| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Editorials |
1 From the Department of Diagnostic Radiology, Yale School of Medicine, LMP 5039, 333 Cedar St, New Haven, CT 06510. Received September 16, 2003; accepted September 25. R.I.W. supported in part by March of Dimes grant HHT–Fy 2002–677 and the Josephine Lawrence Foundation. Address correspondence to S.S. (e-mail: sanjay.saluja@yale.edu).
Index terms: Computed tomography (CT), angiography, 952.12915, 952.12916, 957.12915, 957.12916 • Computed tomography (CT), three-dimensional, 952.12917, 957.12917 • Editorials • Shunts, arteriohepatic • Telangiectasia
Hepatic involvement in hereditary hemorrhagic telangiectasia (HHT) is a complex and intriguing manifestation of the disorder. The prevalence is relatively high in patients with HHT, and yet, the vast majority of patients are asymptomatic. This lack of symptoms occurs because (a) in most cases, the degree of shunt can be well tolerated and compensated for by the heart, and (b) the presence of intrahepatic shunting, while altering greatly the intrahepatic vascular dynamics, does not by itself lead to substantial intrinsic liver parenchymal disease. Hence, the synthetic function of the liver remains preserved. Those who do become symptomatic generally develop heart failure secondary to volume overload and can be treated conservatively with medical regimens to control congestive heart failure. Those who develop symptomatic liver disease, however, have a less favorable course and will ultimately require liver transplantation for palliation.
Three types of vascular shunts may exist in the liver of patients with HHT: hepatic artery to hepatic vein, hepatic artery to portal vein, and portal vein to hepatic vein. Hepatic vein–to–hepatic vein connections have been described but do not constitute a "true" shunt. All three shunts are essentially left-to-right shunts. Characterization of this intrahepatic abnormal shunt anatomy has an important bearing on the manifestation of liver involvement in patients who become symptomatic and also in treatment offered to symptomatic patients.
Like most vascular abnormalities, hepatic involvement lends itself to analysis with dynamic imaging, such as real-time ultrasonography (US), multiphase computed tomography (CT) for generation of CT angiograms of the liver, and conventional angiography. Imaging of hepatic involvement of HHT has mainly comprised Doppler US, which has been widely used at HHT centers in screening for hepatic involvement, and conventional angiography, in an attempt to understand the vascular anatomy and physiology. To our knowledge, there have been no large CT angiographic studies of hepatic vascular abnormalities in this disorder reported to date.
In this issue of Radiology, Stabile Ianora et al (1) from University Hospital, Bari, Italy, report their experience with multi–detector row CT in the assessment of hepatic involvement in patients with HHT. The report represents the first attempt at using multi–detector row CT to visualize and characterize intrahepatic shunts. The authors also offer better insight into the true prevalence of hepatic involvement in this disease and illustrate the spectrum of vascular abnormalities encountered. Interestingly, most of the patients of Stabile Ianora et al were asymptomatic, which (to those unfamiliar with this disease entity) appears unusual, given the extent of disease illustrated; in actuality, this reflects the typical clinical picture of these patients (2).
To understand this disease, the CT findings per se, and the importance of these findings, a brief review of the pathophysiology and clinical aspects of the disease process will be helpful. Histologically, liver involvement can be grouped into three types, as described by Martini in 1978 (3): telangiectasia with fibrosis or cirrhosis, cirrhosis without telangiectasia, and telangiectasia without fibrosis or cirrhosis. It is presently thought that patients with cirrhosis without telangiectasia most likely have chronic posttransfusion hepatitis unrelated to HHT. Macroscopically, the hallmarks of the disease are telangiectasia and arteriovenous malformations. Telangiectases are focal dilations of the postcapillary venules (sinusoids) without preserved capillaries. Arteriovenous malformations are larger lesions with arteriolar-venous connections without intervening capillaries. Indeed, both lesions are believed to be on a continuum with telangiectasia being essentially a very small arteriovenous malformation. Therefore, lesions may enlarge over time and progress from telangiectasia to arteriovenous malformation morphology. These vascular abnormalities are accompanied by fibrous proliferation that leads to a nodular liver. The intervening hepatic parenchyma is preserved within these nodules with its central and portal veins leading to a cirrhosis-like pattern at histologic examination (pseudocirrhosis) and a nodular liver at CT. The preservation of large amounts of normally functioning hepatocytes accounts for the general lack of abnormalities in liver function test results.
The arteriovenous shunts themselves can be of three types: arteriovenous, arterioportal, or portal venous. This has recently been demonstrated angiographically by Hashimoto et al (4). Arteriovenous and arterioportal shunts are common, but in the study of Hashimoto et al, four of seven patients with advanced HHT of the liver demonstrated angiographically visible portal venous shunts. While arteriovenous and arterioportal shunts can lead to unique clinical syndromes, the present-day understanding of portal venous shunts is extremely limited (5). This is partly because they are difficult to visualize and may even be microscopic. This renders them relatively "invisible" with current imaging techniques, and they are often angiographically occult. However, they have immense importance when embolotherapy is contemplated as a possible treatment option for hepatic arteriovenous malformations.
Unlike embolization of HHT arteriovenous malformations in the lung or brain, hepatic artery embolization is associated with substantial morbidity and mortality and often results in fatal hepatic necrosis. Several deaths have been reported after this procedure was performed in the United States, Europe, and, more recently, Japan (6). Portal venous shunts have been postulated as a possible reason why portal circulation alone is unable to sustain the liver parenchyma after therapeutic hepatic artery embolization. Some investigators recommend that hepatic arterial embolization not be performed in these patients, explicitly because the risk of liver necrosis is so great in the presence of portal venous shunts (7). At the present time, even though some groups continue to offer hepatic artery embolotherapy for this disorder (8), there is a growing consensus among most experts that liver transplantation constitutes the definitive treatment for hepatic involvement and that arterial embolotherapy should be avoided (9).
The clinical syndromes associated with hepatic HHT were well described by Garcia-Tsao et al in 2000 (5) on the basis of 19 symptomatic patients with this disease. In theory, all of these syndromes are characterized by increased hepatic blood flow, and, paradoxically, simultaneous relative hepatic ischemia (poverty amidst plenty), since the shunted blood may lead to ischemia of some of the hepatocytes or the peribiliary plexus. Group 1 (most common) consists of patients with severe shortness of breath and high-output congestive heart failure. This clinical profile is a direct manifestation of a hepatic artery–to–hepatic vein shunt. Group 2 consists of patients with clinical manifestations of portal hypertension—that is, splenomegaly, varices, and enlarged portal vein. These patients have an elevated hepatic venous pressure gradient, which is a marker for portal hypertension. These symptoms are presumed to be a direct result of hepatic artery–to–portal vein shunting, as well as periportal fibrosis due to ischemia, both of which would result in portal hypertension. Group 3 is an unusual group that consists of patients with biliary symptoms, including right upper quadrant pain and increased alkaline phosphate levels, and, in the late phase, hyperbilirubinemia. This is a "Caroli-like" disorder that results from biliary strictures with secondary bile cysts. The biliary strictures are presumed to be ischemic in origin with the ischemia in the peribiliary arterial plexus being secondary to ongoing shunting. It has been our experience that CT results do not correlate with the clinical type, and indeed, the clinical manifestations often change with time depending on the type and degree of intrahepatic shunt and/or disease.
Stabile Ianora et al (1) describe CT findings in consecutive patients who presented to their HHT Centre, regardless of whether the patients had either clinical signs and symptoms pertaining to liver disease or abnormalities in their liver function test results. In fact, only four of 70 patients actually had symptoms that could be ascribed to liver disease, and none had abnormal liver function test results. The practice of screening for liver HHT exists in some HHT centers (10). There is no universal consensus on the benefit of liver screening in these patients, especially because the overwhelming majority of patients remain asymptomatic and because there is no preventive therapy or measure that can be offered in response to finding vascular abnormalities in the liver. Nevertheless, most centers that do offer screening use Doppler US to assess hepatic artery diameter or peak systolic velocity to detect underlying liver disease (10,11).
To our knowledge, Stabile Ianora et al (1) are the first to routinely use multiphase CT to accomplish screening. Since CT in general is a far superior modality when compared with US, especially for assessment of vascular abnormalities, it is no surprise that this group has found more vascular lesions in all patients with HHT than those in previous reports. The widely accepted hepatic prevalence prior to the study of Stabile Ianora et al was based on clinical criteria and presence of right upper abdomen bruit and ranged from 8%–31%, but Stabile Ianora et al find the prevalence to be much higher—around 74% of all their patients. This is important new information that contributes to our understanding of the true prevalence of liver involvement in HHT. In the absence of a clinically approved DNA test for HHT, the presence of liver involvement of the type described by Stabile Ianora et al may confirm the diagnosis of disease.
We believe the most important contribution of the report by Stabile Ianora et al is to familiarize radiologists with the CT manifestations and the spectrum of liver vascular malformations in patients with HHT. Since HHT is a rare disease, many radiologists at non-HHT centers are not familiar with its varied CT manifestations. The strikingly impressive and complex vascular malformations in the liver at CT are often not well understood because of unfamiliarity with the disorder. The fact that the patient is asymptomatic despite the presence of such extensive liver disease seems rather paradoxical and confusing. To establish the diagnosis, many physicians and radiologists perform percutaneous liver biopsy and endoscopic retrograde cholangiopancreatography, both of which are unnecessary and potentially dangerous (5). Liver biopsy can be complicated by severe hemorrhage and is contraindicated; hence, it should be avoided totally. Similarly, endoscopic retrograde cholangiopancreatography should be avoided, since it is not necessary to assign the diagnosis. In our experience, biliary cysts and dilatation are common findings at CT in patients with symptomatic liver disease; these findings were not seen in any of the asymptomatic patients in the series of Stabile Ianora et al (1).
The use of three-dimensional imaging is made possible by multi–detector row CT technology and multiphase imaging. The authors have demonstrated that they have been able to identify arteriovenous and arterioportal communications by using a combination of transverse and three-dimensional CT. It is disappointing that despite the use of three-dimensional reformatted images, it is difficult to identify portal venous communications. This is not entirely surprising, however, since it is known that these communications may be microscopic. Identification of them continues to remain a challenge. In the report of Stabile Ianora et al (1), the use of reformatted images enabled identification of hepatic telangiectasia and large confluent vascular lesions in a minority of cases.
In summary, in this excellent report, Stabile Ianora et al (1) define the multitude of CT findings in hepatic HHT so that the diagnosis of this complex disease can be made noninvasively, and biopsy can be avoided. They also establish that CT may be better than US for screening patients with HHT for liver involvement. Their findings suggest that multiphase imaging can help not only in the detection of but also in the characterization of hepatic vascular malformations in HHT. Finally, a new gamut for differential diagnosis (ie, increased hepatic artery diameter with multiple vascular liver shunts and abnormal vascular masses present diffusely in the liver) has been described.
FOOTNOTES
See also the article by Stabile Ianora et al in this issue.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
