Digital Subtraction Cerebral Angiography by Intraarterial Injection: Comparison with Conventional Angiography
Total Page:16
File Type:pdf, Size:1020Kb
593 Digital Subtraction Cerebral Angiography by Intraarterial Injection: Comparison with Conventional Angiography Michael Brant-Zawadzki1 , 2 For 4 months, a prototype digital subtraction system was used to obtain images of Robert Gould2 the cerebral vasculature after intraarterial contrast injections. In 12 instances, the David Norman2 intraarterial injections were recorded with both a digital subtraction unit and conven Thomas H. Newton2 tional direct magnification film-screen system. The digital subtraction and conventional Barton Lane3 film subtraction images were compared and graded for quality and information content by three skilled observers. In addition, quantitative measurements of contrast-detail performance and spatial resolution were obtained on both the digital system and the screen-film imaging chain. In a clinical setting, both the digital subtraction and conven tional film-screen systems provided similar quality images and angiographic informa tion. Contrast-detail curves demonstrated that digital subtraction angiography outper formed conventional film technique for low-contrast objects. Digital subtraction an giography also reduced the time required to obtain the angiogram, markedly reduced film cost, and lowered the contrast agent burden. Digital subtraction imaging of the extra- and intracranial cerebral vasculature after intravenous injections of contrast materi al is becoming an established screening procedure [1-3]. Patient motion, superimposition of multiple vessels, contrast agent burden, and suboptimal resolution [4] restrict the diagnostic utility of this method. Early work has suggested, however, that digital subtraction imaging of intraarteri al contrast injections can provide images of sufficient diagnostic quality to obviate conventional film-screen angiography, thus reducing film cost, and possibly decreasing time required for the examination [5, 6]. For 4 months, we have evaluated intraarterial cerebral digital subtraction angiography (DSA). In selected instances, a comparison of digital and film-screen angiography was made in the same patient. In addition, quantitative measurements of spatial resolution and contrast sensitivity were obtained for both the digital and film This article appears in the November / Decem screen systems. ber 1982 issue of AJNR and the February 1983 issue of AJR. Received April 27. 1982; accepted after revi sion July 9. 1982. Subjects and Methods Presented at the XII Symposium Neuroradiolo gicum, Washington, D.C., October 1982. Equipment 1 Departm ent of Radiology, San Francisco Gen The OSA unit (OF 100, Oi asonics, Inc.) was a prototype, the essential components of eral Hospital, 1001 Potrero Ave., San Francisco, which included a progressively scanned video camera with a 1000:1 signal-to-noise rati o, CA 9411 O. Address reprint requests to M. Brant Zawadzki. interfaced with a 9-inch- (22.9 cm) diameter, triple-mode, cesium iod ide image intensifier (Fluoricon 300, G.E. Medical Systems). Exposures were pulsed using a 0 .6 mm focal spot 2 Department of Radiology, University of Cali fornia Medical Center, San Francisco, CA 94143. x-ray tube (Maxi 100, G.E. Medical Systems). Typical exposure fac tors were 70-85 kVp, 600 rnA , 0.01 3-0.030 sec. Mask mode subtraction images were recorded on a multiformat 3 Department of Radiology, Mt. Zion Hospital, San Francisco, CA 94 120. camera (Matrix In struments). Conventional 1 .8-2 x magnification arteriograms were obtained with Quanta III screens AJNR 3:593-599, November/ December 1982 0195-6108/ 82/0306- 0593 $00.00 and Cronex 4 film (duPont, Inc. ), using a 0.16 mm or 0 .2 mm biased focal spot tube (Maxi © American Roentgen Ray Society 125, G.E. Medical Systems). 594 BRANT-ZAWADZKI ET AL. AJNR :3, November/ December 1982 Clinical Studies systems. A Lucite ph antom was used to analyze the trade-off between spatial resolution and contrast sensitivity. Holes of pro Twenty-two pati ents were studied after conve ntional percuta gressively smaller diameter and depth were drilled through a 2 .54- neous selective catheterization of th e carotid and/or vertebral cm-thick Lucite block, the sm all est holes measuring 0.5 mm (fig. arteries. The patients were 14-82 years old and exhibited a broad 1). Measurements of x-ray intensity transmitted through the solid spectrum of pathology (table 1). A total of 50 angiographic seri alo block of Lucite and through different thicknesses of Lu cite corre grams were obtained usin g a mask mode digital subtraction tech sponding to that thickness remaining after the holes were drilled nique. In 12 instances, images were also recorded on the conven were made using a calibrated ionization chamber (MOH model tional film-screen system. Contrast injecti on rates for th e conven tional angiog rams were standard (8 ml volume, 5 ml /sec rate in the internal carotid artery; 10 ml volume, 8 ml /sec rate in the common 100 carotid artery; 9 ml volume, 6 ml /sec rate in th e vertebral artery). 80 When the patients were studied with DSA, the contrast volume used was typically 60%-80% that of the conventional arteri ographic 50 study. The rate of injecti on was also decreased by a similar factor. Conventional film subtracti ons of selected key images were ob 30 tained. We then matched the digital and conventional subtraction 20 digital images, and three of the coauthors evaluated the matched pairs, unsubtracted III! grading them on a 1 -3 scale (1 = inadequate; 2 = adequate; 3 = Cii excell ent). Three separate categories of informati on were evalu ~ 10 C ated: large vessel (>1 mm) resolution, small vessel «1 mm) reso 0 8 lution (e.g., anterior choroidal, lenticulostriate, etc. ), and image u contrast within the vessels. The points awarded in each category 5 were then summed for both the digital and the conventional film 3 angiograms and compared. In addition, a judgment was passed on each study as to its overall diagnostic quality. 2 Quantitative Measurements 1 0.1 .2 .3 .5 .7 2 3 5 7 10 A conventional lead-bar test pattern was used for measurements Hole diameter, mm of the spatial resolution of both th e digital and screen-film imaging Fig. 2. - Contrast-detail curves for conventional film-screen and OSA im ages obtained by plotting smallest diameter hole visible at any given contrast level (see tex t). Imag ing system performance improves as curve is closer to orig in , since a smaller-sized obj ect can be visualized at a given TABLE 1: Clinical and/ or Radiologic Diagnosis in Patients object contrast level. St udied with DSA Pathotogy No. Patients Tumor 4 TABLE 2: Comparison of DSA and Conventional Angiography Carotid atherosclerosis 3 Vertebrobasilar occlusion ........... .. 3 Totat Score' Tra uma 3 OSA Conventional Aneurysm/ subarachnoid bleed 3 Intraventricular bleed 1 Large vessel resolution 99 100 Crypti c arteriovenous malformation 2 Small vessel resolution 75 83 Spontaneous cavern ous/ carotid fi stula 1 Contrast density 102 95 Other .... ... .. .. 2 Diagnostic study? All All - - ---- Total 22 • Totaled fo r 12 studies by three observers in each category. Scores: 1 = suboptimal; 2 = good: 3 = excellent. Fig. 1.- Contrast-detail phantom. A, Ori ginal film-screen image of Lucite phantom (mag ni fica ti on X l .8) with holes of varying diameter drilled to progres sively shallower depths from left to ri gh t. B, Conventional fil m subtraction shows inc rease of random (q uantu m) noise on image of phantom . A B AJNR:3, November/ December 1982 CEREBRAL DIGITAL SUBTRACTION ANGIOGRAPHY 595 A B c o Fig . 3.- Spontaneous dural arteriovenous communication in 42-year-old compared with conventional subtraction, accounting for slight diHerences in woman with right exophthalmos. Left internal carotid angiography. A, Se number of vessels filled. C, Conventional subtraction angiogram, magnifica lected conventional subtraction of angiogram, AP view (magnification X l .8). tion x 2, in lateral projection. D, DSA, in the lateral projection (reversed Abnormal arteriovenous communication with filling of contralateral (right) polarity). Abnormal dural communication in posterosuperiqr cavernous sinus cavernous sinus and superior ophthalmic vein. e, Matched digital subtraction, region, with retrograde filling of right superior ophthalmic vein and basilar AP view. Comparable detail of anterior choroidal artery, ophthalmic artery, venous plexus. and larger vessels. Digital subtraction was obtained earlier in flow sequence 1015). The subject contrast for each hole depth was defined as a Results percentage difference in measured x-ray intensity transmitted through the original Lucite block as compared with that transmitted Our clinical data summarized in table 2 show that both through the thickness of Lucite remaining after the hole was drill ed. techniques provided eq uivalent large-vessel resolution (> 1 For these measurements, the ionization chamber was placed in the image plane and the x-ray field size was not changed. The diameter mm). Vessels less than 200 ILm in diameter were not as well of the smallest hole visible at each particular depth (contrast) was resolved on the DSA images, as would be predicted from then plotted on a graph (fig. 2). Separate curves were obtained for both the theoretical and measured limits of spati al resolution non subtracted and subtracted digital images and non subtracted of the two systems. Overall image contrast was superior in and subtracted film images. the DSA images despite the lower contrast volume and 596 BRANT-ZAWADZKI ET AL. AJNR:3, November/ December 1982 A 8 Fig . 4.-Encasement of callosomargin al artery due to metastasis to falx in terior cerebral artery at rostrum of corpus call osum . B, DSA (later phase in 60-year-old woman with lung carcinoma. A, Conventional subtraction arterio flow seq uence). En casement of callosomarginal and in cidental narrowing of gram, magnification x 2 (right internal carotid artery injection). Encasement anterior cerebral artery. Comparable small vessel detail. of callosomargin al artery. Incidentall y, mild atherosclerotic narrowing in an- 8 Fig. 5.-Atherosclerosis of precavernous and cavernous segment, right image from DSA study several days later.