Role of shear wave elastography in characterization of hepatic focal lesions

Elastography is a recently developed diagnostic method that aims to evaluate tissue stiffness. Its utility is based on the fact that pathological tissues are generally stiffer than surrounding healthy tissues which often show up as a hard lesion. Shear wave elastography (SWE) is a new technique based on shear waves that has been implemented in diagnostic ultrasound (US) systems. The aim of this study is to study the role of shear wave elastography in differentiation between benign and malignant hepatic focal lesions. The study was conducted on 110 patients (92 males, 18 females) with a mean age of 51.7 years. Age range was from 30 to 70 years; 28 patients were diagnosed with benign lesions, and 82 with malignant lesions. SWE shows that there is a significant difference in stiffness between malignant and benign lesions with p value = 0.002 and with mean ± SD of 10.3 ± 6.31 kPa for the benign lesions and 16.2 ± 9.32 kPa for the malignant group. A cutoff value of 13.24 was selected to differentiate between benign and malignant lesions using the SWE mean providing sensitivity 78.04%, specificity 71.42%, and accuracy 64.2%. SWE is a good tool in the differentiation of benign and malignant hepatic focal lesions.


Background
Liver tumors are tumors or growths on or in the liver. Several distinct types of tumors can develop in the liver because the liver is made up of various cell types. These growths can be benign or malignant. They may be discovered on medical imaging (even for a different reason than the cancer itself) or may be present in patients as an abdominal mass, hepatomegaly, abdominal pain, jaundice, or some other liver dysfunction [1]. Because of its advantageous cost/benefit ratio, widespread availability, and easy execution, ultrasound (US) is the first-line imaging modality in most countries for the initial liver survey and represents the imaging technique that usually detects complex liver lesions [2]. It is generally agreed that no other physical parameter of tissue is changed by pathological or physiological process to a remarkable extent as the tissue elasticity [3]. SWE estimates the speed of shear waves to provide a quantitative estimate of tissue stiffness [4]. SWE has the advantage of being able to image liver tissue stiffness in real time because the shear waves are generated by US impulses [5]. Several features can be described with this technique, including the quantitative evaluation of the lesion elasticity in kilopascals or meters/second and the spatial heterogeneity of stiffness, and moreover, SWE imaging is guided using Bmode images with a higher frame rate. This method can provide more accurate assessment of liver tissue stiffness due to the advantages of SWE and B-mode image guidance [6,7].
The purpose of this study was to evaluate the role of SWE in characterization of hepatic focal lesions.

Methods
This was a prospective study conducted from December 2017 to March 2019 after the approval of protocol from faculty of medicine ethical committee for human research; fully informed written consent was taken from each patient after given an explanation of the procedures and the importance of the study; the confidentiality of the patient's data was guaranteed, and the patients had the right to refuse participation in this study without giving any reason.
The study enrolled 110 patients: 92 male and 18 females with a mean age 51.7 years after fulfillment of inclusion and exclusion criteria. Inclusion criteria were patient presented with at least one hepatic focal lesion greater than 1 cm in diameter and not more than 5 cm deep from the capsule. Exclusion criteria were focal lesion previously managed by intervention radiology, patient with perihepatic ascites, and uncooperative patients.
The patients were subjected to full history taking and laboratory investigations. Real-time ultrasound and shear wave elastogaphy (SWE) was performed using Philips iU22 × MATRIX ultrasound system with a curved array transducer 3-5 MH. The patients were examined in the supine or left lateral position with right arm elevated above the head to improve intercostal access. Gray scale abdominopelvic ultrasonography was initially performed to evaluate the liver size, echogenicity, texture, outline, and focal lesions. We identified hepatic focal lesions for number, site, size, margin, and echogenicity. Then, in SWE scanning, the patient was instructed to hold his or her breath for a few seconds. We located the segment of hepatic focal lesion, and a region of interest (ROI) was adjusted upon the targeted focal lesion; we took from 5 to 8 measurements for quantitative evaluation of stiffness. Then, the machine estimated the velocity of the propagated shear wave in the ROI and automatically translates it to stiffness in kilopascals (KPs). In patient with more than one focal lesion, we ensured that all lesions have the same pathology guided by other modalities like CT and MRI, and then, we analyzed the most accessible one. Multiple successful measurements are obtained, and the results appear in the final report as the average of all measurements. Then, the data obtained by SWE scanning is compared with the results of triphasic CT and raised tumor markers in 92 patients (84%), dynamic MRI in 15 patients (13%), and 3 patients by histopathology to estimate the accuracy, sensitivity, and specificity of SWE in characterization of different hepatic focal lesions.

Statistical analysis
Data were collected, tabulated, and statistically analyzed using a personal computer with Statistical Package of Social Science (SPSS) version 20 [SPSS Inc., Chicago, IL, USA], where the two types of statistics were done.

Descriptive statistics
Quantitative data was expressed as mean ± standard deviation (SD). Qualitative data was expressed as frequency and percentage.

Analytic statistics
Independent samples t test of significance will be used when comparing between two means. Chi-square (χ 2 ) test of significance will be used in order to compare proportions between two qualitative parameters. ROC curve is a graph called a receiver operating characteristic curve; it is a plot of the true positive rate against the false positive rate for different possible cutoffs of diagnostic test or marker. p value ≤ 0.05 was considered to be statistically significant. p value ≤ 0.001 was considered to be highly statistically significant.

Results
Our study was conducted on 110 patients (92 male (83%), 18 female (17%)). The patients' age ranged from 30 to 70 years with a mean age of 50.2 years for the benign lesions and 54.2 years for the malignant; Table 1 shows that the malignant lesions presented more among the male patients aged 50 to 60 years representing 25.5%. Clinical analysis of the clinical presentation of the studied patients revealed that 45 patients (40.9%) were presented by abdominal distention and 40 patients (36.4%) were presented by abdominal pain followed by loss of appetite (13.6%), loss of weight (11.8%), and jaundice (10%) ( Table 2; Fig. 1). The laboratory findings of the studied patients are summarized in (Table 3); the most significant lab investigations in patients with malignant lesions were alphafeto proteins, hepatitis viral markers, and liver function tests (SGPT, SGOT, albumin) with p value of 0.001, 0.001, 0.002, 0.006, and 0.013, respectively. Highresolution ultrasound scanning of the 110 patients yielded the detection of 152 focal lesions. In patient with more than one focal lesion, we ensured that all lesions have the same imaging criteria by other modalities like CT and MRI, and then, we analyzed the most accessible    one. Ultrasound imaging criteria for the liver parenchyma and hepatic focal lesions are listed in Table 4. The cirrhotic liver was significantly related to the presence of malignant lesions with p value < 0.004, while the size of the liver had no statistically significant relationship to the pathology of the lesions with p value 0.6. Also, the presence of splenomegaly and ascites was significantly related to malignancy with p value < 0.02 and 0.019, respectively. Out of the studied 110 patients, 82 patients (74.5%) were finally diagnosed with malignant lesions (70 of them (63.6%) had HCC, 6 patients (5.4%) with cholangiocarcinoma, 6 patients (5.4%) with metastasis) and 28 patients (25.5%) were finally diagnosed with benign lesions (hemangioma in 24 patients (21.8%) and focal fatty infiltration in 4 patients (3.36%)) as seen in Table 5.
All patients were subjected to shear wave elastography scanning. It showed that there was a significant difference in stiffness (p value < 0.002) between malignant and benign lesions with mean ± SD of 10.3 ± 6.31 kPa for the benign lesions and 16.2 ± 9.23 kPa for the malignant group as listed in Table 6. Table 7 and Fig. 2 showed that a cutoff point of 13.24 was selected to differentiate between benign and malignant lesions using the SWE mean providing sensitivity 78.04%, specificity 71.42%, positive predictive value 88.8%, negative predictive value 52.6%, and accuracy 64.2%. We found that the 4 focal fatty infiltration were found to have stiffness mean ± SD values of 12.6 ± 8.8 kPa. The 24 hemangioma lesions were found to have stiffness mean ± SD of 9.5 ± 4.2 kPa (Fig. 3).The stiffness mean ± SD of the 70 HCCs was 15.4 ± 7.2kPa, while the stiffness mean ± SD of the 6 cholangiocarcinoma was 32.5 ± 8.25 kPa. The stiffness of the 6 metastatic lesion mean ± SD was 22.1 ± 0.4 kPa as described in Table 8.

Discussion
The accurate characterization and the differential diagnosis between the different types of FLLs are important aims that all the imaging modalities available today should satisfy [8]. The aim of this study was to assess the accuracy and sensitivity of SWE in characterization of hepatic focal lesions. Shear wave elastography quantitative assessment of focal liver lesions stiffness was done and compared with the final diagnoses obtained by triphasic CT, dynamic MRI, and pathological results. In this work, we found that there was a significant difference in stiffness between benign and malignant groups  with a mean value of 10.32 ± 6.74 for benign lesions and 17.23 ± 8.25 for malignant lesions with a statistically significant difference of p < 0.002. This was consistent with Guibal et al. [8] who reported that there was a significant difference in the stiffness between benign and malignant lesions with means of 18.53 ± 13.5 kPa and 26.9 ± 18.8 kPa for benign and malignant lesions, respectively (with high significant difference p >0.01). Also, Park et al. [9] reported that the stiffness values of malignant lesions were significantly higher than those of benign lesions (p < 0.0001), while other studies have not demonstrated any differences (Ronot et al. [10], Frulio et al. [11], Gallotti et al. [12], and Heide et al. [13]). In this current study, we had 24 cases of hemangiomas which were found to have a mean stiffness of 9.5 ± 4.2 kPa. This was close to the results reported by Guibal et al. [8] who reported a mean stiffness of 13.8 ± 5.  [10,[13][14][15] also described similar observations. In our study, we have 4 focal fatty infiltrations (FFI) which were found to have a mean stiffness of 10.2 ± 5.8 kPa. These results were higher than the values reported by Virchenko et al. [16]. They studied the stiffness values of diffuse hepatic steatosis in 130 patients and reported a mean SWE value of 8.4 kPa for severe steatosis. The difference may be attributed to larger sample size in their study. Ronot et al. [10] also reported similar results regarding diffuse hepatic steatosis with a mean SWE value of 9.6 ± 4.7 kPa.
In the current study, the mean stiffness for 70 hepatocellular carcinomas (HCCs) was 15.4 ± 7.2 kPa. These results are matching with a study carried out by Guibal et al. [8]. They reported a mean stiffness of 14.86 ± 10 kPa for 26 hepatocellular carcinoma which is similar to our study (Fig. 4). Also, our finding was close to those reported by Ronot et al. [10] who reported a mean stiffness value of 19.6 kPa. We also found that cholangiocarcinoma and metastasis were the stiffest of all lesion types. SWE value was 32.5 ± 8.25 kPa for the 6 cholangiocarcinoma. This result was close to Ronot et al. [10] who reported 34.1 ± 7.3 kPa for the two cholangiocarcinoma.
Our result was lower than that reported by Guibal et al. [8] who reported that cholangiocarcinoma were the stiffest of all lesion types with a mean stiffness value of 56.9 ± 25.6 kPa. The difference in both observations may be due to the small sample size in our study. Guibal et al. [8] explained the elevation of the stiffness value of cholangiocarcinoma to their significant fibrotic component. Also, Masuzaki et al. [17] and Heide et al. [13] reported similar observations about the stiffness value. Also, in this current study, a mean stiffness value of 22.1 ± 0.4 kPa was reported for the 6 metastasis cases. This was consistent with Guibal et al. [8]; they studied 53 metastases with mean elasticity values of 21.8 ± 14.6. Guibal et al. [8] reported that the elevation in elasticity values found with SWE likely originates from several factors, including changes in vascular and lymphatic permeability, along with changes in collagen deposition and cellular architecture. Malignant lesions grow in an environment restrained by the surrounding tissue, and this can result in an increase in interstitial pressure [8].
There are several limitations encountered with the use of SWE: 1. The primary cause of failure was the lesion location beyond the penetration limits of SWE generally (> 5 cm), major steatosis, and the proximity to vessels. 2. Other failures were attributed to poor intercostal windows and patient inability to hold their breath long enough to acquire a stable SWE acquisition. 3. Another limitation is that the radiologist who acquired the elastography images was not blinded for the diagnosis, and this could introduce bias.

Conclusion
SWE is a good tool for differentiating between different benign and malignant hepatic focal lesions with high sensitivity, specificity, and accuracy. It shows that there is a significant difference in stiffness between malignant and benign lesions.