271

imitations of Intravenous Digital Subtraction Angiography

Patrick A. Turski ,1 William J. Zwiebel,1 Charles M. Strother,1 Andrew B. CrummY,1 Gastone G. Celesia,2 and 1 Joseph F. Sackett

Eighteen digital subtraction angiography (DSA) examinations [1, 3, 4]. Forty ml meglu mine diatrizoate (Renografin 76, Squibb) were retrospectively evaluated for factors that led to their erro­ was injected at a rate of 14 ml / sec through a 5.3 French pigtail neous interpretation. Overlapping vessels obscured pathologic catheter in serted in th e superi or vena cava via a brachial ve in. The conditions in five cases. In four cases the lesions were not carotid bifurcation was evalu ated in left and right posterior oblique adequately profiled by the DSA projections. Eight lesions were projections with angulation of the patient and x-ray tube of approx­ rendered inconspicuous by misregistration artifacts attributable imately 60°. If the initial oblique views were inadequate, an teropos­ to motion, either from swallowing or from pulsation of vessel terior or repeat obli que views were obtained . When symptoms were walls. One diagnostic error was caused by poor opacification referabl e to the proxim al brachiocephali c vessels, a right posterior from a degraded contrast bolus secondary to low cardiac output. oblique vi ew of the aortic arch was obtained. The examinati on was limited to four injections, thus restricting the number of projections that could be attempted. The size of the image intensifier (1 5.2 cm Digital su btraction angiography (DSA) permits visualization of or 22.9 cm mode) req uired that each anatomic region (aortic arch, vasc ular stru ctures after intravenous injection of contrast material carotid bifurcations, or intracranial vessels) be examin ed sepa­ [1-4]. The method is sensitive, specific, and accurate in its ability rately. to demonstrate pathologic conditions involving the extracranial carotid arteries [5, 6]. The low risk associated with the procedure has resulted in rapid acceptance of the technique for the evaluation Results of vascul ar disease. However, there are situations in which inter­ pretation of the DSA examination results in an incorrect diagnosis. In nine cases DSA led to underestimation of the degree of The purpose of this investigation was to identify factors that may stenosis by at least one grade when compared with the results of contribute to diagnostic error. standard angiography. Three of these lesions were weblike plaques, not visualized by DSA , that proved to be severe stenoses (figs. 1 and 2). Conversely, in five cases the DSA examination suggested Materials and Methods a more severe stenosis th an was demonstrated by standard angiog­ raphy. In the remain ing four cases DSA failed to visuali ze ulcerated The arteriographic im ages of patients referred for standard an­ lesions, two of which were confirmed at su rgery (fig. 3). giography and intravenous DSA between May 1981 and October The diagnostic errors and their causes are summarized in table 1982 were reviewed retrospectively. The patients comprised a 1. Five stenotic lesions involving the common carotid artery or select population in which atherosclerotic lesions were common. proxim al internal carotid artery were incorrectly assessed because Eig hteen DSA examinati ons performed on members of this popu­ th e extern al carotid artery, vertebral artery, or subclavian artery lation were found to have been erroneously interpreted. Sixteen was superimposed on the lesion. Four lesions (two stenoses, two diagnostic errors involved lesions of the carotid bifurcation or ulcerations) were not adequately profil ed by the standard oblique proximal internal carotid artery, one involved the proximal right DSA projections of the carotid bi furcation. Ei ght lesions were ren­ common carotid artery, and one involved the left subclavian artery. dered inconspicuous by misregistration artifacts (in two cases, the All DS A examinations reviewed were considered adequate for di­ cause was motion of the larynx and hyoid bone secondary to agnosis at the time the examination was performed. All lesions swall owin g by the patient; in six cases, pulsations of the vessel wall stu died were atherosclerotic in nature and were c lassified as either resul ted in misregistration of calcified atherosclerotic plaque or stenoses or surface irregularities, which were considered ulcera­ poor registration of the iodin e-containing arterial lumen). tions. Stenoses were graded as minimal (0-30%), moderate (30%- 70%), or severe (70%- 99%). The DSA diagnosis was considered incorrect if subsequent standard angiography changed the grade Discussion of the stenosis (e.g., from minimal to moderate) or revealed previ­ ously undetected surface irregularities suggesting ulcerations. The The simultaneous opacification of the carotid arteries and verte­ DSA examinations that had been erroneously interpreted were then bral arteries inherent in the intravenous DSA technique is a difficult reviewed to identify the causes of the incorrect diagnoses. problem to overcome. Even when the plane of view is varied, The DSA system using a time su btracti on algorithm (also called superimposed vessels may be encountered. In our study, errors in mask mode radiography) has been described in detail elsewhere diagnosis secondary to vessel overlap were found to have occurred

' Department of Radiology, University of Wisconsin Clinical Science Center, 600 Highland Ave., Madison, WI 53792. Address reprint requests to P. A. Turski. ' Department of Neurology, University of Wisconsin, Madison, WI , and Neurology Service, Veterans Administration Hospital, Madison, WI. AJNR 4:271-273, May/ June 1983 0195- 6108/ 83/ 0403- 0271 $00.00 © American Roentgen Ray Society 272 DIGITAL RADIOGRAPHY AND DSA AJNR:4 , May/ June 1983

1A 18 2A 28

Fig. 1 .-A, Intravenous DSA of aortic arch . Left subclavian artery appears free of significant stenosis. Fig . 2.-A, Carotid DSA compromised by po!)r B, Intraarterial DSA reveals high-grade stenosis secondary to vascu lar web (arrow). vascu lar opacification from low cardiac output. C 1- rotid artery appears free of obstruction. B, Standald angiography reveals high-grade stenosis secondary to vascu lar web (arrow).

3A 38 4 Fig . 3. -A, DSA fai ls to show small ulcerations because of vessel overlap, lack of profile view, and Fig . 4. -lntracrani al DSA with incorrect logarI th­ pulsation of vessel wall. B, Standard angiography demonstrates multiple small ulcerations. mic amplification results in loss of iodine signal as carotid artery passes through petrous bon e.

when variations in the opacification of an underlying vessel were tered for DSA limits the number of projections that can be obtainl,d. used to determine the presence or absence of lesions. This method Patients have a reflex urge to swallow during the injectior of is especially unreliable when vascular webs are present. contrast material. This phenomenon occurs even when a nonionic For optimum assessment of degree of severity, lesions should be contrast agent is used [9]. Motion of the larynx from swallow'ng seen in profile. The radiographic projection must be designed in produces subtraction artifacts resulting from misregistration of he such a manner that the x-ray beam is tangent to the vessel wall in iodine-containing im age in relation to the mask image. The dif'er­ question. The most common site for atherosclerotic disease is the ence image is degraded and the carotid bifurcation may be cl e;',rly posterior wall of the carotid bifurcation [7]. This part of the carotid seen in only one projection. If this view does not profile the leflon bifurcation is very difficult to profile with DSA because of superpo­ adequately, abnormalities may remain undetected. Pulsation of the sition of vessels. The inability to obtain a true lateral view of the arterial walls also causes misregistration artifacts th at can obscu re carotid bifurcation with intravenous DSA may lead to incomplete or lesions such as vascular webs and small ulcerations. Althol1 9h incorrect assessment. After reviewing 200 standard arteriograms selection of an alternate mask may improve the quality of the im.lge of the common carotid artery bifurcation, Kasef [8] reported that [2], many of these lesions are so small that they approach the li mit obli que views adequately show the bifurcation in the majority (62%) of spatial resolution possible with current digital equipment. Thf're­ of vessels but that a true lateral view was required for a diagnosti­ fore, small ulcerations cannot be reliably detected by present in ira­ cally adequate display of the bifurcation in the remainder. Unfortu­ venous DSA techniques. The impact of this limitation is perh 1Ps nately, the large amount of contrast material that must be adminis- mitigated by the fact that ulcerated lesions are not always iden li'ied AJNR :4, May/ June 1983 DIGITAL RADIOGRAPHY AND DSA 273

TAB LE 1: Causes of Diagnostic Error in Intravenous Digital perimposed vessels and incompletely profil ed lesions can result in subtraction Angiography diagnosti c errors. Lesions such as vascular webs and small ulcer­ ation s cannot be reli ably identified with current DSA techniques. Diagnostic Error Any in consistencies between the DSA examination and the clinical Cause Underesli- Overesti- Nonvisual- Totals presentation should be resolved with standard angiography. mated mated ized Stenosis Stenosis Ulceration

Vessel overlap: REFERENCES External carotid artery 2 3 Vertebral artery ...... 1. Kruger RA, Mistretta CA, Houk TL, et aL Computerized fluo­ Subclavian artery 1 1 roscopy in real time for noninvasive visuali zation of the cardio­ Lesion not in profile 2 2 4 vascular system. Radiology 1979;130 : 49-57 Misregistration artifact: 2. Christenson PC , Ovitt TW, Fi sher HD, Frost MM, Nudelman S, Swallowing 1 2 Roehrig H. Intravenous angiography using digital video sub­ Pul sation of vessel wall 3 4 traction: intravenous cervicocerebrovascular angiography. Pulsation of calc ified plaque 2 AJNR 1980;1 : 3 79-386, AJR 1980;135: 1145-1152 Degraded contrast bolus 3. Mistretta CA , Crummy AB, Strother CM . Di gital angiography: a perspective. Radiology 198 1 ; 139: 273-276 Totals 9 5 4 18 4. Mistretta CA, Crummy AB . Di agnosis of cardiovascular disease by digital subtracti on angiography. Science 1981 ;214: 761- 765 5. Strother CM , Sackett JF, Crummy AB , et aL Clinical applica­ even with standard angiography [10]. Pu lsation of large calcified tions of computerized fluoroscopy. Radiology 1980;1 36: 781 - atheroscleroti c plaques also produces misregistration artifacts that 783 can obscure severe atheromatous lesions. 6. Ch il cote WA, Modic MT, Pavli cek WA, et aL Digital subtraction In patients with poor cardiac output, the prolonged circulation angiography of the carotid arteries: a comparative stu dy in 100 ti me degrades the contrast bolus [11]. The decrease in vascular patients. Radiology 1981 ;139: 287 -295 contrast results in less definition of th e arteri al wall secondary to a 7. Sundberg J. Locali sation of atheromatosis and calcification in reduced iodine signal. the carotid bifurcation: a post mortem rad iographic investiga­ Other possible causes of diagnostic error, not encountered in tion. Acta Radiol [Oiagn] (Stockh) 1981 ;22: 521-528 th is clin ical group, are incorrect logarithmic amplification and ex­ 8 . Kasef LG . Positional variations of the comm on carotid artery cessive image noise. Logarithmic amplificati on of the iodine signal bifurcati on: implications for digital subtraction angiography. en sures visualization of the entire arterial structure as it passes Radiology 1982;145 :377- 378 th ro ugh regions of dense bone. Insufficient amplification may lead 9. Sackett JF, Mann FA , Turski PA. Digital video angiography­ to loss of th e iodine signal, simulating a vascular occlusion (fig. 4). a study of contrast media. Excerpta Medica (in press) Si mi larly, an inadequate flux of x-ray photons produces excessive 10. Edward s JH , Krichefi II, Ril es TS, Imparato AM . An giographi­ image noise, resulting in decreased spati al resolution and poorly cally undetected ul ceration of the carotid bifurcation as a cause defined lesions. of emboli c stroke. Radiology 1979;132 : 369- 373 In summary, DSA is a safe and accurate method of assessing 11 . Hurst JW. Th e Hea rt. New York: McG raw-Hili , 1982: 1854- patients with extracran ial carotid occlusive disease. However, su- 1855