Shear wave velocity values measured by 2D-shear wave elastography are not different between awake and anesthetized cats without clinically significant hepatic fibrosis

Abstract

Two-dimensional shear wave elastography (2D-SWE) is widely used as a noninvasive method to quantify liver stiffness. In humans, liver stiffness approximates histologic hepatic fibrosis. While histology is the gold standard for diagnosing liver disease, 2D-SWE may be a minimally invasive alternative to biopsy in feline patients. The objectives of this prospective, observational, crossover study were trifold: (1) to assess the feasibility of performing 2D-SWE in awake cats, (2) to determine whether anesthesia altered shear wave velocity (SWV) measurements, and (3) to correlate hepatic stiffness with histologically quantified hepatic fibrosis. Eleven healthy, purpose-bred cats underwent 2D-SWE in awake and anesthetized states. SWV measurements were compared with histologic fibrosis measurements obtained from liver biopsies during the anesthetic period. The mean velocities were not significantly different between awake (1.47 ± 0.18 m/s) and anesthetized (1.47 ± 0.24 m/s) cats. Premedication and anesthetic drugs did not impact mean SWV. There was a higher variability in the SWV values in the awake group. The data points were reliably replicated, with an interquartile range of 0.24 and 0.32 in anesthetized and awake groups, respectively. There was moderate agreement between observers (intraclass correlation coefficient = 0.66). All cats had clinically insignificant fibrosis. There was no correlation between the SWV measurements and the histological fibrosis values. This study demonstrates that 2D-SWE is feasible in awake cats and that the anesthetic protocol employed did not significantly alter mean SWV. This work is the first to histologically validate normal SWV values in cats and show that 2D-SWE cannot differentiate minimal differences in feline hepatic fibrosis.

1 |. INTRODUCTION

B-mode ultrasonography constructs morphological images of organs, however, does not provide quantitative information on the tissue’s elastic properties. Two-dimensional shear wave elastography (2D-SWE) is an emerging imaging technology that can quantify the elastic modulus of a tissue by producing an external acoustic radiation force that causes low frequency shear waves to propagate through tissue. The shear waves propagate in a perpendicular direction relative to the external force. The measured velocity of these waves is related to a specific tissue’s elasticity, through the formula: $E=3\rho v^2$, where $\rho$ is the tissue density and v is the speed of the shear wave.^1,2

Chronic injury to the liver can result in hepatic fibrosis, which is characterized by a progressive accumulation of collagen and extracellular matrix in the hepatic parenchyma. This subsequently leads to distortion of the hepatic architecture.³ Hepatic fibrosis in cats can result from chronic inflammatory disease, with cholangitis/cholangiohepatitis being the most common cause.^4–6 The gold standard for determining the severity of hepatic fibrosis is with liver histology. An increase in collagen connective tissue within the hepatic parenchyma in patients with hepatic fibrosis results in higher SWVs. Shear wave elastography has been used as a noninvasive method to successfully evaluate the stage of hepatic fibrosis in humans, being able to distinguish human patients with no or minimal fibrosis from those with severe fibrosis or cirrhosis.^3,7–9 In human patients with chronic cholangitis, fibrotic peripheral bile ducts are present throughout the liver, forming a parenchyma environment similar to hepatic fibrosis, also causing increased shear wave velocities.¹⁰ 2D-SWE has also been used to evaluate dogs with hepatic fibrosis, documenting higher SWVs in dogs with clinically relevant hepatic fibrosis compared with healthy dogs.¹¹ Additionally, recent reports demonstrating the use of shear wave elastography in dogs with surgical treatment of congenital extrahepatic portosystemic shunts, extrahepatic biliary obstruction, and liver tumors have also been reported.^12–14

The overlap that exists between the ultrasonographic appearance of normal and abnormal tissues on B-mode ultrasound may warrant biopsy, especially for certain hepatic diseases. Limitations of liver biopsy include invasiveness, anesthesia, bleeding, and cost. In veterinary medicine, liver biopsy is not always feasible due to the reasons listed. As performed in humans, shear wave elastography may be a feasible noninvasive technique for the assessment of hepatic fibrosis without biopsy in cats. A standard range for hepatic stiffness has been established for 18 client-owned healthy cats using point shear wave elastography.¹⁵ However, to the best of our knowledge, the use of shear wave elastography to assess hepatic fibrosis has not been validated with liver histology in cats.

The objective of this study was to assess the feasibility of performing 2D-SWE in awake cats, to determine whether anesthesia affected shear wave velocity (SWV) measurements, and to correlate hepatic stiffness with histologically quantified hepatic fibrosis. We hypothesize that cats with normal hepatic histology will have a mean shear wave speed similar to 1.43 m/s (range 1.35–1.51 m/s) reported in a previous study of healthy adult cats and that higher shear speeds will be identified in cats with histologic evidence of hepatic fibrosis.¹⁵

2 |. MATERIALS AND METHODS

2.1 |. Selection and description of subjects

This study was a prospective, observational, crossover design. Eleven, purpose-bred, clinically healthy cats from the LSU, School of Veterinary Medicine, Division of Laboratory Animal Medicine (DLAM) were included. Final decisions for patient inclusion were made by two veterinary internal medicine clinicians, one of whom is a co-author (A.J.) in concordance with the university’s DLAM policy. Signalment consisting of age, breed, sex, and body weight were recorded at the time of recruitment. The population consisted of five spayed females and six castrated males. Their age range was 6–10 years (median: 9 years). Their weight ranged from 4.5 to 8.9 kg (median: 5.8 kg). Complete blood count (CBC) and serum biochemistry results obtained within one year from the start of the study were recorded from the medical records of all included cats. All cats were confirmed clinically healthy based on physical examination, CBC, and serum biochemistry. All animal experimental procedures were approved by the Louisiana State University Veterinary Teaching Hospital Institutional Animal Care and Use Committee.

2.2 |. Data recording and analysis

2.2.1 |. SWV measurement

Two-dimensional shear wave elastography was performed on all cats prior to liver biopsy without sedation or anesthesia. All patients were fasted for 12 h prior to imaging. Measurements were recorded by two observers, a double-board-certified veterinary radiologist (N.R., ECVDI, ACVR) and first-year radiology resident (V.V.) using the same scanner system (GE LOGIQ^™ E10, Chicago, IL) and an 6–15-MHz linear probe. Eleven awake cats were restrained in dorsal recumbency for B-mode imaging. The fur was clipped with margins appropriate for a full abdominal ultrasound. Alcohol and ultrasound gel (Parker Laboratories, Inc., Aquasonic^® 100 Ultrasound Transmission Gel, Fairfield, NJ) were used as coupling media. The probe was placed on the abdomen and the ventromedial region of the liver parallel and caudal to the costal arch was identified for image acquisition. Once a region of liver was identified, shear wave elastography imaging mode was initiated, and a sample box was created. A color map of shear wave propagation was created in real-time within the box, and a 1 cm region of interest (ROI) was positioned inside to indicate the location of measurement (Figure 1). Qualifications for an appropriate ROI were adopted from the “Recommendations for performing liver elastography” described by Ferraioli et al., and modified for veterinary patients.¹⁶ These qualifications included adhering to a strict protocol, patient fasting for at least 4 h prior to the examination, dorsal recumbency, measurements taken 1.5–2.0 cm below the liver capsule and perpendicular to the capsule, and ROI avoidance of large vessels and bile ducts. Once a desired region of the liver was identified, 10 SWV measurements were obtained by each observer, first by a double-boarded veterinary radiologist (N.R., ECVDI, ACVR), immediately followed by the first-year radiology resident (V.V). The interquartile range (IQR) was measured for each set of measurements by the ultrasound system. Each cat was placed under general anesthesia the day after the initial set of SWV measurements were obtained in the awake state. A complete B-mode ultrasound exam of the abdomen was performed for each cat under general anesthesia by the first-year radiology resident (V.V). Subsequently, the same process was repeated as above for acquisition of the 10 SWV measurements each while under general anesthesia. A similar acoustic window for the SWV measurement of the liver was used for all cats.

FIGURE 1.

Representative longitudinal image of the left liver in a cat in “Elasto” mode. The shear wave box is placed in a region of liver deep to the liver capsule. Including a region void of vessels was difficult due to small patient liver size. Real-time shear wave elasto acquisition was initiated, and a shear wave color image appears, allowing observers to place the ROI in a region of uniform color filling.

2.2.2 |. Anesthetic protocols

Anesthetic and analgesic protocols were individualized for each cat. Seven out of eleven cats received a combined protocol of alfaxalone 2 mg/kg intramuscular route, methadone 0.2 mg/kg intramuscular route, and midazolam 0.2 mg/kg intramuscular route for premedication, and alfaxalone 2 mg/kg intravenous route for induction. Four out of eleven cats received a combined protocol of alfaxalone 2 mg/kg intramuscular route, methadone 0.2 mg/kg intramuscular route, and midazolam 0.2 mg/kg intramuscular route for premedication, and alfaxalone 2 mg/kg intravenous route for induction, with an additional dose of glycopyrrolate 0.008 mg/kg IV during the procedure. All of the cats were maintained on isoflurane at 0.5–1 L/min for the duration of the procedure. All cats received a single dose of buprenorphine 0.03 mg/kg intravenous route after the procedure during the recovery period.

2.2.3 |. Hepatic biopsy

Four to five liver biopsies were collected in all cats using an automatic biopsy instrument (Pro-Mag Ultra^™, Argon Medical Devices, Athens, TX) and a 16ga × 10 cm biopsy needle (Pro-Mag^™ Biopsy Needle, Argon Medical Devices, Athens, TX) using ultrasound guidance. Cats were recovered from anesthesia and monitored for 24 h postprocedure for hemorrhage by brief focused abdominal ultrasound examinations 2 and 4 h postoperatively, and by measuring heart rate, mucous membrane color, and CRT. Tissue samples were fixed in 10% neutral buffered formalin, embedded in paraffin, and processed for routine histopathological analysis. Hematoxylin and eosin and Masson’s trichome-stained sections were evaluated for histomorphology and fibrosis. Glass slides were digitized using NanoZoomer digital slide scanner (Hamamatsu, 2.0-HT, Shimokanzo, Iwata City, Shizuoka Pref., 438–0193, Japan). Image analysis was performed using QuPath (open-source software, v.0.3.2).¹⁷ In brief, an area of interest was defined, and threshold values were iteratively set to quantify percentage of fibrosis compared with total tissue area. Figure 2 shows digitized glass slides of liver zone 3 in cat 2 at approximately 5× magnification, with Masson’s Trichrome (Figure 2A), threshold value set to quantify fibrosis (Figure 2B), and threshold value set to quantify total tissue area (Figure 2C).

FIGURE 2.

Digitized glass slides of histology of liver zone 3 in cat 2 at approximately 5× magnification, with Masson’s Trichrome (A), threshold value set to quantify fibrosis (B), and threshold value set to quantify total tissue area (C). The blue stain represents regions of fibrosis (A). The light green color distributes over areas of fibrosis for quantification of hepatic fibrosis (B). The yellow color quantifies the liver parenchyma and excludes regions of fibrosis (C).

2.3 |. Statistical analysis

All data analyses were performed by a Master of Applied Statistics and statistician from the Louisiana State University Department of Veterinary Clinical Sciences (C.L.) using commercial software (JMP Pro 16.1.0, SAS Institute., Cary, NC). Median values of 10 readings from 11 cats, two states (anesthetized and awake), and two observers were used for analysis. Variables were determined from the median values and expressed as the mean ± standard deviation. A mixed analysis of variance model was used to analyze median and IQR of velocity with the state as the fixed effect and each animal, observer, and their interaction as the random effects. The normality of residuals from ANOVA models were examined and confirmed via Shapiro–Wilk test. Interrater reliability was evaluated using the intraclass correlation coefficient (ICC). Statistical significance was at P < 0.05.

3 |. RESULTS

3.1 |. Demographics

Eleven purpose-bred, domestic shorthair cats were included in this prospective study (four spayed females and six neutered males). Median age was 9 years (range: 6–10 years). Biochemical liver enzymes and liver function parameters are detailed in Supplement tables (Supplementary Information S1). On B-mode abdominal ultrasound, one cat had mild peritoneal effusion and urinary bladder debris, and one cat had a left renal cortical infarct and a homogeneous, prominent pancreas. No ultrasonographic abnormalities were noted in the hepatic size, parenchyma, vasculature, or biliary system in any of the 11 cats. Four out of 11 cats were vocalizing, and moderate respiratory motion was present in 11/11 cats during measurements in the awake state. The acoustic window for SWV measurement was limited due to small patient size. A ROI void of liver capsule and vasculature could not be achieved in all cats. In two cats, incomplete sets of 5/10 and 7/10 measurements were obtained during the awake state due to the patient’s temperament.

3.2 |. Shear wave velocity measurements

The mean SWV of the 11 cats were 1.47 m/s regardless of being in either an awake or anesthetized state, ranging from 1.13–2.19 m/s for the anesthetized state (SD = 0.24) and 1.17–1.83 for the awake state (SD = 0.18). Figure 3 shows a moderate amount of variation in the SWV measurements in the awake state. Variability decreased in the anesthetized state with significant overlap in values. Despite the increased variability in the awake cats, the SWV values in the anesthetized and awake cats were not significantly different (P = 0.9730). The IQR was 0.24 ± 0.12 m/s for the anesthetized cats and 0.58 ± 0.22 m/s for the awake cats. The IQR of SWVs in the anesthetized and awake cats were significantly different (P < 0.0001). For the four cats that received the addition of glycopyrrolate during their anesthetic event, there was no correlation between the SWV measurements in these cats with the additional drug. The intraclass coefficient between the two observers was 0.66, indicating moderate agreement.

FIGURE 3.

Shear wave velocity values are plotted against each cat in an awake versus anesthetized state, with each observer represented by different colors. The box-and-whisker plot demonstrates a moderate amount of variation in the velocity values in the awake state. Variability decreased in the anesthetized state. No statistical differences were noted between observers.

3.3 |. Correlation to hepatic biopsy results

One cat was excluded from the histologic analysis due to insufficient tissue (<15 portal tracts) and poor quality of the sample. No abnormalities were identified on liver H&E-stained sections in any cat. Based on review of the Masson’s trichrome stain, no cat had clinically significant fibrosis. The percent of fibrous connective tissue quantified on Masson’s trichrome-stained slides ranged from 2.49% to 7.13% with a median of 4.63% of total tissue area (Table 1). The SWV measurements did not correlate with minimal changes in hepatic fibrous connective tissue (P = 0.26).

TABLE 1.

Mean SWV values for observer 1 (N.R.) and observer 2 (V.V.) in awake and anesthetized states (m/s), and fibrosis measurements (%) for population of 11 clinically healthy cats using QuPath, Open Source software for digital pathology and whole slide image analysis (JavaFX).

Cat	Mean SWV awake state (m/s)	Mean SWV anesthetized state (m/s)	% Fibrosis
1	1.27	1.42	3.09
2	1.40	1.27	5.98
3	1.77	1.55	5.84
4	1.44	1.33	1.85 (excluded)
5	1.34	1.37	5.58
6	1.44	1.37	3.65
7	1.49	1.37	5.34
8	1.39	1.34	7.13
9	1.45	1.66	3.34
10	1.57	1.33	2.49
11	1.57	2.04	4.63

4 |. DISCUSSION

Shear wave elastography is a relatively new ultrasonographic technology that quantitatively assesses tissue stiffness. In this study, we investigated the reproducibility of this emerging diagnostic imaging technique. There was moderate reproducibility between both observer’s measurements (ICC = 0.66). The moderate agreement between observers indicates that 2D-SWE is easy to learn, reliable, and repeatable method for noninvasive assessment of tissue elasticity. This study demonstrated that a first-year diagnostic imaging resident (V.V.) was able to reasonably replicate values obtained by an experienced ECVDI- and ACVR-board-certified radiologist (N.R.), emphasizing that limited training is required to get meaningful data. In addition, the process of obtaining SWV measurements became easier with each cat, indicating that interobserver agreement may have been higher with a larger population due to a progressive learning curve. Elastography is an emerging imaging technique in veterinary medicine. Currently, elastography settings are not standardized across all ultrasound machines that offer this feature which may contribute to differences among users, in combination with the inherent novelty of the technology.

As expected, there was greater variability in SWVs in awake (range 1.17−1.83 m/s) compared with anesthetized cats (range 1.13–2.19 m/s). This supports the use of a standard sedation protocol when performing elastography to limit measurement variability. However, the mean SWV was 1.47 m/s (SD = 0.25) in both groups, indicating that accurate measurements were still obtained in most awake cats. This is in contrast to a previous study that demonstrated that median 2D-SWE measurements of the liver were significantly higher in anesthetized dogs compared with awake dogs.¹⁶ Performing SWE on fully awake cats mimic daily clinical situations. Our result of a mean SWV 1.47 m/s for both groups indicates that if sedation cannot be used due to hemodynamic concerns or present comorbidities for a cat, a reliable SWV measurement of the liver can be obtained as long as a strict protocol according to the World Federation for Ultrasound in Medicine and Biology for obtaining the measurements are followed.¹⁸ In addition, the use of glycopyrrolate in four cats did not have any statistically significant impact on hepatic stiffness. When comparing the velocities obtained from cats that received a dose of glycopyrrolate to mean SWV in the anesthesia group, the authors found no significant difference between groups (P = 0.2991). The IQR for the anesthetized and awake cats was 0.24 and 0.32, respectively. An IQR of <0.30 indicates a reliable data set.¹⁸ The IQR values of 0.24 and 0.32 for the anesthetized and awake cats respectively in this study indicate that a mean velocity of 1.47 m/s (SD = 0.21) can be used as a reliable reference for future shear wave elastography measurements of feline livers.

Although liver histology is the gold standard for the diagnosis and grading of hepatic fibrosis, biopsy is not always feasible due to inherent risks and limitations, such as bleeding, anesthesia, and cost.³ This is the first study comparing 2D-SWE of the liver in clinically healthy cats to hepatic histology. The present study did not demonstrate a correlation between two-dimensional shear wave velocities of the liver and the quantity of hepatic fibrosis in a sample size of 10 clinically healthy cats. Shear wave elastography measurements greater than 2.1 m/s correlate to advanced chronic liver disease in humans.⁹ Although 2D-SWE is effective for the assessment of severe liver fibrosis in human patients, 2D-SWE cannot be used to discern mild liver fibrosis.¹⁹ No cat in this study had clinically relevant histologic hepatic fibrosis nor evidence of liver disease based on biochemical and diagnostic imaging data (Supplemental Tables S1 and S2). Our findings are also consistent with those in healthy dogs and dogs without clinically relevant hepatic fibrosis.¹¹ Dogs with severe hepatic fibrosis, semi-quantitatively assessed using a human grading rubric, have higher shear-wave velocity measurements (2.04 m/s; range, 1.81–2.26 m/s) than healthy dogs (1.51 m/s; range, 1.44–1.66 m/s) and dogs without clinically relevant hepatic fibrosis (1.56 m/s; range 1.37–1.67 m/s).¹¹ We elected to quantify hepatic fibrosis directly using QuPath (open source software, v.0.3.2), rather than applying a subjective human grading scheme with limited applicability to feline hepatic histology. Further, a quantitative measurement is ideal for the subtle changes that were present in this population of clinically healthy cats. Therefore, as in humans, 2D-SWE cannot detect mild differences in low-grade hepatic fibrosis in cats. Determining a correlation between hepatic SWV measurements and moderate or severe histologic hepatic fibrosis in cats is a topic for further studies.

This study had several limitations. First, the guidelines for performing shear wave elastography that were used in this study recommend selecting a region of liver void of capsule and vessels.¹⁸ This was not possible in the present study due to the small liver size in this feline population and therefore a limited acoustic window. This limitation likely caused variation in the velocity values obtained. Shear wave dispersion is the frequency dependence of shear waves in viscous media and dispersion values have been found to decrease with acquisition depth along with shear wave speeds.²⁰ It is possible that the shallow depth of the ROI’s due to small liver size in this study contributed to the variation of values; however, the evaluation of the effects of depth and viscoelasticity was beyond the scope of this study. Second, shear wave elastography and liver biopsy were performed on a small number of cats. It is possible that a lack of correlation between velocity measurements and histologic fibrosis measurements was due to an inadequate patient population and reduced statistical power. Third, this study used clinically healthy cats. The authors expect that if SWV measurements were to be compared with cats with moderate or severe hepatic fibrosis on histopathology, then a stronger correlation would be identified as it is reported in humans with higher grades of hepatic fibrosis.¹⁹

In conclusion, this study established a reference velocity of 1.47 m/s (SD = 0.21) for hepatic stiffness in a sample of 11 healthy cats using 2D-SWE. In support of this study’s hypothesis, this value is similar to the mean SWV reported in a previous study that measured SWV (1.43 m/s; range 1.35–1.51 m/s) in healthy adult cats using point shear wave elastography.¹⁵ Certain anesthetic agents and premedication drugs, in particular glycopyrrolate, had no impact on hepatic stiffness. Although there was greater variability in the measurements within the awake group, the mean velocity obtained was identical to the anesthetized group, indicating that SWV measurements can be obtained in awake cats with accuracy and with moderate interobserver agreement. The present study shows that 2D-SWE is a feasible and reliable, noninvasive imaging modality that can be performed on groups of awake and anesthetized cats. A correlation between SWV measurements and histologic, clinically insignificant hepatic fibrosis was not supported. Further studies with a larger population and higher grades of clinical or histologic hepatic fibrosis are needed to investigate the relationship between two-dimensional shear wave velocities and fibrosis.

DATA ACCESSIBILITY STATEMENT

Data is available from the corresponding author upon reasonable request.