (Siemens Medical Solutions, Munich, Germany), Siemens ACUSON Oxana 2 (Siemens Medical Solutions, Munich, Germany), Aloka Arietta V70 (Hitachi Aloka Medical Systems, Tokyo, Japan) and Mindray Resona 7s (Shenzhen Mindray Bio-Medical Electronic Co., Shenzhen, China). All ultrasound examinations were performed by ultrasound radiologists with at least 10-year liver scan experience.
All patients fasted for at least 8 hours before ultrasound examinations. US examinations were performed in the supine position. A full liver BMUS was performed to locate and characterize the lesion. Using color Doppler mode, the distribution of vascularity within a lesion was detected. If blood flow signals were detected within a lesion, spectral Doppler would then be performed to measure the RI value (difference of peak-systolic and end-diastolic velocity divided by peak-systolic velocity of blood flow). The most obvious artery in CDFI was selected for RI measurement. The RI measurements were performed with the same protocol for all lesions. For multiple lesions, only the largest ones were selected to detect color flow signals and measure RI values.
The analysis software was SPSS version 20.0. Kolmogorov-Smirnov test was performed on continuous data to determine normal distributions. Levene test was performed to detect the homogeneity of variance. One-way ANOVA test and independent sample
From January, 2018 to June, 2021, a total of 576 (173 females, 403 males, age range 20 - 85 years, mean age: 54.82 ± 13.83 years) patients were followed up. Single lesions were detected in 476 patients (82.6%) and multiple lesions were detected in 100 patients (
According to pathological diagnosis, there were 458 malignant lesions and 118 benign lesions. Among all the malignant lesions, there were 347 pathologically confirmed HCC lesions, 51 metastatic lesions, 44 intrahepatic cholangiocarcinoma (ICC) lesions, 5 mixed hepatocellular cholangiocarcinoma lesions, 4 sarcomatoid carcinoma lesions, 3 epithelioid cell malignant lesions, 1 B-cell lymphoma of mucosa-associated lymphoid tissue lesion, 1 sarcoma lesion, 1 multiple plasmacytoma lesion and 1 neuroendocrine carcinoma lesion. Among all the benign lesions, there were 52 focal nodular hyperplasia (FNH) lesions, 16 angiomyolipomas lesion, 23 hemangioma lesions, 9 hepatocellular adenoma lesions, 5 inflammatory lesions, 3 lymphoid tissue lesion, 2 necrotic lesions, 2 pseudolymphoma lesions, 1 accessory spleen lesion, 1 lipoma lesion, 1 paraganglioma lesion, 1 papilloma lesion, 1 hyperplasia tissue lesion and 1 neuroendocrine tumor lesion.
Most of malignant FLLs were hypoechoic (69.87 %,
Color blood flow signals were detected in 576 lesions, but only 432 (75%) lesions allowed RI measurement. In malignant lesions, intralesional signals were obtained in 80.57%, perilesional signals in 12.66% and both intralesional and perilesional signals in 6.77% of lesions. In benign lesions, intralesional signals were obtained in 78.81%, perilesional signals in 11.86% and both intralesional and perilesional arterial signals in 9.32% of lesions.
While comparing the RI values of benign and malignant lesions, the mean RI of benign lesions was significantly lower than that of malignant lesions (0.54 ± 0.09
RIs for different focal liver lesions
Pathological diagnosis | Proportion n/all (%) | RI (range) | RI (mean ± SD) |
---|---|---|---|
Malignant lesions | 332/458 (72.5%) | 0.43–1.00 | 0.71 ± 0.12 |
Hepatocellular carcinoma | 254/347 (73.2%) | 0.43–1.00 | 0.72 ± 0.12 |
Liver metastasis | 30/51 (58.82%) | 0.52–1.00 | 0.70 ± 0.10 |
Intrahepatic cholangiocarcinoma | 35/44 (79.5%) | 0.51–1.00 | 0.68 ± 0.10 |
Mixed hepatocellular Cholangiocarcinoma | 3/5 (60%) | 0.68–0.82 | 0.73 ± 0.08 |
Sarcomatoid carcinoma | 3/4 (72.5%) | 0.69–0.72 | 0.71 ± 0.02 |
Epithelioid cell tumor | 3/3 (100%) | 0.45–0.59 | 0.51 ± 0.07 |
Benign lesions | 100/118 (84.7%) | 0.35 – 0.84 | 0.54 ± 0.10 |
Focal nodular hyperplasia | 48/52 (92.3%) | 0.40–0.69 | 0.54 ± 0.08 |
Angiomyolipoma | 15/16 (93.8%) | 0.35–0.72 | 0.51 ± 0.09 |
Hemangioma | 14/23 (60.9%) | 0.35–0.66 | 0.51 ± 0.09 |
Hepatocellular adenoma | 8/9 (88.9%) | 0.42–0.66 | 0.51 ± 0.08 |
Inflammatory pseudotumors | 4/5 (80%) | 0.58–0.84 | 0.68 ± 0.12 |
Taking RI = 0.615 as a cutoff value to differentiate malignant and benign lesions (RI >0.615: malignant; RI <0.615: benign), the sensitivity, specificity, positive predictive value and negative predictive value were 82.80%, 81.00%, 81.34% and 82.48%, respectively.
All the HCC lesions were classified into 4 groups (group 1: ≤2 cm, group 2: 2–5 cm, group 3: 5–10 cm, group 4: >10 cm). We found that the RI values of large lesions (group 4: >10 cm) were significantly lower than those of small lesions (group 1: ≤2 cm, group 2: 2–5 cm). However, no significant difference was observed in other groups (Tab. 2, Fig. 5, Fig. 6, Fig. 7, Fig. 8).
Comparison of resistance index in hepatocellular carcinoma (HCC)
Group | Number | Size of HCC (cm) | RI (mean ± SD) |
---|---|---|---|
1* | 63 | < 2 | 0.73 ± 0.12 |
2# | 113 | 2–5 | 0.75 ± 0.12 |
3 | 63 | 5–10 | 0.70 ± 0.11 |
4*, # | 15 | >10 | 0.61 ± 0.13 |
Group 1, 4: Significant difference (
Group 2, 4: Significant difference (
Based on the differences in liver background, all the HCC lesions were classified into 3 groups (group 1: normal background, group 2: fatty liver background, group 3: cirrhosis background). We found that the RIs of HCC in normal liver background were significantly lower than those of cirrhosis liver background. However, no significant difference was observed in other groups (Tab. 3).
Comparison of resistance index in hepatocellular carcinoma in different liver background
Group | Background | Number | RI (mean ± SD) |
---|---|---|---|
1* | normal | 82 | 0.69 ± 0.12 |
2 | Fatty liver | 16 | 0.75 ± 0.11 |
3* | cirrhosis | 156 | 0.73 ± 0.12 |
Group 1, 3: Significant difference (
All the FNH lesions were classified into 4 groups (group 1: ≤2 cm, group 2: 2–5 cm, group 3: 5–10 cm, group 4: >10 cm). A comparison of RI values in different groups showed no significant differences (Tab. 4).
Comparison of resistance index in focal nodular hyperplasia lesions (FNH)
Group | Number | Size of FNH (cm) | RI (mean ± SD) |
---|---|---|---|
1 | 3 | <2 | 0.47 ± 0.05 |
2 | 28 | 2–5 | 0.56 ± 0.07 |
3 | 16 | 5–10 | 0.54 ± 0.08 |
4 | 1 | >10 | 0.49 |
The type and distribution of blood flow signals were correlated with the histology of FLLs. Previously, the presence of both intra- and peritumoral arterial flow was strongly suggestive of malignancy, while the presence of intralesional venous flow was remarkably suggestive of benignancy. Intratumoral flow could be detected in 73.9% of malignant lesions and in 25% of benign lesions(12). Although arterial flow signal was more frequently detected in malignant lesions, there were some typical color flow signs specific for benign FLLs(11,13). Central feeding artery is a typical sign of FNH lesion and the spoke-wheel sign has a high diagnostic value(14,15). However, in the present study we focused on RI values rather than CDS-architecture.
RI represents the resistance of distal vessels(16,17). Combining the ultrasound features and the RI parameter would be more helpful in the differential diagnosis of benign and malignant liver tumors(8,12). In 2014, Nahar
In our study, we evaluated the RI values of HCC lesions with different size. We found that RIs in large FLLs were lower than those of small lesions. The reasons may be correlated with pathological features, such as histological pattern, pseudocapsular growth type, and absence of necrotic areas. As the lesion grows, the formation of arteriovenous shunt and the destruction of pseudocapsules may reduce the resistance of vessels within HCC lesions.
The main limitation is the retrospective nature of this study. Due to the retrospective nature, some of the information may be incomplete. Besides, there may be some factors affecting the RI measurement, such as the localization of the lesion in different liver regions. Further prospective clinical study will be necessary to validate the RI diagnostic criteria in differentiating between benign and malignant FLLs. In the future, besides typical patterns of perfusion, CEUS could be a better imaging method for detecting blood flow in FLL.
Color Doppler examination is a valuable ultrasound imaging method for detecting blood flow signals of FLLs. The RI parameter would be helpful in differentiating malignant and benign FLLs.