| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Letters to the Editor |
Department of Imaging, Imperial College of Science, Technology & Medicine, Hammersmith Hospital, Du Cane Road, London W12 ONN, United Kingdom
Department of Imaging, Imperial College of Science, Technology & Medicine, Charing Cross Hospital, London, United Kingdom
Hepatobiliary Service, Memorial Sloan-Kettering Cancer Center, New York, NY
Editor:
We read with interest the article by Dr Vogl and colleagues in the July 1998 issue of Radiology (1) on the effects of hepatic arterial embolization in patients with unresectable hilar cholangiocarcinoma. Dr Vogl and colleagues concluded that their "data show that preoperative [arterial] embolization of affected hepatic lobes results in a marked increase in the volume of noninvolved liver parenchyma," which will in some patients "allow reclassification of the case from an initially technically unresectable state to a resectable condition."
We argue, however, that the published data leave room for serious doubt as to the validity of these conclusions. There are a number of comments that we would like to make.
1. Hepatic arterial embolization by itself does not induce liver atrophy (or hypertrophy of the lobe that is not embolized) unless sinusoidal perfusion is compromised either because an embolic agent is used that produces a very distal occlusion or because the portal venous flow is also diminished or occluded (2,3). Coils, as used by Dr Vogl and colleagues for hepatic arterial embolization, produce a localized occlusion, but, almost immediately, collateral vessels develop that reconstitute the more distal intrahepatic branches. Although the coils will reduce the perfusion pressure (in the short term) to the intrahepatic vessels within the embolized lobe, they will not cause liver atrophy or compensatory hypertrophy; any observed change in liver size must therefore be due to other factors.
2. Dr Vogl and colleagues stated in their introduction that "various methods have been proposed to increase the regeneration and functional capacity of unaffected liver segments before surgery. In particular, venous and arterial embolizations have been performed." None of the cited references mention the use of arterial embolization to produce lobar atrophy or compensatory hypertrophy; the reference regarding arterial embolization (4) is a comparative study of hepatic resection and hepatic arterial embolization for the treatment of hepatocellular carcinoma and is not relevant to this topic.
3. Unlike proximal hepatic arterial embolization, portal venous occlusion is a potent cause of hepatic atrophy. It is extraordinary that Dr Vogl and colleagues failed to mention the degree of portal venous compromise by the tumor present in the diseased hepatic lobes that were embolized. Hilar cholangiocarcinomas commonly constrict and/or occlude the main portal venous branches, and it is extremely likely that this occurred to some extent in several (if not all) of the patients who, as stated in the article, had initially inoperable tumors. Is it not much more likely that it was the combination of portal venous involvement and the subsequent hepatic arterial embolization that induced hypertrophy of the nonembolized hepatic lobe? Dr Vogl and colleagues should have determined whether there was any relationship between the degree of portal venous obstruction and the percentage of volume change that subsequently occurred within the left lobe of the liver.
It is also very important to know whether there was any evidence of disease progression between the initial and final computed tomographic (CT) scans in terms of portal venous involvement. Obviously, if some of the patients demonstrated a constricted but patent portal venous branch on their first scan that was occluded on the follow-up scan, this alone, regardless of the arterial embolization, might well explain any observed changes in liver volume.
4. Dr Vogl and colleagues initially used radiologic criteria for diagnosing unresectability. Following embolization, however, they used surgical criteria (at laparotomy or laparoscopy). It is important to know whether the CT scans obtained after hepatic arterial embolization were critically analyzed to see how many of these tumors would still have been reported as demonstrating unresectability. Readers of this article do not know whether the outcome of surgery would have been the same if all the patients had proceeded directly to laparotomy without hepatic arterial embolization.
5. Figures 1 and 2 of the article may cause readers concern as to the accuracy of the liver volume measurements. In figure 1a, the caudate lobe was not included in the volume measurement of the left lobe of the liver before hepatic arterial embolization. In figure 1b, however, the caudate lobe was included in the measurement after hepatic arterial embolization. It is not surprising, therefore, that an "increase" in liver volume was demonstrated in the patient.
Dr Vogl and colleagues mentioned that the right lobe of the liver increased in volume after hepatic arterial embolization and that this was, they believed, due to enlargement of the caudate lobe. In both figures 1 and 2, however, the caudate lobe appears to have been included in the left liver lobe volume measurements. In these two patients at least, therefore, the caudate lobe could not have been the cause of the observed right liver lobe enlargement.
6. Perhaps most important, it appears that Dr Vogl and colleagues did not take into account the progression of disease that occurred between the initial and final CT examinations after hepatic arterial embolization. Biliary tract obstruction, particularly when it involves both the right and left hepatic ducts, may cause an increase in liver volume (5). Only two patients underwent drainage of their left hepatic ducts, and yet all had bilateral duct involvement. Many of these patients will therefore have developed worsening biliary obstruction in this lobe (and in the right lobe if it is not drained). This is seen in figure 2, in which the left ducts are clearly larger on the follow-up scan than on the initial (pre–hepatic arterial embolization) scan. This progression of disease might well, at least partly, explain the observed increase in the volume of the right and left lobes.
7. It is surprising, given the unresectability of the tumors in all these patients on the basis of their initial radiologic examination findings, that only five patients required biliary stent placement due to "severe jaundice." This implies that the remaining patients were either anicteric or only mildly jaundiced and that they, therefore, had a reasonable functional hepatic reserve. It is unusual, in our experience, for patients with unresectable tumors of the hilar bile duct not to require some form of biliary drainage. This is unless the reason for unresectability is bilateral vascular involvement by tumor (eg, right and left portal venous involvement, hepatic arterial involvement in one lobe and portal venous involvement in the other), which was presumably not the case in the patients in this article, as such vascular involvement would not have been improved by embolization.
8. Dr Vogl and colleagues should have given their reasons for diagnosing unresectability in these patients. For example, why was the tumor shown in figure 2 believed to be unresectable? The CT scan obtained before hepatic arterial embolization demonstrates hypertrophy of the left lateral segments and atrophy of the right and quadrate lobes. It is admitted that only a single image from this study was seen, but it appears that the lateral segments of the left lobe of the liver represented at least 30% of the total liver volume (probably much more in terms of functional liver tissue since the extensive tumor involved the remainder of the liver on the CT scan and hepatic arteriogram).
9. In their discussion, Dr Vogl and colleagues suggested that arterial embolization may be advantageous when compared with portal venous embolization "in that it results in simultaneous reduction of tumor growth in the embolized right hepatic lobe." From where did the authors obtain the information that arterial embolization causes a reduction in the size of hilar cholangiocarcinomas? These tumors are almost invariably poorly vascular, and there is no evidence, as far as we know, that arterial embolization (particularly when it is performed only with coils) has any effect on their growth. The authors should have provided evidence from their own data that this does occur or should have provided a suitable reference.
In light of our previous comments, we do not agree that the data presented by Dr Vogl and colleagues have shown that a marked increase in the volume of the nonembolized hepatic lobe will result after hepatic arterial embolization. We believe that there are concerns as to the accuracy of the volume measurements that are based on the submitted images, and there are other factors that the authors did not discuss that we believe are more likely to have had a greater influence on any real lobar volume changes.
References
Department of Radiology, Institute of Diagnostic and Interventional Radiology, University Hospital, Johann Wolfgang Goethe-Universität Frankfurt am Main, Theodor-Stern-Kai 7, 60590 Frankfurt, Germany
1. As of this writing, to our knowledge there are only a few publications on the effects of hepatic arterial occlusion and liver atrophy and hypertrophy (1). Most of the articles published thus far analyze the effect of sinusoidal perfusion on liver atrophy and hypertrophy.
Arterial coil embolization so far has been used only in a preoperative setting for hepatocellular carcinomas, as cited in our article (2). In their comments, Dr Jackson and colleagues state that hepatic arterial embolization by itself does not induce liver atrophy. We believe that our data support the theory of atrophy induced by arterial occlusion, which is still the subject of several animal studies that are being performed at different institutions.
Dr Jackson and colleagues also state that coils, as used in our study, produce only a localized occlusion. This is definitely not true and can be proved angiographically. In our series, total occlusion of the embolized arterial segments, even in the distal parts, was seen at the angiography performed 30 minutes after successful coil embolization in control subjects prior to removal of the introducer sheaths. Of course, this effect is seen only after appropriate placement of the coils. The development of collateral vessels is in our opinion an effect that occurs after several weeks or months. The development of collateral vessels as an immediate or midterm response to vessel occlusion so far, to our knowledge, has not been described.
Results of our ongoing study show a volume increase of the nonembolized liver segments, which, together with the results of the liver function tests, documents hypertrophy of the liver. The exact morphologic changes that take place in this setting are currently being evaluated at our institution.
2. As mentioned previously, to our knowledge no larger studies have been conducted that have used arterial embolization in a preoperative setting for the treatment of cholangiocarcinoma. Our article is of relevance, as it describes the technique used for the transarterial hepatic coil embolization.
3. Of course, all CT scans obtained prior to and after embolization were thoroughly reviewed. There was no evidence of involvement of the portal vein. In addition, there was evidence of compression of portal venous structures within the left lobe on the late follow-up CT scans. Furthermore, we were able to differentiate between tumor progression, portal involvement, and liver hypertrophy. However, this may be difficult if one uses an inadequate imaging protocol.
4. All points made by Dr Jackson and colleagues are well explained in the Materials and Methods of our article. A patient's candidacy for surgery was determined on the basis of exact analysis of all the helical CT scans and liver function test results by a panel of experienced radiologists and surgeons. The patient's tumor was classified as not resectable when liver function test results indicated low liver parenchymal reserve and when the CT examination revealed a low volume of the left liver lobe. The decision on the surgical technique was always made by the surgeon. Proceeding directly to surgery was not possible for all patients in our study.
5. It is true that the caudate lobe was accidentally included in segmentation in figure 1b. We checked the volumetric data for this patient through reanalysis and recalculation of the results of the CT. The data are correct; the image was obtained from doing a series of measurements to determine caudate lobe enlargement. Yet, even on this image, it is evident that there is a relevant increase in the volume of the left liver lobe.
6. We did take the progression of disease into account. In our opinion, the progression of disease within the 44 days between embolization and surgery is not responsible for the marked increase in liver volume. With regard to Dr Jackson and colleagues' concerns about figure 2, we state that there is no evidence of progression of disease or biliary obstruction if all the images are analyzed. The section in figure 2d is slightly more cranial compared with the section in figure 2c; nevertheless, we chose these two images because they were the most alike.
7. The placement of a biliary stent was not successful in all patients who required it. In the other patients, the placement of a biliary stent was not necessary, as they were either anicteric or only mildly jaundiced. Hepatic function was impaired in all patients. The indication for the placement of some kind of biliary drain in patients with hilar cholangiocarcinomas varies in the study findings published so far. There was no evidence of involvement of either the proximal hepatic artery or the portal vein in any patient in our study.
8. Again, we need to refer Dr Jackson and colleagues to the Materials and Methods in our article. The determination of resectability cannot be made on the basis of a single CT image. The volumetric data for the patient in figure 2 show that there was a 42% volume increase after embolization. One always has to include the entire CT examination to determine liver volume and morphologic changes because of the tumor.
9. In our opinion, any occlusion of an arterial tumor supply results in the reduction of tumor growth, if not destruction. This also applies to poorly vascularized tumors.
It is true that, compared with other liver malignancies, cholangiocellular carcinomas reveal some differences in tumor vascularity and perfusion since they are also supplied via the portal vessels. The effect of arterial coil embolization does not differ from that achieved by other embolization materials if the coils are placed appropriately.
Because hilar cholangiocarcinomas of Bismuth grade III also show an involvement of at least parts of portal blood vessels, they do subsequently receive a greater amount of their blood supply via arterial vessels. Therefore, an arterial embolization should also induce a reduced tumor growth. Only one patient in our series revealed a slight tumor progression, as proved at laparoscopic exploration. In this case, tumor progression was already suspected during the follow-up helical CT examination. In the other three patients in whom surgery could not be performed, there was no evidence of tumor progression, but the location of the tumor made surgical resection impossible. We did not supply data on tumor growth after arterial coil embolization because it is still a subject of research at our institution and is not of primary interest in the setting of the presented study.
Overall, we thank Dr Jackson and colleagues for their different points of view on this subject. Nevertheless, we think that our data do support the theory that arterial coil embolization leads to a significant increase in liver volume.
Finally, we believe that the benefits of arterial embolization over portal venous embolization in the reduction of tumor growth and achievable volume increase still need to be evaluated in a randomized study.
References
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| RADIOLOGY | RADIOGRAPHICS | RSNA JOURNALS ONLINE |
