Paediatric focal nodular hyperplasia: A case study of typical contrast-enhanced ultrasound findings with quantitative analysis and correlated with magnetic resonance imaging

Introduction

Focal nodular hyperplasia (FNH) is a benign liver lesion composed of hepatocytes, occurring in histologically normal liver. FNH is the second most common benign liver lesion after a cavernous haemangioma. It has a prevalence of 0.9–3% and results from an underlying vascular malformation.^1–4 It is often an incidental finding, as the patient is usually asymptomatic. As the lesion is of no clinical significance, a prompt accurate diagnosis can avoid unnecessary intervention and long-term follow-up. In symptomatic cases or cases with atypical imaging appearances, biopsy or surgery may be considered. ¹

With any focal liver lesion, diagnosis primarily relies on imaging, with ultrasound (US) often the first-line technique for the detection and initial characterisation. Conventional US techniques are limited for accurate characterisation of a focal liver lesion, often necessitating further imaging work up with contrast-enhanced CT or MRI. Contrast-enhanced ultrasound (CEUS) is a valuable technique for assessing a focal liver lesion, ⁵ and is well suited and establish the diagnosis of FNH. US contrast agents using harmonic imaging techniques have increased accuracy for the diagnosis of FNH, demonstrating a characteristic enhancement pattern, similar to CT and MRI, but with the advantage of real-time visualisation and superior temporal resolution.^4,6–8 CEUS examination is advantageous, particularly for the paediatric population, as there is no ionising radiation, no deployment of nephrotoxic contrast agents and a more patient-friendly technique compared with both CT and MR imaging. Although the use of US contrast agent for paediatric focal liver lesions has been approved by the U.S. Food and Drug Administration, ⁹ it is still used ‘off-license’ in Europe, driven by the advantages of CEUS in the paediatric population. ¹⁰ The recommended dose of SonoVue™ is calculated according to body weight, 0.03 mL/kg (maximum 2.4 mL per injection). A commonly used dosing scheme for CEUS in children is 0.6 mL for those under six years old, 1.2 mL in children aged between 6 and 12 years and 2.4 mL for those older than 12 years. ¹¹

We present a case of a paediatric FNH where CEUS demonstrated the classic enhancement pattern of this lesion in real-time and correlate these findings with MRI to illustrate the value of real-time assessment of CEUS over MRI, where a limited number of contrast-enhanced phases are acquired. Moreover, the value of quantitative analysis of CEUS is illustrated through the analysis of this case and discussion of the relevant literature.

Case report

A 15-year-old female presented to the emergency department with right upper quadrant abdominal pain, which was sharp, non-radiating, lasted for a few seconds and resolved spontaneously without medication. A clinical history and examination were unremarkable. Routine blood tests were normal. An incidental focal liver lesion was demonstrated on conventional US, measuring 3.0 × 3.3 cm in segment V of the liver (Figure 1(a)). Colour Doppler demonstrated only limited vascularity within the central part of the lesion, with no characteristic pattern (Figure 1(b)).

OPEN IN VIEWER

Figure 1.

B-mode image (a) showing a well-defined slightly hyperechoic focal liver lesion (arrow). Colour Doppler image (b) showing only mild vascularity within the central part of the lesion.

CEUS examination was performed using a standard departmental protocol following the administration of 2.4 mL of SonoVue™ (Bracco SpA, Italy) to further characterise the focal liver lesion, which demonstrated centrifugal enhancement with a characteristic spoke-wheel pattern on the early arterial phase (Figure 2(a)). Late arterial phase showed homogeneous enhancement of the focal liver lesion, higher than that of the adjacent liver parenchyma (Figure 2(b) and (c)). The portal venous phase demonstrated persistent enhancement with no wash out, isoenhancing and inconspicuous compared to the adjacent liver parenchyma (Figure 2(d)).

OPEN IN VIEWER

Figure 2.

Contrast-enhanced ultrasound: early arterial phase imaging demonstrates a spoke-wheel pattern of enhancement within the mass (arrow, a). The subsequent image in arterial and late arterial phase showing centrifugal homogeneous enhancement of the lesion, higher than that of adjacent liver parenchyma (arrow, b and c). Venous phase showing no wash out of the lesion, which now appears isoechoic to adjacent liver parenchyma and is inconspicuous (arrow, d).

The focal liver lesion enhancement profile was objectively quantified using the commercially available software Vuebox® (Bracco Suisse SA, Geneva, Switzerland) which generated the time–intensity curve analysis and parametric images, confirming the subjective observations. The time–intensity curve objectively confirmed the earlier and increased enhancement of the lesion compared with the liver parenchyma, while no wash out was observed in the portal venous phase (after 60 seconds), where the focal liver lesion showed higher enhancement than the liver (see Figure 3(a)). Parametric images were produced showing various enhancement parameters. The parametric image of wash-in rate of the focal liver lesion area objectively showed that the central part of the lesion had a higher wash-in rate than the periphery, in keeping with the centrifugal pattern of enhancement, typical for FNH (Figure 3(b)). This is also confirmed in the parametric image of rise time (Figure 3(c)), where the central part shows a lower rise time and hence quicker enhancement. Although this is straightforward in large lesions, the use of quantitative software might be particularly valuable for small lesions which enhance very quickly and cannot be reliably assessed subjectively.

OPEN IN VIEWER

Figure 3.

Multiparametric CEUS examination findings: screenshots from the quantification software Vuebox®. Image (a) time–intensity curves of the lesion (green curve) and the adjacent liver parenchyma (yellow curve) showing increased early enhancement compared to the liver parenchyma and no wash out in the venous phase. Parametric image (b) showing the wash-in rate of the lesion’s area. The central part of the lesion has a higher wash-in rate than the periphery, in keeping with the centrifugal pattern of enhancement. Parametric image of rise time (c), the central part shows a lower rise time and hence quicker enhancement.

MRI with dynamic gadolinium contrast enhancement was performed after six months as part of a multidisciplinary decision to aid the diagnosis and to assess for a change in lesion size and compression of the adjacent hepatic veins because the patient was symptomatic. MRI did not demonstrate an increase in the size of the lesion (Figure 4(a)) or vascular compromise. Similar to CEUS, dynamic contrast-enhanced T1-weighted sequences showed homogeneous arterial enhancement of the lesion (Figure 4(b)) and persistent enhancement without wash out on the venous and hepatobiliary phase (Figure 4(c) and (d)). Static T2-weighted sequence demonstrated a small central hyperintense scar (Figure 4(e)). Unlike the detailed evaluation achieved by CEUS, the early arterial spoke-wheel pattern of enhancement was not demonstrated on MRI given the limited number and the static nature of images acquired in standard contrast-enhanced phases. The patient had conservative clinical follow-up due to recurring symptoms of abdominal pain. A biopsy was not obtained as the diagnosis was confidently established based on imaging alone.

OPEN IN VIEWER

Figure 4.

Magnetic resonance imaging. Precontrast (arrow, a) demonstrates a 3 × 3.3 cm iso- to mildly hypointense mass in liver segment V, compared to normal liver parenchyma. Arterial phase (arrow, b) demonstrates homogeneous intense enhancement, which is isointense on the venous phase, (arrow, c) and does not wash out on the delayed (hepatobiliary) phase (arrow, d). T2 imaging demonstrates a central hyperintense scar (arrow, e).

Discussion

The imaging appearances of FNH reflect the distinct histological characteristics of this entity. FNH may be difficult to detect on conventional US, being slightly hypo-, iso- or hyperechoic to the adjacent liver and well-circumscribed, as a result of the histological resemblance of this lesion to liver parenchyma. A hypoechoic halo may be seen, corresponding to compressed liver parenchyma. Mild mass effect with blood vessel displacement may be the only suggestion to the presence of a lesion when the FNH is isoechoic to the adjacent liver. Larger lesions may appear heterogeneous in echogenicity. One of the histological hallmarks of FNH is the central scar containing fibrous tissue with malformed arteries. The arteries travel along fibrous septa and divide into smaller capillaries that are connected to sinusoids. Blood drains into the hepatic veins via large venous branches, while FNH does not contain portal veins.

The central scar can be visualised in some cases as a hyperechoic area,^1,4 while CEUS typically appears non-enhancing given that microbubbles cannot pass into the interstitial space from the vascular space, being exclusively an intravascular agent. Colour and power Doppler US techniques may reveal the presence of a central feeding artery, branching in smaller arteries in the centre of the lesion and with a typical stellate or ‘spoke-wheel’ distribution, considered a characteristic finding on flow visualisation techniques. This pattern is estimated to be present in two-thirds of FNH, dependent on the imaging technique’s sensitivity and technical factors. ⁴

MRI detects the FNH central scar and provides information both for tissue characteristics and vascular profile of the lesion. FNH is usually iso- to hypointense on T1 and iso- to hyperintense on T2-weighted images. The central scar is characteristically hyperintense on T2-weighted images. FNH enhances homogeneously and intensely in the arterial phase except for the central scar which enhances in the delayed phase.^1,3

CEUS is being used increasingly for the characterisation of focal liver lesions and is proven useful for the diagnosis of FNH. The distinction between FNH and other benign or malignant hypervascular focal liver lesions is important to avoid unnecessary biopsy, surgery and follow-up. ¹ Contrary to hepatocellular adenoma, blood flow in FNH has a centrifugal direction; a finding readily visualised with CEUS. ¹² FNH may demonstrate the centrifugal pattern of enhancement both subjectively and quantitatively, by analyzing the time–intensity curve and using parametric imaging. Quantitative analysis of CEUS consists of the quantification of signal generated by the microbubbles within a predefined region-of-interest, similarly to what happens with CT or MRI perfusion studies. This can be performed with either software built-in the US device or off-line with commercially available packages such as VueBox® (Bracco Suisse SA, Geneva, Switzerland). Quantitative CEUS has a two-fold application: (i) generation of time–intensity curves where the intensity of the signal returned by the microbubbles is plotted against time and (ii) the generation of parametric images where the various parameters are visualised in colour maps. Parameters typically calculated on this type of analysis include peak enhancement, wash-in rate, rise time and wash out rate. Quantitative analysis achieves the optimal potential of CEUS as it provides objective results with better interobserver agreement and is readily interpretable. It is also very useful in the setting of small or heterogeneous lesions where the pattern of enhancement cannot be readily assessed because of rapid filling of the lesion. Parametric images objectively present detailed information about the enhancement pattern of every lesion. Studies using parametric analysis of CEUS have shown that the use of quantitative variables like the rise time, time to peak and mean transit time have good diagnostic accuracy in diagnosing FNH and is equal to the subjective evaluation by experienced radiologists. It has also been used for differentiating hepatocellular carcinoma from metastases, with good results.^1,7,13,14 Furthermore, quantitative CEUS has been used for the investigation of liver metastases, showing excellent intra- and interobserver reproducibility, as well as for applications beyond the liver such as skeletal muscles affected by systemic sclerosis. Thanks to the technique’s inherent advantages, quantitative CEUS is a promising tool and is endorsed by the European Federation of Societies of Ultrasound in Medicine and Biology, issuing relevant guidelines. ¹⁵

Conclusion

The spatial and temporal resolution of CEUS within its field of view is superior to conventional imaging techniques, including MRI for the characterisation of focal liver lesions. CEUS can be used for the initial diagnosis and follow-up of focal liver lesions, especially in young children with the benefits of being relatively inexpensive, easily accessible and no ionising radiation. It is a useful technique for the differentiation of benign from malignant lesions and has the potential to establish a diagnosis of FNH, based on the typical enhancement pattern, similar to that observed on MRI, but observed in real time and with optimal contrast, spatial and temporal resolution. Using CEUS in the paediatric patient is particularly important in order to ‘image gently’ ¹⁶ aimed at reducing radiation exposure in the imaging of children, with a much better temporal appreciation of the enhancement dynamics of a focal liver lesion. In this case, CEUS could accurately diagnose FNH in the paediatric patient. Nevertheless, it is true that in the setting of a high risk patient, including those with a history of malignancy or chronic liver disease, a more extensive imaging protocol including CT for staging would have to be considered.