Contrast-Enhanced Ultrasound of Congenital and Infantile Hemangiomas: Preliminary Results From a Case Series
Abstract
OBJECTIVE. The contrast-enhanced ultrasound (CEUS) imaging features of hepatic vascular tumors in infants, including infantile hemangioma (IH) and congenital hemangioma (CH), are not well reported. Frequent inaccurate use of lesion terminology in the literature has created diagnostic confusion. The purpose of this study is to describe the CEUS features of IH and CH.
MATERIALS AND METHODS. Ten patients, ranging in age from 8 days to 16 months, with hepatic vascular tumors were included for retrospective analysis. Gray-scale ultrasound, color Doppler ultrasound, and CEUS features were reviewed, and interobserver kappa coefficients were calculated. Final diagnoses were clinically determined by a pediatrician with expertise in vascular anomalies except in one patient who underwent surgical excision.
RESULTS. Of the 10 patients, five had CHs and five had IHs. All 10 lesions were hyperenhancing in the early arterial phase. In the portal venous phase, four of five (80%) CHs showed hyperenhancement relative to normal liver parenchyma, whereas four of five (80%) IHs showed isoenhancement. In the late phase, washout of contrast material was seen in three of five (60%) IHs, whereas one IH remained isoenhancing and one IH was hyperenhancing. None of the CHs showed late washout. Interobserver kappa coefficients for CEUS features ranged from 0.60 to 1.00.
CONCLUSION. Except for the CEUS feature portal venous phase enhancement (κ = 0.60), good to excellent (κ = 0.74–1.00) agreement about CEUS features of IHs and CHs was observed. A significant proportion of IHs (60%) showed washout at delayed phase imaging, which has also been reported with malignancies. Recognition of the overlap in imaging appearance of these two entities is vital to preventing misdiagnosis of malignancy.
Imaging with contrast-enhanced ultrasound (CEUS) for the evaluation of focal liver lesions in children has increased in popularity since the U.S. Food and Drug Administration approval of sulfur hexafluoride lipid-type A microspheres (Lumason, Bracco Diagnostics). CEUS has multiple advantages over CT and MRI [1]: CEUS does not use ionizing radiation, is easily portable, and is readily available. Additionally, because patient sedation is rarely necessary, CEUS can usually be performed immediately after traditional ultrasound, sooner than CT or MRI can be scheduled in most cases. Furthermore, the safety profile of CEUS in pediatric patients is excellent, with the largest study of pediatric patients showing very few adverse events, none of which qualified as serious [2]. These advantages have quickly made CEUS the first-line imaging test to definitively diagnose a newly discovered focal liver lesion or multiple similar-appearing lesions at UPMC Children's Hospital of Pittsburgh, rather than MRI.
Despite the increased use of CEUS in the pediatric population, most of the studies in the literature that describe the CEUS appearance of focal liver lesions were performed in the adult population. One of the more common hepatic lesions encountered in the pediatric patient population is infantile hemangioma (IH); congenital hemangioma (CH) is less common than IH. Both of these lesions are vascular tumors that are unique to infants. However, there is often confusion between these vascular tumors and the lesion still colloquially called “hemangioma” in almost all CEUS studies of adults and many reviews of pediatric patients in the literature. Hemangioma is actually best described as a venous malformation based on the current reference standard classification system developed by the International Society for the Study of Vascular Anomalies (ISSVA) [3–5]. The ISSVA classification reflects the significant progress that has been made in the understanding of vascular anomalies since the seminal description by Mulliken and Glowacki [6] and accounts for the histologic and behavioral differences among lesions, including distinguishing between vascular tumors and vascular malformations. Despite this progress in knowledge, there is still confusion regarding the terminology used to describe vascular anomalies, resulting in inaccuracy both in clinical practice and in the research setting [3, 4, 7]. The traditional term “hemangioendothelioma” to describe IH or CH has been abandoned [5]. IH and CH are vascular tumors, whereas venous malformations are not [5]. Most importantly, using the correct terminology is crucial because management of pediatric vascular lesions may hinge on an accurate imaging diagnosis [8, 9].
The purpose of this study is to describe the CEUS appearance of the pediatric hepatic vascular tumors IH and CH using a series of cases in neonates and infants from UPMC Children's Hospital of Pittsburgh. We assess the interreader reliability of CEUS findings and maintain strict adherence to the terminology from the most recent 2018 revision of the classification of vascular lesions from the ISSVA.
Materials and Methods
Patient Selection
This study was an institutional review board–approved retrospective review of CEUS examinations performed between October 1, 2016, and July 25, 2018. A search of the PACS was performed and yielded 67 children who underwent CEUS to evaluate indeterminate focal liver lesions; 10 patients with a final diagnosis of a vascular tumor were included in our study group.
Sonography Technique
All patients underwent a complete diagnostic liver ultrasound examination with gray-scale and color Doppler imaging using standard windows, followed by CEUS. The initial indication for CEUS was to determine lesion benignity or malignancy before any additional clinical or diagnostic imaging workup. Examinations were performed using an ultrasound machine (Logiq E9 or E10, GE Healthcare) with a curved transducer (9 L or C1–6, GE Healthcare) depending on lesion location. CEUS was performed after IV administration of sulfur hexafluoride lipid-type A microspheres using a dose of 0.03 mL/kg, which was followed by a saline flush. All patients were monitored for immediate adverse reactions according to department protocol.
In accordance with American College of Radiology CEUS Liver Imaging Reporting and Data System (LI-RADS) guidelines [10], cinematic images of the lesion in question were recorded for the first 30 seconds, capturing the arterial and portal venous phases of contrast enhancement. Intermittent scanning was then performed with attention to the lesion every 30 seconds until contrast material washed out from the background liver parenchyma, which was a mean of approximately 6 minutes after contrast administration. Recorded images were saved to the institutional PACS.
Imaging Review
Images were deidentified, and two subspecialty-trained board-certified pediatric radiologists with 2 years of liver CEUS experience independently reviewed the gray-scale ultrasound, color Doppler ultrasound, and CEUS images while blinded to the diagnosis and all other clinical information.
Lesion echogenicity relative to background liver on gray-scale, Doppler, and contrast-enhanced images was recorded in addition to various ancillary features. When multiple lesions were present, the lesion in the center of the FOV was selected for evaluation. On gray-scale images, lesions were rated as hyper-, iso-, or hypoechoic relative to background liver parenchyma; on color Doppler images, lesions were rated as hyper-, iso-, or hypovascular compared with background liver parenchyma. Lesion intensity on CEUS during the early arterial, late arterial, portal venous, and delayed phases of enhancement was rated as hyper-, iso-, or hypoenhancing compared with the enhancement of the surrounding liver parenchyma. Ancillary features regarding gray-scale, early arterial, late arterial, portal venous, and delayed phases were also recorded.
Ancillary imaging features were also assessed. Generalized hepatomegaly was determined subjectively by evaluating for the presence of bulbous margins of the liver and extension of the right hepatic lobe beyond the inferior margin of the right kidney. Localized enlargement was determined as enlargement of the liver in the immediate vicinity of a lesion secondary to mass effect. Aortic caliber above and below the origin of the celiac artery was also measured when it was visible at both locations. A caliber change was considered to be significant when the difference of the anteroposterior diameter of the aorta was greater than 2 mm.
After independent imaging interpretation, consensus on lesion intensity was made in all cases.
Reference Standard
The final determination of lesion diagnosis in nine cases was made by a pediatrician who specializes in pediatric vascular anomalies. This was accomplished by review of all clinical imaging study reports, laboratory results (including α-fetoprotein value, complete blood cell counts, and thyroid function test results), clinical notes, ancillary conditions (e.g., heart failure, hypothyroidism, coagulopathy), and any follow-up after initial imaging. In one case, surgical resection of the lesion was performed secondary to congestive heart failure, and the diagnosis was confirmed pathologically.
Statistics
Statistical analysis was performed using the R software package (version 3.4.1, The R Foundation), including the calculation of descriptive statistics, the Mann-Whitney test and interobserver kappa coefficient. For comparing means of continuous variables from two groups, such as lesion size and patient age, a Mann-Whitney test was performed. Interobserver agreement was considered moderate if the kappa coefficient ranged from greater than 0.40 to 0.60; good, if ranging from greater than 0.60 to 0.80; or perfect, if ranging from greater than 0.80 to 1.00 [11]. To test the difference in proportions between groups, we used a Fisher exact test. A significance threshold of p ≤ 0.05 was used.
Results
Recruitment and Contrast Safety
Clinical characteristics including patient age at CEUS and gestational age did not significantly differ between the IH and CH cohorts (Table 1). No adverse reactions to sulfur hexafluoride lipid-type A micro-spheres occurred.
| Characteristic | IH | CH | p |
|---|---|---|---|
| Demographic characteristic | |||
| Patient age (mo), mean ± SD | 4.0 ± 1.9 | 7.3 ± 7.7 | 0.40 |
| Term, > 37 weeks, % of patients | 60 | 80 | 0.58 |
| Lesion, maximum diameter (mm), mean ± SD | 9.5 ± 6.3 | 33.0 ± 30.0 | 0.26 |
| No. of lesions per patient, mean ± SD | 12.7 ± 10.0 | 1.0 ± 0 | 0.05 |
| Clinical characteristic, % of patients | |||
| Cutaneous lesions | 100 | 0 | 0.001 |
| Heart failure | 0 | 20 | 0.99 |
| Consumptive hypothyroidism | 40 | 20 | 0.99 |
| Final clinical outcome | |||
| Observation | 40 | 80 | |
| Treatment with propranolol | 60 | 0 | |
| Surgical resection | 0 | 20 |
Lesion Characteristics and Ancillary Findings
Final clinical diagnosis was consistent with either rapidly involuting CH (n = 5) or IH (n = 5). There were no cases of malignancy, including no cases of angiosarcoma, hepatoblastoma, or metastasis. The mean number of hepatic lesions per patient was higher in the IH group (mean number of lesions, 12.7 ± 10.0 [SD]; range, 1–50 lesions) than the CH group (mean number of lesions, 1.0 ± 0 lesion; range, 1 lesion per patient) (p = 0.05). Mean lesion diameter was higher in the CH group (33.0 ± 30.0 vs 9.5 ± 6.3 mm) (p = 0.20).
The most common gray-scale appearance of IH was multiple hypoechoic lesions for which there was 100% reader agreement. IHs were hypervascular at color Doppler imaging (80%) or were similar to the background liver (20%). At CEUS, there was rapid peripheral discontinuous or heterogeneous early arterial enhancement with rapid complete fill-in; five of five (100%) IHs showed hyperenhancement. All IH cases showed sustained enhancement in the portal venous phase with no evidence of contrast washout. Consensus determination was that four of five (80%) IHs were isoenhancing and one (20%) was hyperenhancing in the portal venous phase. In the delayed phase, consensus determination was that one (20%) IH was hyperenhancing, one (20%) was isoenhancing, and three of five (60%) IHs lesions showed contrast washout. Of the IH lesions with washout, two of five (40%) IHs had mild washout, and one (20%) IH showed moderate delayed phase washout. None of the IH lesions had marked washout. Figure 1 shows the gray-scale and CEUS appearances of a patient with IH. Video S1, which can be viewed in the AJR electronic supplement to this article available at www.ajronline.org, shows the enhancement pattern of this case of IH.
Fig. 1A —5-month-old female infant who underwent contrast-enhanced ultrasound (CEUS) to further evaluate multiple liver lesions found at gray-scale ultrasound.
A, Multiple infantile hepatic hemangiomas appear hypoechoic on gray-scale ultrasound image. Deepest lesion (short arrow) was best seen on all phases of CEUS and was used for image analysis by both readers with perfect agreement; however, other lesions (long arrows) are also visible.
Fig. 1B —5-month-old female infant who underwent contrast-enhanced ultrasound (CEUS) to further evaluate multiple liver lesions found at gray-scale ultrasound.
B, Still CEUS images obtained from real-time cinematic clip (Video S1, which can be viewed in AJR electronic supplement to this article available at www.ajronline.org at 9 (B), 10 (C), and 12 (D) seconds after contrast administration show peripheral discontinuous early arterial enhancement with rapid centripetal complete fill-in of lesion (arrow) before background liver parenchyma enhancement.
Fig. 1C —5-month-old female infant who underwent contrast-enhanced ultrasound (CEUS) to further evaluate multiple liver lesions found at gray-scale ultrasound.
C, Still CEUS images obtained from real-time cinematic clip (Video S1, which can be viewed in AJR electronic supplement to this article available at www.ajronline.org at 9 (B), 10 (C), and 12 (D) seconds after contrast administration show peripheral discontinuous early arterial enhancement with rapid centripetal complete fill-in of lesion (arrow) before background liver parenchyma enhancement.
Fig. 1D —5-month-old female infant who underwent contrast-enhanced ultrasound (CEUS) to further evaluate multiple liver lesions found at gray-scale ultrasound.
D, Still CEUS images obtained from real-time cinematic clip (Video S1, which can be viewed in AJR electronic supplement to this article available at www.ajronline.org at 9 (B), 10 (C), and 12 (D) seconds after contrast administration show peripheral discontinuous early arterial enhancement with rapid centripetal complete fill-in of lesion (arrow) before background liver parenchyma enhancement.
Fig. 1E —5-month-old female infant who underwent contrast-enhanced ultrasound (CEUS) to further evaluate multiple liver lesions found at gray-scale ultrasound.
E, Still delayed phase CEUS image shows mild hypoenhancement (contrast washout) of lesions (arrows).
With the exception of arterial phase hyperenhancement, the imaging appearance of CH was overall more variable than that of IH, as outlined in Table 2. On gray-scale ultrasound, 60% of CH lesions were heterogeneous and 40% were hyperechoic. The heterogeneous lesions were large (range, 3.1–8.6 cm) and had large peripheral hyperenhancing vessels with areas of central nonenhancement, similar to the previously reported MRI appearance of CH [12]. The more homogeneous-appearing lesions were smaller (range, 8–9 mm) than the heterogeneous lesions and showed diffusely enhancing appearance at CEUS. Size differences between the heterogeneous and hyperechoic CH lesions did not reach statistical significance (p = 0.73). None of the enhancing areas within CH lesions showed subsequent hypoenhancement relative to background liver parenchyma in the portal venous or delayed phases (i.e., no washout occurred in any CH lesion). Figure 2 shows the typical CEUS appearance of CH. Video S2, which can be viewed in the AJR electronic supplement to this article available at www.ajronline.org, shows the arterial phase appearance of a very large exophytic CH in a patient who underwent resection because of congestive heart failure.
| Ultrasound Appearance and Reader | IH (n = 5 Lesions) | CH (n = 5 Lesions) | ||||||
|---|---|---|---|---|---|---|---|---|
| Reader Determinations About IHs (% of IH Lesions) | κ Coefficient | p | ConsensusAboutIHs (% of IH) | Reader Determinations About CHs (% of CH) | κ Coefficient | p | Consensus About CHs (% of CH Lesions) | |
| Gray-scale appearance | 1.00 | 0.002 | Hypoechoic in 100% | 1.00 | 0.002 | Hypoechoic in 100% | ||
| Reader 1 | Hypoechoic in 100% | Hypoechoic in 100% | ||||||
| Reader 2 | — | — | ||||||
| Color Doppler appearance | 1.00 | 0.002 | Hypervascular in 80%, background in 20% | 0.74 | 0.02 | Hypervascular in 80%, background in 20% | ||
| Reader 1 | Hypervascular in 80%, background in 20% | Hypervascular in 80%, background in 20% | ||||||
| Reader 2 | — | Hypervascular in 60%, background in 40% | ||||||
| CEUS appearance | ||||||||
| Early arterial phase | 1.00 | 0.002 | Hyperenhancing in 100% | 1.00 | 0.002 | Hyperenhancing in 100% | ||
| Reader 1 | Hyperenhancing in 100% | Hyperenhancing in 100% | ||||||
| Reader 2 | — | — | ||||||
| Late arterial phase | 1.00 | 0.002 | Hyperenhancing in 100% | 1.00 | 0.002 | Hyperenhancing in 100% | ||
| Reader 1 | Hyperenhancing in 100% | Hyperenhancing in 100% | ||||||
| Reader 2 | — | — | ||||||
| Portal venous phase | 0.60 | 0.06 | Background in 80%, hyperenhancing in 20% | 1.00 | 0.002 | Hyperenhancing in 80%, background in 20% | ||
| Reader 1 | Background in 100% | Hyperenhancing in 80%, background in 20% | ||||||
| Reader 2 | Background in 60%, hyperenhancing in 40% | — | ||||||
| Delayed phase | 0.85 | 0.0001 | Hyperenhancing in 20%, background in 20%, hypoenhancing in 60% | 1.00 | 0.002 | Hyperenhancing in 80%, background in 20% | ||
| Reader 1 | Hypoenhancing in 100% | Hyperenhancing in 80%, background in 20% | ||||||
| Reader 2 | Hyperenhancing in 20%, hypoenhancing in 80% | — | ||||||
Note—Dash (—) indicates reader agreement. Background indicates enhancement similar to that of background liver. CEUS = contrast-enhanced ultrasound.
Fig. 2A —16-month-old female infant with solitary congenital hepatic hemangioma.
A, Gray-scale ultrasound image shows solitary liver lesion with heterogeneous echogenicity (arrow). Most of peripheral lesion is subtly hyperechoic, whereas center of lesion is hypoechoic.
Fig. 2B —16-month-old female infant with solitary congenital hepatic hemangioma.
B, Contrast-enhanced ultrasound (CEUS) image obtained at 20 seconds after IV contrast administration shows peripheral discontinuous early arterial enhancement (arrow).
Fig. 2C —16-month-old female infant with solitary congenital hepatic hemangioma.
C, CEUS images obtained at 25 (C) and 50 (D) seconds after IV contrast administration show gradual heterogeneous incomplete fill-in of center of lesion (arrow).
Fig. 2D —16-month-old female infant with solitary congenital hepatic hemangioma.
D, CEUS images obtained at 25 (C) and 50 (D) seconds after IV contrast administration show gradual heterogeneous incomplete fill-in of center of lesion (arrow).
Fig. 2E —16-month-old female infant with solitary congenital hepatic hemangioma.
E, CEUS image obtained at 1 minute 52 seconds after IV contrast administration shows sustained enhancement of peripheral aspect of congenital hemangioma (arrow), which is slightly hyperenhancing compared with background liver, and unchanged central nonenhancement.
Ancillary imaging findings are described in Table 3. Assessment of ancillary imaging findings outside the lesion included assessment of hepatomegaly and aortic caliber. Diffuse hepatomegaly was seen in one of five (20%) IH cases and zero of five (0%) CH cases. Focal enlargement of the liver contour due to mass effect was seen in zero of five (0%) IH cases and two of five (40%) CH cases. The aortic caliber above and below the celiac artery origin was confidently imaged in four of 10 (40%) patients, one patient with IH patients and three patients with CH. No significant tapering of the aorta was seen in the three patients with CH as defined by a change in aortic caliber of 2 mm or less. In a single case of IH, there was a change in aortic caliber of more than 2 mm.
| Ancillary Findings | Percentage of IH Lesions (n = 5) | Percentage of CH Lesions (n = 5) | κ Coefficient | p |
|---|---|---|---|---|
| Gray-scale ultrasound appearance | ||||
| Presence of calcifications or cystic changes | 20% | 40% | 1.00 | 0.0016 |
| CEUS | ||||
| Early arterial phase | ||||
| Description of hyperenhancement | Peripheral and discontinuous in 80%, heterogeneous in 20% | Peripheral and discontinuous in 40%, heterogeneous in 60% | 1.00 | 0.0016 |
| Late arterial phase | ||||
| Description of hyperenhancement | Peripheral and discontinuous in 80%, heterogeneous in 20% | Peripheral and discontinuous in 40%, heterogeneous in 60% | 1.00 | 0.0016 |
| Delayed phase (washout) | ||||
| If hypoenhancinga, to what degree? | Mild in 40%, moderate in 20% | NA | 1.00 | 0.0016 |
Note—CEUS = contrast-enhanced ultrasound, NA = not applicable.
a
One IH case (20%) is not included because it was not discernable from background liver on this phase (i.e., it was isoenhancing).
Clinical and Imaging Outcomes
Adjunct imaging with gadolinium-enhanced MRI was performed for diagnosis in three of the 10 (30%) patients. CT was performed in one patient (10%) with CH who ultimately underwent tumor resection; however, CT findings were nondiagnostic because of contrast infiltration at the injection site. This patient subsequently underwent contrast-enhanced MRI (one of the three patients mentioned earlier). Unfortunately, single-phase contrast-enhanced MRI in this patient was thought to be suboptimal for evaluation of the tumor. Therefore, CEUS was performed for further lesion characterization and preoperative planning and was considered diagnostic. In a second patient with CH, contrast-enhanced MRI was performed at age 1 month; however, because the lesion had not involuted, CEUS was requested at age 12 months for additional lesion evaluation. One of the 10 (10%) patients with innumerable IHs and some areas of intralesional calcifications underwent MRI after CEUS; in that case, MRI findings were diagnostic of IH.
Of the five patients with IH, three (60%) underwent medical therapy with oral propranolol. Of the CH cohort, one (20%) patient underwent surgical resection for heart failure. All patients underwent follow-up ultrasound to resolution or significant reduction in lesion size except one patient with CH who underwent surgical resection. Imaging follow-up for the other nine patients ranged from 4 to 32 months depending on whether the lesion resolved on follow-up.
Four of five (80%) patients with IH had complete resolution of the lesions on imaging. In one of five (20%) patients with IH, the lesions had significantly decreased in size and number at 13 months of follow-up. One of five (20%) patients with CH had complete resolution on follow-up ultrasound. In three of five (60%) patients with CH, the lesions had significantly decreased in size on follow-up ultrasound ranging from 13 to 19 months after the initial CEUS examination.
Interobserver Variability
Generally, there was perfect agreement (κ coefficient = 1.00) between observers across core imaging findings (Table 3). Three categories generated disagreement between the readers: lesion appearance on color Doppler imaging, portal venous phase CEUS, and delayed phase CEUS.
For lesion appearance on color Doppler imaging, one of five (20%) patients with CH was deemed to have a hypervascular lesion by reader 1 but not by reader 2. Consensus review determined that the lesion was hypervascular on color Doppler ultrasound. Agreement was 100% for IHs (hypervascular, 80% of IH lesions; background, 20% of IH lesions).
The appearance of IHs on portal venous phase CEUS generated reader disagreement for one of five (20%) cases. In that case, reader 1 said there was isoenhancement, and reader 2 said there was hyperenhancement. This lesion was deemed to be hyperenhancing by both readers at consensus determination.
For the appearance of IH on delayed phase CEUS, there was disagreement with one of five (20%) patient with multiple IHs. It was subsequently discovered that the readers were comparing two different lesions, because both lesions were well seen during all phases of CEUS. After consensus determination about the specific lesion being evaluated, there was agreement regarding the imaging appearance (hyperenhancing). In another case, both readers initially agreed that the lesion was hypoechoic. However, at consensus determination, the lesion was actually no different than background liver. Thus, the final determination in this case (background) was different than the initial interpretation (hyperenhancing) for both readers.
Discussion
In children, the two most commonly encountered hepatic vascular anomalies are benign tumors: IH and CH, with the former much more common than the latter [3]. IHs are typically multifocal or diffuse and are not fully developed at birth; rather, IHs develop a few weeks to months after birth. IH lesions show reactivity to glucose transporter type 1 (GLUT1) at immunostaining may cause hypothyroidism due to production of type 3 iodothyronine deiodinase, which may impact metabolism, growth, and brain development [12]. IHs may also cause high output heart failure due to shunting [13]. Similar to IH cases described in the literature, IH cases in our series were associated with cutaneous lesions, and all IH lesions showed hyperenhancement in the arterial phase [14]. CHs are solitary lesions that are present and are largest in relation to patient size at birth and thus may be seen at prenatal imaging. CHs are negative at GLUT1 staining, contain capillaries intermixed with malformed vessels, and contain foci of extramedullary hematopoiesis [12]. Three types of CHs occur: rapidly involuting, partially involuting, or noninvoluting. All CH lesion subtypes have completed the proliferative phase by birth, with no subsequent cellular proliferation. However, intralesional hemorrhage is possible, which can cause increased lesion size [12]. Like IH, CH may cause congestive heart failure due to shunting. A few case reports of the use of CEUS in patients with CH describe features similar to those seen in our case series [1, 15].
Although IH and CH are benign lesions, the differential diagnosis includes malignancies such as metastatic neuroblastoma and hepatoblastoma [16]. Clinical assessment remains vital in the workup of these patients, and CEUS might play an important imaging adjunct. CEUS has been shown to have excellent diagnostic accuracy for distinguishing benign from malignant lesions in studies of adults [16, 17]. Appropriate imaging triage is imperative, not only to guide patients to appropriate management, but also to rapidly exclude or diagnose malignancy. Additionally, we found a high degree of reader agreement in interpreting CEUS features.
Consensus determination revealed that 60% of IH lesions showed mild-to-moderate hypoenhancement (washout) on delayed phase imaging. Contrast washout is hypothesized to occur because IH lesions are true vascular neoplasms rather than vascular malformations containing large vascular channels. However, this theory is speculative, and we did not specifically study the mechanism for contrast washout in this case series. Delayed phase contrast washout was not present in the CH group. Previous studies evaluating focal liver lesions at CEUS have reported that hypoenhancement in the portal venous and delayed phases is concerning for malignancy. Specifically, although metastases have a variable appearance in the arterial phase, studies in the literature suggest that marked early washout in the portal venous phase (< 60 seconds) is characteristic of metastases [17]. None of the vascular lesions in this study had marked washout or portal venous phase washout. Thus, CEUS can be used to exclude metastasis in this patient population, which may help alleviate caregiver anxiety when a lesion or multiple lesions are discovered in an infant. Depending on the clinical setting, this knowledge may also eliminate the need for additional imaging studies when there will be no change in treatment decisions. In contrast to hepatic metastases, hepatocellular carcinoma (HCC) characteristically has subtle washout that occurs in the delayed phase rather than portal venous phase [17]. This type of washout was seen in 60% of IH cases in our study [13–15]. Given the overlap of washout appearance and timing of IH with HCC, there could be confusion at CEUS. However, IH and HCC occur in very different patient populations (infants vs adolescents), and clinical findings can be used to easily differentiate IH from HCC.
Both IH and CH lesions showed peripheral, nodular discontinuous enhancement, a feature that is also described with venous malformations in adults. However, venous malformations, which are vascular malformations with large vascular channels, do not have delayed phase washout [17], which was present in most of the IH cases in our case series. IH lesions have been described to have confluent rather than discontinuous peripheral and centripetal enhancement at dynamic contrast-enhanced MRI [12]. The discontinuous peripheral enhancing appearance in the early arterial phase in four of five patients with IH in our study may be explained by differences in soft-tissue, spatial, and temporal resolution of ultrasound compared with MRI; small gaps inconspicuous at MRI may be visible at ultrasound. However, this theory is speculative.
Limitations of our study must be recognized. First, our single-center experience yielded only a small sample size (five patients with IH and five patients with CH), which impairs the generalizability of our findings. Further studies with a larger number of patients are needed to confirm our findings. Second, although final clinical diagnosis was made by a pediatrician with expertise in pediatric vascular anomalies, pathologic confirmation of diagnosis was obtained in only one of 10 cases in this retrospective case series. This lack of pathologic confirmation is because IH and CH do not commonly require biopsy for diagnosis unless there are atypical clinical or imaging features [3]. Finally, the determination of the degree of washout in the delayed phase on CEUS was ultimately subjective, because each reader was allowed to select his or her interpretation of washout intensity (e.g., mild vs moderate).
Conclusion
With the exception of the CEUS feature of portal venous phase enhancement (κ = 0.60), there was good to excellent (κ = 0.74–1.00) interobserver agreement for the CEUS features of IH and CH. Consensus interpretation revealed that IH and CH are hypervascular in the arterial phase; that IH and CH have sustained enhancement in the portal venous phase (unlike metastases); and that a significant proportion of IHs (60%) showed contrast washout at delayed phase imaging, which has also been reported with HCC. Recognition of the overlap in imaging appearance of IH and HCC is vital to preventing misdiagnosis of malignancy.
Supplemental Content
References
1.
Anupindi SA, Biko DM, Ntoulia A, et al. Contrast-enhanced US assessment of focal liver lesions in children. RadioGraphics 2017; 37:1632–1647
2.
Yusuf GT, Sellars ME, Deganello A, Cosgrove DO, Sidhu PS. Retrospective analysis of the safety and cost implications of pediatric contrast-enhanced ultrasound at a single center. AJR 2017; 208:446–452
3.
Restrepo R, Palani R, Cervantes LF, Duarte AM, Amjad I, Altman NR. Hemangiomas revisited: the useful, the unusual and the new. Part 2. Endangering hemangiomas and treatment. Pediatr Radiol 2011; 41:905–915
4.
Kollipara R, Odhav A, Rentas KE, Rivard DC, Lowe LH, Dinneen L. Vascular anomalies in pediatric patients: updated classification, imaging, and therapy. Radiol Clin North Am 2013; 51:659–672
5.
Shamir SB, Kurian J, Kogan-Liberman D, Taragin BH. Hepatic imaging in neonates and young infants: state of the art. Radiology 2017; 285:763–777
6.
Mulliken JB, Glowacki J. Hemangiomas and vascular malformations in infants and children: a classification based on endothelial characteristics. Plast Reconstr Surg 1982; 69:412–422
7.
Hassanein AH, Mulliken JB, Fishman SJ, Greene AK. Evaluation of terminology for vascular anomalies in current literature. Plast Reconstr Surg 2011; 127:347–351
8.
Mulligan PR, Prajapati HJ, Martin LG, Patel TH. Vascular anomalies: classification, imaging characteristics and implications for interventional radiology treatment approaches. Br J Radiol 2014; 87:20130392
9.
Wassef M, Blei F, Adams D, et al.; ISSVA Board and Scientific Committee. Vascular anomalies classification: recommendations from the International Society for the Study of Vascular Anomalies. Pediatrics 2015; 136:e203–e214
10.
American College of Radiology website. CEUS LI-RADS v2017 CORE. www.acr.org/-/media/ACR/Files/RADS/LI-RADS/CEUS-LI-RADS-2017-Core.pdf?la=en. Published 2017. Accessed January 11, 2020
11.
Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics 1977; 33:159–174
12.
Merrow AC, Gupta A, Patel MN, Adams DM. 2014 Revised classification of vascular lesions from the International Society for the Study of Vascular Anomalies: radiologic-pathologic update. RadioGraphics 2016; 36:1494–1516
13.
Wildgruber M, Sadick M, Müller-Wille R, Wohlgemuth WA. Vascular tumors in infants and adolescents. Insights Imaging 2019; 10:30
14.
Hughes JA, Hill V, Patel K, Syed S, Harper J, De Bruyn R. Cutaneous haemangioma: prevalence and sonographic characteristics of associated hepatic haemangioma. Clin Radiol 2004; 59:273–280
15.
Piorkowska MA, Dezman R, Sellars ME, Deganello A, Sidhu PS. Characterization of a hepatic haemangioma with contrast-enhanced ultrasound in an infant. Ultrasound 2018; 26:178–181
16.
Gnarra M, Behr G, Kitajewski A, et al. History of the infantile hepatic hemangioma: From imaging to generating a differential diagnosis. World J Clin Pediatr 2016; 5:273–280
17.
Burrowes DP, Medellin A, Harris AC, Milot L, Wilson SR. Contrast-enhanced US approach to the diagnosis of focal liver masses. RadioGraphics 2017; 37:1388–1400
FOR YOUR INFORMATION
A data supplement for this article can be viewed in the online version of the article at: www.ajronline.org.
Information & Authors
Information
Published In
Copyright
© American Roentgen Ray Society.
History
Submitted: August 18, 2019
Accepted: September 28, 2019
Version of record online: January 22, 2020
Keywords
Authors
Metrics & Citations
Metrics
Citations
Export Citations
To download the citation to this article, select your reference manager software.