| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1
Department of Radiology, Stanford University Medical Center, 3000 Pasteur Dr.,
H-3646, Stanford, CA 94305-5642.
2
Department of Surgery, Stanford University Medical Center, Stanford, CA
94305-5642.
Received May 1, 2001;
accepted after revision June 20, 2001.
Address correspondence to D. Y. Sze.
Abstract
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
SUBJECTS AND METHODS. Thirty patients were referred for chemoembolization of unresectable hepatocellular carcinoma. Initial selective chemoembolization plans were formulated on the basis of diagnostic biphasic CT or MR imaging. Ultrafast CT hepatic arteriography was performed using a multidetector CT scanner and selective contrast material injection into the hepatic artery. The entire liver was scanned in a single breath-hold of approximately 20 sec with a slice thickness of 1 mm. Lesions and their arterial supplies were identified, and these data were immediately used to formulate a final plan for chemoembolization.
RESULTS. Hypervascular masses were detected in 29 patients. In 16 (53%) of the patients, preprocedural CT or MR imaging underestimated the number of lesions. In nine (30%) of these 16 patients, the additional lesions were detected only on CT hepatic arteriography, not on conventional angiography. CT hepatic arteriography findings had a major impact on planning the way in which chemoembolization treatment was performed. In three of the nine patients, the previously undetected lesions were treated with additional superselective chemoembolization. In the other six patients, chemoembolization was performed less selectively than originally planned.
CONCLUSION. Primarily because of the high sensitivity of multidetector CT hepatic arteriography in revealing small and multifocal hepatomas, findings of this modality frequently alter treatment plans involving selective administration of chemoembolic material.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Chemoembolization of the liver may also be limited by comorbidities, especially cirrhosis and hepatic failure. Superselective chemoembolization is gaining acceptance as a way to minimize collateral damage to the functioning liver and to maximize tumoricidal efficacy [4], but increased selectivity could result in incomplete treatment, particularly of small lesions that are difficult to detect. The sensitivity of non-invasive diagnostic imaging, including CT, MR imaging, and sonography, is 70-90% for larger lesions but falls to less than 50% for lesions measuring less than 1 cm in diameter [5,6,7,8]. Improved sensitivity is possible with more invasive means, such as CT arterial portography, intraoperative sonography, and CT hepatic arteriography with direct hepatic arterial contrast material injection. Earlier CT hepatic arteriography protocols were limited by greater slice thickness and slow scanner speed [9,10,11]. In this study, we investigated the feasibility of high-resolution ultrafast CT hepatic arteriography using a multidetector CT scanner [12] and evaluated the impact that these scans had on the planning and the performance of chemoembolization.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
All patients underwent multiphasic CT before referral for chemoembolization, except four patients who underwent contrast-enhanced MR imaging because of azotemia. Only 13 patients had their diagnostic CT scans obtained at our institution; the remaining 17 were referred from other institutions and received their imaging workups before referral. The quality of these outside scans varied, but all were obtained on helical scanners using a biphasic protocol.
Of the 30 patients, four had previously undergone chemoembolization, two
had undergone surgical resections, and one had received an orthotopic liver
transplant. One patient had been treated with proton beam irradiation, and
another with percutaneous ethanol ablation. All had recurrent tumor. The other
21 patients had received no prior treatment for hepatocellular carcinoma.
Twenty-three patients had elevated serum 
-fetoprotein levels, and
overall the mean and median levels were 6,022 and 179 ng/mL, respectively.
Only 11 patients had biopsy-confirmed hepatocellular carcinoma. The others
were presumed to have hepatocellular carcinoma because of typical imaging
characteristics of the disease—a hypervascular mass and two of the
following three criteria: cirrhosis, viral hepatitis, or an elevated level of
-fetoprotein (>150 ng/mL, or increasing >20 ng/mL per month).
Patients were presented at a multidisciplinary tumor board and judged to have
unresectable tumors on the basis of tumor characteristics or comorbidities or
both.
Preliminary plans for selective hepatic arteriography and chemoembolization were prepared after examination of diagnostic CT or MR imaging. The planning entailed localizing each hypervascular lesion and prospectively recording the segmental location. Plans emphasized using the greatest selectivity possible to minimize collateral damage to the healthier portions of the liver.
After giving informed consent, patients underwent flush aortography and complete hepatic angiography, usually using a Simmons or Rosch hepatic catheter with a side hole (Beacon Tip Torcon NB; Cook, Bloomington, IN) with the tip positioned in the proper or common hepatic artery. Anteroposterior and 30° right anterior oblique images were obtained, and the hypervascular masses identified were matched with those seen on previous diagnostic CT or MR imaging. Any additional masses detected were recorded.
In patients with anomalous arterial supply, a single catheter with custom sideholes or two separate catheters from bilateral femoral accesses were used. For patients with replaced left hepatic arteries and dominant masses in the left lobe, selective catheterization of the replaced left hepatic artery was performed with a customized catheter with five sideholes punched at the level at which the catheter traversed the celiac axis to allow opacification of the proper hepatic artery. Similarly, for patients with replaced left hepatic arteries and dominant masses in the right lobe, such a catheter was placed with the tip in the proper hepatic artery with sideholes in the celiac axis to allow contrast material to flow into the left gastric artery. For patients with replaced right hepatic arteries, separate catheters were placed in the replaced right and proper hepatic arteries through bilateral femoral access.
Patients were then transferred by gurney to the multidetector CT (LightSpeed QX/i; General Electric Medical Systems, Milwaukee, WI). During a single breath-hold lasting up to 28 sec, diluted contrast material containing 140 mg/mL iodine (Omnipaque 140; Nycomed, Princeton, NJ) was injected into the hepatic arterial catheter for the duration of the scan. The rate of injection was 1.5 ± 0.5 mL/sec, adjusted according to the flow rate of the hepatic artery observed during digital subtraction angiography (DSA) and to minimize reflux into the gastroduodenal artery. After a 7-sec delay, scanning was performed from cephalad to caudad at 1.25-mm collimation, 4-row detection, and a pitch of 6. Images were reconstructed at 1-mm intervals. Images were immediately reviewed by an interventional radiologist and compared with the prior diagnostic study; additional hypervascular lesions were identified and recorded. Arterioportal shunts were differentiated from small hepatocellular carcinomas by the shape and location of lesions [13,14,15]. A final plan for chemoembolization was then formulated on the basis of the findings. The total contrast material load from complete hepatic angiography and CT hepatic arteriography averaged 110 mL.
Patients with confirmed hypervascular masses were then treated by chemoinfusion of an emulsion containing ethiodated poppyseed oil (Ethiodol; Savage, Melville, NY), cisplatinum, and adriamycin. Maximum doses were 20 mL of Ethiodol, 50 mg of cisplatinum, and 50 mg of adriamycin. Arterial supply to identified lesions was subselected using a microcatheter (Rapid Transit; Cordis, Warren, NJ), and the emulsion was infused until stasis was achieved. In patients in whom flow was particularly fast or persisted after initial chemoinfusion, Gelfoam (Upjohn, Kalamazoo, MI) slurry embolization was added during or after chemoinfusion. In seven patients, equivocal lesions that were identified on CT hepatic arteriography were not visible on DSA and did not obviously sequester Ethiodol; these patients underwent unenhanced CT 24-72 hr after chemoembolization to evaluate Ethiodol sequestration.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Of the 30 patients, hypervascular lesions characteristic of hepatocellular carcinoma were revealed on CT hepatic arteriography in 29 patients. The mean number of lesions detected on CT hepatic arteriography in patients with countable lesions was 2.58 (range, 1-18 lesions), which was greater than the mean number detected on DSA—2.04 (range, 1-9 lesions) (Fig. 1A,1B,1C,1D,1E). The additional lesions found in these patients were as large as to 1.7 cm in diameter, but most were small (mean, 0.9 cm) (Fig. 2A,2B,2C,2D,2E,2F,2G,2H). However, some of the additional lesions identified through either method were satellite lesions adjacent to larger masses and were in the same vascular distribution as the parent lesion. Therefore, the identification of these additional lesions did not necessarily alter the chemoembolization plans. Surprisingly, only four patients had anomalous hepatic arterial anatomy, two with a replaced left and two with a replaced right hepatic artery.
|
|
|
|
|
|
|
|
|
|
|
|
|
CT hepatic arteriographic results exactly matched the diagnostic CT or MR imaging results in only 11 patients (36.7%). Of the 19 patients in whom the results did not match, 16 had additional lesions identified, and three had fewer lesions identified. In two patients, CT hepatic arteriography revealed that ill-defined hypervascular lesions seen on diagnostic CT probably represented perfusion abnormalities fed by the portal vein rather than by the hepatic artery. One of these patients received chemoembolization only for the identified but unresectable tumor, which represented a simpler treatment than that originally planned. The other patient, found now with only a solitary lesion and mild comorbidities, was reevaluated by surgeons and deemed a surgical candidate, and received no chemoembolization at all.
In the other patient in whom fewer lesions were identified, no hypervascular masses were identified on CT hepatic arteriography at all. This patient had a recurrent 2.5-cm hypervascular mass in segment IVa approximately 7 months after a chemoembolization procedure. DSA and CT hepatic arteriography both showed multiple segmental arterial occlusions, presumably caused by prior chemoembolization [16] and no direct arterial supply to the tumor from the hepatic artery. Parasitized supply was identified on DSA from the left internal thoracic artery and was treated by superselective chemoembolization (Fig. 3A,3B,3C,3D).
|
|
|
|
Parasitized supply was also identified from an omental branch of the gastroduodenal artery in two other patients. In one patient, the gastroduodenal artery arose from the right hepatic artery distal to the tip of the catheter, and this supply was depicted on a CT hepatic arteriogram. In the other patient, the omental branch was not included in the CT hepatic arteriography contrast material injection, resulting in incomplete opacification of the tumor. The parasitized omental supply was identified on the DSA hepatic arteriogram in both patients, and the arteries were selected and treated by chemoembolization.
Sixteen patients had additional lesions identified. Additional lesions were identified not only on CT hepatic arteriography but also on DSA in seven patients. Because DSA is performed routinely as part of a chemoembolization procedure, CT hepatic arteriography had no net effect on the treatment of these patients. However, in nine patients (30.0%), the additional lesions were not detected on DSA, only on CT hepatic arteriography. Additional subselective angiography was performed in three patients until these other lesions were found, and subselective chemoembolization was then performed. In the other six patients, chemoembolization was administered in a less selective fashion than originally planned in order to include the segment(s) containing the previously undetected lesions, and these lesions did sequester Ethiodol at the time of chemoembolization.
Twenty of the patients also had enhancing lesions on CT hepatic arteriography suspected to be arterioportal shunts. These were generally peripheral, small, wedge-shaped, and frequently ill defined. On DSA, the lesions were not seen as discrete, round lesions, and parallel hepatic artery and portal vein branches were sometimes opacified. In seven of these patients, equivocal lesions suspected to be arterioportal shunts were also treated by chemoembolization. These patients underwent unenhanced CT 24 to 72 hr after chemoembolization treatment, and no Ethiodol sequestration was detected in these equivocal lesions. In two other patients, however, a lesion originally interpreted to be a benign arterioportal shunt was eventually found to be larger at follow-up CT, indicating that, in retrospect, these suspected arterioportal shunts were probably small malignancies.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
The improved speed and spatial resolution of CT scanners have made feasible high-definition CT angiography. CT angiography has become a new standard in certain applications, including evaluation of aortic stentgrafts and screening for pulmonary emboli. Speed of scanning has increased approximately threefold through the use of multirow detection [12]. Current designs allow simultaneous acquisition of four helical trajectories, resulting in the capability of scanning at a resolution of 1-mm slice thickness at the speed of approximately 1 cm (10 slices) per sec. A healthy liver can be scanned in 20-25 sec, easily performed in one breath-hold, and a small cirrhotic liver can be scanned even more quickly.
Emerging data provide supporting evidence that detection of small and multiple lesions is improved using multidetector technology [18], but even the most aggressive IV contrast material bolus will be diluted and scattered by the time it reaches the hepatic artery; isolating the arterial from the portal phase completely may be difficult or impossible. It is probably for this reason that DSA can often detect hypervascular masses that were missed on contrast-enhanced CT. Combining the advantages of each, CT hepatic arteriography maintains the very high lesion-to-background contrast of DSA and adds the three-dimensional spatial resolution of CT, resulting in extremely high sensitivity for hypervascular lesions [17, 19, 20]. Sensitivity is especially improved for small lesions and for lesions adjacent to and obscured by others. The first attempts at CT hepatic arteriography were severely limited by scanning speed [10, 11], but with the advent of multiple detector technology, a CT hepatic arteriographic study can be performed in the same amount of time as—and with less total iodine than—a DSA. The impact that CT hepatic arteriography may exert on the treatment of patients with hepatocellular carcinoma reflects this improved sensitivity to multiple and small lesions. In 30% of the patients in this study, the actual chemoembolization procedure performed was altered because of the results of CT hepatic arteriography. In another 23%, CT hepatic arteriography matched the results of DSA, and both were more sensitive than CT. The treatment plans resulting from CT hepatic arteriography exactly matched those resulting from diagnostic multiphasic CT in only one third of patients. Identification of additional lesions can alter the plans for selective chemoembolization and can also affect the determination of resectability.
The clinical importance of detecting such small tumors remains unknown. Most hepatocellular carcinomas are slow-growing tumors, and as long as a patient is compliant and his or her physician is conscientious about follow-up, these small tumors can be treated when they grow to exceed a threshold of detectability. However, even small tumors have associated morbidity, mortality, and metastatic potential [21]. The most encouraging statistics for survival after treatment of hepatocellular carcinoma continue to come from surgical resection in qualifying candidates whose 5-year survival rates average approximately 20% [22,23,24,25,26]. These numbers apply to patients with single lesions, Child-Pugh class A liver function, and minimal comorbidities. Even in patients who do not qualify for surgical resection, complete treatment of all malignant lesions, regardless of size, would seem to be optimal before the relentless progression of cirrhosis precludes aggressive therapy.
Because most patients with hepatocellular carcinoma have underlying cirrhosis and hepatic dysfunction, treatments are designed to be hepatocyte-sparing [27]. For surgery, the goal of sparing hepatocytes translates to smaller segmental or nonanatomic resections [28], and for chemoembolization, it translates to subselective treatment through microcatheters. The drawback of hepatocyte-sparing therapies, though, is that primary tumors may be incompletely treated, and distant lesions may be entirely untreated. Even in patients who undergo curative resection, the rate of recurrence in patients with cirrhosis is extremely high, approaching 100% in 5 years [22,23,24,25]. The rate of recurrence after chemoembolization is also very high, to the point that chemoembolization is considered by some to be "tumor maintenance" rather than definitive oncolytic therapy. The high rates of recurrence are frequently attributed to suboptimal treatments, new metastases, or new primary tumors in the high-risk substrate or field defect of a cirrhotic liver. The frequent identification of multiple additional lesions through CT hepatic arteriography thus offers another explanation: these recurrences may represent growth of tumors that previously went undetected, either because of their small size or suboptimal imaging.
Several drawbacks of CT hepatic arteriography were evident in this study. As might be expected, increased sensitivity is accompanied by potentially decreased specificity. Indeed, 66.7% of our patients revealed hypervascular lesions that were not stereotypical for hepatocellular carcinoma and were suspected of being arterioportal shunts [13,14,15]. The criteria for diagnosis of arterioportal shunts include a small size, wedge shape, and peripheral subcapsular location. Unfortunately, some hepatocellular carcinomas also exhibit macroscopic arterioportal shunting, and some benign arterioportal shunts may be large, central, or rounded. Only two of our patients (6.7%) had lesions interpreted as arterioportal shunts that later proved to be malignant, but longer follow-up will certainly reveal additional false-negatives. In addition, other neoplasms, including adenomas and regenerating nodules, may also be hypervascular, and the successful obliteration of these lesions may have no clinical impact at all.
Another drawback of CT hepatic arteriography is the difficulty in detecting lesions fed by anomalous or parasitized extrahepatic arterial supply [29]. Two of our patients had replaced left hepatic arteries, and injection of contrast material into custom catheters with sideholes cut at the level of the celiac axis was necessary to perfuse both left and right lobes. Resultant enhancement was suboptimal in the lobe opacified via the sideholes. More complete enhancement could have been achieved by placing catheters into both hepatic arteries either concomitantly or sequentially, but this procedure would have required either placing two catheters through the celiac axis or forcing the patient to make two trips to the CT scanner. In the patients with the replaced right hepatic arteries, complete CT hepatic arteriography was accomplished with bilateral femoral access.
Tumors perfused exclusively by extrahepatic arterial supply are also missed on CT hepatic arteriography. This drawback is particularly relevant in patients who have previously undergone chemoembolization and may have chemotherapy-induced vasculitis or occlusions [16]. Detection of these lesions and complete delineation of blood supply to tumors even partially fed by extrahepatic collaterals require that contrast material be administered less selectively, either IV or directly into the aorta. When parasitized supply is identified, these arteries can also be treated with selective chemoembolization. When vasculitis is identified on DSA, CT hepatic arteriography may not yield sufficient information to warrant the additional radiation, contrast material, and effort.
Lastly, the main disincentive to performing CT hepatic arteriography at our institution is the required coordination of moving the patient from the angiography suite to the CT suite and back while keeping a catheter positioned in the hepatic artery. In institutions in which chemoembolization is the primary function of the angiography suite, custom installations have been devised so that the patient may remain on a procedure table that may be used with both a CT scanner and an adjacent C-arm angiography unit [30]. Similar combined CT and angiography units are becoming commercially available.
In conclusion, the findings of ultrafast high-resolution CT hepatic arteriography using multidetector CT may substantially alter the plans for selective chemoembolization treatment of hepatocellular carcinoma that were based on the findings of standard preprocedural biphasic CT or MR imaging and intraprocedural DSA. This influence of multidetector CT hepatic arteriography is primarily due to the modality's increased sensitivity to small and multifocal lesions. In addition, CT hepatic arteriograms may have an impact on determination of surgical resectability. However, differentiating small malignant lesions from arterioportal shunts and other benign lesions remains difficult, and the survival benefit of treating small malignant lesions remains unknown.
Acknowledgments
The authors thank Laura Logan and the Lucas Center 3D Laboratory for image
processing.
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
This article has been cited by other articles:
|
|
 |
|
  B. C. Meyer, B. B. Frericks, M. Voges, M. Borchert, P. Martus, J. Justiz, K.-J. Wolf, and F. K. Wacker Visualization of Hypervascular Liver Lesions During TACE: Comparison of Angiographic C-Arm CT and MDCT Am. J. Roentgenol., April 1, 2008; 190(4): W263 - W269. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
1 cm) suspected of being hepatocellular carcinoma. Patient was treated by
two sessions of lobar chemoinfusion, covering entire liver.
