Contrast-enhanced ultrasonography findings using a perflubutane-based contrast agent in patients with early hepatocellular carcinoma

European Journal of Radiology 83 (2014) 95–102

Contents lists available at ScienceDirect

European Journal of Radiology journal homepage: www.elsevier.com/locate/ejrad

Contrast-enhanced ultrasonography findings using a perflubutane-based contrast agent in patients with early hepatocellular carcinoma Kazushi Numata a,∗ , Hiroyuki Fukuda a , Haruo Miwa a , Tomohiro Ishii a , Satoshi Moriya a , Masaaki Kondo a , Akito Nozaki a , Manabu Morimoto a , Masahiro Okada b , Shigeo Takebayashi c , Shin Maeda d , Akinori Nozawa e , Masayuki Nakano f , Katsuaki Tanaka a a

Gastroenterological Center, Yokohama City University Medical Center, 4-57 Urafune-cho, Minami-ku, Yokohama, Kanagawa 232-0024, Japan Department of Radiology, Kinki University Faculty of Medicine, 377-2 Ohno-Higashi, Osaka-Sayama, Osaka 589-8511, Japan Department of Radiology, Yokohama City University Medical Center, 4-57 Urafune-cho, Minami-ku, Yokohama, Kanagawa 232-0024, Japan d Division of Gastroenterology, Yokohama City University Graduate School of Medicine, 3-9 Fukuura, Kanazawa-ku, Yokohama, Kanagawa 236-0004, Japan e Department of Pathology, Yokohama City University Medical Center, 4-57 Urafune-cho, Minami-ku, Yokohama, Kanagawa 232-0024, Japan f Pathological Department, Ofuna Chuo Hospital, 6-2-24 Ofuna, Ofuna, Kanagawa 247-0055, Japan b c

a r t i c l e

i n f o

Article history: Received 20 June 2013 Received in revised form 19 September 2013 Accepted 21 September 2013 Keywords: Contrast-enhanced ultrasonography Early hepatocellular carcinoma Contrast-enhanced CT Vascularity

a b s t r a c t Objective: We evaluated the contrast-enhanced ultrasonography (US) imaging features of early hepatocellular carcinomas (HCCs) and compared these findings with those obtained using contrast-enhanced computed tomography (CT). Subjects and methods: Forty-three patients with 52 early HCCs with a mean maximal diameter of 15.6 mm were enrolled in this retrospective study. After confirming the location of the target lesion using fusion imaging combining conventional US and hepatobiliary phase of contrast-enhanced magnetic resonance (MR) imaging with gadolinium ethoxybenzyl diethylenetriaminepentaacetic acid, we evaluated findings of contrast-enhanced US using a perflubutane-based contrast agent. The contrast-enhanced US detection rates for hyper-vascularity in early HCCs were compared with those obtained for contrast-enhanced CT. Results: Transient hypo-vascularity subsequent to iso-vascularity during arterial phase and isovascularity during portal and post-vascular phases were the predominant contrast-enhanced US findings seen for 25 (48.1%) of the 52 early HCCs. Nine (17.3%) showed iso-vascularity during all three phases, while 1 (1.9%) showed hypo-vascularity during all three phases. The remaining 17 (32.7%) showed partial or whole hyper-vascularity during arterial phase, iso-vascularity during portal phase, and iso- or hypovascularity during post-vascular phase. The detection rate for the hyper-vascularity of early HCCs using contrast-enhanced US (32.7%, 17/52) was significantly higher than that obtained using contrast-enhanced CT (21.2%, 11/52) (P < 0.05 by McNemar test). Conclusion: Hypo-vascularity, iso-vascularity, and hyper-vascularity were observed during the arterial phase of contrast-enhanced US in 50.0%, 17.3%, and 32.7% of the early HCCs, respectively. Contrastenhanced US was more sensitive than contrast-enhanced CT for the detection of hyper-vascularity in early HCCs. Of note, early HCCs might not exhibit the early arterial enhancement that is generally considered to be a typical finding for HCCs. © 2013 Elsevier Ireland Ltd. All rights reserved.

1. Introduction In general, hepatocellular carcinoma (HCC) develops in a multistep fashion from a dysplastic nodule (DN) to early HCC and, finally,

∗ Corresponding author. Tel.: +81 45 261 5656; fax: +81 45 261 9492. E-mail addresses: [email protected], [email protected] (K. Numata). 0720-048X/$ – see front matter © 2013 Elsevier Ireland Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ejrad.2013.09.025

to advanced HCC [1–3]. The characteristic imaging appearance of advanced HCC is a hyper-vascular lesion with washout during the portal venous phase or delayed phase of contrast-enhanced computed tomography (CT) and a well-defined margin on conventional ultrasonography (US). However, most early HCCs appear as hypovascular or iso-vascular lesions using contrast-enhanced CT [4] and have ill-defined margins on conventional US [2,5]. Advanced HCC lesions often show microvascular invasion, despite their small size [6]. The presence of microvascular invasion suggests that the

96

K. Numata et al. / European Journal of Radiology 83 (2014) 95–102

prognosis of these lesions may be less favorable than that of early HCC. Early HCC includes invasion of the portal tract but not invasion of the portal vessel [2,6]. Curative treatment, such as ablation or resection, in patients with early HCC remains the best hope for longterm survival. Therefore, the ability to detect and diagnose early HCC is essential. Conventional US is the recommended modality for the surveillance of HCC in patients with chronic liver disease. Especially, patients with cirrhosis caused by infection with hepatitis B or C should be screened at 6-month intervals [7]. However, in cases with advanced cirrhosis, conventional US detects many hypo- or hyper-echoic areas with ill-defined margin in the liver parenchyma [8,9], making the targeting of early HCC lesions difficult. Therefore, few reports have focused on the imaging characterization of early HCC using contrast-enhanced US [10–12]. Recently, “fusion imaging”, which fuses US images with multiplanar reconstructed CT or magnetic resonance (MR) images on a single screen in real time has become available [13,14]. Fusion imaging combining conventional US and the hepatobiliary phase of contrast-enhanced MR imaging using gadolinium ethoxybenzyl diethylenetriaminepentaacetic acid (Gd-EOB-DTPA; Primovist; Bayer Schering Pharma AG, Berlin, Germany) as the reference standard can help to select targeted HCC lesions, especially early HCC lesions that appear as hypo- or hyper-echoic lesions on conventional US [13]. Biopsies from these early HCC lesions can be performed precisely under guidance using this fusion imaging [13]. After confirmation of the location of the target lesion, the vascularity of each hepatic lesion can then be evaluated using contrast-enhanced US. Contrast-enhanced US using a perflubutane microbubble contrast agent (Sonazoid® ; Daiichi Sankyo, Tokyo, Japan) in the low mechanical index (MI) mode is more sensitive than gadoxetate disodium-enhanced MR imaging for the assessment of the arterial hyper-vascularity of hepatic nodules, including HCC and DN [15]. Moreover, Sonazoid-enhanced US in the high MI mode allowed us to evaluate lesions located deep within the liver as well as hyper-echoic lesions using two-dimensional or threedimensional US [16]. In the present study, after confirming the location of the target lesions using fusion imaging (combining conventional US and the hepatobiliary phase of contrast-enhanced MR with Gd-EOB-DTPA), we evaluated the contrast-enhanced US findings for early HCC and compared these findings with those obtained using contrast-enhanced CT.

2. Subjects and methods 2.1. Patients Institutional review board approval and informed consent from all the patients were obtained for this retrospective study. At our institution, 122 consecutive patients with early HCC underwent conventional US, contrast-enhanced MR with Gd-EOB-DTPA, contrast-enhanced CT, fusion imaging combining conventional US and the hepatobiliary phase of contrast-enhanced MR with GdEOB-DTPA, contrast-enhanced US, and a tumor biopsy between January 2010 and December 2012. Hepatocyte uptake of Gd-EOBDTPA is poor in patients with severe liver dysfunction, making it difficult to detect HCC lesions during the hepatobiliary phase of contrast-enhanced MR with Gd-EOB-DTPA. Therefore, we excluded patients with Child–Pugh grade C liver cirrhosis. Iso-intense lesions observed during the hepatobiliary phase of contrast-enhanced MR with Gd-EOB-DTPA and iso-echoic nodules on conventional US were also excluded because of the difficulty in achieving adequate fusion images to provide guidance for biopsies of the lesions. Finally, 43 patients with 52 hepatic lesions were diagnosed as early HCCs and were eligible for inclusion in the present study.

Table 1 Clinical characteristics of the subjects with early HCC lesions enrolled in this study. Characteristics No. of patients Single lesion/two lesions/three lesions Age (mean, range, years) Sex Male/female Etiology of HCC Hepatitis B/Hepatitis C/Hepatitis B and C/Alcohol abuse Child–Pugh classification Class A/B Diameters of lesions (mean, range, mm) (n = 52)

43 35/7/1 70, 55–84 28/15 4/37/1/1 40/3 15.6, 10–28

The patient population included 28 men and 15 women (mean age, 69.7 ± 7.9 [SD] years; age range, 55–84 years). The mean maximal diameter of the early HCCs was 15.6 ± 4.0 mm (range, 10–28 mm). Forty patients had Child–Pugh class A cirrhosis of the liver, and 3 patients had Child–Pugh class B cirrhosis. The cirrhotic background consisted of hepatitis B virus infection in 4 patients, hepatitis C virus infection in 37 patients, both hepatitis B and C virus infection in 1 patient, and alcohol abuse in 1 patient (Table 1).

2.2. Pathological diagnosis After confirming the location of the HCC lesions using fusion imaging combining conventional US and the hepatobiliary phase of contrast-enhanced MR with Gd-EOB-DTPA, at least 2 samples were taken from each of the nodules using a 21-gauge fine needle biopsy (SONOPSY; Hakko, Tokyo, Japan) or an 18-gauge biopsy needle (Biopty-Cut; Bard, Covington, GA) under US guidance to ensure an accurate histological diagnosis [13]. A biopsy of the adjacent liver tissue was also performed to compare the histological findings between the hepatic nodules and the adjacent liver. The International Consensus Group for Hepatocellular Neoplasia (ICGHN), which includes pathologists and clinicians from 13 countries, recently arrived at a consensus regarding the pathologic criteria for early HCC [2,17]. The ICGHN stated that the presence of stromal invasion, tumor cell invasion into the intratumoral portal tracts [18], should be recognized as the most important pathologic finding for the diagnosis of early HCCs. First, the presence of hypercellularity, characterized by an increased nuclear cytoplasmic ratio and cell atypia (such as deformity of the nuclei), was evaluated using hematoxylin–eosin staining (Figs. 1I and 2J), and architectural alterations in the thin trabecular structure and acinus were evaluated using silver impregnation. The diagnosis of stromal invasion is subjective and may require the assistance of Victoria blue staining (Figs. 1J and 2K) [19] and immunohistochemical stains (cytokeratin 7) [19,20] to differentiate it from pseudo-invasion. All the lesions were diagnosed strictly according to the ICGHN pathologic criteria by a liver pathologist who was blinded to the clinical data (A.N., the chief pathologist at our institution). A second liver pathologist (M.N., auditing pathologist), who is a member of the ICGHN, was asked to diagnose the lesions independently. Six lesions with hyper-cellularity and architectural alterations in the thin trabecular structure and acinus, but no proof of stromal invasion despite the use of Victoria blue staining to clarify the presence of stromal invasion, were excluded from the study. After a consensus meeting, 43 patients with 52 lesions were diagnosed as having early HCC. The histological diagnoses of the adjacent liver tissue were chronic hepatitis in 17 of the 43 patients and cirrhosis in the remaining 26 patients.

K. Numata et al. / European Journal of Radiology 83 (2014) 95–102

97

Fig. 1. A 73-year-old man with an early HCC lesion (maximum diameter, 14 mm) in segment III; (A) fusion image combining conventional US (left side) and the hepatobiliary phase of contrast-enhanced MRI with Gd-EOB-DTPA (right side). The hepatobiliary phase of contrast-enhanced MRI with Gd-EOB-DTPA shows a hypo-intense area in segment III (arrowhead). The fusion image shows the targeted early HCC lesion, which appears as an ill-defined hypo-echoic lesion on conventional US (arrowhead). Thus, we are able to perform precise biopsies of this lesion under the guidance of the fusion image. A pathological examination revealed an early HCC; (B) fusion image combining conventional US (left side) and the arterial phase contrast-enhanced CT (right side). A lesion corresponding to the hypo-echoic lesion seen using conventional US (arrowheads) appeared as an iso-attenuation area during arterial phase contrast-enhanced CT. This lesion also appeared as an iso-attenuation area during the hepatic venous and equilibrium phases; (C) 3D ultrasound images in coronal plane, which can be translated from front to back, with a slice distance of 1.5 mm on arterial phase contrast-enhanced 3D US showing hypo-vascularity in the central areas of the lesion (arrows); (D–E) arterial phase contrast-enhanced US images show transient hypo-vascularity in the central areas of the lesions (arrows); (F–H) arterial, portal, and post-vascular phase contrast-enhanced US images show continuous iso-vascularity throughout the lesion. The arrowheads seen in Figs. C–H indicate the margin of the lesion; (I) hematoxylin–eosin staining reveals slight hyper-cellularity with increased nuclear cytoplasmic ratio and deformed nuclei. However, stromal (portal tract) invasion cannot be seen clearly; (J) victoria blue staining, showing elastic fibers surrounding the portal tract in blue, reveals stromal (portal tract) invasion compatible with a diagnosis of early HCC. The arrowheads indicate the portal tract, and cancer cells are present within the portal tract.

98

K. Numata et al. / European Journal of Radiology 83 (2014) 95–102

2.3. Imaging methods 2.3.1. Contrast-enhanced CT images Contrast-enhanced CT imaging was performed using two devices: (1) a 16-channel multidetector scanner (Aquilion 16; Toshiba Medical, Tokyo, Japan) with a tube voltage of 120 kV, a tube current set at the automatic milliampere exposure setting, a reconstruction section and an interval thickness of 5 mm, a pitch of 15, and a gantry speed of 0.5 s per rotation; and (2) an 80-channel multidetector scanner (Aquilion PRIME; Toshiba Medical, Tokyo, Japan) with a tube voltage of 120 kV, a reconstruction section and interval thickness of 5 mm, a pitch of 65, and a gantry speed of 0.5 s per rotation. A nonionic contrast agent (iopamidol [Iopamiron 300 or 370; Bayer HealthCare, Osaka, Japan]) was injected. Patients weighing less than 70 kg received 300 mg of iodine per milliliter, whereas those weighing 70 kg or more received 370 mg of iodine per milliliter. After a power injector (Dual Shot GX; Nemoto Kyorindo, Tokyo, Japan) was used to inject 100 mL of iopamidol at 3 mL/s through a catheter placed in the antecubital vein, the scanning time in the arterial phase was confirmed using an automatic bolus-tracking program (RealPrep; Toshiba Medical). The trigger point for starting arterial phase scanning was set at an attenuation of 230 HU from the baseline attenuation of the abdominal aorta at the celiac artery level and an additional start delay of several seconds (Aquilion 16, 10 s; Aquilion PRIME, 12 s). Hepatic venous phase scanning was performed 70 s after contrast injection, and equilibrium phase images were acquired 180 s after injection. 2.3.2. MR imaging MR imaging was performed using a 1.5-T whole-body imager (Avant; Siemens Medical System, Erlangen, Germany). Only hepatobiliary phase images were obtained after an i.v. bolus injection of 0.1 ␮mol/kg of Gd-EOB-DTPA flushed with 10 mL of sterile saline solution from the antecubital vein. The scan delay time was 20 min after the initiation of contrast injection, and the images were obtained using fat-suppressed volumetric interpolated breath-hold examination (FS VIBE) T1-weighted sequences (TR, 6.2 ms; TE, 3.15 ms; flip angle, 20◦ ; band width, 260 Hz/pix; matrix, 166 × 320; acquisition time, 20 s) and a fast low angle shot (FLASH) T1weighted sequence (TR, 115 ms; TE, 4.76 ms; flip angle, 70◦ ; band width, 260 Hz/pix; matrix, 192 × 256; acquisition time, 20 s × 3). In addition, turbo spin-echo (TSE) pace respiratory-triggered T2-weighted sequences and echo planar imaging (EPI) diffusionweighted sequences were also obtained. 2.3.3. US imaging 2.3.3.1. Conventional US (B-mode) images. First, we assessed the detection of HCC lesions using the LOGIQ E9 or LOGIQ S8 ultrasound system (GE Healthcare, Milwaukee, WI, USA) with native tissue harmonic gray-scale imaging using a convex probe with a frequency of 2–5 MHz and a micro-convex probe with a frequency of 2–5 MHz (hereafter referred to as conventional US). 2.3.3.2. Fusion imaging combining conventional US and the hepatobiliary phase of contrast-enhanced MR with Gd-EOB-DTPA. The fusion imaging system was composed of a position-sensing unit mounted on an ultrasound unit, a magnetic field transmitter, and two sensors connected to a transducer bracket. The transmitter was placed on a stand so that the area being scanned was within the range of the transmitter. Using a position sensor attached to the transducer, the fusion feature allowed us to import pre-acquired volumetric MR digital imaging and communication in medicine (DICOM) data and to register the ultrasound image with this pre-acquired volumetric MR DICOM data. To fuse the images correctly, a registration procedure is needed. Thus, registration was performed by defining common anatomical planes. First, a common anatomical landmark,

such as the portal vein bifurcation, was identified on the real-time US image and the MR image; these points were then marked on the respective images. After successful registration of the US and MR images, the results of real-time ultrasound scanning could then be viewed simultaneously with the corresponding multiplanar reconstruction slice from the pre-acquired volumetric MR DICOM data. The procedures described above were also used to perform fusion imaging, in which conventional US and contrast-enhanced CT images were combined. 2.3.3.3. Contrast-enhanced imaging software suitable for the evaluation of hepatic lesions. Contrast-enhanced US using low MI contrast mode at a low MI (0.2–0.3) cannot remove the influence of the hyper-echogenicity of hyper-echoic hepatic lesions, and this hyperechogenicity disturbs the accurate evaluation of the vascularity of hepatic lesions. In particular, whether hyper-echoic lesions exhibit hypo-vascularity during the post-vascular phase cannot be adequately evaluated (Fig. 2G). However, contrast-enhanced US with a high MI contrast mode (coded harmonic angio [CHA] at a high MI [0.7–1.0]) can eliminate the background B-mode findings, such as hyper-echogenicity, enabling the observation of hyper-echoic lesions (Fig. 2H). A low MI contrast mode also cannot evaluate lesions located more than 10 cm from the skin surface because of the attenuation of the ultrasound beam. Thus, we used contrast-enhanced US with a high MI mode to evaluate the vascularity of hyper-echoic lesions or lesions located between 10 and 12 cm from the skin surface [21]. For most lesions located more than 12 cm from the surface, vascularity cannot be evaluated accurately even using a high MI contrast mode. We also used a high MI contrast mode to obtain contrast-enhanced threedimensional (3D) US images using an autosweep 3D function [22,23]. Contrast-enhanced 3D US enabled us to evaluate subtle changes in vascularity throughout a lesion at a glance using multiple parallel planes with an adjustable inter-plane distance in each dimension (Fig. 1C). Contrast-enhanced 3D US can provide three different views, such as three orthogonal dimensions, accompanied by 3D images rendered using the maximum intensity or average intensity combined with the surface mode simultaneously (Fig. 2I). The main imaging software used for the contrast mode of the LOGIQ E9 or LOGIQ S8 utilizes a low MI mode, whereas the main imaging software used for the contrast mode of the LOGIQ 7 utilizes a high MI mode. As mentioned above, in cases with hyper-echoic lesions and lesions located between 10 and 12 cm from the skin surface and in cases that were examined using contrast-enhanced US with a high MI mode or contrast-enhanced 3D US with a high MI mode, immediately after the locations of the HCC lesions were confirmed using LOGIQ E9 or LOGIQ S8, we switched to the LOGIQ7 ultrasound system (GE HealthCare; Milwaukee, WI, USA). Using the LOGIQ7 ultrasound system with a 3.5-MHz convex probe or 4D3CL volume probe, we then re-confirmed the location of the HCC lesions. 2.3.4. Contrast-enhanced US procedures The procedures used for both contrast-enhanced US with a low MI mode and that with a high MI mode were essentially the same. A 0.2-mL dose of Sonazoid was injected into an antecubital vein at 0.2 mL/s via a 24-gauge cannula followed by 2 mL of 5% glucose after the Sonazoid injection. Contrast-enhanced US images were acquired during three contrast phases, consisting of an arterial phase (about 10–50 s after injection), a portal phase (about 80–120 s after injection), and a post-vascular phase (about 10 min after injection). Contrast-enhanced US using a low MI mode at a low MI (0.2–0.3) can provide a real-time evaluation of tumor enhancement at 12 frames per second during all three phases. On the other hand, when contrast-enhanced US was performed using the CHA mode

K. Numata et al. / European Journal of Radiology 83 (2014) 95–102

99

at a high MI (0.7–1.0), the lesion was scanned at 2–8 frames per second after the injection. We usually used 8 frames per second to observe the tumor vessels by eliminating microbubbles in the microvessels but not in relatively large vessels, such as the tumor vessels and portal veins, thereby prolonging the observation time for the tumor vessels during the arterial phase. A rate of 2 frames per second was used to observe tumor enhancement as a result of microbubble destruction within and around the tumor during the three phases. The transmission power was 70–100%, and the MI ranged from 0.7 to 1.0. Using the CHA mode at a high MI at 2 frames per second with the focus point just beneath the lesion, we manually scanned the whole lesion to destroy any microbubbles within or around the tumor. We called this method “high MI intermittent imaging” [16]. This procedure enabled the evaluation of tumor vascularity in hyper-echoic nodules or lesions located deep within the liver (between 10 and 12 cm from the skin surface) during the three phases [16]. The process of Sonazoid-enhanced 3D US includes two steps: data acquisition and image reconstruction. Thanks to equipment development, a volume probe that automatically scans with an internal sectorial mechanical tilt movement to obtain data is now available. This type of probe provides a convenient and fast means of data acquisition in the contrast-enhanced phase. The scanning parameters, including the volume angle and number of scanning frames, need to be adjusted before the examination. After data acquisition, the raw data are stored on a hard disk, and image reconstruction can be performed at any time. For 3D images, the enhancement of the volume of interest can be visualized in three orthogonal dimensions, and in each dimension, the enhancement of the volume of interest can be presented in multiple parallel planes with an adjustable inter-plane distance. 2.3.5. Image analysis 2.3.5.1. Evaluation of echogenicity of early HCC using conventional US. The echogenicity of the lesions was noted as being either hypo-echoic or hyper-echoic, compared with the surrounding liver parenchyma. We also evaluated the presence and absence of a hypo-echoic rim, mosaic pattern, and a nodule-in-nodule pattern. 2.3.5.2. Evaluation of intensity of early HCC during hepatobiliary phase of contrast-enhanced MR imaging with Gd-EOB-DTPA. The signal intensity of the lesions was recorded as hypo-intense or hyper-intense, compared with the surrounding liver parenchyma. 2.3.5.3. Evaluation of vascularity of early HCC on contrast-enhanced CT. We defined arterial enhancement as being present when a lesion showed high-attenuation on postcontrast arterial phase images, compared with precontrast images. In partial highattenuation-type early HCCs, part of the lesion (less than half)

Fig. 2. A 78-year-old man with an early HCC lesion (maximum diameter, 10 mm) in segment V; (A–D) non-enhanced CT, arterial, hepatic venous, and equilibrium phase contrast-enhanced CT images show a low-attenuation area compared with the surrounding liver parenchyma (arrowhead); (E) fusion image combining conventional US (left side) and hepatobiliary phase of contrast-enhanced MRI with Gd-EOB-DTPA (right side): an HCC lesion located in segment IV is easily recognizable as a hypo-intense area using the hepatobiliary phase of contrast-enhanced

MRI with Gd-EOB-DTPA (black arrowhead). The fusion images enabled conventional US to detect this hyper-echoic early HCC lesion easily (white arrowhead); (F) arterial phase contrast-enhanced US with a high MI mode image shows hyper-vascularity throughout the lesion; (G) post-vascular phase contrast-enhanced US with a low MI mode image shows iso-vascularity throughout the lesion; (H) post-vascular phase contrast-enhanced US with a high MI mode image shows hypo-vascularity throughout the lesion. The arrowheads seen in Figs. F–H indicate the margin of the lesion; (I) 3D ultrasound images in the post-vascular phase show hypo-vascularity throughout the lesion in coronal plane, which can be translated from front to back (lower right), in sagittal plane from left to right (upper right), and in transverse plane from up to down (lower left), and accompanied by 3D images rendered using the average intensity and the surface mode (upper left). The arrowhead indicates the lesion; (J) hematoxylin–eosin staining reveals slight hyper-cellularity with an increased nuclear cytoplasmic ratio, deformity of the nuclei, and stromal invasion (arrowheads); (K) Victoria blue staining shows stromal (portal tract) invasion compatible with a diagnosis of early HCC. Compared with hematoxylin–eosin staining, Victoria blue staining reveals elastic fibers, enabling the portal tract to be discriminated more clearly, as shown by the blue lines. The arrowheads seen in Figs. J and K indicate the portal tract, and cancer cells are present within the portal tract.

100

K. Numata et al. / European Journal of Radiology 83 (2014) 95–102

showed high-attenuation compared with the surrounding liver parenchyma, indicating a partially increased intranodular arterial blood supply. In high-attenuation-type early HCCs, the whole lesion or a large part of it (more than half) showed high-attenuation compared with the surrounding liver parenchyma, indicating an increased intranodular arterial blood supply. Attenuations of the whole lesions were also recorded as low- or as iso-attenuation compared with the surrounding liver parenchyma on the three phases of contrast-enhanced CT images. 2.3.5.4. Evaluation of vascularity of early HCC on contrast-enhanced US. We retrospectively evaluated the images for the enhancement patterns during the arterial phase and classified the patterns into five categories relative to the enhancement pattern in the surrounding liver parenchyma as follows: hypo-vascularity (the whole lesion), transient hypo-vascularity (the whole or a portion of the lesion) and subsequent iso-vascularity (the whole lesion), iso-vascularity (the whole lesion), partial hyper-vascularity (less than half of the lesion), and hyper-vascularity (the whole lesion or more than half of the lesion). In the portal phase, the enhancement patterns of the lesions were classified into two categories as follows: hypo-vascularity (the whole lesion), and iso-vascularity (the whole lesion). The enhancement patterns of the lesion during the post-vascular phase were classified into three categories: iso-vascularity (the whole lesion), partial hypo-vascularity (hypovascularity in less than half of the lesion and iso-vascularity in the remaining area), and hypo-vascularity (the whole lesion). 2.3.5.5. Image evaluation. The image evaluation was performed independently by four experienced radiologists who were blinded to the final diagnoses; each of the radiologists had at least 5 years of clinical experience performing sonography, CT, and MRI. The first and second radiologists reviewed all the sonographic images including the conventional US, fusion imaging, and contrastenhanced US findings recorded on still images and cine clips. The third and fourth radiologists reviewed all the contrast-enhanced CT and contrast-enhanced MR images using a commercially available viewer system or a picture archiving and communication system (Synapse; Fujifilm Medical, Tokyo, Japan). Each of the two groups of readers met to arrive at a consensus for the image evaluation.

Table 2 Contrast-enhanced CT findings for early hepatocellular carcinoma lesions (n = 52). Arterial phase

Hepatic venous phase

Equilibrium phase

No. of lesions (%)

Iso Low Iso Partial high Partial high Partial high High High

Iso Low Low Low Iso Iso Iso Iso

Iso Low Low Low Iso Low Iso Low

13 (25.0%) 19 (36.5%) 9 (17.3%) 4 (7.7%) 2 (3.8%) 3 (5.8%) 1 (1.9%) 1 (1.9%)

High: high-attenuation (the whole or more than half of the lesion) in the arterial phase; Partial high: partial high-attenuation (less than half of the lesion) in the arterial phase; Iso: iso-attenuation (the whole lesion) in the arterial, hepatic venous, and equilibrium phases; Low: low-attenuation (the whole lesion) in the arterial, hepatic venous, and equilibrium phases

Fifty-one (98.1%) of the 52 HCCs appeared as hypo-intense areas (Figs. 1A and 2E), and 1 (1.9%) HCC appeared as a hyper-intense area. 3.3. Findings of contrast-enhanced CT Table 2 shows the contrast-enhanced CT findings for the early HCCs. Thirteen (25.0%) of the 52 early HCCs appeared as isoattenuation areas during all three phases; these lesions were not detected using contrast-enhanced CT (Fig. 1B). Nineteen (36.5%) appeared as low-attenuation areas during all three phases (Figs. 2A–D). Nine (17.3%) appeared as iso-attenuation areas during the arterial phase and as low-attenuation areas during both the hepatic venous and equilibrium phases of contrast-enhanced CT. Nine (17.3%) appeared as partial high-attenuation areas during the arterial phase. Four of these 9 lesions appeared as low-attenuation areas during both the hepatic venous and equilibrium phases. Two of these 9 lesions appeared as iso-attenuation areas during both the hepatic venous and equilibrium phases. Three of these 9 lesions appeared as iso-attenuation areas during the hepatic venous phase and as low-attenuation areas during the equilibrium phases. The remaining two (3.8%) appeared as whole high-attenuation areas during the arterial phase, iso- (n = 1) or low-attenuation (n = 1) areas during the hepatic venous phase, and low-attenuation areas during the equilibrium phase.

2.4. Statistical analysis The concordance between the contrast-enhanced US and the contrast-enhanced CT images for the detection of hyper-vascularity within the early HCCs was evaluated using the McNemar test. A value of P < 0.05 was considered to indicate a statistically significant difference. The SPSS version 19.0 software package (IBM SPSS, Inc., Chicago, IL, USA) was used for the statistical analysis. 3. Results 3.1. Findings of conventional US Twenty-six (50.0%) of the 52 early HCCs appeared as hypoechoic areas (Fig. 1A) and the remaining 26 (50.0%) early HCCs appeared as hyper-echoic areas (Fig. 2E). One (1.9%) of the 52 early HCC lesions had a hypo-echoic rim. None of them exhibited a mosaic pattern. Three (5.8%) of the 52 early HCC lesions exhibited a nodule-in-nodule pattern. 3.2. Findings of hepatobiliary phase of contrast-enhanced MRI with Gd-EOB-DTPA All 52 early HCCs were recognized during the hepatobiliary phase of contrast-enhanced MR with Gd-EOB-DTPA.

3.4. Findings of contrast-enhanced US Table 3 shows the contrast-enhanced US findings for the early HCCs. Transient hypo-vascularity with subsequent iso-vascularity during the arterial phase and iso-vascularity during the portal and Table 3 Contrast-enhanced US findings for early hepatocellular carcinoma lesions (n = 52). Arterial phase

Portal phase

Post-vascular phase

No. of lesions (%)

Hypo Hypo-isoa Iso Partial hyper Partial hyper Hyper Hyper

Hypo Iso Iso Iso Iso Iso Iso

Hypo Iso Iso Partial hypob Iso Iso Hypo

1 (1.9) 25 (48.1) 9 (17.3) 5 (9.6) 3 (5.8) 6 (11.5) 3 (5.8)

Hyper: hyper-vascularity (the whole lesion or more than half of the lesion) during the arterial phase; Partial hyper: partial hyper-vascularity (less than half of the lesion) during the arterial phase; Iso: iso-vascularity (the whole lesion) during the arterial, portal, and post-vascular phase; Hypo: hypo-vascularity (the whole lesion) during the arterial, portal, and post-vascular phase. a Hypo-iso: transient hypo-vascularity (the whole or a portion of the lesion) with subsequent iso-vascularity (the whole lesion) during the arterial phase. b Partial hypo: the hypo-vascularity area accounts for less than half of the lesion and the remaining area exhibits iso-vascularity during the post-vascular phase.

K. Numata et al. / European Journal of Radiology 83 (2014) 95–102 Table 4 Detection rates of hyper-vascularity for early hepatocellular carcinoma lesions (n = 52) between contrast-enhanced US and contrast-enhanced CT. Lesions

Contrast-enhanced US

Detected Not detected P-valuesa

33% (17/52) 67% (35/52) P < 0.05

Contrast-enhanced CT 21% (11/52) 79% (41/52)

a According to McNemar test comparing the rates of the detection of hypervascularity of early hepatocellular carcinomas using contrast-enhanced US and contrast-enhanced CT.

post-vascular phases were the predominant contrast-enhanced US findings seen for 25 (48.1%) of the 52 early HCCs (Figs. 1C–H). Nine (17.3%) showed iso-vascularity during three phases. One (1.9%) showed hypo-vascularity during all three phases. Eight (15.4%) showed partial hyper-vascularity during the arterial, iso-vascularity during the portal, and iso- (n = 3) or partial hypovascularity (n = 5) during the post-vascular phase. All three patients who had a nodule-in-nodule pattern on conventional US showed partial hyper-vascularity during the arterial, iso-vascularity during the portal, and partial hypo-vascularity during the post-vascular phase. The remaining 9 (17.3%) showed whole hyper-vascularity during the arterial, iso-vascularity during the portal, and iso(n = 6) or hypo-vascularity (n = 3) (Figs. 2F and H) during the post-vascular phase. Thus, hypo-vascularity, iso-vascularity, and hyper-vascularity were observed during the arterial phase of the contrast-enhanced US findings for 50.0%, 17.3%, and 32.7% of the early HCCs, respectively. 3.5. Detection rates of hyper-vascularity in early HCCs for contrast-enhanced US and contrast-enhanced CT The detection rate for the hyper-vascularity of early HCC lesions using contrast-enhanced US (33%, 17/52) was significantly higher than that using contrast-enhanced CT (21%, 11/52) (P < 0.05 by McNemar test) (Table 4). 4. Discussion Early HCC is generally hypo-vascular and has ill-defined margins [2,5]. Thus, it has a somewhat vague outline on conventional US. Recently, Sano et al. reported that the hepatobiliary phase of contrast-enhanced MR with Gd-EOB-DTPA is the only technique that successfully depicts early HCCs, compared with contrastenhanced CT, CT during arterial portography, and CT during hepatic arteriography [4]. To enable the histological diagnosis of hepatic lesions that exhibit unclear margins on conventional US but that are clearly detectable during the hepatobiliary phase of contrastenhanced MR with Gd-EOB-DTPA, we used fusion imaging as a reference [13]. Under the guidance of fusion imaging combining conventional US and the hepatobiliary phase of contrast-enhanced MR with Gd-EOB-DTPA, we were able to perform biopsies of these lesions and to obtain histological diagnoses of early HCC. Only a few reports concerning the radiographic findings of early HCC have been published [4,10–12,24]. Rhee et al. reported that none (0%) of 19 lesions showed typical arterial enhancement in the arterial phase and washout in both the portal venous phase and the equilibrium phase on contrast-enhanced CT, but that 3 (16%) of the 19 lesions showed typical arterial enhancement in the arterial phase and washout in the portal venous, hepatic venous, and equilibrium phase on contrast-enhanced MR with Gd-EOB-DTPA [24]. Kudo et al. reported that 12 (57%) of 21 early HCCs had no increased arterial supply during the pure arterial phase but had portal supply during the portal phase and showed persistent enhancement during the post-vascular phase of Sonazoid-enhanced US [10]. Four (19%) of the 21 early HCCs showed a slightly increased arterial

101

supply during the pure arterial phase and an almost persistent enhancement during the post-vascular phase. The remaining 5 (24%) nodules among the early HCCs presented with partial arterial enhancement within the hypo-vascular nodule during the pure arterial phase followed by an iso-vascular pattern during the portal phase and partial washout within the iso-vascular nodules during the post-vascular phase [10]. Giorgio et al. reported that 12 (86%) of 14 early HCCs showed an inhomogeneous and reticular pattern of hyper-vascularization during the arterial phase and no washout during the portal and late phase using SonoVue (Bracco, Milan, Italy) [11]. According to their figures, the cases with an inhomogeneous and reticular pattern of hyper-vascularization showed hypo-vascularity, compared with the surrounding liver parenchyma [11]. In the present study, transient hypo-vascularity with subsequent iso-vascularity during the arterial phase and isovascularity during the portal and post-vascular phases were the predominant contrast-enhanced US findings seen in 48.1% (n = 25) of the 52 early HCCs. We think that these contrast-enhanced US findings are similar to the ‘inhomogeneous and reticular pattern of hyper-vascularization’ images reported by Giorgio et al. [11]. Early HCCs showing hypo-vascularity during the arterial phase may be due to reductions in both the arterial and portal supplies to the lesion [25]. Overall, 17.3% (n = 9) of the early HCCs showed iso-vascularity during all three phases. Hyper-vascularity during the arterial phase of contrast-enhanced US was observed in 32.7% (n = 17) of the early HCCs. We believe that our results for the hyper-vascularity of early HCCs are similar to those report by Kudo [10]. However, in our study, typical HCC findings, such as hyper-vascularity during the arterial phase with hypo-vascularity during the late phase, were observed in only 5.8% (n = 3) of early HCCs. Early HCCs have variable degrees of arterial and portal venous supply, and these combined alterations may exhibit various contrast-enhanced US findings. Moreover, in the present study, the detection rate for the hyper-vascularity of early HCCs using contrast-enhanced US was significantly higher than that using contrast-enhanced CT. Contrast-enhanced US with its real-time evaluation of tumor vascularity can reliably detect arterial hyper-vascularity, whereas contrast-enhanced CT with a pre-determined scanning delay can miss the timing of arterial hyper-vascularity. Consequently, contrast-enhanced US has the advantage of enabling continuous observation for the detection of subtle changes in vascularity in early HCCs [12]. Using CT during hepatic arteriography, Hayashi et al. demonstrated that borderline lesions with partially increased intranodular arterial supply had a high risk of transforming into typical HCCs within a short period [26]. Therefore, it is important to evaluate the presence of small hyper-vascular areas within the lesions. Contrast-enhanced US is easy to perform, with a low cost and no exposure to ionizing radiation unlike contrast-enhanced CT and CT during hepatic arteriography. If subtle hyper-vascularity within early HCCs can be confirmed using contrast-enhanced US, these early HCCs with hyper-vascularity could be curatively treated using RFA because of the absence of portal vein invasion during this stage. Our study had several limitations. First, the pathological diagnosis of early HCC was based on criteria obtained using needlebiopsied specimens. Therefore, it was difficult to perform a precise biopsy of small hyper-vascular areas visible on contrast-enhanced US images that might have corresponded to dedifferentiated HCC areas. This limitation could have resulted in an underestimation of the histological grades in cases with HCC lesions with internal histological heterogeneity. Second, iso-intense lesions on the hepatobiliary phase of contrast-enhanced MR with Gd-EOB-DTPA and iso-echoic nodules on conventional US were excluded from this study because of the difficulty in obtaining fusion images and in performing biopsies of the lesions. However, Kobayashi

102

K. Numata et al. / European Journal of Radiology 83 (2014) 95–102

et al. reported that few of the iso-intense borderline lesions on the hepatobiliary phase of contrast-enhanced MR with Gd-EOBDTPA progressed to hyper-vascular HCC [27]. Third, we used normal-type contrast-enhanced CT and compared the results with contrast-enhanced US. If a low-tube-voltage had been used for the contrast-enhanced CT studies, we might have detected a higher proportion of hyper-vascular early HCCs [28]. Fourth, the image classification for all the modalities was subjective. Further objective evaluation of contrast-enhanced US is needed to confirm the diagnostic performance for early HCCs. In conclusion, early HCCs showed various vascular findings when observed using contrast-enhanced US. In cases with hypervascular early HCCs, contrast-enhanced US was more sensitive than contrast-enhanced CT for detecting hyper-vascularity within the lesions. We believe that this modality can be used to evaluate the precise vascularity of early HCCs. Early HCCs might not exhibit the early arterial enhancement that is generally considered to be a typical finding for HCCs, and this important exception should be carefully noted. Appendix A. Supplementary data Supplementary material related to this article can be found, in the online version, at http://dx.doi.org/10.1016/ j.ejrad.2013.09.025. Conflicts of Interest Conflicts of Interest The authors and their institutions have no conflicts of interest to report (including financial or personal relationships) that might have inappropriately influenced their actions and that would have occurred within 3 years of the work being submitted References [1] Kudo M. Multistep human hepatocarcinogenesis: correlation of imaging with pathology. J Gastroenterol 2009;44(S19):112–8. [2] International Consensus Group for Hepatocellular Neoplasia. Pathologic diagnosis of early hepatocellular carcinoma: a report of the International Consensus Group for Hepatocellular Neoplasia. Hepatology 2009;49:658–64. [3] Kitao A, Zen Y, Matsui O, Gabata T, Nakanuma Y. Hepatocarcinogenesis: multistep changes of drainage vessels at CT during arterial portography and hepatic arteriography—radiologic-pathologic correlation. Radiology 2009;252:605–14. [4] Sano K, Ichikawa T, Motosugi U, et al. Imaging study of early hepatocellular carcinoma: usefulness of gadoxetic acid-enhanced MR imaging. Radiology 2011;261:834–44. [5] Kojiro M. Focus on dysplastic nodules and early hepatocellular carcinoma: an Eastern point of view. Liver Transpl 2004;10:S3–8. [6] Nakashima Y, Nakashima O, Tanaka M, Okuda K, Nakashima M, Kojiro M. Portal vein invasion and intrahepatic micrometastasis in small hepatocellular carcinoma by gross type. Hepatol Res 2003;26:142–7. [7] Bruix J, Sherman M. Management of hepatocellular carcinoma: an update. Hepatology 2011;53:1020–2.

[8] Numata K, Tanaka K, Kiba T, et al. Nonresectable hepatocellular carcinoma: improved percutaneous ethanol injection therapy guided by CO2 -enhanced sonography. Am J Roentgenol 2001;177:789–98. [9] Numata K, Isozaki T, Ozawa Y, et al. Percutaneous ablation therapy guided by contrast-enhanced sonography for patients with hepatocellular carcinoma. Am J Roentgenol 2003;180:143–9. [10] Kudo M, Hatanaka K, Inoue T, Maekawa K. Depiction of portal supply in early hepatocellular carcinoma and dysplastic nodule: value of pure arterial ultrasound imaging in hepatocellular carcinoma. Oncology 2010;78(S1): 60–7. [11] Giorgio A, Calisti G, di Sarno A, et al. Characterization of dysplastic nodules, early hepatocellular carcinoma and progressed hepatocellular carcinoma in cirrhosis with contrast-enhanced ultrasound. Anticancer Res 2011;31:3977– 82. [12] Kim TK, Lee KH, Khalili K, Jang HJ. Hepatocellular nodules in liver cirrhosis: contrast-enhanced ultrasound. Abdom Imaging 2011;36:244–63. [13] Kunishi Y, Numata K, Morimoto M, et al. Efficacy of fusion imaging combining ultrasonography and hepatobiliary phase of contrast-enhanced MR image with gadolinium-ethoxybenzyl-diethylenetriamine-pentaacetic acid for detecting small hepatocellular carcinoma. Am J Roentgenol 2012;198:106–14. [14] Numata K, Fukuda H, Morimoto M, et al. Use of fusion imaging combining contrast-enhanced ultrasonography with a perflubutane-based contrast agent and contrast-enhanced computed tomography for the evaluation of percutaneous radiofrequency ablation of hypervascular hepatocellular carcinoma. Eur J Radiol 2012;81:2746–53. [15] Sugimoto K, Moriyasu F, Shiraishi J, et al. Assessment of arterial hypervascularity of hepatocellular carcinoma: comparison of contrast-enhanced US and gadoxetate disodium-enhanced MR imaging. Eur Radiol 2012;22:1205–13. [16] Numata K, Luo W, Morimoto M, et al. Contrast-enhanced ultrasound of hepatocellular carcinoma. World J Radiol 2010;28:68–82. [17] Desmet VJ. East-West pathology agreement on precancerous liver lesions and early hepatocellular carcinoma. Hepatology 2009;49:355–7. [18] Nakano M, Saito A, Yamamoto M, Doi M, Takasaki K. Stromal invasion and blood vessel wall invasion in well differentiated hepatocellular carcinoma. Liver 1997;17:41–6. [19] Kobayashi S, Kim SR, Imoto S, et al. Histopathological diagnosis of early HCC through biopsy: efficacy of Victoria blue and cytokeratin 7 staining. Dig Dis 2012;30:574–9. [20] Park YN, Kojiro M, Tommaso LD, et al. Ductular reaction is helpful in defining early stromal invasion, small hepatocellular carcinomas, and dysplastic nodules. Cancer 2007;109:915–23. [21] Takizawa K, Numata K, Morimoto M, et al. Use of contrast-enhanced ultrasonography with a perflubutane-based contrast agent performed one day after transarterial chemoembolization for the early assessment of residual viable hepatocellular carcinoma. Eur J Radiol 2013;82:1471–80. [22] Luo W, Numata K, Morimoto M. Clinical utility of contrast enhanced threedimensional ultrasound imaging with Sonazoid: findings on hepatocellular carcinoma lesions. Eur J Radiol 2009;72:425–31. [23] Luo W, Numata K, Morimoto M, et al. Differentiation of focal liver lesions using three-dimensional ultrasonography: retrospective and prospective studies. World J Gastroenterol 2010;16:2109–19. [24] Rhee H, Kim MJ, Park YN, Choi JS, Kim KS. Gadoxetic acid-enhanced MRI findings of early hepatocellular carcinoma as defined by new histologic criteria. J Magn Reson Imaging 2012;35:393–8. [25] Hayashi M, Matsui O, Ueda K, et al. Correlation between the blood supply and grade of malignancy of hepatocellular nodules associated with liver cirrhosis: evaluation by CT during intraarterial injection of contrast medium. Am J Roentgenol 1999;172:969–76. [26] Hayashi M, Matsui O, Ueda K, et al. Progression to hypervascular hepatocellular carcinoma: correlation with intranodular blood supply evaluated with CT during intraarterial injection of contrast material. Radiology 2002;225:143–9. [27] Kobayashi S, Matsui O, Gabata T, et al. Relationship between signal intensity on hepatobiliary phase of gadolinium ethoxybenzyl diethylenetriaminepentaacetic acid (Gd-EOB-DTPA)-enhanced MR imaging and prognosis of borderline lesions of hepatocellular carcinoma. Eur J Radiol 2012;81:3002–9. [28] Marin D, Nelson RC, Schindera ST, et al. Low-tube-voltage, high-tube-current multidetector abdominal CT: improved image quality and decreased radiation dose with adaptive statistical iterative reconstruction algorithm—initial clinical experience. Radiology 2010;254:145–53.