Downloaded from www.ajronline.org by 193.255.88.62 on 01/09/15 from IP address 193.255.88.62. Copyright ARRS. For personal use only; all rights reserved

293

Diagnosis of Portal Vein Thrombosis: Value of Color Doppler Imaging

Franklin

N. Tessler1

Brian J. Gehring1 Antoinette S. Gomes1 Rita R. Perrella1

Nagesh Ragavendra1 Ronald W. Busuttil2 Edward

G. Grant1

This study was undertaken to determine the accuracy of color Doppler imaging in the diagnosis of portal vein thrombosis. Two hundred fifteen patients were studied with color Doppler imaging to determine patency of the main portal vein. Sonographic findings were confirmed in 75 patients, aged 19 to 66 years Correlation with angiography was obtained in 13 patients, and surgical correlation was obtained in the remaining 62. Nine patients had portal vein thrombosis on the basis of these gold standards. Sonograms were classified as showing either patency or thrombosis, depending on the ability to show color flow within the main portal vein. Agreement between sonography and angiography or surgery was found in 69 patients (61 patent, eight thrombosed). One patient with a patent portal vein at sonography was found to have a thrombosed vessel at surgery, whereas five patients without portal venous flow at sonography had patent vessels at angiography (one patient) or surgery (four patients). Overall sensitivity and specificity for detection of portal vein thrombosis were 89% and 92%, with an accuracy of 92%, a false-negative rate of 0.11, a negative predictive value of 0.98, and a positive predictive value of 0.62. We postulate that the majority of errors in our study occurred in vessels that, although patent, had only sluggish flow, which could not be resolved because of technical limitations. We conclude that color Doppler imaging is a valuable screening procedure for the assessment of portal vein patency. If the sonogram shows a patent portal vein, no further studies are required. However, a lack of demonstrable flow does not always indicate thrombosis, and other imaging studies should be performed for confirmation. AJR

157:293-296,

Portal real-time

August

1991

vein thrombosis (PVT) can be diagnosed noninvasively or duplex Doppler sonography [1 -9], CT [7], or MR

does not provide

sonography

information

about flow dynamics

by using either [1 0]. Gray-scale

in the portal vein.

Instead, the diagnosis of PVT is based on the demonstration of echogenic thrombus and other morphologic features [3]. Unfortunately, if the thrombus in hypoechoic, or if the lumen appears relatively echogenic despite vascular patency, portal venous thrombosis may be missed or falsely diagnosed [9]. Use of Doppler sonography, with its ability to detect and characterize blood flow, can improve diagnostic

accuracy Received February 7, 1 991 vision March 28, 1991. 1 Department of Radiological

;

accepted

after

Sciences,

re-

UCLA

School of Medicine, CHS 77-132, 10833 Le Conte Ave., Los Angeles, CA 90024-6904. Address reprint requests to F. N. Tessler. 2 Department of Surgery, UCLA

School

of Med-

a small

[5, 9]. However, volume

to obtain

necessary

duplex

Doppler

0361 -803X/91/1

in both

the

©

American

10833

Le Conte

572-0293

Roentgen

Ray Society

Ave.,

Los

color-coded

(CDI) provides

Angeles, CA 90024-1721.

BR-272,

Doppler

time;

sonography

therefore,

a global impression

By superimposing imaging

of interest,

CHS

duplex given

only measures

sampling

of flow within

an apparently normal Doppler signal may be obtained patients who have occlusion of the portal vein [5].

region

icine,

at any

lower

used for evaluation

a rapid,

thereby

imaging and

information noninvasive

overcoming

from

the portal from

on a gray-scale means

many

extremities

of the portal

venous

[1 2, 1 3].

limitations

system;

In two

however

is

vessels

in

image, color Doppler

[1 1]. CDI has been used to diagnose upper

sites

vein. Moreover,

collateral

of mapping

of the

flow within

multiple

recent

blood

flow

within a and thrombosis

of gray-scale

venous studies,

independent

CDI

was

confirmation

294

TESSLER

Downloaded from www.ajronline.org by 193.255.88.62 on 01/09/15 from IP address 193.255.88.62. Copyright ARRS. For personal use only; all rights reserved

of the sonographic findings was not performed [1 4, 15]. The purpose of our study was to determine if CDI could be used to diagnose PVT. The results of angiography and surgery were used as the gold standards.

Subjects

and

ET AL.

AJR:157,

flow.

Because

of CDI in detecting Between September 1988 and May 1 990, CDI was used to evaluate the hepatic vasculature in 21 5 patients. Color flow imaging of the portal vein was performed by using commercially available equipment (Ultramark 9, Advanced Technology Laboratories, Bothell, WA) and 2.25- and 3.0-MHz phased-array transducers. Real-time sonography was used first to locate the portal vein at the level of the porta hepatis. The color flow system was then activated, the region of interest was placed over the portal vein, and the scanning parameters were adjusted until consistent, nonrandom color appeared within the vessel (Fig. 1). Whenever flow could not be shown despite careful manipulation of CDI controls (Fig. 2), a duplex Doppler examination of the portal vein was performed also. If the portal vein could not be seen with real-time sonography, color and spectral Doppler imaging of the porta hepatis were performed to search for venous flow. CDI of the hepatic veins, the inferior vena cava, and the hepatic artery was performed

also in selected cases, depending the CDI studies knowledge

were

performed

of angiographic

Angiographic

or

on the indication by experienced

for the study. All operators,

without

findings.

surgical

confirmation

of

portal

vein

patency

or

thrombosis was available in 75 of the 21 5 patients (34 men, 41 women; age range, 19 to 66 years, mean 45 years). These patients

made up the final study group. Seventy sonograms as part of a workup for hepatic transplantation,

two

were performed

patients.

angiographically

because of transplant failure, one prior to portosystemic shunt surgery, and two after shunt procedures. Angiographic correlation was obtained in 13

with

The

high-dose

portal

venous

portal

system

arteriography

was

evaluated

with

selective

celiac

and/or

splenic artery and superior mesenteric artery injections. Tolazoline 25 mg (Ciba-Geigy, Summit, NJ) was administered intraarterially for the superior mesenteric artery injections. digital subtraction techniques were used

Delayed films and film or when needed to evaluate

1991

the portal venous system. The interval between CDI and angiography ranged from 0 to 334 days, with a median of 2 days. Surgical correlation was obtained in 62 patients. The interval between CDI and surgery ranged from 2 to 327 days, with a median of 90 days. The sonograms and angiograms were classified as showing either patency or thrombosis of the main portal vein; no attempt was made to grade

Methods

August

our

sole

criterion

our

purpose

PVT, absence for diagnosing

was

of color PVT.

to determine

within

Although

the

the portal cases

were

accuracy

vein was reviewed

retrospectively, this classification was based entirely on the appearance of the portal vein at the time of the examination. Surgical correlation was obtained through a review of surgical reports and summary sheets, which clearly indicated the condition of the portal vein at surgery.

Results

In 1 3 patients

in whom

CDI was compared

with angiogra-

phy,

the sensitivity for detection of PVT was 1 00%, with a specificity of 88% and a false-negative rate of zero. In the group of 62 patients with surgical correlation, sensitivity and specificity were 75% and 93%, respectively, with a falsenegative rate of 0.25. In all, CDI agreed with the gold-standard examination in 69 patients (61 patent, eight thrombosed) and disagreed in six patients (five with a false-positive diagnosis of thrombosis and one with a false-negative diagnosis). Overall sensitivity was 89%, specificity was 92%, and the falsenegative rate was 0.1 1 . Accuracy for diagnosis of PVT in all 75 patients was 92%, with a negative predictive value of 0.98 and a positive predictive value of 0.62. In 1 2 patients in whom angiography was compared with surgery, sensitivity and specificity were 1 00% and 90%, with a false-negative rate of

zero. To assess

the possible

effect

of an excessive

delay

be-

tween CDI and the gold-standard examination, we performed a separate analysis in a subset of 24 patients in whom 1 month or less elapsed between sonography and angiography or surgery. In this group, sensitivity for detection of PVT was 1 00%, specificity was 89%, and the false-negative rate was

zero. One patient in whom sonography showed a patent portal vein had a thrombosed vessel found during surgery; however, 9 months had elapsed between CDI and surgery. Another patient in whom the portal vein was patent on CDI and

nonvisualized

on angiography

had a patent vessel at surgery.

Discussion

Several features of PVT on real-time sonography have been described, including echogenic thrombus, cavernous trans-

Fig. 1.-65-year-old woman with a patent portal vein, proved surgically. Oblique color Doppler sonogram of porta hepatis shows portal vein (P) filled with color. Because flow is directed toward transducer (into liver), it Is depicted in red. Blue color indicates flow in a branch directed away from transducer. Hepatic artery (arrowhead) is seen in cross section, anterior to portal vein.

formation of the portal vein, collateral formation, and venous enlargement [3]. However, Van Gansbeke et al. [3] were unable to show intraluminal thrombus in 33% of their patients with PVT. In another series comparing duplex Doppler sonography with angiography, Nelson et al. [5] had four patients with PVT; none of thefour had intraluminal thrombus identified with real-time sonography, and the diameters of the portal vein were normal in all four. In our nine patients with proved PVT, portal vein echogenicity varied from hypoechoic with internal echoes to echogenic. Significantly, the portal vein

AJA:157,

August

COLOR

1991

DOPPLER

OF

PORTAL

Downloaded from www.ajronline.org by 193.255.88.62 on 01/09/15 from IP address 193.255.88.62. Copyright ARRS. For personal use only; all rights reserved

could not be located by using real-time sonography in four patients in this group. However, the portal vein may not be visualized when gray-scale imaging is used, even in the

absence of thrombosis [1]. Duplex Doppler sonography, with its ability to show flow, can provide direct evidence of portal vein patency [7]. However, Parvey et al. [1 6] characterized duplex Doppler sonography of the portal vein as “a prodigious technical undertaking” and speculated might be overstating

that CDI might be helpful. Although this the difficulty somewhat, duplex Doppler

VEIN

THROMBOSIS

295

had an impact in cases in which CDI and the gold-standard procedure both showed PVT (eight patients). However, the mean interval between CDI and angiography was only 3.4 days in five patients with angiographic correlation. Of three patients with surgically proved thrombosis, two had an interval between examinations of only 21 days; 4 months elapsed

between

CDI and surgery

that the probability that these patients is low.

An excessive

interval

in the remaining thrombosis

patient.

developed

may have contributed

We think

after

CDI

in

to a discrep-

In the study by Nelson et al. [5], seven of eight patients

with

ancy between CDI and surgical findings in one patient who had a patent portal vein on CDI and a thrombosed vessel at

what

had

surgery

sonography

has

were

a variety

described

of pitfalls

and

as “inadequate

technical

Doppler

limitations.

tracings”

patent vessels at angiography. CDI combines gray-scale and spectral Doppler imaging and has been used to diagnose thrombosis in the extremities [12, 13], in portosystemic shunts [17], and in the hepatic veins [1 8]. In a recent study by RaIls [15], CDI was used to detect thrombosis

criteria validated

of the main

similar to ours. independently

surgery. For many patients between

CDI

and

vein

in 30 patients

Unfortunately,

portal

the findings

by any correlative

or

surgery.

which patent

not skewed examinations

sensitivity

would

was the only false-negative

have

been

been excluded,

1 00%,

with

the overall

a false-negative

rate

of zero.) Apparent portal venous flow seen on CDI in this patient may have originated in a collateral vessel in the setting of cavernous transformation. Care must be taken not to mistake

the hepatic artery is characterized by a pulsatile spectral Doppler signal, in contrast to the relatively monophasic flow pattern seen in collateral veins. Although we did not specifically attempt to determine the accuracy of angiography in the diagnosis of PVT, we were able to compare angiography and surgery in 12 patients. Angiography failed to visualize a portal vein in one patient in whom CDI and surgery both showed a patent vessel. Unfortunately, even the angiographic gold standard is not perfect. Although CDI may obviate angiography, the techniques are

or by

time elapsed However,

the

because patients with long were included. In cases in

both CDI and the gold-standard procedure showed vessels (61 of 75 patients), we felt that the interval

was irrelevant. Similarly, we did not consider relevant in five patients in whom the diagnosis of on the CDI findings was incorrect. Arteriography was performed after CDI, and it is unlikely that completely resolved in these patients. The interval

later. (This

Had this patient

by using

results in our patients with a delay of 1 month or less were very similar to those in the overall group. This suggests that

the results were intervals between

9 months in our series.

were not

technique

in our series, considerable angiography

result

the interval PVT based or surgery thrombosis might have

not

a tortuous

equivalent,

hepatic

despite

artery

the

(Fig.

3)for

temptation

collateral

to

think

“sonographic angiography” Angiography directly strates vascular patency by showing a lumen filled trast medium, whereas CDI can only imply patency ing flow.

Although

vessels

patent, the converse

Fig. 2.-61-year-old woman with portal vein thrombosis proved angiographically. A, Oblique color Doppler sonogram of porta hepatis. Lack of color indicates absence portal vein (P), implying thrombosis. Hepatic artery is seen anterioriy (arrowhead). B, venous phase angiogram shows thrombus (arrowheads) in portal vein.

of flow in

with

flow

vessels;

of

CDI

as

demonwith conby show-

on CDI can be considered

is not always true.

Fig. 3.-36-year-old man with false-positive diagnosis of portal vein thrombosis. Oblique color Doppler sonogram of perth hepatis. No color is seen in portal vein (P), Implying thrombosis. Multiple, tortuous hepatic artery loops (arrowheads) are seen anteriorly. Main portal vein was found to be patent during surgery.

TESSLER

296

CDI has technical limitations and pitfalls that can lessen accuracy and lead to erroneous results [1 9]. In our study, five of six errors made using CDI were false-positives; that is, CDI

Downloaded from www.ajronline.org by 193.255.88.62 on 01/09/15 from IP address 193.255.88.62. Copyright ARRS. For personal use only; all rights reserved

failed

to show

flow

in portal

veins

later

shown

to be patent.

We were unable to rescan any of these patients, five of whom had surgery. Any of a number of technical factors (such as inappropriate color gain, output, sensitivity, and gray-scale vs color

write

priority

settings)

could

have

resulted

in an artifac-

tual lack of flow. However, in our patients without portal vein flow on CDI, we were always able to show concomitant flow in other vessels such as the inferior vena cava. It is therefore unlikely that significant missettings of the CDI scanning parameters went unnoticed. Other technical factors could have led to a false-positive diagnosis of PVT in our series. CDI may be incapable of resolving slow flow in some instances [1 9], depending on the equipment used, the condition of the patient, and the experience of the operator. In CDI of the portal vein, the difficulty is compounded by the depth of the vessel and by the poor clinical status of many patients referred for examination. We postulate that the majority of errors in our series were related to vessels that, although patent, had only sluggish or absent flow. In his study, RaIls [1 5] diagnosed “static flow” in a portal vein branch that showed no flow on CDI but was opacified on contrast-enhanced CT. We conclude that CDI, with its high negative predictive value, is a valuable screening procedure for detecting PVTif the portal vein is shown to be patent, no further studies are required. In patients in whom CDI shows no flow in the portal vein, other imaging studies should be performed to confirm the diagnosis of PVT.

ET AL.

AJA:157,

1986:14:554-557 3. Van Gansbeke D, Avni EF, Delcour graphic

features

of portal

C, Engelholm L, Struyven J. Sonovein thrombosis. AJR 1985:144:749-752

4. Garra BS, Shawker TH, Doppman JL, Sindelar WF. Comparison

of angiog-

raphy and ultrasound in the evaluation of the portal pancreatic carcinoma. JCU 1987;15:83-93 5. Nelson AC, Lovett KE, Chezmar JL, et al. Comparison sonography and angiography in patients with portal 1987;149:77-81

system

6. Koslin DB, Borland LL. Duplex Doppler examination venous system. JCU 1987:15:675-686

P, Meire H, Williams A. Assessment portography

and

ultrasound

of

venous

in

of pulsed Doppler hypertension. AJR

of the liver and portal

7. Miller VE, Borland LL. Pulsed Doppler duplex sonography and CT of portal vein thrombosis. AJR 1985;145:73-76 8. Harkanyi Z, Temesi M, Varga G, Weszelits V. Duplex ultrasonography in portal vein thrombosis. Surg Endosc 1989;3:79-82

9. Alpem 10.

1 1. 12.

13.

MB, Rubin JM, Williams

DM, Capek

P. Porta hepatis:

duplex

Doppler US with angiographic correlation. Radiology 1987;162:53-56 Bisset GSI, Strife JL, Balistreri WF. Evaluation of children for liver transplantation: value of MR imaging and sonography. AJR 1990;155:351-356 Grant EG, Tessler FN, Perrella AR. Clinical Doppler imaging. AiR 1989;1 52:707-717 Rose SC, Zweibel WJ, Nelson BD, et al. Symptomatic lower extremity deep venous thrombosis: accuracy, limitations, and role of color duplex flow imaging in diagnosis. Radiology 1990;175:639-644 Grassi CJ, Polak JF. Axillary and subclavian venous thrombosis: follow-up evaluation with 1990;175:651-654

14. Fraser-Hill

color

Doppler

flow

US

and

venography.

Radiology

MA, Atri M, Bret PM, Aldis AE, Illescas FF, Herschom

Intrahepatic portal venous system: variations demonstrated and color Doppler US. Radiology 1990:177:523-526

15. RaIls PW. Color Doppler sonography venous system. AJR 1990:155:517-525

of the hepatic

with

SD.

duplex

artery and portal

16.

Parvey HA, Eisenberg AL, Giyanani V, Krebs CA. Duplex sonography of the portal venous system: pitfalls and limitations. AJR 1989;152:765-770 17. Grant EG, Tessler FN, Gomes AS, et al. Color Doppler imaging of portosystemic shunts. AiR 1989:154:393-397

18. Grant EG, Perrella AR, Tessler

of arterial

1991

scanning. Clin Radiol 1988:39:381-385 2. Freling NJM, Schuur KH, Haagsma EB, van der Moor J. Ultrasound as first imaging modality in superior mesenteric and portal vein thrombosis. JCU

REFERENCES 1 . Aaby N, Karani J, Powell-Jackson portal vein patency: comparison

August

19.

FN, Lois J, Busuttil

syndrome: the results of duplex 1989;152:377-381 Mitchell DG. Color Doppler imaging: Radiology 1990:177:1-10

and

color

principles,

RW. Budd-Chiari

Doppler limitations,

imaging.

AJR

and artifacts.

Diagnosis of portal vein thrombosis: value of color Doppler imaging.

This study was undertaken to determine the accuracy of color Doppler imaging in the diagnosis of portal vein thrombosis. Two hundred fifteen patients ...
732KB Sizes 0 Downloads 0 Views