Graefe's Arch Clin Exp Ophthalmol (1995) 233:135-139 © Springer-Verlag 1995

Antonio Mendivil Victoria Cuartero Maria P. Mendivil

Received: 13 September 1993 Revised version received: 19 July 1994 Accepted: 1 September 1994

A. Mendivil. V. Cuartero M. P. Mendivil Department of Ophthalmology, Ram6n y Cajal Hospital, Alcalfi de Henares University, Madrid, Spain A. Mendivil c/Arrieta, 8, E-28013 Madrid, Spain

Color Doppler imaging of the ocular vessels

Abstract • Background: Color Doppler imaging allows for simultaneous two-dimensional anatomical imaging and Doppler evaluation of blood flow velocity. • Methods: We examined 40 normal eyes (17 males and 23 females, aged 16 to 57 years) with this technique. The color Doppler unit used in this study had a 5.0-MHz PLF503 ST phased-array scanning head. Each vessel examination was repeated 10 times during a single session. • Results: The following peak flow velocities were found:

Introduction Recent advances in ultrasonography have made it possible to evaluate orbital blood flow under real-time and physiological conditions [9, 17, 19, 21, 22]. Color Doppler probes are capable of demonstrating numerous vascular structures in the eye. Ocular vessels were difficult to sample reliably with pulsed Doppler only, because they could not be distinguished on gray-scale imaging alone. Although color Doppler improves our ability to depict vascular structures in the eye, spectral Doppler remains essential, as it defines changes in vascular impedance and provides information about flow characteristic in individual eyes [4, t3]. Color Doppler imaging (CDI) incorporates the use of two-dimensional B-scan with color Doppler. This technology allows the examination with minimal discomfort and risk of blood flow in the orbit, which cannot be seen with B-scan alone. The velocity of the blood flow can also be evaluated [9, 17, 19]. A Doppler instrument generates a continuous or pulsed beam of ultrasound that can be used to detect a change in frequency of the sound wave

central retinal artery, 12.5_+ 2.4 cm/ s; central retinal vein, 4.4_+ 0.49 cm/s; posterior ciliary arteries, 14.4 + 2.6 cm/s; vorticose veins, 5.5_+ 1.0 cm/s; ophthamic artery, 36.9 _+7.0 cm/s. Ophthalmic artery systolic and end-diastolic velocities declined as a function of age; however, these changes were not significant (systolic: r = - 0.24; diastolic, r = - 0.22). • Conclusion: This noninvasive technique allows quantitative assessment of blood flow velocity in these vessels.

caused by movement of an echo source. In ophthalmology, the targets in the body are moving acoustic interfaces such as vessel walls and red blood cells. By converting the Doppler shift into velocity information, one can quantify the amount of flow through a vascular structure and derive precise information on the character of the flow. When a frequency shift due to blood flow toward the transducer occurs, it is usually displayed on the screen by a red color. Conversely, when blood flows away from the probe, the Doppler response is typically indicated by a blue color on the B-scan display [19, 22]. There are not many reports in the literature of applications of CDI in the study of the eye [4, 13]. In this paper, we present our CDI findings in the ophthalmic and central retinal vessels in 40 normal individuals.

Patients and methods The Toshiba Sonolayer SSA-270 color Doppler unit was used in this study, with a 5.0-MHz PLF-503 ST phased-array scanning head. The CDI examination was performed with the probe placed

136

on the closed eyelids using methylcellulose as a coupling gel. The subject was reclined with head close to the screen of the instrument. The duration of the Doppler spectral analysis ranged from 30 to 60 s per eye, and the ocular examination lasted approximately 15 rain. All examinations were performed and images interpreted by the same experienced individuals. The systolic and end-diastolic blood flow velocity in the ophthalmic artery (OA), central retinal artery (CRA), posterior ciliary arteries (PCA), central retinal vein (CRV) and vorticose veins (VV) were measured in all of the volunteers. Peak systolic velocity measurements were obtained by placing the sample volume in the highest-velocity flow stream with the angle correction cursor parallel to the colorencoded lumen. The vascular examination was repeated for each vessel 10 times during a single session. The blood flow is depicted in either blue or red depending on direction. When spectral analysis was performed, the estimated in situ peak temporal average (SPTA) intensity fell well below the safety limit established by the American Institute of Ultrasound in Medicine (AIUM) [14]. We reduced the time of exposure to pulsed Doppler examination as much as possible; color Doppler was used first to provide a survey of the area of interest and identify appropriate blood vessels. Once optimally visualized by color flow, specific ocular vessels could then be rapidly interrogated by pulsed Doppler. Limiting the pulsed Doppler portion of the examination is particularly important as color imaging uses much less power than pulsed Doppler and, hence, is probably safer. The CDI machine features artifacts similar to those seen with other sonographs. These can be divided into imaging artifacts (acoustic enhancement, acoustic shadowing, beam width artifact, reverberations, etc.) and Doppler artifacts (range ambiguity artifact, signal dropout, mirroring, etc.). Some experience is needed before using CDI in order to minimize these artifacts. Forty-three individuals were examined using the above-described fashion. All were healthy and without ocular or systemic vascular disease. Three volunteers were excluded because of technical difficulties (unclear definition of spectral display). The remaining 40 subjects, 17 males and 23 females, aged 16-57 years, showed acceptable graphic representation of Doppler information.

Fig. 1 Color Doppler imaging (CDI) of the ophthamlic artery, showing initial high steep maximum systolic peak and mid-diastolic peak flow. The ophthalmic artery nasal to the optic nerve is easily identified because it comes toward the transducer

Results The OA, CRA, CRV, VV and PCA were visualized and studied in all subjects. The OA, which is a branch of the internal carotid artery, can be traced up inferolateral to the optic nerve. After entering the orbit, it continues to lie inferolateral to the nerve for a short distance but usually (in 80% of cases) crosses over the nerve and runs medially. This crossing over on the part of the OA was easily demonstrated in some volunteers. The OA could be identified on the basis of its pulsatility, its position parallel to the optic nerve and its flow toward the transducer. The flow velocity waveform is similar to that of the internal carotid artery showing a step maximum systolic peak, often a dicrotic notch, and low diastolic velocity (Fig. 1). The first important intraorbital branch of the OA is the CRA, which enters the optic nerve approximately 12 mm behind the globe. The C R A and accompanying C R V can be seen within the a n t e r i o r 2 m m of the optic nerve shadow. T h e C R V is distinguished by the contin-

Fig. 2 CDI of the central retinal artery (CRA) and vein (CRV) parallel to each other within the optic nerve shadow. Observe lag between CRA systolic peak (up) and subsequent CRV puls peak

(down)

uous venous flow and the C R A by the pulsatile arterial flow (Fig. 2). The OA has a host of other orbital ramifications: the two major PCA (one on the nasal side, which branches off first and in 80% of people has the same point of origin as the CRA, and the other on the temporal side), the lacrimal artery, the ethmoidal arteries, numerous muscular arteries, and the supratrochlear artery [11, 12]. The short posterior ciliary arteries, about 20 in number, and the two long posterior ciliary arteries are all derived from the PCA. On either side of the optic nerve slightly posterior to the CRA, the short and long posterior ciliary arteries can be identified. The Doppler spectrum of

137

Table 1 Velocitiesin ocular vessels (cm/s, n = 40)

Maximum systolic Minimum diastolic blood flow, blood flow, mean ,+SD (range) mean ,+ SD (range) Ophthalmic artery Central retinal artery Central retinal vein

36.9 ,+7.0 (29~49) 12.5 -+2.4 (9-17) 4.37 ,+ 0.49

(~5)

Posterior ciliary artery 14.4 ,+2.6 (lO-2t) Vorticose vein 5.5 ,+ 1.0 (5-8)

Fig. 3 CDI of a posterior ciliary artery (PCA), lateral to the optic nerve. The diastolic flow in the PCA is greater than in the CRA, probably reflecting low vascular resistance

10.6 _+4.3 (6-20) 4.0 _+0.72 (3-5) 2.62 _+0.49

(2-3)

4.75,+ 1.7 (3-8) 3.62_+0.87 (2-5)

subjects with paired observations, no statistically significant difference was seen between the first and second measurements.

Discussion

Fig. 4 CDI of a vorticose vein showing typical venous flow

the PCA showed velocity-time spectra very similar to those of the CRA. The end-diastolic flow in the PCA was, however, higher, reflecting the low-resistance vascular channels of the choroid (Fig. 3). The VV showed typical continuous venous flow (Fig. 4). The results of the study are summarized in Table 1. The systolic and end-diastolic peak flow velocities in the OA declined with increasing age of the patients (systolic r = - 0 . 2 4 ; diastolic r = -0.22), but these differences were not significant. The ratio of systolic peak velocity to en-diastolic peak velocity rose as a function of age and of the systolic-diastolic blood pressure ratio in the OA (for age r=0.20; for systolic-diastolic blood pressure ratio r=0.19), No significant difference was noted between right and left eyes or between male and female subjects. In

CDI can demonstrate the relative direction and velocity of blood flow in color, superimposed on a conventional gray-scale ultrasound image that depicts stationary tissue. The addition of color Doppler sonography has made the detection of vessels easier and the correction of the Doppler angle more accurate, resulting in more rapid and precise acquisition of data [9, 17, 19]. The status of the flow in the ocular vessels is routinely obtainable along with anatomic information about the structures that they supply. Indications for CDI include: peripheral vascular scanning, carotid artery imaging and flow analysis [5], pediatric echocardiography [3], liver/gall bladder scanning, urinary tract/ retroperineal area scanning, gynecology scanning [8], pediatric abdominal scanning, fertility studies scanning, thyroid scanning, breast scanning, abdominal aortic scanning [2], pancreas/spleen scanning, prostate/testicular/penile scanning [18], fetal scanning, male pelvic tumor detection scanning and as an adjunct to vascular surgery [20]. CDI of normal [4, 13] and pathologic eyes has been performed [7, 10]. The frequencies used in ophthalmology are generally in the range 5-10 MHz. Lower frequencies penetrate deeply into the orbit but have poor resolution, while higher frequencies have fine resolution but may penetrate only part way into the orbit [19, 22]. We obtained sufficient resolution of the minute structures in the eye and orbit with a frequency in the range of 5 M H z (7.5 MHz is the optimal frequency for examining most orbital vasculature). It is not possible to calculate blood volume measurements, because CDI cannot accurately assess the diameters of the vessels. Retinal vessel flow was detected in all patients, but not measured because

138

of the low definition of the flow-velocity waveform. The short and long posterior ciliary arteries were identified, but distinction among them were not possible. The calculated velocities in the OA, CRA, CRV and PCA in our 40 volunteers did not differ significantly from those found by other authors [1, 6, 13]. We have not found published in the literature blood flow velocities in the VV of healthy volunteers. We found that the blood flow velocities declined with age, but these changes were not significant. The normal characteristic low-resistance waveform was seen for the PCA (with gradual descent from peak systolic velocity and relatively high diastolic velocity. Mean peak systolic and end-diastolic velocities in the PCA were greater than in the CRA. Flow in the CRV was depicted in blue, showing that it runs from proximal to distal, with a typical venous flow patterns. The fast Fourier transform spectrogram of the pulsed Doppler sampled in the colored flow showed either typical arterial flow for arteries or typical venous flow for veins. Isolated velocity measurements (peak systolic, end-diastolic, etc.) provide little information about the vascular impedance; they are usually of interest in ocular examination for the information they give regarding stenoses. For these reasons, the ratio between systolic and diastolic flow is used; it gives an estimate of the distal arteriolar bed and is angle-independent. The pulsatility index (PI) is defined as (peak systolic velocity -end-diastolic velocity) divided by mean velocity. This measurement may be most exact because it takes into account the entire frequency spectrum during each cardiac cycle; lower indices represent increased diastolic flow and decreased vascular impedance. The PCA have significantly lower impedance to flow than the CRA (PI lower). Differentiation between the short and long posterior ciliary arteries and distinction among them has not been possible. This trouble is usual in most studies with CDI. The range of error estimated with Doppler flow data of the maximum systolic velocity is about 18% [6]. This

is the reason why measurement of velocity in units more precise than centimeters per second is not worth-while in the ocular vessels. CDI can demonstrate the dilated, arterialized superior ophthalmic vein (SOV) with high-velocity blood flow toward the transducer, characteristic of carotid cavernous sinus fistula. This capability may eliminate the need for computed tomography and magnetic resonance imaging in the examination of suspected arteriovenous fistulas. After embolization CDI can also verify occlusion of the fistula. Presumed dural cavernous arteriovenous malformation also shows dilated SOV with blood flow reversal, dilated preseptal high-blood-flow shunts, and thickened extraocular muscles [7]. CDI may also evaluate orbital varix and vascularity in intraocular tumors, providing in vivo cross sections of the vascular topography of tumors in real time. For example, the vasculature of melanomas play a role in the response to irradiation; well-perfused tissue is more radiosensitive than poorly perfused tissue [10]. CDI will probably become an important tool in the evaluation of patients with signs or symptoms of orbital disease. The technique is harmless, independent of the status of the ocular median, and noninvasive. The SPTA is the mean value of the ultrasonic intensity, measured at that point in the beam where there is maximum ultrasonic intensity. This parameter varies with the area being examined and the type of scan being performed. In the guideliner established by the United States the Food and Drug Administration on the maximum permissible intensity for Doppler imaging, the suggested SPTA is 17 mW/cm 2. Nevertheless, 100 mW/ cm 2 is the limit established by the AIUM [14]. Only with very high ultrasound energies (about 100 W/cm2), several orders of magnitude more than the power output used in this study, did Lizzi et al. produce minimal choroidal lesions in rabbits [15, 16]. In conclusion, CDI is a new modality for the study of ocular and orbital hemodynamics and allows detailed analysis of the blood flow in the eye.

References 1. Berger RW, Guthoff R, Helmke K, Winkler P, Draeger J (1989) Doppler sonographische Befunde der arteria and vena centralis retinae. Fortschr Ophthalmol 86:334-336 2. Bezzi M, Mitchell DG, Needleman L, Goldberg BB (1988) Iatrogenic aneurysmal protal-hepatic venous fistula. J Ultrasound Med 7:457-461 3. Duncan WJ (1988) Color Doppler in clinical cardiology. Saunders, Philadelphia, pp 1-5

4. Erickson SJ, Hendrix LE, Massaro BM, et al. (1989) Color Doppler flow imaging of the normal and abnormal orbit. Radiology 173:511-516 5. Erickson SJ, Mewissen MW, Foley WD, et al. (1989) Color Doppler evaluation of arterial stenoses and occlusions involving the neck and thoracic inlet. Radiographics 9.389-406 6. Feke GT, Tagawa H, Deupree DM, Goger DG, Sebag J, Weiter J (1989) Blood flow in the normal human retina. Invest Ophthalmol Vis Sci 30:5865

7. Flaharty PM, Lieb WE, Sergott RC, Bosley TM, Savino PJ (1991) Color Doppler imaging: a new noninvasive technique to diagnose and monitor carotid cavernous sinus fistulas. Arch Ophthalmol 109:522-526 8. Fleischer AC (1991) Ultrasound imaging 2000: assessment of uteroovarian blood flow with transvaginal color Doppler sonography. Potential clinical applications in infertility. Fertil Steril 55:684

139

9. Grant EG, Tessler FN, Perrella RR(1989) Clinical Doppler imaging. Am J Radiol 52:707-717 10. Guthoff RF, Berger RW, Winkler P, Helmke K, Chumbley LC (1991) Doppler ultrasonography of malignant melanomas of the uvea. 109: 537-541 1I. Hayreh SS (1974) The choriocapillaries. Albrecht van Graefes Arch Klin Exp Ophthalmol 192:165-179 12. Hayreh SS (1974) Submacular choroidal vascular pattern. Experimental fluorescein fundus angiography studies. Albrecht von Graefes Arch Klin Exp Ophthalmol 192:181196

13. Lieb WE, Cohen SM, Merton DA, et al. (1991) Color Doppler imaging of the eye and orbit: technique and normal vascular anatomy. Arch Ophthamol 109: 527-531 14. Lizzi FL, Mortimer AJ (1988) Bioeffects considerations for the safety of diagnostic ultrasound. J Ultrasound Med 7 [Suppl]:l-38 15. Lizzi FL, Packer AJ, Coleman DJ (1978) Experimental cataract production by high frequency ultrasound. Ann Ophthalmol 10:934-942 16. Lizzi FL, Coleman DJ, Driller J, Fanzen LA, Leopold M (1981) Effects of pulsed ultrasound on ocular tissue. Ultrasound Med Biol 7:245-252 17. Merritt CR (1987) Doppler flow imaging. J Clin Ultrasound 15:591-597 18. Moddleton WD, Thorne DA, Melson GL (1989) Color Doppler ultrasound of the normal testis. Am J Roentgenol 152: 293-297

19. Powis RL (1988) Color flow imaging: understanding its science and technology. J Diagn Med Sonogr 4:234-245 20. Rubin JM, Hatfield MK, Chandler WF, Black KL, DiPietro MA (1989) Intracerebral arteriovenous malformations: intraoperative color Doppler flow imaging. Radiology 170:219-222 21. Scoutt LM, Zawin ML, Taylor KJW (1990) Doppler ultrasound. II. Clinical applications. Radiology 174:309-319 22. Taylor KJW, Holland S (1990) Doppler ultrasound. I. Basic principles, instrumentation, and pitfalls. Radiology 174:297-307