Remember me
ORIGINAL ARTICLE
Year : 2023 | Volume
: 50
| Issue : 1 | Page : 55-60
Role of compression sonoelastography in aiding differentiation of benign and malignant solid hepatic lesions
Archana Bala1, Rajagopal Kadavigere2, K Prakashini2, Ramakrishna Narayanan3
1 Department of Radiology, Shri Sathya Sai Medical College and Research Institute, Chengalpet, Tamil Nadu, India
2 Department of Radio-Diagnosis and Imaging, Kasturba Medical College, Manipal University, Manipal, Karnataka, India
3 Department of Radiology, Nizam's Institute of Medical Sciences, Hyderabad, Telangana, India
Correspondence Address:
Rajagopal Kadavigere
Department of Radio-Diagnosis and Imaging, Kasturba Medical College, Manipal University, Manipal, Karnataka
India
Source of Support: None, Conflict of Interest: None
CheckDOI: 10.4103/jss.jss_135_22
Background: The liver is primarily or secondarily involved by numerous vascular, metabolic, infectious, and neoplastic processes resulting in formation of focal liver masses, and the detection of such focal liver lesions is frequently accomplished with sonography. However, the categorization a liver mass as benign or malignant on ultrasound has always been a diagnostic dilemma. Objective: This study aimed to assess if the addition of compression sonoelastography to conventional B-mode ultrasound aided in diagnostic accuracy of the focal hepatic lesions. Materials and Methods: We evaluated B-mode characteristics of 52 liver lesions followed by calculation of their strain values on compression sonoelastography. The lesions were categorized as benign or malignant by ascertaining a cutoff strain value and the comparison was made with the histopathological diagnosis/contrast-enhanced computed tomography characteristics of the lesions. Results: The mean strain index value of malignant hepatic lesions (2.12 ± 1.06) was statistically higher than the benign lesions (0.92 ± 1.06) with 2-tailed P = 0.002. The sensitivity, specificity, and positive and negative predictive values of compression sonoelastography in diagnosing a malignant pathology were 74.4%, 88.9%, 94.6%, and 46.7%, respectively, and the additional evaluation of B-mode features yielded higher sensitivity (95.4% vs. 83.7%) and negative predictive value (75% vs. 46.7%). Conclusion: Compression sonoelastography is an efficient and beneficial complementary tool to B-mode imaging in evaluating solid liver lesions.
Keywords: Benign, focal, liver lesion, malignant, sonoelastography, ultrasound
Ultrasound is usually the primary screening and surveillance imaging modality employed and it continues to be at the forefront of imaging abdominal organs such as liver since its inception. The sensitivity and specificity of gray scale ultrasound in diagnosing malignant focal liver lesions ranges from 76.9% to 80.6% and 75% to 90.5%, respectively,[1],[2] and the real challenge of ultrasound vests in the optimal identification of patients with malignant appearing focal liver lesions for further investigation with computed tomography/magnetic resonance imaging (CT/MRI) or with interventional procedures like biopsy. However, the specificity of ultrasound in diagnosis of benign lesions such as hemangioma is quite higher amounting to almost 98.9%.[2] The main issue is to warrant proper diagnosis, so that definite management guidelines can be provided.
Ultrasound elastography is an advanced illustration of palpation, furnishing measurements related to stiffness in addition to anatomical images and its techniques may be broadly divided into strain and shear-wave methods.[3] Strain elastography is useful for the appreciation of the physical characteristics of the tissue by estimating axial displacement of tissue caused by mechanical stress in real-time, which is generally applied externally with the transducer by the operator. The elastogram is computed from data of the change of signals before and after compression and is displayed in a split-screen mode along with the B-mode image from which strain indices are calculated.[4]
To our knowledge various studies are available to demonstrate the usage of shear wave elastography to demonstrate liver fibrosis[5],[6],[7],[8],[9],[10],[11],[12],[13],[14] and also to characterize focal liver lesions[9],[10] but very few studies are available to demonstrate the role of real-time/compression ultrasound elastography in evaluation of focal solid liver lesions. Hence, in this study, we analyzed the ability of real-time sonoelastography in the dynamic analysis to characterize and differentiate between the benign and malignant liver lesions and to assess the ability of elastography to add to the diagnostic value of conventional gray scale ultrasound.
Materials and MethodsThis prospective study was conducted over a period of 12 months in the department of radiodiagnosis in a tertiary health care hospital, after the study protocol was approved by the institutional ethics committee. Informed written consent was obtained from all patients before the procedure.
B-mode ultrasound evaluation followed by compression ultrasound elastography was performed on 52 patients with focal liver lesions detected on routine ultrasonography/CT/MRI imaging.
All the ultrasound and strain elastographywere done on an Aplio XG–Toshiba ((Toshiba medical systems corp., Japan) model apparatus and the freehand real-time elastographic examinations were performed by transabdominal approach with a convex transducer probe of central frequency 3.5MHz.
The liver lesions were examined by subcostal approach and lesions that were approachable only through intercostal approach were excluded, for lack of accessiblity. Patients with gross ascites and deep seated lesions (depth >8 cm), where the adequacy of compression was questionable were also excluded from the study. Apart from the above mentioned exclusion criteria, pregnant patients and patients with no normal adjacent liver parenchyma for comparison of a focally involved region of interest were also excluded.
A total of 52 patients with a median age of 56 years with liver lesions were evaluated of which 43 were malignant and 9 were benign on histopathological confirmation/contrast-enhanced CT (CECT) typical enhancement pattern. The lesions that were evaluated by elastography are summarized in [Table 1]. The B-mode ultrasound characteristics of the lesions were evaluated and the lesions were categorized as benign or malignant based on the imaging characteristics in [Table 2].
Diagnosis was achieved through confirmative imaging and histopathological evidence viz CECT for benign lesion and ultrasound-guided biopsy for malignant lesion. Appropriate treatment protocols were devised and were offered to the patients.
ResultsThe results of our study was as follows:
Of the 52 patients included, majority of them had malignant lesions accounting for 80.76% and 17.30% had benign lesions. A subgroup analysis was done to identify the characteristics of the lesions, which showed metastases to be the predominant lesion (n = 28) and hepatocellular carcinoma with second highest prevalence (n = 14). There was one case of cholangiocarcinoma. The primary sites of all the metastatic lesions, which were included in the study, are summarized in [Table 3. On the other hand 8/52 patients had hemangiomas and 1/52 had regenerative nodules of chronic Budd-Chiari syndrome.
Imaging and statistical analysis
The mean diameter of benign lesions included in the study (5.82 cm) was more than that of malignant lesions (3.85 cm). The maximal depth measured perpendicularly from skin surface to the lesion was 6.8 cm in a case of metastasis with unknown primary and the lesion with least depth was a hepatocellular carcinoma at a depth of 0.7 cm. Maximal lesions to which elastography were applied were in segment IV of left lobe (30.77%) and least of the lesions were in segment VI of right lobe (1.92%). In case of multiple lesions distributed in both lobes, lesions from preferably left lobe were taken for elastographic evaluation because of ease of subcostal approach in case of left lobe lesions. The segmental distribution of lesions is given in [Figure 1].
Figure 1: Pie-chart showing the segmental distribution of lesions chosen for elastography. Maximal lesions to which elastography were applied were in segment IV of left lobe (30.77%) and least of the lesions were in segment VI of right lobe (1.92%)The fourteen patients with histopathologically proven diagnosis of hepatocellular carcinomas included in the study had no obvious features of cirrhosis on ultrasound/CT. The liver parenchyma of patients with other lesions also did not show features of cirrhosis and the patients revealed no relevant history to suggest cirrhosis. The mean strain index value of all malignant lesions (2.12 ± 1.06) was significantly higher than the mean strain index value of the benign lesions (0.92 ± 1.06) with a 2-tailed P = 0.002 [Figure 2]. The distribution of strain index values for all the benign and malignant lesions is illustrated in [Table 4]. The least strain index is 0.42, seen in a case of hemangioma and maximal strain index is 6.05, seen in a case of hepatocellular carcinoma.
Figure 2: Box plot of strain indices (distribution of strain indices left side – benign and right side – benign)Receiver operating curve were generated and the best cutoff for mean strain index was ascertained, which was taken as 1.32 with a sensitivity and specificity of 83.7% and 79.8%, respectively [Figure 3]. On the cross-table analysis of compression ultrasound elastography in characterizing focal liver lesions into benign and malignant in comparison with the final diagnosis, positive and negative predictive values of 94.6% and 46.7%, respectively, were obtained [Table 5]. On further comparison, the dual diagnosis made on B-mode imaging as well as elastography by cross table analysis yielded a sensitivity, specificity, and positive and negative predictive value of 95.34%, 66.66%, 93.18%, and 75%, respectively [Table 6], and the likelihood ratio of a lesion being malignant was 2.79 times higher when combined B-mode and elastographic analysis was obtained.
Figure 3: ROC curve generated showed mean strain index taken as 1.32 to have a sensitivity and specificity of 83.7% and 79.8%, respectively. ROC: Receiver operating curveThe grey scale image and corresponding compression sonoelastogram and axial contrast enhanced computed tomography (CECT) images from three patients with hemangioma [Figure 4], hepatocellular carcinoma [Figure 5] and hepatic metastasis [Figure 6] are shown.
Figure 4: Hemangioma. (a) Gray-scale image of the liver showing a hyperechoic hemangioma in the left lobe of the liver (arrow). (b) Corresponding compression sonoelastogram from the same patient. The region of interest 1 is in the lesion (arrow) and region of interest 2 is in the adjacent liver parenchyma (dotted arrow) at the same depth. The mean strain index calculated from the relaxation wave was 0.72 (arrow head) in this lesion. (c and d) Axial CECT from the same patient in arterial (c) and equilibrium phases (d) Showing peripheral puddling of contract in arterial phase in the hemangioma (arrow in c) with progressive centripetal filling in (block arrow in d). CECT: Contrast-enhanced computed tomography
Figure 5: Hepatocellular carcinoma. (a) Gray-scale image of the liver showing an illdefined heterogeneously hypoechoiec lesion in the left lobe of the liver (arrow). (b) Corresponding compression sonoelastogram in the same lesion is displayed with the region of interest 1 within the lesion (block arrow) and region of interest 2 in the adjacent liver parenchyma (dotted arrow) at the same depth. The mean strain index of the lesion calculated from the relaxation wave is displayed as 2.31 (arrow head). (c and d) Axial CECT from the same patient in arterial (c) and porto-venous (d) showing heterogeneous peripheral enhancement in arterial phase with relative washout seen as isodensity to surrounding parenchyma in portovenous phase. CECT: Contrast-enhanced computed tomography
Figure 6: Hepatic metastasis. (a) Gray-scale image of the liver showed multiple well-defined randomly distributed hypoechoiec lesion is both lobes of liver (not shown) with one of the lesion in subcapsular location chosen for elastography (arrow). (b) Corresponding compression sonoelastogram in the same lesion is displayed with the region of interest 1 within the lesion (block arrow) and region of interest 2 in the adjacent liver parenchyma (dotted arrow) at the same depth. The mean strain index of the lesion estimated is displayed as 2.09 (arrow head). (c) Axial CECT from the same patient in porto-venous phase shows multiple hypoattenuating lesions in the liver (small arrow). There is also an ill-defined hypodense lesion involving distal body the tail of pancreas (long arrow) which was the site of primary as biopsy revealed adenocarcinoma of pancreas. (d) Axial CECT from the same patient in arterial phase shows the subcapsular lesion chosen for elastography (dotted arrow) among multiple other similar sized hypoattenuating lesions in the liver. CECT: Contrast-enhanced computed tomography Discussion The principal advantages of ultrasonography in abdominal imaging include a very good cost/quality ratio and also the ability to provide dynamic character (”real time”) to imaging. To add to the diagnostic accuracy of ultrasonography, the supplementary use of real time/compression elastography, which measures tissue displacement in one-dimensional plane and conveys new information about internal tissue architecture to differentiate between benign and malignant focal liver lesions, has been proposed.
While sonograms impart visual information related to the local acoustic backscatter energy from tissue components, elastograms aim at estimating the local strain, Young's moduli or Poisson's ratios.[11] In compression elastography the images are semi quantitative and do not directly depict the elasticity.[12],[13],[14] The study of liver elastography dates back to as early as 1980's but the first clinical report of elastography for the liver was done by Sandrin et al.[15] Although there have been numerous studies on illustration of diffuse liver fibrosis by elastography, there are only a few studies published regarding evaluation of focal liver lesions by elastography.
The study conducted by Gheorghe et al. has confirmed that the real time elastography is accurate enough in detecting regenerating hepatocarcinoma nodules. The study was conducted on patients with cirrhosis with small, under 3 cm, subcapsular nodules.[16] Another study done by Kato et al.[17] has also shown good results of the real time elastography in differentiating hepatocellular carcinoma and hepatic metastases. The authors have divided the tumors according to a new system called elasticity type of liver tumor from Type A (lesions which has even strain) to Type D (lesions having no strain). According to this, most hepatocellular carcinomas were classified into the Type B, while most metastases were included into Type C and D.
In our study, the mean strain index was taken as 1.32, any value below the reference range was considered as benign and above was considered as malignant. This yielded a sensitivity of 83.7% and specificity of 79.8%. These statistical values when compared to a similar study done by Onur et al.,[18] were high and significant. This can be attributed to the inclusion of all lesions which were more than a centimeter in at least one diameter which is not clear in the latter's study. Another reason for high sensitivity and specificity in our study can be due to the exclusion of cirrhosis patients as cirrhosis can alter the elasticity of the liver parenchyma.
In various other studies using acoustic radiation force impulse elastography (ARFI) to evaluate focal liver lesions such that by Davies and Koenen[19] and Shuang-ming et al.,[9] by using a cutoff of 2.2–2.5 m/s, the sensitivity and specificity for malignancy ranged from 89% to 97% and 95%–100%, respectively. However, when combining B-mode imaging and strain index value on elastography to differentiate the characteristics of the lesion, we were able to achieve a higher sensitivity of 95.34% which was comparable to the results from ARFI elastography.
The limitations of our study can be attributed to the relatively small sample size. However, other confounding factors like cirrhosis were excluded which helped us in obtaining a better result in regard to sensitivity and specificity. To the best of our knowledge this is the first study to implement combined evaluation of B-mode imaging with strain elastography, which has yielded better results in characterization of liver lesions.
ConclusionOur study has demonstrated that strain elastography mirrors the stiffness of focal solid liver lesions with a significant degree of accuracy. The inclusion of elastography to B-mode imaging increases the diagnostic efficacy, which can be beneficial in view of ease of use. However, further prospective studies with larger study population are needed to prove its efficacy.
Financial support and sponsorship
Nil.
Conflicts of interest
There are no conflicts of interest.
References
Comments (0)