Clinical Setting,
Patients and BRTO procedures
This prospective,
cohort study was approved by our local institutional review board and included all consecutive patients with gastric varices (Fig.
1A) undergoing BRTO at a single academic hospital.
Twenty consecutive patients (Table 1,
mean age = 62.4 years,
range 31-81,
9 males and 11 females) with gastric varices were included in the study cohort between July 2005 and November 2008.
At first the portosystemic collaterals were assessed using multislice (MS) computed tomography (CT) (Sensation Cardiac,
Siemens,
Erlangen,
Germany) (Fig.
1B) or magnetic resonance (MR) imaging (Intera Archieva,
Philips,
Netherlands) with several postprocessing techniques including multiplanar reconstructions and volume renderings.
Prior to BRTO,
conventional arterioportography was performed to evaluate hemodynamics and to exclude specific causes of portal hypertension such as massive arterioportal shunt or portal venous thrombosis [4].
If the efferent vein was not delineated,
FACT during arterial portography (FACT-AP) was added [5,
6].
The BRTO procedure was performed according to the method of Kanagawa et al [1].
Briefly,
a 5.2 or 6 French balloon catheter (11- or 20-mm balloon diameter (Clinical Supply,
Gifu,
Japan)) was inserted from the right internal jugular or femoral vein and wedged into the left adrenal vein.
First,
balloon-occluded left adrenal venography (BOAV,
Fig.
1C) was performed with the balloon inflated by air or diluted contrast medium after the administration of effervescent salts to expand the stomach.
Then,
repeat BOAV during FACT acquisition (Fig.1D-F [BOAV-FACT]) was performed to confirm that the contrast medium opacification had reached the target gastric vessels.
If not,
additional interventions including downgrading and coaxial advancement of a microcatheter were performed and FACT was repeated until full opacification of the target varices was obtained.
Downgrading [7] consists of selective occlusion of the venous collaterals using absolute ethanol or coils for patients with varices classified as grade 3 or higher in Hirota’s classification [2].
After obtaining opacification of the target varicose veins,
a mixture of 10 mL of 10% ethanolamine oleate (EO) (Oldamin,
MOCHIDA Pharmaceutical,
Tokyo,
Japan) and 10 mL of contrast medium (Iomeron 400,
Eisai co.,
Tokyo,
Japan) or a foam made of 2 mL of 3% polidocanol (Polidocasklerol,
ZERIA Pharmaceutical,
Tokyo,
Japan) and 8 mL of air using a pumping method via a three-way stopcock was injected into the target varices (Fig.
1G-H),with confirmation of appropriate location of the sclerosing agent by additional FACT (BRTO-FACT).
At least 6 hours later the catheter was retrieved after confirming variceal thrombosis by the absence of blood return on aspiration and by injecting small amount of contrast media under fluoroscopy and observing hemostasis even after balloon deflation.
Thereafter,
procedure-related complications were recorded during subsequent hospital stay and defined as major if they involved more than standard post-BRTO care and/or lead to prolonged hospital stay.
The patients were followed up by CT (Fig.
1I) and endoscopy (Fig.
1J).
FACT Technique
The procedures were performed in an interventional radiology suite equipped with a commercially available ceiling-mounted DSA unit (AXIOM Artis FD system and DynaCT,
software version VB30E,
Siemens,
Erlangen,
Germany).
Several series of 3D rotational flat panel detector angiographic images were obtained during breathhold that covered 200 degrees of circular trajectory for 4 seconds.
For FACT-AP,
FACT image acquisition was initiated 20 seconds after the injection of iodinated contrast medium (Iomeron 350; Eisai,
Tokyo,
Japan) at a flow rate of 3-5 mL/sec for 5 seconds.
For the other FACT examinations,
image acquisitions were obtained immediately after the conventional DSA using manual injection of iodinated contrast during balloon occlusion.
Acquired images (240 images; frame rate,
24.8/sec; 0.8 degree of increment for each frame; resolution,
1,024 pixels) were transferred to a workstation (Leonardo,
Siemens),
where 3D CT-like images were reconstructed in approximately two minutes.
Data Analysis
The visibility of the target gastric varices on BOAV without and with FACT (1; definitely invisible,
2; probably invisible,
3; undetermined,
4; probably visible,
5; definitely visible) was independently recorded by two attending interventional radiologists,
each of whom had more than ten years of clinical experience in academic practice (J.K.,
Y.K.).
After independent interpretations,
the differences in assessment between both radiologists were resolved by consensus and the diagnostic confidence level regarding whether the target varices were opacified or not (0: undetermined,
1: probable,
2: definite) on BOAV without and with FACT was compared.
To evaluate the level of interobserver agreement with regard to the visibility of the target gastric varices,
a Kendall W test was performed.
Kendall W coefficients between 0.5 and 0.8 were considered to indicate good agreement,
and coefficients higher than 0.8 were considered as excellent agreement.
To determine statistical differences in the diagnostic confidence level between FACT-BOAV and conventional BOAV,
Wilcoxon signed rank test statistics were calculated.
Finally,
the behavior of the operators was analyzed in consensus with respect to procedural technique during BRTO and the influence of FACT on operators’ strategy/technique was graded in three categories: category 1 if no relevant additional information was gathered,
category 2 if more confidence was obtained by performing BOAV-FACT but without leading to technical/strategical modifications during BRTO,
and category 3 if FACT provided additional information that influenced the operator’s decision making process enough to modify the steps in the procedural technique.