Pitfalls in a Sonographic Diagnosis of Liver Abscess in Children

Background:The purpose of this article is to identify the pitfalls of sonography in the diagnosisof liver abscesses, hematomas, and hepatic tumors, which appear similar and therefore aredifficult to differentiate from each other.

Methods:Cases were collected at the China Medical University Hospital between January 2008and January 2010. Liver abscesses were initially diagnosed by sonograph in selected patientswho were younger than 18 years.

Results:There were 15 patients in whom a liver mass was diagnosed by ultrasound, but 6 ofthem were excluded from further study because of failure to meet any of the screeningcriteria. Nine patients with a mean age of 11.3 years (range 5-17 years) were initially suspected to have liver abscesses by ultrasound and were enrolled in the study. These ninepatients were identified as follows: five with liver abscess, one with liver hematoma, one withhepatic lymphoma, one with perihepatic abscess, and one with undifferentiated liver sarcoma.Ultrasonography alone was sufficient for diagnosis in five patients. Four patients requiredabdominal CT scanning to confirm final diagnosis.

Conclusion:Different liver lesions may present sonographic images similar to those of liverabscesses. Therefore, it is suggested that patients in whom liver abscesses were diagnosedby ultrasound undergo further evaluation if the clinical condition is less likely.

Copyright©2012, Taiwan Pediatric Association. Published by Elsevier Taiwan LLC. All rightsreserved.

1. Introduction

Abdominal ultrasound is a common medical procedure tolocate abdominal lesions and evaluate gastrointestinalfunction. Additionally, the procedure will show clearimages of the liver. Abdominal ultrasound is used onpatients who are clinically suspected of having a liverabscess. The ultrasonic appearance of a typical liverabscess has been described as having the following criteria:acoustic enhancement, abscess wall, peripheral halo, septation, and internal debris.1,2However, some intrahepaticlesions, such as a necrotic hepatic tumor or hematoma,could present with a sonographic appearance similar to thatof liver abscess. For this reason, we reviewed nine patients’sonographies with presentations similar to that of liverabscess and attempted to identify the pitfalls by anyspecific features or methods.

2. Materials and Methods

Between January 1, 2008 and January 1, 2010, a retrospective ultrasound-based study of 15 patients youngerthan 18 years was conducted. These patients wereadmitted to China Medical University Hospital with aninitial diagnosis of liver mass by sonography.

Liver abscess is screened by using the following criteria:acoustic enhancement, abscess wall, peripheral halo, septation, and internal debris.1,2Six patients were excludedbecause of failure to meet any of the screening criteria.Nine of 15 patients, on initial suspicion of liver abscess(meeting at least one of the five criteria), were enrolled inour study. Another imaging study was indicated if thepatient’s clinical course was not compatible with manifestations of liver abscess (Figure 1). In our study, we chosecomputed tomography (CT) as the secondary screeningtool. However, one of the patients in the unconfirmed liverabscess group had been scheduled for serial ultrasonography instead of CT, because his family informed us of thepatient’s trauma history in time. The clinical characteristics of these nine patients are recorded inTable 1. Resultsfrom their imaging studies, microbiologic and pathologicfindings, and blood examinations are presented inTable 2.The mean age was 11.3 years (range 5-17 years). An ultrasound machine (HP Sonos 5500, Stockton, CA, UnitedStates) was used with gray-scale ultrasound and 3-5 MHztransducers to examine these patients.

3. Results

This series comprised five liver abscesses (Figure 2), oneliver hematoma, one hepatic lymphoma, one perihepaticabscess, and one undifferentiated liver sarcoma. In all ofthese cases, sonographic findings demonstrated variousdegrees of features of liver abscess, as presented inTable 3. Therefore, the tentative diagnosis of liver abscesswas made initially. Fever and abdominal pain were thechief complaints (77.8%). Confirmed diagnoses needed tobe further proved by CT in cases 5, 6, 7, and 9 due tounusual clinical presentations or imaging findings. Biopsywas performed for cases 6 and 7. The lesion size ranged from 24 mm to 65 mm in diameter on the initial scan.Although diagnostic abdominal CT was not necessary forcases 1, 3, and 4, we still arranged abdominal CT toperform CT-guided percutaneous catheter drainage (PCD).

Hypoechoic masses were present in all nine patients.Four patients revealed homogeneous echogenicity, whereasheterogeneous echogenicity was discovered in fivepatients. Internal echoes were seen in seven patients;septation was noted in cases 5 and 7.

Six of the nine patients (cases 1, 3, 4, 5, 7, and 9)showed variable degrees of distal acoustic enhancement.Acoustic enhancement was not related to the size or degreeof echogenicity of the mass. A sharp echogenic wall wasseen in five patients (cases 2, 3, 5, 7, and 9). A peripheralhalo was present in cases 2 and 6.

Case 6 was a large B-cell lymphoma with some sonographic characteristics of liver abscess, such as multipleheterogeneous hypoechoic lesions with a peripheral halo(Figure 3A). There was no vascularity noted on colorDoppler imaging. According to the patient’s history, endstage renal disease had been diagnosed and the patienthad previously received renal transplantation. Because ofimmunosuppression therapy, which may lead to thepossibility of oncogenesis (sirolimus/tacrolimus/mycophenolate: lymphoproliferative disease), abdominal CT(Figure 3B) and biopsy were performed to differentiatediagnosis of hepatic malignancy. The abdominal CTdemonstrated multiple hypovascular tumors in the liver and spleen, and the pathologist confirmed the diagnosis of Bcell lymphoma.

Case 7 was a hepatic undifferentiated sarcoma. Thesonographic findings included echogenic wall, heterogeneous hypoechoic lesion with septation inside, and acousticenhancement (Figure 4A). No significant vascularity wasseen on color Doppler imaging. These findings were similarto those of a liver abscess. However, persistent abdominalpain and fatigue were noted after days of antibiotictreatment. A final diagnosis of sarcoma was confirmed bypathology. As reviewed, previous trauma history andabscess-like sonography had led to an initial misdiagnosis.However, a normal white blood cell count and C-reactiveprotein (CRP) level, combined with unusual clinical courserefractory to antibiotic treatment, led to doubt of thetentative diagnosis of liver abscess. CT revealed a heterogeneous mass containing multiple cystic spaces separatedby hyperdense septations of varying thickness3(Figure 4B).

Case 8 was a hematoma, and the sonography was similarto that of a liver abscess initially. Because the patientconcealed a traffic accident from us, as many children willdo, a liver abscess could not be ruled out initially. Inaddition, fever, leukocytosis, and abscess-like sonogramwere highly compatible with liver abscess presentation.When the family hinted the possibility of trauma, a followup abdominal ultrasound on the next day and another 6days later were performed. Hepatic hematoma wasconfirmed due to a typical change in the ultrasonicappearance with time (Figure 5).

Case 9 was a perihepatic abscess, probably endoscopicretrograde cholangiopancreatography-related iatrogenictrauma. The lesion was located along the falciform ligament and looked like a liver abscess between the S2 and S4lobes. While performing CT-guided PCD, we found theabscess located in the perihepatic space (Figure 6A). AfterPCD, some clear bile-stained fluid was drained and thevalue of amylase of drainage fluid was low (8U/L). Wereviewed the initial sonographic images, which showeda perihepatic lesion communicated with the abscess-likelesion between the S2 and S4 segment (Figure 6B).

4. Discussion

The etiologies of an intrahepatic abscess can have severaldifferent causes. The first is intraabdominal sepsis withspread by direct extension and/or portal bacteremia,usually due to appendicitis or diverticulitis. Second, biliarytract disease, such as cholecystitis or ascending cholangitis,can cause intrahepatic abscess. Third, it may be caused bytrauma and generalized septicemia with bacterial seedingvia the hepatic artery.

Ultrasound imaging is traditionally used as a first-linetool for suspicion of liver abscesses. In addition, it providesthe available clinical information in most instances. Ultrasound is the preferred imaging technique for the detectionof liver abscess because of its lower cost, noninvasivenature, and availability. Five of the nine patients (55.6%) were correctly identified as having liver abscess initially.The key conditions that must be differentiated fromhepatic abscess on ultrasound are necrotic neoplasm,hematoma, and complicated (hemorrhagic or inflammatory) cyst.

The gray scale of sonography in patients with hepaticpyogenic abscess varies with disease progression. Duringthe acute stage, the shape of the lesion is usually irregular.It is composed of an accumulation of neutrophils in an areaof liquefaction necrosis of tissues. The echogenicity isvariable due to high protein and lipid content or smallbubbles of gas. It can be either misdiagnosed as a solid massor overlooked altogether. During the subacute and chronicstages, extensive fibroblastic proliferation and vascularizedconnective tissue are present. As abscesses mature, thelesion becomes more rounded and hypoechoic, with debrisin the middle and thick walls on the outside. Therefore,wall enhancement usually appears in subacute and chronicabscesses and is generally not visible in the acute inflammatory lesions.

Ultrasonic features of a necrotic hepatic neoplasm,which simulate those of an abscess, include sonolucency,acoustic enhancement (if liquefaction has occurred), andan irregular, echo-poor wall. Moreover, multiple lesions ordifferent echo patterns favor a malignant process.

Case 6 was initially thought to be multiple fungal liverabscess because of sonographic features, which appearedin the form of a multiple bull’s eye pattern. According toprevious studies, lymphomatous involvement of the liveris more often secondary to non-Hodgkin lymphoma, a largeB-cell lymphoma in our case, than to Hodgkin lymphoma.8The typical sonographic findings of hepatic lymphomainclude discrete anechoic or hypoechoic nodules, whichwere compatible with our case. The peripheral halo is attributed to an inflammatory reaction and increased bloodflow seen in tumors. The thickness of the halos in hepaticabscesses and tumors may be similar.2On a sonographicbasis alone, it may be difficult to distinguish liver abscessesfrom lymphomatous metastasis to the liver; therefore,another imaging modality is required.

Case 7 showed the typical ultrasound findings of liverabscess, such as internal echoes, acoustic enhancement,and wall enhancement, and confused our tentative diagnosis. In our study, high white blood cell count and elevatedCRP were noted in six patients, and one patient had a highwhite blood cell count without an elevated CRP. Cases 6and 7 had malignancy without high blood white cell countor elevated CRP. Therefore, whenever abdominal sonography shows liver abscess-like lesions without high whiteblood cell count or elevated CRP, further evaluations fordiagnosis of malignancy should be added.

In case 9, we ruled out the possibility that the intrahepatic abscess had ruptured into the perihepatic areabecause the drainage fluid is too clear to be like that froman abscess. In this case, we could not differentiate perihepatic fluid from intrahepatic abscess by sonographyalone. However, CT more clearly ascertained that theabscess-like lesion was located along with the falciformligament and communicated with the perihepatic area.

Traditionally the location of hepatic abscess and thepatient’s cooperation affect the accuracy of ultrasound.11However, we cannot be sure that a lesion is an abscess,even when the lesion is compatible with sonographicfeatures of abscess and is located in the S4 or S5 area wherethere is a relatively easier ultrasound approach. We canfind the high sensitivity of ultrasound for diagnosing liverabscesses, and the range is from 66% to 90% in previousreports.12-14Similar to other reports, the sensitivity ofultrasound was high (100%) in our study, but the falsepositive rate was also rather high (55.6%). According toour experience, ultrasound is suggested to be the first-linediagnostic tool for suspicion of liver abscess because of its high sensitivity. However, an additional imaging study, suchas CT, should be arranged if any of the following conditionsexists. First, past history or drug history may induceanother liver disease. Second, patients present unusualclinical manifestations or progressive courses. Third,patients’ conditions are refractory to adequate treatment.Fourth, the lesion located along the falciform ligamentshould be evaluated carefully.

Conventional ultrasonography has the advantage ofaccurately visualizing the biliary tree and distinguishingsolid from cystic structures, whereas CT has the advantageof visualizing the posterior and superior aspects of the liver.Although magnetic resonance imaging can detect liverabscesses, it is much less useful because it cannot guidepercutaneous aspiration. Contrast-enhanced ultrasoundmay be used to enhance both the detection andcharacterization of a pyogenic liver abscess. Relative toconventional ultrasound, contrast-enhanced ultrasound isable to reveal peripheral rim enhancement and central lowreflective debris in the arterial phase. It will revealparenchymal hypoperfusion in the vicinity of the abscess inthe venous phase.15CT identifies a relatively homogeneousliver abscess better than does conventional ultrasound.However, contrast-specific sonography brings out evenmore differences.

No matter how experienced an ultrasound operator is,the diagnosis of liver mass will not be complete withoutadditional confirmation study, such as microbiologicconfirmation for liver abscess or histology for neoplasm.Unless hematoma is suspected from clinical history andcompatible sonographic findings, abscess or neoplasmcannot be diagnosed by ultrasound alone.

5. Conclusion

Ultrasonography is a useful screening tool when liverabscess is suspected because of its high sensitivity andnoninvasive properties. However, it should not be thoughtof as the gold standard diagnostic tool. Whenever thesonographic diagnosis of liver abscess is doubtful or theclinical presentations are less likely, further evaluationsuch as CT or even contrast-enhanced ultrasound should bedone to confirm the diagnosis.