333

Evaluation of Cerebral Hemispheric Contrast Transit with Intravenous Digital Subtraction Angiography

Joachim F. Seeger,1 Raymond F. Carmody, Janice R. L. Smith, Theron W. Ovitt, and Kevin McNeill

Thirty-six patients with varying degrees of stenosis of one and with minimal or no stenosis on th e opposite side with intrave­ internal carotid artery were studied using intravenous digital nous DSA. Our intravenous technique has been described [6]. subtraction angiography, assessing relative hemispheric wash­ Images were obtained with a speciall y designed system using an in and washout of contrast medium. Ipsilateral delayed contrast image intensifier with 14, 10, and 6 in ch (35.6, 25.4, and 15.2 cm) transit was seen in 16 of 18 patients with a 70% or greater field capabilities. Th e initial arch aortogram covered a fi eld from th e carotid artery stenosis. Decreased hemispheric peak contrast aorti c arch to just beyond the common carotid artery bifurcations. den sity was also seen in most of these patients. With further Subsequent oblique views of th e neck, to demonstrate the common improvements in the computer program and faster imaging rates, carotid bifurcations optimall y, in cluded th e distal common carotid intravenous digital subtraction angiography has the capability of arteries and bifurcations, as well as th e en tire internal caroti d providing valuable physiologic data along with anatomic infor­ arteries, including their cavernou s segments. The vertebral arteries mation in patients with suspected cerebral ischemia. also were generall y seen in their en tirety. The fin al anteroposterior (or posteroanterior) head images were collimated to th e skull pe­ ri phery wi th a lead template and comprised th e distal internal carotid In an attempt to identify or quantify changes in cerebral perfusion and vertebral arteries and the entire intracranial vasculature, in­ in patients with transient ischemi c attacks (TIAs), studies of cere­ cluding the peripheral middle cerebral artery territori es. Images brovascular transit times have been perform ed in nuclear medicine were obtained at a rate of one/ sec with a 512 x 512 matrix and at with radionuclide angiography, using either visual or computer­ two/ sec with a 512 x 256 matrix. assisted analysis [1 , 2]. Similar principles have been used with Th e digital head images were acquired in a linear mode and rapid , sequential (dynamic) computed tomographic (CT) scanning could be amplified in ei ther a linear or logarithmic fash ion. Logarith­ of the brain after intravenous bolus infusion of iodinated contrast mic amplification was generall y optimal for demonstratin g vessels medium [3, 4], taking advantage of the more precise intracranial passing through the dense skull base, but th e linear mode was anatomic detail of CT. However, neither technique provides what often superior in areas of more uniform background density, such many consider to be the most important information, namely, the as the peripheral cerebral vascular territori es. The relative wash in morphology of the arteries in the neck and head. This information and washout of contrast materi al in the distal internal carotid arteries can only be supplied by cervicocerebral angiography. However, and middle cerebral artery territories were evalu ated by visual angiography carries a small but definite ri sk and a significant examination, from initial opacificati on through the venous phase. expense and generally is not used as a screening examination. More recently, in an attempt to increase the sensitivity of the study The recent advent of intravenous digital subtraction angiography in cases with subtle differences on visual inspection, time-density (DSA ) has provided a new, safe, and relatively simple means of cu rves have been derived from computer-defined regions of interest assessing, on an outpatient basis, the anatomy of the aortic arch obtained from identical areas in each middle cerebral artery terri­ and great vessel origins, the common and intern al carotid arteri es, tory. We are currently im proving the program to all ow th e examin er th e vertebral arteries, and th e major intracranial arteri es [5, 6]. It more flexibility in defining the contours of the regions of interest. can be used as a screening procedure and in many instances has replaced conventional angiography as the definitive preoperative examination in patients with TIAs. Since an intravenous injection Results simultaneously delivers contrast material to all the vessels leading Eighteen pati ents had a measured 70% or greater diameter to the brain, intravenous DSA should also be able to provide stenosis of one distal common or proximal intern al carotid artery at physiologic transit time information similar to radionuclide angiog­ intravenous DSA. Sixteen of these showed a delay of contrast raphy. The objective of this preliminary study was to determin e if transit distal to the stenosis compared with the unaffected side, but intravenous DSA can demonstrate changes in contrast transit on three could not be studied beyond the arterial phase because of th e side of a carotid stenosis. image-deg rad ing motion artifacts. The other 13 patients showed a relative delay in washin and washout of contrast material in th e middle cerebral artery territory on the side of stenosis, even though Subjects and Methods there was a variable degree of demonstrable coll ateral circulation We studied 36 consecutive patients with a 50% or greater through th e circle of Willis. Three patients had computer-generated diameter stenosis of one common and/ or internal carotid artery time-density curves that confirmed the contrast delay and also

I All authors: Department of Rad iology. University of Arizona Health Sciences Center. Tu cson. AZ 85724. Address reprint requ ests 10 J. F. Seege r. AJNR 4:333-337, May / June 1983 0 195-6108/ 83/ 0403-0333 $00.00 © Ameri can Roentgen Ray Soc iety 334 NEW CONTRAST AGENTS AJNR:4, May/ June 1982

Fig. 1 .- Case 1. A, Oblique view of neck. A high-grade stenosis of left internal carotid artery origin (arrow). B, Serial one/ sec images show relative delay in both opacification and clearing of left intern al carotid and left middle cerebral arteries, and decreased contrast density in lef; convexity region compared with right. Note lack of collateral flow to left middle cerebral artery Left posterior cerebral artery (arrow) should not be mistaken for A 1 segment, which presumabl, is hypoplastic.

A

1 \, 6

3

B

showed a decrease in peak contrast density compared with th e wash in and washout on the side of stenosis occurred in an asyml - unaffected side. A fourth case showed a delay but an in crease in tomati c patient. peak contrast density on th e side of th e stenosis. Only one of th e 18 cases with less th an 70% stenosis (a measured 66%) showed an observable, albeit ve ry subtl e, relati ve delay in contrast transit, Representative Case Reports confirmed by computer-generated time-density curves. Case 1 Since nearl y all patients were stud ied for transient cerebral ischemia, and because of the lim ited numbers in thi s preliminary A 70-year-old man noted a bruit in th e left neck wh il e lyin g in bed report, no good correlation was fou nd between th e degree of at night. He gave no hi story of TIAs or oth er evidence of cerebr" ­ delayed transit or the decrease in peak contrast density and clinical vascular in suffic iency. An abnorm al left oculopleth ysmog ram led a symptoms. In fact, one of th e most clear-cut examples of delayed intravenous DSA, which demonstrated a hi gh-grade stenosis at lie AJNR:4, May/ June 1983 NEW CONTRAST AGENTS 335

Fig. 2.-Case 2. A, Serial paired im­ ages show technique of obtaining re­ gions of interest from right (left of pair) and left (right of pair) middle cerebral artery territories for time-density meas­ urements. Visual in spection alone shows 1 2 delay in contrast transil and decreased peak cortical ··stain ·· on right. B, Time- I density curves derived from A show de- lay as well as lower peak on right. .. 3 4 I· I

5· 6 J

orig in of the left internal carotid artery (fig. 1 A) . Delayed opacifica­ Case 2 tion and clearing of the left internal carotid and left middle cerebral artery were demonstrated on serial head images (figs. 1 B and 1 C). A 65-year-old man who had undergone a recent left carotid The st udy also provided additional important intracranial anatomic endarterectomy was seen after 1 day of several attacks of left body in formation by reveali ng a very hypoplastic proximal left anterior weakness and numbness. A focal stenosis at the right internal cerebral artery, which prevented significant right-to-Ieft coll ateral carotid artery origin was beli eved to be present but was difficult to circulation into the proxim al middle cerebral artery. Although con­ define with intravenous DSA due to vertebral artery superimposition. jectural, it could be argued that this patient was at a higher risk for However, serial head images clearly showed a relative delay of in farction because of this variation in the circle of Willis. Left carotid contrast flow on th e right (fig. 2A), and subsequent computer­ endarterectomy was performed without incident. generated time-density measurements showed both a delay and a 336 NEW CONTRAST AGENTS AJNR:4, May/ June 1983

Fig. 3. - 75-year-old man with clini­ cally suspected vertebrobasilar in suffi­ ciency. Four serial paired images show lefl internal carotid and left vertebral ar­ tery occlusion. Visual inspection shows simultaneous opacification of middle cerebral arteri es but slightly less density over left midconvexity (image 4). Com­ puter-generated time-density curves were identical. More adjustable regions of interest conforming to middle cerebral territories would probably provide more accurate data.

slightly lower peak density (fig . 2B) . The patient underwent unevent­ ful right carotid endarterectomy.

Discussion

It is clear that this technique does not evaluate cerebral blood flow but only compares transit of contrast material through the internal carotid arteries and th e major arteries and veins of the two cerebral hemispheres. Also, delayed opacification does not nec­ essarily reflect reduced fl ow, since norm al flow to an organ may be delayed if it must arrive by a circuitous route. For example, with common carotid artery occlusion, the intern al carotid artery may remain patent, receiving excell en t external-to-internal coll ateral flow at the level of the common carotid artery bifurcation. The dynamic intravenous DSA study will show a delayed opacification of the intern al carotid artery and hemispheric arteries and veins on the side of occlu sion, even th ough the overall flow through that internal carotid artery is not reduced. Nevertheless, we believe this tech­ nique can provide clinically useful hemodynamic information in patients with carotid artery stenosis. We already have seen several cases in which the severity of a carotid stenosis, which was difficult to grade anatomically on th e DSA study of the neck, was clarified by the transit study. At this time, intravenous DSA is not as sensitive as radionuclide angiography, which can demonstrate a delay in hem ispheric perfusion even with 50% carotid stenosis [2]. More rapid imag ing (four to six frames/ sec) will undoubtedly increase the sensi ti vity of this technique. Intravenous DSA has the additional advantage of being able to Fig. 4.-65-year-old man after a right cerebral infarct 3 weeks befor ~ define the pattern and degree of interh em ispheri c cross circulati on. with residual left hemiparesis. Upper images show occlusion of ri ght intern,,' carotid artery (arrows). Lower images show some collateral fl ow to ri g ~ ' As shown in case 1, it is hoped th at this information may help middle cerebral artery but markedly decreased convexity " stain " comparel identify th ose patients with carotid stenosis who are at greatest risk with the left. for stroke. However, considerable further investigation is neces­ sary. Even in this preliminary series, we have noted a much higher frequency of symmetric, " compensated," hemispheri c contrast flow in patients with unilateral carotid occlusion (fig. 3) th an in those with We are not certain if the relatively lower peak contrast density 0 high-grade stenosis, who almost invariably show a delay of arteri al th e sid e of stenosis seen in several of our patients refl ects dr­ filling and clearing on the side of the stenosis. As reported by Awad creased hemispheric perfusion. Although this phenomenon is USl et al. [7], symmetri c hemispheri c opacification does not necessarily usually seen most strikingly in patients with a completed stroke (fi, imply normal hemispheric perfusion. A symmetri c pattern can be 4), it can also be demonstrated in patients with TIAs (fig. 2A) an i seen in patients with abnormal xenon-1 33 inhalation cerebral blood was even seen in one asymptomatic patient (Figs. 1 B and 1 C). flow stu dies. We are currently comparin g the results of our intra­ With further refinements in technology and technique, it is 0 1 r venous DSA transi t time studies with xenon inhalation studies in hope th at a single intravenous DSA examination may in most case ; patients with unilateral carotid stenosis. provide suffic ient morphologic and hemodynami c informatic 1 AJ NR:4, May/ June 1983 NEW CONTRAST AGENTS 337

needed in the radiologic workup of the patient with unilateral TIAs, assisted radionuclide angiography in transient ischemic attack decreasing the need for some " noninvasive " examinations and, and prolonged reversible ischem ic neurological deficit. Stroke more significantly, red ucing the need for " invasive" cerebral an­ 1981 ; 12: 829-834 giography. It must be emphasized that the success of this technique 3. Norman D, Axel L, Berninger WH, et al. Dynami c computed heavily depends on the dedication of all concerned to obtain the tomography of th e brain. Techniques, data analysis, and ap­ highest quality studies possible, with intimate physician involvement plications. AJNR 1981 ;2: 1-12 in the conduct and interpretation of the procedures. Good or ex­ 4. Axel L. Cerebral blood flow determination by rapid-sequence cellent anatomic studies of the cervical and cerebral vessels and computed tomography. Radiology 1980; 137: 679-686 useful dynamic information were obtained in 90% of the cases in 5. Modic MT, Weinstein MA, Chilcote WA, et al. Digital subtraction our preliminary series. angiography of the intracranial vascular system: comparative study in 55 patients. AJNR 1981 ;2: 527 -534 6. Seeger JF, Weinstein PR , Carmody RF, et al. Digital video REFERENCES subtraction angiography of the cervical and cerebral vascula­ 1. Armas R, Kirkwood JR, Zubi S. The posterior cerebral radio­ ture. J Neurosurg 1982;56: 1 73-1 79 nuclide angiogram for detection of carotid artery disease. Clin 7. Awad I, Little JR, Modic MT, et al. Intravenous digital subtrac­ Nucl Med 1981 ;6: 19-22 tion angiography: an index Of coll ateral cerebral blood fl ow in 2. Buell U, Scheid KF, Lanksch W, et al. Sensitivity of computer internal carotid artery occlusion. Stroke 1982;13:469-472