The utility of Xenon-133 liver scan in the diagnosis and management of nonalcoholic fatty liver disease

Nonalcoholic fatty liver disease (NAFLD) encompasses a wide spectrum of liver damage and is the most common cause of chronic liver diseases in Western countries. Although a relatively common condition affecting approximately 20% of the general population, NAFLD is especially prevalent in obese individuals, a figure likely to rise as obesity rates in Western countries continue to increase. Liver biopsy remains the gold standard diagnostic method; however, its invasive nature, among other factors, has prompted the need to develop less invasive, alternative methods to quantify hepatic fat and determine disease severity. Xenon-133 liver scanning is one such method that has been in use for more than 10 years in the evaluation of patients with suspected NAFLD. This study compared Xenon-133 liver scan with other currently used, invasive and noninvasive methods of liver assessment.

Abstract

BACKGROUND:

Nonalcoholic fatty liver disease (NAFLD) is an important and common condition affecting approximately 20% of the general population. Given the limitation of radiological investigations, diagnosis often requires a liver biopsy.

OBJECTIVE:

To compare Xenon-133 (Xe-133) liver scanning with ultrasonography in the diagnosis of NAFLD.

METHODS:

From January 2003 to February 2007, 258 consecutive patients with suspected NAFLD underwent Xe-133 liver scanning at Royal Victoria Hospital (Montreal, Quebec). Of these, 43 patients underwent ultrasonography and liver biopsy for the evaluation of NAFLD. Patients with other liver diseases and significant alcohol consumption were excluded. Two nuclear medicine physicians assessed liver Xe-133 uptake and measured the grade of steatosis using a standardized protocol. The degree of steatosis was determined from biopsy specimens assessed by two hepatopathologists.

RESULTS:

NAFLD was identified by liver biopsy in 35 of 43 patients (81.4%). Xe-133 scan demonstrated 94.3% sensitivity (95% CI 81.4% to 98.4%) and 87.5% specificity (95% CI 52.9% to 99.4%) for the presence of NAFLD. The positive and negative predictive values for detection of steatosis by Xe-133 scan were 97.1% (95% CI 85.1% to 99.8%) and 77.8% (95% CI 45.3% to 93.7%), respectively. The positive and negative likelihood ratios were 7.54 (95% CI 1.20 to 47.26) and 0.07 (95% CI 0.02 to 0.26), respectively. Two patients with NAFLD (5.7%) who had a negative Xe-133 scan result had histologically mild steatosis (<10%). The grade of steatosis on liver biopsy was highly correlated with the results of the Xe-133 scan (r=0.87; P<0.001). The sensitivity and specificity of ultrasound in diagnosing steatosis were 62.9% and 75%, respectively.

CONCLUSION:

Xe-133 liver scan proved to be a safe, reliable, non-invasive method for diagnosing and quantifying hepatic steatosis, and was superior to ultrasound.

Nonalcoholic fatty liver disease (NAFLD) is a common clinicopathological condition characterized by mild elevation of liver enzyme levels and significant fat deposition in the hepatocytes of a patient without a history of significant alcohol use (1). NAFLD refers to a wide spectrum of liver damage including simple steatosis. Nonalcoholic steatohepatitis (NASH), with or without cirrhosis (1–3), lies at the most severe end of the NAFLD continuum.

NAFLD is the most common cause of chronic liver diseases in Western countries. NAFLD affects 10% to 39% of the general population in various countries. The prevalence increases to 57.5% (4) to 74% in obese individuals (5,6). NAFLD affects 2.6% of children (7) and 22.5% (7) to 52.8% (8) of obese children. Approximately 26% of adults in the United States are obese and, as this percentage continues to rise, the prevalence of NAFLD is expected to increase concomitantly (9).

NASH is a common explanation for abnormal liver test results in blood donors, and it is the cause of asymptomatic elevation of aminotransferase levels in up to 90% of cases once other causes of liver disease are excluded (10). It is the most common cause of abnormal liver test results among adults in the United States (11).

Recent studies have found that NASH is a strong predictor of cardiovascular disease and may play a central role in the cardiovascular risk of the metabolic syndrome (12).

Several studies have demonstrated that hepatic steatosis – in the absence of inflammation (ie, NAFLD) – does not progress to cirrhosis (2,9). On the other hand, nearly 20% of patients with NASH develop cirrhosis within five to 10 years (2,13,14), and more develop fibrosis slowly, with late-onset cirrhosis (2,14).

Pathological examination of liver specimens remains the gold standard for establishing a diagnosis of NAFLD (15), and is currently the only means of staging the disease and, thus, providing important prognostic information. Serial biopsies are also used to determine the effect of medical treatment (3,15).

Although considered to be the standard, liver biopsy is an invasive procedure associated with morbidity and, rarely, mortality (16). Many hepatologists are reluctant to perform a biopsy on asymptomatic patients. Because of its patchy nature, hepatic steatosis may be missed or underestimated at biopsy, and missing such a diagnosis would result in a delay in diagnosis (17). Furthermore, given the lack of effective medical treatment for patients with NAFLD, a liver biopsy may not be necessary to diagnose asymptomatic patients with clear risk factors (eg, diabetes mellitus, hyperlipidemia and obesity) (2).

Therefore, it is necessary to develop a noninvasive method to quantify hepatic fat, and to determine how this method correlates with histological fat quantification and disease severity.

Ultrasonography often reveals a hyperechoic texture or a bright liver because of diffuse fatty infiltration (1). However, this is a non-specific finding that cannot stage the disease, and should not be used to diagnose NAFLD. Other imaging modalities, such as computed tomography (CT) and magnetic resonance imaging (MRI), are useful in demonstrating hepatic steatosis, at least when fat accumulation is moderate to severe (18–21). A recent study reported the overall sensitivity and specificity for ultrasound (US), CT and MRI in detecting pathologically confirmed fatty liver to be 56% and 82%, respectively (22). Another study showed that magnetic resonance spectroscopy can be used to measure hepatic triglyceride content and, thus, could quantify hepatic fat and grade hepatic steatosis (23,24). However, these tests may underestimate less-severe steatosis and cannot distinguish between simple steatosis and NASH. In addition, CT is associated with radiation exposure, and MRI is costly and not readily available for mass screening. Similar to US, focal fat infiltration may present diagnostic difficulties on CT (25,26). Accordingly, there is a need to develop simple, inexpensive, reproducible, noninvasive and safe tests that can accurately diagnose and grade hepatic steatosis.

Xenon-133 (Xe-133) gas is highly fat-soluble and, therefore, concentrates in fatty tissues. Xe-133 has been used to detect lipoid tumours in various parts of the body (27). Previous studies have demonstrated Xe-133 uptake by fatty liver (28–31). In fact, several investigators have quantified radioactive xenon uptake in fatty livers using either visual grading or hepatic retention ratio and correlated this with the amount of fat in hepatic cells (32–35). Xe-133 gas is inexpensive and safe, and is associated with a very low radiation risk (36).

Since early 2002, Xe-133 liver scanning has been routinely used at the Royal Vitoria Hospital (Montreal, Quebec) for the evaluation of all patients with suspected NAFLD, and the institution is proficient in using this mode of liver scanning and interpreting the results. However, the usefulness of Xe-133 liver scan in the diagnosis and management of NAFLD has not been well studied. Accordingly, the present study was conducted to compare Xe-133 liver scan with US for the diagnosis of NAFLD.

METHODS

Patient selection

A review of the computerized nuclear medicine database for Xe-133 liver scanning revealed 258 consecutive patients who underwent a Xe-133 liver scan as part of their initial work up for NAFLD between January 2003 and February 2007. At the same time, all of these patients underwent ultrasonographic liver examinations. These patients were closely followed, and their baseline demographic information and prediagnosis history were recorded. Of these, 43 patients underwent liver biopsy after Xe-133 and US liver scans. All liver biopsies were interpreted by two experienced liver pathologists, with differences reconciled by consensus.

Approval to review the medical records and images was granted by the institutional review board.

The following criteria for patient enrollment were used:

1. Abnormal liver enzyme levels or clinical features suggestive of diffuse liver disease.

2. Documented history of minimal alcohol intake (≤20 g/day).

3. Other causes of liver disease, such as chronic viral hepatitis, autoimmune hepatitis, hemochromatosis, Wilson disease, alpha-1 antitrypsin deficiency, primary biliary cirrhosis and drug-induced liver disease, were excluded.

Extensive clinical and laboratory data including age, sex, body mass index (BMI) and history of diabetes mellitus or hyperlipidemia were recorded. The laboratory evaluations included a serum hepatic profile (aspartate aminotransferase, alanine aminotransferase, total bilirubin, alkaline phosphatase, total protein and albumin levels, and international normalized ratio), serum glucose, lipid profile, viral serologies for hepatitis B and C, autoantibodies (antinuclear antibody, antismooth muscle antibody, antimitochondrial antibody), serum protein electophoresis, iron profile (ferritin, transferrin saturation), and ceruloplasmin and alpha-1 antitrypsin levels.

BMI was calculated using the standard formula: weight (kg)/(height [m²]). Obesity was defined as a BMI ≥30 kg/m² for both men and women.

Liver US assessment

All patients underwent US examinations as part of their initial evaluation for NAFLD. The examinations were interpreted by one of two experienced US radiologists according to standard examination protocols. High-resolution US systems equipped with 3.5 MHz transducers were used for all patients.

The sonographic criteria for grading diffuse fatty infiltration of the liver was as follows: grade 0 (normal) – normal liver echo texture; grade 1 (mild) – slight, diffuse increase in fine echoes in hepatic parenchyma, normal visualization of diaphragm and intrahepatic vessel border; grade 2 (moderate) – moderate, diffuse increase in fine echoes with slightly impaired visualization of intrahepatic vessels and diaphragm; grade 3 (severe) – marked increase in fine echoes with poor or nonvisualization of the intrahepatic vessel borders, diaphragm and posterior portion of the right lobe.

Xe-133 liver scan assessment

All Xe-133 scans were performed in the same nuclear laboratory and interpreted by two specialists with more than five years experience using this technique, and following a standard protocol.

Xe-133 liver scan was performed using a digital gamma camera (Maxxus F23, GE Healthcare, USA) linked to a minicomputer. The patients rebreathed 155 MBq (5 mCi) of Xe-133 for 5 min from a closed-circuit rebreathing system – the equilibrium phase. The system was then opened to room air and patients began a 10 min washout phase. All counting data were acquired in frame mode (128×128 matrix) at 1 min per frame in the hepatic region, and both lung bases in anteroposterior view with the patient supine for 5 min during the equilibrium phase and at 3 min per frame for 10 min during the washout phase.

The assessment of the Xe-133 liver scan by the nuclear medicine specialist included qualitative assessment based on the visual grading, which is the visualization of hepatic uptake of the Xe-133 and quantitative assessment using hepatic xenon retention compared with the intensity of the lung at the end of the washout phase.

Hepatic steatosis was graded from 0 to 3: grade 0 corresponded to no radioactivity noted in hepatic area at any time, while grade 3 corresponded to radioactivity comparable with the greatest activity in the lung (Figures 1A to 1D).

Figure 1).

Xenon-133 liver scans of normal (A), mild (B), moderate (C) and severe (D) hepatic steatosis

Indication for liver biopsy

All biopsies were performed following evaluation by US and Xe-133 scans of the liver. The indications were either for diagnosis of NAFLD/NASH in patients who continued to show abnormal liver tests despite lifestyle changes, or those whose initial workup was inconclusive. Pathological assessment was performed based on the grading and staging system developed by Brunt et al (37). The degree of fatty infiltration was assessed and graded based on the percentage of involved hepatocytes: mild (up to 33%); moderate (33% to 66%); and severe (>66%, as in Table 1). The grades and stages of NASH were assessed using the same protocol (37). Briefly, fibrosis was staged as follows: stage 0 – no fibrosis; stage 1 – zone 3 perisinusoidal fibrosis, focal or extensive; stage 2 – perisinusoidal fibrosis, focal or extensive, and periportal fibrosis, focal or extensive; stage 3 – bridging fibrosis, focal or extensive; and grade 4 – cirrhosis.

TABLE 1.

Demographic and clinical data of 43 patients who underwent liver biopsy, and Xenon-133 and ultrasound scans

Median age, years (range)	56 (36–80)
Male/female, n/n (% female)	28/15 (34.9)
Body mass index, kg/m2, mean (95% CI)	29 (23–45)
Diabetes, %	23
Histology
Grade of steatosis, n
No steatosis	8
Mild (<33%)	11
Moderate (33%–66%)	15
Severe (>66%)	9
Stage (3–4), n (%)	13 (30.2)

Statistical analysis

The sensitivity, specificity, positive and negative predictive values, likelihood ratios and accuracy of the Xe-133 and US liver scans for prediction of steatosis were calculated, with the assumption that the histological examination is the gold standard for diagnosis of hepatic steatosis. Pearson’s product-moment coefficients were calculated to determine the correlation of the grades of steatosis between histology and Xe-133 liver scan findings.

RESULTS

Clinical and demographic data for the 43 patients analyzed are presented in Table 1. A diagnosis of NAFLD was made in 35 of 43 patients (81.4%) at liver biopsy (Table 2). The other eight patients were found to have no fat on liver biopsy. Xe-133 liver scan was found to have 94.3% sensitivity (95% CI 81.4% to 98.4%) and 87.5% specificity (95% CI 52.9% to 99.4%) (Table 3), compared with 62.9% sensitivity (95% CI 55.8% to 66.9%) and 75.0% specificity (95% CI 44.0% to 92.6%) for US (Tables 4 and 5).

TABLE 2.

Xenon-133 liver scan versus liver biopsy in the diagnosis of steatosis (n=43)

		Liver histology		Total
		Steatosis present	Steatosis absent
Xenon-133 liver scan	Steatosis	33	1	34
	No steatosis	2	7	9
	Total	35	8	43

Data presented as n

TABLE 3.

Performance characteristics for the detection of steatosis on Xenon-133 liver scan (n=43) compared with liver biopsy

		Sensitivity	Specificity	Predictive value		Likelihood ratio (95% CI)
				Positive	Negative	Positive	Negative
Liver histology	Any steatosis (n=43)	94.3 (81.4–98.4)	87.5 (52.9–99.4)	97.1 (85.1–99.8)	77.8 (45.3–93.7)	7.54 (1.20–47.26)	0.065 (0.017–0.26)
	Steatosis >33% (n=32)	100 (92.3–100)	87.5 (64.5–87.5)	96.0 (88.6–96.0)	100 (73.7–100)	8.00 (2.60–8.00)	0.000 (0.000–0.12)

Data presented as % (95% CI) unless otherwise indicated

TABLE 4.

Performance characteristics for detection of steatosis on ultrasound scan (n=43) compared with liver biopsy

		Sensitivity	Specificity	Predictive value		Likelihood ratio (95% CI)
				Positive	Negative	Positive	Negative
Liver histology	Any steatosis (n=43)	62.9 (55.8–66.9)	75.0 (44.0–92.6)	91.7 (81.3–97.5)	31.6 (18.5–39.0)	2.51 (0.995–8.99)	0.49 (0.495–0.36)
	Steatosis >33% (n=32)	80.0 (71.2–85.5)	71.4 (39.9–90.9)	90.9 (80.9–97.1)	50.0 (28.0–63.7)	2.80 (1.19–9.44)	0.28 (0.16–0.72)

Data presented as % (95% CI) unless otherwise indicated

TABLE 5.

Ultrasound (US) scan versus liver biopsy in the diagnosis of steatosis (n=43)

		Liver histology
		Steatosis present	Steatosis absent	Total
US liver histology	Steatosis	22	2	24
	No steatosis	13	6	19
	Total	35	8	43

Data presented as n

The positive and negative predictive values of Xe-133 liver scan for the detection of steatosis were 97.1% (95% CI 85.1% to 99.8%) and 77.8% (95% CI 45.3% to 93.7%), respectively. The positive likelihood ratio and negative likelihood ratio were 7.54 (95% CI 1.20 to 47.26) and 0.07 (95% CI 0.02 to 0.26), respectively.

The Xe-133 and US scans demonstrated 100% and 80% sensitivity, respectively, in detecting >33% steatosis, with a positive predictive value of 96% and 90.9%, respectively. The two patients with NAFLD who had negative Xe-133 scan results had histologically very mild steatosis (<10%).

In addition, Xe-133 scan was highly accurate in detecting steatosis in obese patients (n=17, BMI ≥30 kg/m²), with sensitivity of 93.75% and specificity of 100% (Table 6).

TABLE 6.

Steatosis in Xenon-133 scans versus liver biopsy in 17 patients with body mass index ≥30 kg/m²

Liver biopsy	Sensitivity	Specificity	Predictive value
			Positive	Negative
Steatosis	93.75	100	100	50

Data presented as %

There was strong correlation between grades of steatosis using histology and Xe-133 scan (r=0.87; P<0.001) (Figure 2). All patients with advanced fibrosis (stage 3 or 4) (30.2%) were positive on Xe-133 scan, and all patients (76.7%) with moderate or severe steatosis on Xe-133 had histologically defined NASH (Figure 3).

Figure 2).

Correlation of grade of steatosis on Xenon-133 (Xe-133) scan versus biopsy in 43 patients (r=0.87; P<0.001). No Number

Figure 3).

Comparison of grade of steatosis on Xenon-133 (Xe-133) scan versus grade of nonalcoholic steatohepatitis on biopsy in 43 patients. No Number

DISCUSSION

Despite the inherent limitations of retrospective studies, the present study is the first to examine the utility of Xe-133 scan in the evaluation of NAFLD patients. It should be noted that the population analyzed in our study was representative of a highly selected cohort typical of a specialized tertiary care referral centre, and that the conclusions of the study cannot be applied to populations encountered in primary care settings. In addition, the relatively small number of subjects recruited reflected the difficulty of obtaining liver biopsy samples in otherwise asymptomatic patients, and the trend toward avoiding liver biopsy in the absence of a definitive treatment for NAFLD.

In the present study, we found that Xe-133 liver scan is an accurate test for the quantification of hepatic fat and the diagnosis and grading of steatosis, and is superior to US. Although no CT scan or MRI data were available for comparison, Xe-133 scan is more accurate for diagnosis of hepatic steatosis than historically reported for these other two modalities (20). The present study showed that Xe-133 scanning could reliably rule in or rule out the presence of moderate to severe hepatic steatosis. Despite this, and compared with other imaging modalities, Xe-133 scan was more accurate in diagnosing mild grades of steatosis.

Because the majority of NAFLD patients are obese, the accuracy of US and noncontrast-enhanced CT in diagnosing steatosis was significantly worse with increasing BMI (20,36). In the present study, Xe-133 scan was able to detect steatosis in this population with high accuracy. In addition, all patients with steatosis and advanced fibrosis (stage 3 to 4) had a positive Xe-133 scan, unlike US, in which steatosis and fibrosis can have similar appearance (38–41).

No studies have examined the degree of intra- and inter-observer agreement in the visual grading of Xe-133 retention at the end of the washout phase, which is the method used at our institution. This may be a limitation of our study. However, a previous report showed very high degree of intra- and inter-observer agreement on the severity of steatosis using the Xe-133 hepatic retention ratio method (34).

One major limitation of Xe-133 scan is that it only detects fat; therefore, it is not expected to distinguish between different subtypes of NAFLD, nor provide information about liver morphology. Therefore, combining Xe-133 scan with US is a useful, noninvasive method of establishing the diagnosis of NAFLD. Although it correlated well with the histological grade of NASH, it does not predict the stage well. It may be useful to assess how it performs when combined with either US or perhaps with newer noninvasive methods, such as tissue elastography or biochemical assays for the assessement of fibrosis. It is probable that individuals with an abnormal Xe-133 scan and non-invasive positive test for fibrosis would be the most likely to benefit from a liver biopsy. This should be assessed in future studies.

SUMMARY

Xe-133 liver scan is inexpensive, simple, readily available, safe and a more accurate method of diagnosing hepatic steatosis compared with other imaging modalities. The value of a liver biopsy for the diagnosis of NAFLD in routine clinical practice remains controversial, especially in the presence of a generally good prognosis for most patients with NAFLD, the lack of an established form of effective therapy, and the risks and costs associated with the biopsy. Therefore, once ongoing alcohol use (≥20g/day to 30 g/day) and other common causes of liver disease are excluded by clinical and laboratory evaluation, Xe-133 scan with ultrasound scan may be a useful noninvasive way to establish a diagnosis of NAFLD.